IAVSD 2025

Aug 18, 2025, 12:00 AM → Aug 22, 2025, 11:59 PM Europe/Warsaw

Shanghai

Jun Wang (Tongji University)

Mon, August 18

7:30 AM → 8:30 AM Registration

Note: The main conference sessions will be held at the Yifu Building, Tongji University, 1239 Siping Road, Shanghai throughout the symposium. Parallel sessions will be organised in the classrooms adjecent to the Yifu Building. Each oral presentation will be 30 minutes long and each parallel session will include 3 or 4 presentations. Guidance for presenters will be available via the conference website soon.

8:30 AM → 9:00 AM Opening Ceremony

9:00 AM → 10:00 AM State of the Art I

Conveners: Tim Gordon, Wanming ZHAI

9:00 AM Motion Comfort: What about Human Body Dynamics in Road and Rail Vehicles?

Speakers: Chen Shen (Section of Rail Systems and Monitoring, TU Delft, Netherlands), Georgios Papaioannou (Department of Cognitive Robotics, TU Delft, Netherlands), Malte Rothhämel (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Riender Happee (Department of Cognitive Robotics, TU Delft, Netherlands)

10:00 AM → 10:30 AM Plenary I

Conveners: Tim Gordon, Wanming ZHAI

10:00 AM Negotiation Analysis of Electric Multiple Units (EMUs) Under Traction/Braking Conditions

Paper ID: 266

The Sichuan-Tibet railway features lots of helix lines, where horizontal curves and gradients are coupled on certain sections to accommodate topographical constraints. Therefore, it is necessary to study the dynamic performance of trains passing through curves under traction/braking conditions. Research indicates that, compared to coasting conditions, traction and braking have a significant impact on dynamic performance, with the traction torque notably exacerbating the trend of deteriorating curve negotiation performance, while the breaking torque sometimes optimize the operation safety of EMUs. The trend of changes in wheel load reduction rate and wheel axle lateral force with respect to traction/braking force is nearly linear. Furthermore, the impact of torque distribution between the front and rear axles of a bogie on vehicle’s safety and steering on curved tracks is studied. The simulation results prove that safety and wear performance are generally satisfactory by distributing torque to the rear axle, and the larger the torque value is, the more obvious the optimization effect is. And Optimized torque distribution strategy can also improve stability of EMUs.

Speakers: Dao Gong (College of Transportation, Tongji University, China), Jinsong Zhou (College of Transportation, Tongji University, China), Yiyang Song (College of Transportation, Tongji University, China), Yuheng Gu (College of Transportation, Tongji University, China), Zhanfei Zhang (College of Transportation, Tongji University, China)

10:30 AM → 11:00 AM Coffee Break

11:00 AM → 12:30 PM Rail 1: Track System I

Convener: Michael Craft (MxV Rail, USA)

11:00 AM Direct and Indirect Calculation of the Track Decay Rate Based on the Analytical Method

Paper ID: 215

Rolling noise constitutes a major component of the total noise in railway systems and plays a critical role in environmental noise pollution, particularly at medium to high train speeds. One of the key factors influencing rolling noise is the track decay rate (TDR), which is a crucial parameter for understanding and mitigating this type of noise. Accurate determination of the TDR is essential for estimating the noise radiation from the tracks. Typically, the TDR is obtained experimentally using the impact hammer method (direct approach) or the energy iteration method during train pass-by (indirect approach). However, both methods are complex and resource intensive. Furthermore, studies have revealed significant discrepan-cies between the TDR values obtained from these two methods. To address these challenges, this paper investigates the feasibility of calculating the TDR employ-ing the theoretical modeling and validates the accuracy of the proposed models through experimental measurements.

Speakers: Markus Hecht (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany), Qiuyong Tian (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany), Yifan Yang (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany)

11:30 AM A Novel Method for Determining Track Irregularity Thresholds on Long-Span High-Speed Railway Bridges

Paper ID: 22

This study proposes an improved method for evaluating track irregularities on long-span high-speed railway bridges. Traditional chord measurement method (CMM) often relies on a fixed chord length and neglect ultra-long-wavelength ir-regularities induced by static bridge deformation (SBD). To address these limita-tions, a multi-chord approach is developed, incorporating both 5 m and 80 m chords to effectively capture short- and long-wavelength irregularities. A refined train–track–bridge coupled dynamic model, which includes carbody flexibility, is used to determine CMD thresholds based on ride comfort and safety criteria. Simulation results identify sensitive wavelength bands and define CMD limits for both lateral and vertical directions. Additionally, the method introduces a correc-tion mechanism to isolate the influence of SBD on CMD, allowing for accurate evaluation under varying bridge deformation amplitudes. The proposed approach offers a more comprehensive framework for static acceptance of track geometry, enhancing its applicability to complex long-span bridge environments.

Speakers: Ruoyu Li (Southwest Jiaotong University, China), Shengyang Zhu (Southwest Jiaotong University, China), Wanming Zhai (Southwest Jiaotong University, China)

12:00 PM Rail corrugation identification using a time-frequency normalized superposition method of multi-source detection data

Paper ID: 131

The identification method of rail corrugation has been widely studied based on axle box acceleration (ABA), wheel-rail force (WRF), and vehicle interior noise (VIN), while there is little research on the rail corrugation identification through data fusion for ABA, WRF, and VIN. The traditional approach employs time-frequency analysis to calculate the energy ratio of characteristic frequency bands associated with rail corrugation from single-source data, with identification conducted through threshold-based detection. However, this approach is affected by the frequency transfer characteristics of sensors, leading to an imbalance in the energy ratio evaluation metric for characteristic frequency bands. To address this problem, we propose an improved rail corrugation identification algorithm, named time-frequency normalized superposition method (TFNSM), which enhances identification accuracy by taking more information into consideration. Mileage alignment of multi-source data is an important prerequisite for data fusion. ABA and WRF data belong to the detection data of the high-speed comprehensive inspection train (HCIT), with high and consistent mileage accuracy. The self-developed portable onboard device can not only collect the VIN, but also time-synchronously detect the three-directional acceleration and angular velocity of the vehicle body. Among them, the longitudinal acceleration of the vehicle body is used for integral calculation to estimate the train speed and mileage information, and derivative dynamic time warping (DDTW) is adopted to estimate the mileage deviation between the angular velocity and the track curve, thereby correcting the estimation error of train speed and mileage. Then, the TFNSM algorithm is adopted to fuse the multi-source data after mileage alignment and calculate the energy ratio of the characteristic frequency band. High-speed railway line testing demonstrates significant enhancement in mileage alignment accuracy for VIN data, with near-zero mileage errors observed in curved sections. When the energy ratio threshold was set to 0.3, the identified rail corrugation locations matched the field conditions, and the estimated rail corrugation wavelengths (101.5 mm and 126.5 mm) were consistent with measurements obtained using the rail corrugation trolley. The experimental results indicate that the proposed approaches can achieve improved accuracy for rail corrugation identification.

Speakers: Fei Yang (Infrastructure Inspection Research Institute, China Academy of Railway Sciences, China), Jianli Cong (Infrastructure Inspection Research Institute, China Academy of Railway Sciences, China), XUEGENG MAO (Infrastructure Inspection Research Institute, China Academy of Railway Sciences, China), Xianfu Sun (Infrastructure Inspection Research Institute, China Academy of Railway Sciences, China), Zilong Wei (Infrastructure Inspection Research Institute, China Academy of Railway Sciences, China)

11:00 AM → 12:30 PM Rail 1: Vehicle Dynamics I

Convener: Jinsong Zhou (Tongji U, China)

11:00 AM Dynamic performance evolution of high-speed trains considering different service states of yaw damper

Paper ID: 243

The high-speed trains operating at a higher speed represent a key development direction for railway transportation, making it imperative to thoroughly ex-plore their dynamic performance transformation laws in the context of elevat-ed speeds and complex operational environments. The performance of yaw damper is significantly influenced by environmental temperature and the damping valve working state, has a considerable impact on the dynamic char-acteristics of the high-speed trains. This study investigates the dynamic per-formance evolution of high-speed trains with elevated operation speed under varying service states of yaw damper. A high-speed train dynamic model in-corporating a comprehensive physical parameter model of the yaw damper, has been developed considering different temperatures and damping valve states. The findings indicate that the high-speed trains maintain good stability with elevated speeds under different service states of yaw damper and low equivalent conicity conditions. As the equivalent conicity increases, the dy-namic performance of the high-speed train undergoes a substantial enhance-ment across a range of temperatures and yaw damper working states. The in-sights gained from this study offer valuable implications for the design, opti-mization and maintenance of the high-speed trains operating at elevated speeds.

Speakers: Chenchen Jiang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Liang Ling (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Wanming Zhai (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Zheshuo Zhang (Intelligent Transportation System Research Center, Hangzhou City University, China)

11:30 AM Dynamic Characteristic Research of Primary Suspension under Medium and High Frequency Vibration

Paper ID: 65

This study investigates the dynamic characteristics of the steel spring and hydraulic damper in railway vehicle primary suspension systems under medium-to-high frequency excitations. Experimental tests and finite element simulations were performed to analyze the modal frequencies, vibration modes, and damping properties of the spring under both no-load (AW0) and full-load (AW3) conditions. The results indicate that the first-order natural frequency of the spring concentrates near 85Hz, with a nonlinear stiffness increase observed under AW3 conditions attributable to coil contact. Subsequently, dynamic testing of the primary suspension was conducted. The dynamic stiffness of the steel spring exhibited significant amplification near its natural frequency, followed by a progressive decline with increasing frequency. Meanwhile, the energy dissipation efficiency of the hydraulic damper exhibited a frequency-dependent decrease, particularly when the steel spring resonated with external excitation at approximately 85 Hz. This resonance condition resulted in inadequate damping performance, facilitating vibration energy transfer to the bogie frame and consequently accelerating structural fatigue accumulation. This study reveals the dynamic response mechanism of the primary suspension system under medium-high frequency excitation, which is of great significance for improving the operational stability and safety of vehicles.

Speakers: Huanyun Dai (State Key Laboratory of Rail Transit Vehicles System, Southwest Jiao tong University, China), Li Xin (State Key Laboratory of Rail Transit Vehicles System, Southwest Jiao tong University, China), Suqin Wang (State Key Laboratory of Rail Transit Vehicles System, Southwest Jiao tong University, China)

12:00 PM Influences of Rail Vehicle Suspension Dynamics to Train Tractive Energy Savings

Paper ID: 53

This study investigates the influence of rail vehicle suspension dynamics on train tractive energy consumption, focusing on energy dissipation through vehicle vibrations. Vibrational energy, while relatively small compared to other resistance forces, is persistent and accumulates over time, impacting overall energy efficiency. A multibody dynamic simulation model was developed to analyze energy consumption in freight wagons with and without primary suspension rubber pads under various speeds (20–80 km/h) and track conditions (FRA classes 3–6 and smooth track). The results demonstrate that incorporating rubber pads in the suspension system reduces energy consumption by damping vibrations, with energy savings diminishing as track conditions worsen. This research highlights the potential of small design modifications, such as rubber pads in three-piece bogie suspensions, to enhance energy efficiency, contributing to decarbonization efforts in rail transport by addressing previously overlooked vibrational energy dissipation in tractive resistance.

Speakers: Colin Cole (Centre for Railway Engineering, Central Queensland University, Australia), Maksym Spiryagin (Centre for Railway Engineering, Central Queensland University, Australia), Qing Wu (Centre for Railway Engineering, Central Queensland University, Australia), Roger Buckley (Aurizon, Brisbane, QLD 4006, Australia), Sanjar Ahmad (Centre for Railway Engineering, Central Queensland University, Australia)

11:00 AM → 12:30 PM Rail 1: Wheel-Rail Interaction I

Convener: Gang Shen (Tongji University, China)

11:00 AM Wheel‒rail impact loads induced by wheel flats – Simulation model using data from 3D laser-scanned wheel tread defects

Paper ID: 341

Wheel–rail impact loads induced by discrete wheel tread irregularities, such as wheel flats, can cause severe damage to both vehicles and tracks. In Swe-den, regulations mandate the removal or reprofiling of wheels if the wheel flat exceeds 60 mm in length, independent of axle load and train speed. Wheel impact load detectors measure the peak loads generated by out-of-round wheels. If the measured loads surpass the alarm threshold of 350 kN, a train speed reduction is enforced until the vehicle with the damaged wheel is decoupled from the train. However, the impact load generated by a specific wheel flat is influenced by several factors, including vehicle speed, axle load, the lateral position of the rolling circle relative to the wheel tread damage, and the dynamics of the coupled vehicle–track system. Using three-dimensional (3D) laser-scanning of wheel tread geometry as input, this study compares the peak loads generated by two wheel flats of different lengths and depths through simulations of vertical dynamic (high-frequency) vehi-cle–track interaction in the time domain. The findings demonstrate that the trajectory of the wheel–rail contact, influenced by the ratio of wheel flat depth to length, is a more significant parameter than the length of the wheel flat.

Speakers: Jens Nielsen (Department of Mechanics and Maritime Sciences/CHARMEC, Chalmers University of Technology, Sweden), Klara Mattsson (Trafikverket, Sweden), Michele Maglio (Trafikverket, Sweden), Tore Vernersson (Department of Mechanics and Maritime Sciences/CHARMEC, Chalmers University of Technology, Sweden)

11:30 AM Robustness Issues of Wheel-Rail Contact Models in Railway Vehicle Dynamics Simulations

Paper ID: 89

In railway vehicle dynamic simulations, robustness issues in wheel–rail contact algorithms may lead to artificial wheel–rail impacts, thereby affecting the accuracy of simulation re-sults. In this work, the phenomenon of numerical robustness issues caused by the wheel-rail contact algorithm in vehicle dynamic simulations is investigated. First, the manifesta-tions of robustness issues in wheel–rail contact algorithms and their impacts on vehicle dynamic simulation results and wheel–rail operate safety assessments are presented. Then, three common factors that contribute to the lack of robustness in wheel–rail contact mod-els are discussed. Finally, corresponding mitigation strategies are proposed to address these issues.

Speaker: Yu Sun (College of Transportation Engineering, Nanjing Tech University, China)

12:00 PM A Simple Onboard Methodology to Measure the Wheel Rail Contact Forces

Paper ID: 225

This paper proposes a simplified methodology for measuring wheel-rail contact forces in railway vehicles using a reduced number of strain gauges, in order to lower costs and enable a widespread use of load measuring wheelsets. A new signal processing algorithm that uses Fourier series of the strain signal to estimate the contact forces is presented. The method is implemented in a test-rig environ-ment and is validated experimentally under a variety of combinations of vertical and lateral forces and contact positions. The results demonstrate a high level of accuracy, even when just a single gauge is used. This cost-effective solution has potential applications in vehicle monitoring, maintenance optimization, and dy-namic performance analysis, making it a promising tool for the railway industry.

Speakers: Asier Alonso (CAF I+D, JM Lardizabal, Spain), Luis Baeza (Universitat Polit`ecnica de Val`encia, I2MB, Spain Rail 1 Room 114, S), Xabier Perez (CAF I+D, JM Lardizabal, Spain)

11:00 AM → 12:30 PM Road 1: Learning-based-Control

Convener: Lu Xiong (Tongji U, China)

11:00 AM Adaptive Data-Driven Predictive Control for Path Following in Autonomous Driving

Paper ID: 30

This paper introduces a data-driven model predictive control~(DMPC) algorithm tailored for path tracking in autonomous vehicles. The method enables direct controller design using measured data, omitting the need for explicit model information. Leveraging the conventional unit step response modeling approach, the model coefficients undergo dynamic updates with online data, capturing real-time system dynamics. Comparative analysis against the actual vehicle demonstrates the superior modeling accuracy achievable through the proposed online update mechanism. Furthermore, trajectory tracking experiments on the real vehicle affirm the enhanced control performance of the proposed DMPC over traditional MPC approaches.

Speakers: Haoqi Hu (School of Automotive Studies, Tongji University, China), Hong Chen (College of Electronic and Information Engineering, Tongji University, China), Lin Zhang (College of Electronic and Information Engineering, Tongji University, China), Qiang Meng (School of Automotive Studies, Tongji University, China)

11:30 AM Learning-Based Self-Tuning of LTV-MPC Drift Control: Bridging Simulation and Real Vehicle Implementation

Paper ID: 50

Drift control has become an interesting and significant topic in the fields of active safety and autonomous driving. High nonlinearities and system dynamic uncertainties make drift control particularly challenging, while traditional model-based methods usually face difficulties in adapting to dynamic environmental changes. To address this issue, we propose an algorithm that utilizes Reinforcement Learning (RL) to adaptively adjust the parameters of a Linear Time-Varying Model Predictive Controller (LTV-MPC) for drifting. First, the algorithm is validated on a high-fidelity simulation platform, demonstrating that in constant-radius drift scenarios with sudden road friction changes, stable drift tracking control can still be achieved. Furthermore, real-world tests are conducted on a full-size B-class rear-wheel-drive electric vehicle. After training, the proposed algorithm can reduce tracking error by over 50\%. The proposed self-tuning framework can adapt to dynamic environment changes and optimize the control performance of extreme drifting tasks.

Speakers: Daofei Li (Zhejiang University, China), Jinyuan Wei (Zhejiang University, China)

12:00 PM Powerslide Assistance Systems: A Comparison of Analytical and Data-Driven Control Approaches

Paper ID: 347

"This study compares two control approaches for powerslide stabilisation: one based on PD-control and the other on Reinforcement Learning.
Both control schemes are designed as driver assistance systems for all-wheel drive battery electric vehicles with individually driven front and rear axles, operating in closed-loop with a human driver.
The human driver provides steering inputs to follow a path, while the assistance system controls front and rear axle torques to track a predefined vehicle sideslip angle trajectory.
The approaches are analysed in terms of performance and robustness, including the simulation-to-reality transfer."

Speakers: Florian Jaumann (TU Wien, Institute of Mechanics and Mechatronics, Austria), Johannes Edelmann (TU Wien, Institute of Mechanics and Mechatronics, Austria), Manfred Ploechl (TU Wien, Institute of Mechanics and Mechatronics, Austria), Manuel Eberhart (TU Wien, Institute of Mechanics and Mechatronics, Austria), Michael Unterreiner (CARIAD SE, Vehicle Energy, Motion & Body, Germany), Tobias Schuster (TU Wien, Institute of Mechanics and Mechatronics, Austria)

11:00 AM → 12:30 PM Road 1: Motion Planning

Convener: Matteo Massaro (Università di Padova)

11:00 AM Green Light Optimal Speed Advisory for Public Transport: an MPC-based Strategy

Paper ID : 242

This paper presents a Model Predictive Control (MPC)-based Green Light Optimal Speed Advisory (GLOSA) system tailored for urban public transport, specifically buses. Unlike conventional GLOSA systems developed for private vehicles, the proposed approach addresses bus-specific challenges such as limited acceleration, frequent stops, and variable dwell times. The control algorithm integrates real-world traffic light plans and statistical dwell time data to compute optimal speed trajectories that reduce stops at red lights and improve energy efficiency. The Optimal Control Problem (OCP) minimises longitudinal acceleration and jerk while maximising the travelled distance, subject to constraints reflecting motor characteristics, traffic light phases, and bus stop locations. The system is implemented using an effective point-mass vehicle model in MATLAB-Simulink with the ACADO toolkit. Real-world data from Milan’s public transport system is used to model the scenarios as well as the OCP parameters, demonstrating its potential to reduce energy consumption while maintaining service reliability.

Speakers: Daniele Vignarca (Department of Mechanical Engineering, Politecnico di Milano, Italy), Davide Tavernini (Centre for Aerodynamics, Aerospace and Automotive Engineering, University of Surrey, U.K.), Edoardo Sabbioni (Department of Mechanical Engineering, Politecnico di Milano, Italy), Stefano Arrigoni (Department of Mechanical Engineering, Politecnico di Milano, Italy), Umberto Montanaro (Centre for Aerodynamics, Aerospace and Automotive Engineering, University of Surrey, U.K.)

11:30 AM CAPF: A Convolution-Based Artificial Potential Field for Autonomous Driving

Paper ID : 263

We propose a convolution-based artificial potential field method for autonomous driving, which integrates the vehicle contour and heading angle into the environment. This allows the vehicle to be equivalently treated as a point, making it more compatible with the field-based method. The gravitational potential field is formulated using the Dijkstra distance rather than the Euclidean distance, effectively mitigating the risk of local optima. By leveraging the concept of "image convolution" from artificial intelligence, we use vehicles with varying heading angles as kernels and perform multiple convolutions with the environment to generate feature maps. These feature maps are then used to construct repulsive potential fields tailored to different vehicle heading angles. Additionally, we incorporate a torque repulsive potential field that is proportional to the yaw rate to ensure trajectory smoothness. Ablation and comparison experiments validate that the proposed motion planning method yields safer and smoother trajectories compared to traditional approaches.

Speakers: Chaojie Zhang (Department of Control Science and Engineering, Tongji University, China), Jun Wang (Department of Control Science and Engineering, Tongji University, China), Xichao Wu (Department of Control Science and Engineering, Tongji University, China)

12:00 PM Transferable Human-Like Trajectory Planning Method with Reconfigurable Artificial Potential Fields

Paper ID : 337

With the scarcity of naturalistic driving datasets support, experience transfer becomes a viable strategy in addition to supplementing a large amount of transportation data for autonomous driving. To achieve a lightweight transfer method of driving experience, the human-like trajectory planning method based on reconfigurable artificial potential fields is proposed. Firstly, the data-driven trajectory planning methods’ fitting degradation issues in the scenarios lacking data support is discussed and analyzed, explaining the necessity for introducing experience transfer. Secondly, based on the idea of deconstructing the impact of traffic elements on driving awareness, the reconfigurable Artificial Potential Field framework is proposed to achieve experience transfer. Finally, the performance of experience transfer was tested and verified. Simulation results demonstrate that the proposed method significantly reduces human-like trajectory planning similarity errors in unseen roundabout scenarios when compared to multilayer perceptron planners. The proposed model has been trained in straight parallel lane scenarios, can be transferred and applied to previously unseen scenarios by fully leveraging the scenario reconfigurable characteristics. The integration of limited natural driving data with the interpretable structure of Artificial Potential Fields facilitates lightweight knowledge transfer, offering a novel approach to improving generalization in human-like trajectory planning under data-scarce conditions.

Speakers: Haifeng Du (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Jiabao Tan (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Konghui Guo (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Xinjie Zhang (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Xu Nan (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China)

11:00 AM → 12:30 PM Road 1: Steer-by-Wire I

Convener: Lars Drugge (KTH, Sweden)

11:00 AM Exploring Learning Rate of Remote Drivers: A Simulator-Based Study with Realistic Feedback and Delays

Paper ID : 84

Remote driving is emerging as an important backup for automated vehicles, yet the adaptability process and learning rate of remote drivers remain unexplored. In this study, we present a simulator-based experiment designed to evaluate how drivers adapt to remote driving environments characterized by realistic delays and feedback. The experiment employs a high-fidelity driving simulator that replicates real-world remote driving conditions—including motion cueing, steering force, and auditory feedback—while introducing controlled driving feedback delays. Participants navigate a dynamic scenario incorporating changing curvature roads, slalom manoeuvres, lane changes, and parking tasks over 10 training rounds. Both objective performance metrics (e.g., time consumption, lane following deviation, velocity, lateral acceleration) and subjective assessments (e.g., trust, controllability, familiarity, workload) are collected and analysed using repeated measures ANOVA. Results indicate that drivers rapidly adapt to the remote driving environment, achieving stable familiarity and reduced mental workload within the first 4-5 rounds. These findings provide valuable insights for designing effective training protocols and improving remote control tower systems.

Speakers: Jonas Mårtensson (Integrated Transport Research Lab, KTH Royal Institute of Technology, Sweden), Lin Zhao (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Malte Rothhämel (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Mikael Nybacka (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden)

11:30 AM A Generic Steering Model for Driver-in-the-Loop Simulator Applications

Paper ID : 90

Speakers: Matthias Becker (MdynamiX AG, Hessstr. 89, Germany), Peter Pfeffer (Munich University of Applied Sciences, Germany)

12:00 PM Mitigation of input time-delay effects in by-wire steering structure through observation-mapped state prediction

Paper ID : 210

This paper presents a state prediction method tailored for steer-by-wire systems, aiming to mitigate the destabilizing effects of unavoidable input time delays caused by the by-wire signaling process. The method employs a high-gain observer to estimate higher-order angular states of the steering mechanism. Based on these estimates, a state prediction function is formulated to compensate for the delay. The approach is integrated with an LQR controller that explicitly incorporates actuator lag dynamics. The effectiveness of the proposed method is validated through high-fidelity co-simulations, demonstrating its applicability to similar by-wire systems, such as teleoperated vehicles.

Speakers: Hangyu Lu (Shanghai Jiao Tong University, China), Shuhan Liu (State Key Laboratory of Mechanical System and Vibrations, China), Xiaodong Wu (Shanghai Jiao Tong University, China)

12:30 PM → 1:30 PM Lunch

1:30 PM → 3:00 PM Rail 2: Pantograph-Catenary Systems I

Convener: Liang Ling (Southwest Jiaotong U, China)

1:30 PM Development of a Validated Neutral Section Model for Pantograph-Catenary Interaction Dynamics

Paper ID: 349

Neutral sections are key components in railway overhead electrification systems, enabling safe transitions between differing electrical supplies. This study focuses on short neutral sections (NS), specifically the NSR25 design from Arthur Flury. A detailed finite element model of the neutral section is constructed here using a combination of CAD data, literature review, and industry technical information. This includes mechanical components such as the isolator rods, spring-droppers and lever arms. Three scenarios are ana-lysed using the state-of-the-art PantoCat software, namely a baseline model without NS, a model with NS (WNS), and a NS with defective spring-droppers (DNS). Dynamic analyses at 160 km/h reveal that the inclusion of NS in good conditions marginally increase peak contact force and standard deviation. When defective NS configurations are used, substantial peak forces and contact losses are observed. The elasticity of the catenary system and maximum contact wire uplift at the supports are also assessed, showing reduced mechanical compliance in the WNS, which are especially evident in the defective NS. These findings provide insight into how the nominal NS design has a low impact on pantograph-catenary interaction performance. However, the results also highlight the importance of proper maintenance procedures for these discrete features in order to maintain optimal contact quality and minimize damaging the train current collection assets.

Speakers: Jose Rebelo (Institute of Railway Research, University of Huddersfield, UK), Jose Santos (IDMEC, Instituto Superior T´ecnico, Universidade de Lisboa, Portugal), João Pombo (Institute of Railway Research, University of Huddersfield, UK), Pedro Antunes (Institute of Railway Research, University of Huddersfield, UK)

2:00 PM A New Method for Predicting the Fatigue Life of Catenary Dropper Using Image Recognition Technology

Paper ID: 19

Based on a ground-based fixed-point monitoring system for catenary droppers, the dynamic responses of the droppers were gathered during the pantograph's traversal of the catenary system. Utilizing image recognition technology, the precise positions of both ends of the droppers were identified and converted into spatial coordinates through edge detection algorithms. Based on these coordinate data, the compression curves of the droppers were further calculated. By applying the correlation between compression and stress in the droppers, these curves were transformed into corresponding stress values, serving as crucial input parameters for forecasting the fatigue life of the droppers.

Speakers: Hongbo Kou (Infrastructure Inspection Research Institute, China Academy of Railway Sciences Co., Ltd., China), Jing Lu (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Jing Wang (Infrastructure Inspection Research Institute, China Academy of Railway Sciences Co., Ltd., China), Longfei Kong (Locomotive & Car Research Institute, China Academy of Railway Sciences Co., Ltd., China), Yongming Yao (Infrastructure Inspection Research Institute, China Academy of Railway Sciences Co., Ltd., China), Zhipeng Yang (Infrastructure Inspection Research Institute, China Academy of Railway Sciences Co., Ltd., China)

2:30 PM Simplifying railway catenary structures by cancelling catenary stagger

Paper ID: 257

"The railway catenary system is a network of overhead wires used to supply electric power to trains via pantographs. To maintain good contact between pantograph and catenary, the catenary has a complex structure consisting of many wires and acces-sories, as shown in Fig.1. The catenary is laterally offset from the track centre line by steady arms at each pole, which is known as stagger. The stagger is used to distribute wear evenly across the pantograph contact strips. However, implementing stagger requires precise calculations to design the amount and pattern of the stagger, com-plex structures at each registration arm, specialized equipment during installation and maintenance, and regular inspections to keep the stagger within an allowable range. In long-term operation, the complexity of the structures and the large number of components reduce the catenary reliability and increase the time and cost of catena-ry repair.
To make railway transport more affordable and reliable, it would be beneficial to simplify the catenary structure and minimize the number of catenary components. The study proposes a concept to simplify the catenary structure by cancelling the catenary stagger. This allows for more structural tolerance to catenary alignment in design and construction, and helps to simplify the structure of the registration arms and steady arms (even cancelling the structure of steady arms). From the pantograph side, it can help to reduce the width of the pantograph head to lower the weight and reduce air resistance. However, it is necessary to make sure that the cancellation of the catenary stagger neither jeopardizes the pantograph-catenary dynamics nor leads to excessive wear concentration on the pantograph contact strips.
Numerical studies on the pantograph-catenary dynamic performance and wear pattern of contact strips without stagger are performed. A 3D pantograph-catenary dynamic model [1] is used to study the pantograph-catenary dynamics. For the zero-stagger catenary, the catenary is aligned in the centre line of a straight track while all other parameters are kept unchanged. An existing wear prediction model of the con-tact strips [2] is used to investigate the wear pattern. The pantograph is assumed to run at a constant speed. To avoid wear concentration for the catenary without stag-ger, a sinusoidal lateral motion of the pantograph base is added. In reality, a panto-graph lateral movement in relation to the carbody exists. The lateral movement can be caused by sway motion due to soft lateral stiffness of pantograph and bogie hunt-ing excitation, or imposed on purpose by actuators. (For some EMU tilting trains there is an active tilting system incorporated into the pantograph to counter the carbody’s tilt).
Based on the study, the pantograph-catenary dynamics and wear pattern of pan-tograph contact strips with and without catenary stagger are compared. Fig. 2(a) compares the pantograph-catenary contact forces with and without stagger (for zero-stagger catenary, the amplitude of the pantograph sinusoidal lateral movement is 0.1 m). The results show that cancelling the stagger does not deteriorate the pantograph-catenary dynamics. Fig. 2(b) shows the wear rate of the pantograph contact strips under the catenary with zero stagger. The wear rate of the contact strip is regular in the middle part, but peaks appear on the two sides of the contact strip. These peaks can be reduced when the lateral movement is carefully designed and controlled. To avoid frequent replacement of the contact strips, the thickness of the strips can be increased when the pantograph width is significantly reduced without increasing the amount of material needed.
Even though there are some difficulties left for future development, e.g. wind sta-bility, catenary design in curves, detailed pantograph design and control, etc., the study shows a possibility to significantly change the railway catenary structure by cancelling the catenary stagger. In this way, the catenary systems can be significantly simplified in design, construction and maintenance, which has the potential to make railway electrification cheaper and more reliable than today."

Speakers: Bastian Schick (KTH Royal Institute of Technology, Sweden), Sebastian Stichel (KTH Royal Institute of Technology, Sweden), William Zhendong Liu (KTH Royal Institute of Technology, Sweden), Yan Xu (Kunming University of Science and Technology, China)

1:30 PM → 3:00 PM Rail 2: Vehicle Dynamics II

Convener: Luis Baeza (Universitat Politècnica de València, Spain)

1:30 PM Study on Dynamic Behaviors of High-Speed Train Vehicles with a Cracked Wheelset Axle

Paper ID: 320

Due to the complex operating environment and loading conditions, railway axles are inevitably subjected to fatigue and impacts from foreign objects, which may initiate cracks. Consequently, understanding the dynamic behaviour of railway vehicles in the presence of a cracked axle is of significant interest. This paper develops a coupled vehicle–track dynamic model that incorporates a cracked axle, aiming to investigate the effects of axle cracks on the dynamic performance of high-speed railway vehicles. A validated vehicle–track coupled dynamic model is formulated, accounting for the flexibility of both the track system and the wheelset. A novel methodology is proposed to model the crack breathing effect within the axle, enabling simulation of the dynamic behaviour at the crack interface and its impact on overall vehicle dynamics. Furthermore, the developed model is used to study the influence of a cracked axle under the condition of wheel polygonal wear. Results indicate that the presence of a crack leads to variations in axle bending stiffness, causing fluctuations in the wheel–rail contact force, particularly at the wheel rotation frequency. This can amplify the peak impact forces associated with wheel polygonization. Additionally, the cracked axle introduces multiple frequency components, which may serve as key indicators for the early detection of axle cracks.

Speakers: Bo Peng (School of Mechanical Engineering, Southwest Jiaotong University, China), Maoru Chi (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Ningrui Yang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), XINGWEN WU (School of Mechanical Engineering, Southwest Jiaotong University, China), Zheng Guan (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

2:00 PM Carbody Sway Behaviour of Train negotiating Tunnels due to Vortex Induced Vibrations

Paper ID: 355

This work focuses on the vortex induced vibrations (VIV) of passenger cars passing through tunnels in mountainous areas. In response to the complex tun-nel environment and multi-source excitation conditions, the train-tunnel sys-tem dynamic model is established using fluid-structure coupled method. The nonlinear lateral dynamic behavior of VIV is well reproduced. Parametric studies are conducted to propose potential solutions to relieve the vortex in-duced vibrations. It is concluded that the frequency of VIV is around 1.5Hz which is close to carbody rigid mode. In the scenario of low damping ratio of system, the resonance between VIV and carbody rigid modes could happen. To install the in-train dampers or decrease the secondary stiffness can relieve the vortex induced vibrations.

Speakers: Chen Yang (CRRC Puzhen Vehicle Co.Ltd., China), Jing Zeng (Southwest Jiaotong University, China), Lai Wei (Southwest Jiaotong University, China), Xiaoping Jia (Southwest Jiaotong University, China)

2:30 PM Study on low-frequency hunting of high-speed vehicle at medium speed range

Paper ID: 312

"Typically, a rail vehicle will exhibit hunting vibrations when the operational speed goes up to a speed above which the vehicle system becomes unstable, and this speed is known as the critical speed. However, in real service, high-speed vehicles can experience low-frequency hunting motion in an intermediate range of speeds, rather than at the maximum operational speed. This significantly dete-riorates the ride comfort and increases the safety risk and maintenance cost for both vehicle and infrastructure due to increased wear. This type of low-frequency hunting vibration at medium speed range remains under-theorized and demands effective mitigating solutions.
Experimental tests were conducted to identify this special low-frequency hunting motion and a vehicle dynamics model was developed to replicate the vehicle instability at medium speed range through both linear and nonlinear analyzing approaches, which facilitates a deep understanding for this low-frequency hunt-ing vibration, and provides solutions to mitigate this phenomenon."

Speakers: Bin Fu (State Key Laboratory of Mechanical Transmission for Advanced Equipment, Chongqing University, China), Binbin Liu (Dipartimento di Meccanica, Politecnico di Milano, Italy), Gang Chen (CRRC Changchun Railway Vehicle Co. LTD., China), Stefano Bruni (Politecnico di Milano, Italy), Xin Ding (CRRC Changchun Railway Vehicle Co. LTD., China)

1:30 PM → 3:00 PM Rail 2: Wheel-Rail Interaction II

Convener: Edwin Vollebregt (Vtech CMCC, NL)

1:30 PM When Are Non-Elliptic and Conformal Wheel-Rail Contact Models Required for Multibody Simulations?

Paper ID: 277

The non-elliptic index and conformal index are proposed as quantitative measures to assess the influence of non-Hertzian wheel-rail contact on vehicle dynamics within multibody simulation frameworks. The assessment is conducted using both online and offline approaches, employing Simpack and CONTACT, and involves comparisons for a single wheelset-track system and a complete vehicle-track model under various running conditions. Numerical simulation results demonstrate that both non-ellipticity and con-formality of the contact patch contribute to contact force deviations, and their effects are coupled. Deviations observed in the online mode are smaller than those in the offline mode, primarily due to differences in the equivalent penetrations applied in Simpack and CONTACT. The proposed indices pro-vide a statistical basis for evaluating the significance of non-Hertzian con-tact, offering a practical means to determine whether such effects can be neglected in dynamic simulations.

Speakers: Bin Fu (State Key Laboratory of Mechanical Transmission for Advanced Equipment, Chongqing University, China), Binbin Liu (Dipartimento di Meccanica, Politecnico di Milano, Italy), Edwin Vollebregt (Vtech CMCC, Rotterdam, The Netherlands), Qinghua Guan (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Stefano Bruni (Politecnico di Milano, Italy)

2:00 PM Efficient Modelling of Train-Track Interaction Using Floating Frame of Reference Formulation

Paper ID: 354

Extending the idea of the floating frame of reference formulation (FFRF), the Moving Mode Method (MMM) is introduced to model track flexibility utilizing multiple mode shapes that move along the track with the vehicle. It offers improved computational efficiency due to the few elastic coordinates used to describe vehicle-track dynamic interaction and the shorter length of the precalculated mode shapes over the lengthy rail employed in the simula-tion. In this work, a three-dimensional (3D) track model based on the MMM is developed and the wheel-rail interaction on a specific type of short-wavelength singular defect on the rail surface is employed for comparative analysis against beam model.

Speakers: Aki Mikkola (LUT University, Finland), Chen Shen (Delft University of Technology, The Netherlands), Jose Escalona (University of Seville, Spain), Xinxin Yu (Tampere University, Finland)

2:30 PM Simulation of the Wheel–Rail Contact Temperature Induced by Braking

Paper ID: 213

High friction is essential for transmitting braking forces between the wheel and rail. However, adverse conditions will reduce adhesion, leading to wheel slip or lock-up and severe damage like flats. This paper investigates this phe-nomenon using a hybrid simulation approach that integrates a wagon-track interaction model with a 3-D finite element (FE) thermal analysis. This model, incorporating temperature-dependent material properties, predicts that wheel lock-up causes a rapid surface temperature accumulation to near-ly 1000 °C. This extreme temperature exceeds the martensite-to-austenite transformation value and could induce surface damage like wheel flats.

Speakers: Claudio Colao (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany), Hecht Markus (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany), Qiuyong Tian (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany), Yifan Yang (Institute of Land and Sea Transport Systems, Technische Universit¨at Berlin, Germany)

1:30 PM → 3:00 PM Road 2: Diagnostic, Fault-management and Maintance

Convener: Johannes Edelmann (TUWien, Austria)

1:30 PM Lifetime Performance Maintenance for X-by-wire Chassis: An Integrated Control and Diagnosis Approach

Paper ID: 152

"In autonomous vehicles with X-by-wire chassis, actuator degradation reduces control accuracy and safety. Traditional estimation methods require strong excitation, yet normal driving conditions often provide weak excitation amid noise, hindering real-time fault diagnosis. To address this issue, this paper proposes an Integrated Control and Diagnosis Approach (ICDA) that enables full-lifecycle actuator state monitoring while ensuring control performance. The proposed ICDA consists of a dual model predictive controller and a least square based estimator to integrate actuator status monitoring with active safety
control. The effectiveness of the proposed ICDA framework is verified by taking active rear wheel steering as an example, and the experimental results show that the approach has significant advantages in improving the monitoring accuracy and real-time performance under weak excitation conditions, as well as under noisy conditions, which provides an important support for the intelligence and safety of vehicles."

Speakers: Chen Tang (Institute of Intelligent Vehicles, Tongji University, China), Dong Wang (Institute of Intelligent Vehicles, Tongji University, China), Haoran Lv (Institute of Intelligent Vehicles, Tongji University, China), Weidong Liu (Institute of Intelligent Vehicles, Tongji University, China), Xiong Lu (Institute of Intelligent Vehicles, Tongji University, China)

2:00 PM Fault-Tolerant Predictive Control for Distributed Drive Vehicles Based on the Koopman Operator

Paper ID: 180

This paper addresses the issue of yaw stability control for distributed drive vehicles in the event of drive wheel failure by proposing a mechanism-data hybrid fault-tolerant method. The core concept is to utilize a model identification and control allocation strategy driven by a mechanism-data hybrid approach, thereby eliminating the need for explicit Fault Detection and Isolation (FDI) modules and achieving recognition and tolerance of various faults. Furthermore, in response to the impact of internal uncertainties and external disturbances on the accuracy of fault detection, a high-precision chassis data model is constructed using the Koopman operator, enabling real-time accurate representation of unmodeled errors and disturbances.This approach reduces dependency on precise explicit modeling of vehicle dynamics, implementing fault-tolerant control for drive wheel failures through a data-driven strategy. Based on this, the proposed method has been validated via simulation using the co-simulation platform of Matlab/Simulink and CarSim. Simulation results demonstrate that the proposed Koopman Model Predictive Control (MPC) fault-tolerant algorithm effectively ensures lateral stability and driving safety for distributed drive vehicles under various fault conditions. The integration of the Koopman operator into the fault-tolerant control framework represents an innovative step towards advanced automotive fault management, offering a robust solution that enhances vehicle performance and safety in complex driving environments without requiring detailed knowledge of specific fault characteristics.

Speakers: Chen Tang (Tongji University, China), Haoran Lv (Tongji University, China), Lu Xiong (Tongji University, China), Wuji Dai (Tongji University, China)

2:30 PM Towards a data-driven approach to objectify subjective assessments by quantifying vehicle handling

Paper ID: 216

Subjective assessments are key to evaluating vehicle handling, yet how they arise from it is still not fully understood. To address this, the paper presents a data-driven method to quantify vehicle hand-ling by learning a compact numerical representation (embedding) from time series data. Unlike previous approaches that focus on selected objective measures, the proposed method aims to capture the full vehicle handling. A set of requirements is defined to ensure that the embeddings are valid and meaningful, and metrics based on established objective measures are introduced. Following an analysis of driver-vehicle interaction for feature selection, a neural network combining 1D convolution and attention is trained. The data is taken from simulated arbitrary circuits rather than predefined test maneuvers. The model estimates understeer and roll gradients with moderate accuracy. The results suggest that it is feasible to learn embeddings that reflect relevant aspects of vehicle handling and provide a foundation for further analysis or comparison.

Speakers: Kai Storms (Technical University of Darmstadt, Institute of Automotive Engineering, Germany), Steven Peters (Technical University of Darmstadt, Institute of Automotive Engineering, Germany), Tobias Hoyer (Technical University of Darmstadt, Institute of Automotive Engineering, Germany)

1:30 PM → 3:00 PM Road 2: Motion Comfort I

Convener: Derong Yang (Volvo Cars, Sweden)

1:30 PM Modeling Head-Neck Dynamics under Lateral Perturbations Using MPC to Mimic CNS postural stabilization strategy

Paper ID: 285

Automated vehicles will allow occupants to engage in non-driving tasks, but limited visual cues will make them vulnerable to unexpected movements. These unpredictable perturbations create a “surprise factor,” forcing the central nervous system to rely on compensatory postural adjustments, which are less effective, and are more likely to trigger sensory conflicts. Since the head is a key reference for sensory input (vestibular and vision), models accurately capturing head-neck postural stabilization are essential for assessing AV comfort. This study extends an existing model predictive control-based framework to simulate head-neck postural control under lateral perturbations. Experimental validation against human data demonstrates that the model can accurately reproduce dynamic responses during lateral trunk perturbations. The results show that muscle effort combined with partial somatosensory feedback provides the best overall dynamic fit without requiring corrective relative and global head orientation integrators for posture.

Speakers: Chrysovalanto Messiou (Delft University of Technology, The Netherlands), Georgios Papaioannou (Delft University of Technology, The Netherlands), Riender Happee (Delft University of Technology, The Netherlands)

2:00 PM EKASTOS: Individualized Human Body Model for Seated Occupants in Automated Vehicles

Paper ID: 286

Whole-body vibration significantly influences ride comfort, particularly in automated vehicles, where limited visual cues and unpredictable movements introduce a “surprise factor” that challenges postural stabilization. Traditional seat-to-head transmissibility analyses often rely on simplified, linear assumptions that may not fully capture human dynamics under multi-axis perturbations. This study presents Ekastos, a computationally efficient, full-body model with individualized anthropometry and active postural control, developed in the Simscape MATLAB environment to address these limitations. Ekastos builds upon a previously published model, the efficient human model (EHM), and extends it by individualizing it through 31 anthropometric parameters, and adding articulated upper limbs. Validation against experimental data shows that Ekastos without arms generally provides a better fit than the original EHM model across most frequency response functions. While the inclusion of upper limbs did not consistently improve the fit, configurations with fully actuated arms showed some improvement, particularly in head roll and yaw responses during lateral perturbations. These findings suggest that upper limbs and modeled mobility may contribute to whole body vibration predictions, but a comprehensive assessment of their impact on Ekastos’ dynamic response is limited by the constraints of the current optimization method.

Speakers: Chrysovalanto Messiou (Department of Cognitive Robotics, Delft University of Technology), Georgios Papaioannou (Department of Cognitive Robotics, Delft University of Technology), Riender Happee (Department of Cognitive Robotics, Delft University of Technology)

2:30 PM Seat-to-head-transmissibility of seated occupants while being driven and engaged in non-driving related tasks

Paper ID: 311

Automated driving is a major technological advancement with benefits like improved safety, reduced environmental impact, and increased accessibility. However, motion comfort remains a critical challenge that could hinder AV adoption. Consumers value AVs for enabling engagement in non-driving tasks (NDRTs) and productive use of travel time. Yet, design aspects like handing over control, backward seating, limited road view, and NDRTs can increase discomfort and motion sickness (MS). These factors also affect postural stability, as unpredictable maneuvers induce excessive head and body motion. Notably, no studies have examined seat-to-head vibration transmission during NDRTs, highlighting the need for further research. This paper will present results from a human participants experiment for evaluating seat-to-head transmissibility while engaged in NDRTs, and while being driven.

Speakers: Chrysovalanto Messiou (Delft University of Technology, Netherlands), Farjam Tajdari (Delft University of Technology, Netherlands), Georgios Papaioannou (Delft University of Technology, Netherlands), Riender Happee (Delft University of Technology, Netherlands)

1:30 PM → 3:00 PM Road 2: Steer-by-Wire II

Convener: Makoto Yamakado (Kanagawa, Japan)

1:30 PM Steering Feel on Driving Simulators through the Implementation of Hardware-in-the-Loop and Model-in-the-loop Methods

Paper ID: 49

"Steering systems are often only tested in the real vehicle but progressively the hardware-in-the-loop (HiL) testing with virtual environments establishes an objective evaluation process. On HiL test benches, subjective driver assessments are limited. This research topic is using a new approach to compare advanced steering feel in the dynamic driving simulator. The aim is to develop the steering feel on the simulator,
by either using the feedback from the HiL test bench or directly from the force feedback motor as a model-in-the-loop (MiL) system. The HiL principle is applied on a test bench to further assess system conditions, while having the driver evaluating the steering feeling. To realize this, the steering column connection is virtualized by transferring the rotation angle and torque between the motors on the steering gear and the
feedback unit. This allows a seamless transition between HiL and MiL and can be used to compare steering models to the real hardware. This approach has to proof that both hardware and software components can be validated in a controlled environment."

Speakers: Johannes Hendewerk (MdynamiX AG, Hessstr. 89, Germany), Kilian Joerg (MdynamiX AG, Hessstr. 89, Germany), Korbinian Thaler (MdynamiX AG, Hessstr. 89, Germany), Peter Pfeffer (Munich University of Applied Science, Competence Center Vehicle Dynamics, Germany)

2:00 PM Feedback Torque Modelling for Steer-by-Wire Systems Using Hierarchical Local Model Trees to Emulate EPS Behaviour

Paper ID: 287

To replicate the realistic steering feel of electric power steering (EPS) systems in Steer-by-Wire (SbW), this paper proposes a steering torque modeling ap-proach based on the Hierarchical Local Model Tree (HILOMOT). Steering data were collected from open-loop tests conducted on a hardware-in-the-loop EPS test bench. A set of vehicle motion parameters, including vehicle speed, steering angle, angular velocity, yaw rate and lateral acceleration, was selected as model inputs through feature importance analysis. The HILOMOT model and a benchmark backpropagation (BP) neural network were trained and evaluated. Results show that the HILOMOT model achieves lower prediction errors and higher correlation coefficients compared to the BP network. The model was further validated using Double Lane Change (DLC) and Slalom maneuvers, yielding root mean square errors of 0.22 Nm and 0.19 Nm, respectively, with 95.6% and 100% of the predicted torque values falling within the commonly reported upper perceptual threshold of 0.58 Nm. These findings indicate that HILOMOT is a promising approach for modeling nonlinear steering behavior and replicating steering torque in SbW systems.

Speakers: Qiao Zhang (Chair of Automotive Engineering, Technische Universit¨at Berlin, Germany), Steffen Mueller (Chair of Automotive Engineering, Technische Universit¨at Berlin, Germany)

2:30 PM Serial Gaussian Process Internal Models to Account for Steering Torque Feedback in a Learning Driver Model

Paper ID: 314

An approach is presented which uses an architecture of serial Gaussian Processes (GPs) to represent the human learning of vehicle dynamics (including steering system dynamics) in a driver model. Using serial GPs in this way allows the dimensionality of the input space to be reduced, which improves the data efficiency and learning rate of the GPs, ultimately leading to a more computationally efficient driver model. The driver model is implemented in a model predictive control (MPC) framework, which uses the GPs to predict future vehicle states based on the driver's control input of handwheel torque. The variance of the GPs' predictions can be used in the MPC cost function to influence the degree of adventure or caution in the driving behaviour. The performance of the proposed architecture is compared to a single GP internal model, demonstrating that the serial GP architecture is able to achieve similar path following performance with significantly reduced computational cost.

Speakers: David Cole (Department of Engineering, University of Cambridge, UK), Harry Fieldhouse (Department of Engineering, University of Cambridge, UK)

3:00 PM → 3:30 PM Coffee Break

3:30 PM → 5:00 PM Rail 3: Optimization, Maintenance and Monitoring I

Convener: Binbin Liu (Politecnico di Milano, IT)

3:30 PM Torque-Modulation-Based Method for Rapid Identification of Rail Sections with Low Friction Levels

Paper ID: 272

Accurate measurement of the coefficient of friction (CoF) between wheel and rail is desirable for railway operations, as it determines maximum traction and braking forces, directly influencing operational timetables and track utilisation. Conventional measurement approaches—including laboratory twin-disc machines and field-deployed hand-operated tribometers—exhibit significant scaling limitations that compromise correlation with actual in-service friction conditions experienced by operational trains. While torque-modulation-based approaches (i.e. identifying CoF based on the phase difference between the wheel torque and angular velocity) for direct and train-borne CoF measurement demonstrate potential, existing methods require multiple torque modulation steps and thus extensive measurement distances, and near-saturation torques with high friction forces. This study presents an innovative CoF classification method utilising single-step torque modulation operating far below saturation conditions. The methodology induces controllable oscillation of wheel rolling velocity by applying traction and braking torques with slightly different amplitudes to the front and rear wheelsets. This effectively improves the detectability of the phase difference between the applied wheel torque and wheel angular velocity at low torque levels across varying CoF conditions. The proposed method is tested using co-simulations with a multibody dynamics model in VI-Rail and a torque control system in MATLAB/Simulink. The results demonstrate the effectiveness and efficiency of the method for real-time friction classification.

Speakers: Gokul J. Krishnan (Delft University of Technology, The Netherlands), Rolf Dollevoet (Delft University of Technology, The Netherlands), Zhen Yang (Delft University of Technology, The Netherlands), Zili Li (Delft University of Technology, The Netherlands)

4:00 PM Energy-Saving Optimization Strategy for Train Speed Trajectory Based on Improved Non-dominated Genetic Algorithm

Paper ID: 288

With the rapid development of railway transportation, train energy-efficient optimization has become an important challenge for the railway industry. This paper addresses the energy-efficient optimization problem of urban rail transit trains and proposes a multi-objective speed trajectory optimization method based on an im-proved NSGA-II algorithm. A dynamic speed adjustment method based on train operating positions is proposed to generate maximum speed operating curves, and an adaptive sampling interval encoding strategy, progressive comfort constraint handling mechanism, and time error-based adaptive genetic operator control method are designed to achieve multi-objective coordinated optimization of energy consumption, punctuality, and comfort. Case study results show that the improved algorithm has significant improvements in convergence performance and algorithm stability compared to traditional methods, achieving a 4% reduction in energy consumption per unit distance and a 100% comfort compliance rate, validating the effectiveness and engineering practicality of the algorithm.

Speakers: HongZhe Li (College of Electronic and Information Engineering, Tongji University, China), Jinsong Kang (College of Transportation, Tongji University, China), Xinyue Li (College of Transportation, Tongji University, China)

4:30 PM Railway vehicle comfort optimization using stochastic comfort metrics

Paper ID: 244

"Secondary suspensions for passenger railway vehicles are designed, almost unavoidably, with air springs connected to reservoirs via flow restrictors. The diameter of the flow restrictor has a major influence on suspension behaviour and, consequently, it comes as no surprise that diameter optimisation has received considerable attention to optimise comfort.
The authors have argued that comfort indexes should be regarded as stochastic variables and, therefore, with random indexes, suspension optimisation may target differ-ent probability function statistics, such as mean, standard deviation, a selected percentile, skewness…, or any combination thereof.
The computational cost when using the simplified dynamic model alongside the compound comfort inference recipe is not unaffordable, making it possible to evaluate several statistics (objective functions) for a wide, yet feasible, range of restrictor diameters. Results will show the compound comfort mean, standard deviation, 90th percentile, and skewness, as a function of front and rear suspension restrictor diameters, for several rail quality levels and train speeds. These functions are compared to that representing the deterministic Mean Comfort index from standard EN–12299. It will be shown that optimum diameters are significantly influenced by the type of comfort metric used."

Speakers: Antonio Javier Nieto (Department of Mechanical Engineering, University of Castilla La Mancha, Spain), Carmen Ramiro (Department of Mechanical Engineering, University of Castilla La Mancha, Spain), Daniel Rodríguez (Department of Mechanical Engineering, University of Castilla La Mancha, Spain), Eduardo Palomares (Department of Mechanical Engineering, University of Castilla La Mancha, Spain), Publio Pintado (Department of Mechanical Engineering, University of Castilla La Mancha, Spain), Ángel Luis Morales (Department of Mechanical Engineering, University of Castilla La Mancha, Spain)

3:30 PM → 5:00 PM Rail 3: Vehicle Dynamics III

Convener: Jens Nielsen (Chalmers University of Technology, Sweden)

3:30 PM Data-Driven Differentiable Adaptive Ensemble Kalman Filter for Enhanced Vehicle State Estimation

Paper ID: 305

Straddle-type monorail vehicles offer significant advantages in urban rail transit, such as high maneuverability and strong adaptability to complex terrains. However, these vehicles are subjected to severe impact loads during operation, particularly at beam connection points, which can result in high stress concentrations and lead to fatigue damage in critical components. This study proposes a Data-Driven Differentiable Adaptive Ensemble Kalman Fil-ter (DDAEKF) for real-time state estimation of monorail bogies. By integrat-ing recursive Bayesian filtering with deep learning-based state transition and observation models, the proposed method enhances the accuracy of state es-timation under dynamic operating conditions and lays the foundation for subsequent fault diagnosis. The model is validated through numerical simu-lations and a 1:8 scale test rig, achieving stable and robust tracking of vehi-cle states over a 2000-second evaluation period. The results demonstrate that DDAEKF effectively estimates displacement and velocity while mitigating uncertainties in orientation estimation. This method provides a reliable state monitoring approach to improve the safety and operational efficiency of straddle-type monorail systems.

Speakers: Jiarui Hu (College of Transportation, Tongji University, China), Kaiying Zhang (College of Transportation, Tongji University, China), Lihui Ren (College of Transportation, Tongji University, China), Maozhenning Yang (College of Transportation, Tongji University, China), Xiaoguang Ma (College of Transportation, Tongji University, China), Yuanjin Ji (College of Transportation, Tongji University, China)

4:00 PM Developing an AI-infused framework for the automatic generation of vehicle system dynamics models

Paper ID: 293

Vehicle system dynamics modelling and simulations are widely used throughout a fleet lifecycle to support design, operation, and maintenance. These models and simulations are typically implemented in specialized software and require extensive do-main knowledge to be developed and used, hence their usability is often reduced to design/approval stages and remains limited in day-to-day operations within railway companies. Making these models more accessible could enable widespread use of physics-based digital twins to improve fleet operations. Recent advances in deep learning show promising applications for automatically generating technical content, particularly in engineering contexts. For instance, SIMULINK models have been successfully generated, debugged, and tested using transfer learning on large-scale language models such as GPT-2 [1]. Similarly, transformer-based models have demonstrated success in generating optimal configurations of mechanical components and interfaces for given design problems [2].

This paper proposes a framework for developing vehicle system dynamics models using artificial intelligence (AI) generative tools, with GENSYS Multibody Dynamics software [3] as a case study. The framework aims to create a prompting interface for interacting with physical system models and including their elements, enabling do-main engineers to reconfigure models and interpret results without requiring specialized programming skills. This capability is particularly valuable for Digital Twin applications, where rapid model adaptation and results interpretation are essential.

2 Methodology
In this paper, transfer learning is used to adapt a pre-trained large language model to the domain of vehicle dynamics modelling, inside the context of a specialized soft-ware tool. The framework encompasses three primary modules: vehicle configuration and design parameters, operational scenario definition, and analysis specification. The approach involves: (1) collecting and pre-processing a dataset of existing GENSYS simulation input files, focusing on these three modules to create comprehensive training datasets, (2) simplifying these models to capture essential dynamics while removing implementation-specific details, (3) fine-tuning a pre-trained language model to understand the relationship between natural language prompts and vehicle model structures, and (4) developing a validation process to ensure generated models are both syntactically correct and physically meaningful. Post-processing scripts then transform the model outputs back into GENSYS-compatible formats.

3 Expected Results
The framework will be demonstrated through an exemplary Vehicle Dynamics Generative Model (VDGM) implementation, enabling a new paradigm for interacting with vehicle multibody dynamics models and simulations. Users will be able to prompt the system to generate vehicle models, define operational condition scenarios, and specify desired analysis outputs. The paper will present the framework's effectiveness in generating valid and useful vehicle dynamics models, evaluating both the technical accuracy of the generated models and their practical utility in real-world applications.

This work contributes to bridging the gap between complex vehicle dynamics multi-body modelling tools and their practical application in industrial settings, potentially improving the utilization of simulation throughout the vehicle lifecycle. The frame-work provides a foundation for developing more accessible and user-friendly inter-faces for vehicle dynamics modelling, supporting the broader adoption of digital twin technologies in railway operations.
References
[1] S. L. Shrestha and C. Csallner, “SLGPT: Using transfer learning to directly generate simulink model files and find bugs in the simulink toolchain,” ACM International Conference Proceeding Series, pp. 260–265, 2021, doi: 10.1145/3463274.3463806.
[2] Y. Etesam, H. Cheong, M. Ataei, and P. K. Jayaraman, “Deep Generative Model for Mechanical System Configuration Design,” 2024, [Online]. Avail-able: http://arxiv.org/abs/2409.06016
[3] DEsolver, “GENSYS.” [Online]. Available: https://www.gensys.se/"

Speakers: Colin Cole (Centre for Railway Engineering, Central Queensland University, Australia), Esteban Bernal (Centre for Railway Engineering, Central Queensland University, Australia), Maksym Spiryagin (Centre for Railway Engineering, Central Queensland University, Australia), Qing Wu (Centre for Railway Engineering, Central Queensland University, Australia)

4:30 PM Rail Vehicle Dynamics and Track Irregularity Monitoring using Artificial Intelligence

Paper ID: 17

"Due to increasing rail traffic and infrastructure utilisation, continuous monitoring of vehicle and track condition by vehicle sensors is becoming increasingly important. However, most existing research attempts to solve a specific problem, e.g. estimating track irregularities from axle-box accelerations. In this work, we show that multi-body simulation can be used to generate a vehicle-specific dataset that allows continuous and simultaneous estimation of dynamic quantities such as accel-
erations, forces and track irregularities using a neural network with, for example, only axle-box accelerations as input. Our results demonstrate that the neural network can accurately predict the above quantities in one model with high accuracy."

Speakers: Christian Schindler (Institute for Rail Vehicles, RWTH Aachen University, Germany), Michael Ebbers (Institute for Rail Vehicles, RWTH Aachen University, Germany), Philipp Leibner (Institute for Rail Vehicles, RWTH Aachen University, Germany)

3:30 PM → 5:00 PM Rail 3: Wheel-Rail Interaction III

Convener: Xinxin Yu (Tampere U, Finland)

3:30 PM Effect of Unsteady Wheel-Rail Contact on the Prediction of Rail Corrugation and Curve Squeal

Paper ID: 218

This study investigates the influence of unsteady tangential contact modelling on the prediction of rail corrugation and curve squeal. A frequency-domain unsteady contact formulation is integrated into a linearised wheel–rail interaction model and applied to a tight-radius metro curve scenario. While steady contact models are sufficient to identify corrugation-prone wavelengths, unsteady contact models improve the prediction accuracy of the wear rate. In the case of curve squeal, the unsteady approach captures additional high-frequency instabilities that are not detected using steady contact models. The findings highlight the critical role of unsteady contact effects in accurately capturing dynamic wheel–rail phenomena, particularly at higher frequencies.

Speakers: Egidio Di Gialleonardo (Politecnico di Milano, Department of Mechanical Engineering, Italy), Federico Castellini (Politecnico di Milano, Department of Mechanical Engineering, Italy), Juan Giner-Navarro (I2MB, Universitat Polit`ecnica de Val`encia, Spain), Leonardo Faccini (Politecnico di Milano, Department of Mechanical Engineering, Italy), Luis Baeza (I2MB, Universitat Polit`ecnica de Val`encia, Spain), Roberto Corradi (Politecnico di Milano, Department of Mechanical Engineering, Italy)

4:00 PM Experimental study on modal coupling characteristics and mechanism of heavy-haul train-track dynamic interaction

Paper ID: 295

The modal features of the heavy-haul train-track coupled system are the important property that determines its vibration characteristics. To ensure the safety and reliability, a good modal coupling relationship between the key components of the heavy-haul train-track coupled system is essential. This paper presents an experimental study on the modal coupling characteristics and mechanism of the heavy-haul train-track dynamic interaction. A series of field tests are carried out to obtain the modal parameters and vibration characteristics of the typical heavy-haul train and track systems under different operating conditions. For the heavy-haul train, the main components of the HXD1 heavy-haul locomotive and the C80 freight wagon are tested, including the car-body, bogie frame, etc. For the track system, the rail and sleeper of different track structures are tested, including the straight track and curved track. Through the analysis of extensive field test data, the modal coupling characteristics and mechanism of the heavy-haul train-track dynamic interaction is systematically studied. The results show that the modal coupling vibration of the components can be excited under certain operating conditions and the vibration with a special frequency can be amplified. The relevant conclusions provide data support for the research on modal decoupling and forward design of the heavy-haul train-track coupled system.

Speakers: Chunlei Zhao (CRRC Qiqihar Rolling Stock Co., Ltd, China), Jianhua Wang (Guoneng Shuohuang Railway Development Co., Ltd., China), Liang Ling (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Peibin Jiang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Wanming Zhai (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Xuancheng Yuan (CRRC Zhuzhou Locomotive Co., Ltd, China)

4:30 PM Modelling of Railway Wheelsets in Turnouts Incorporating Real Contact Geometry

Paper ID: 219

"This study summarizes a comprehensive method for solving the geometric wheel–rail contact problem, applicable to both mainline tracks and turnouts. The proposed approach ensures accurate identification of all potential contact regions, along with precise computation of the contact plane and distances between undeformed wheel and rail surfaces. Unlike conventional methods that rely on solving nonlinear equations and may miss critical contacts, this method uses a closed-form analytical solution to fully detect intersections—modeling the wheel as a set of frusta of cones, and the rail as linear segments. This formulation supports parallel processing on CPUs or GPUs.
To address elastic contact, the method employs Variational Theory to determine traction distributions. The contact plane is derived through Principal Component Analysis, and the rail running surface is modeled using mapped sequences of measured profiles from real-world turnouts.
Static simulations of a wheelset over a crossing panel reveal key behaviors such as sudden contact shifts, variations in overall contact stiffness, and differences in wheel radii. These lead to increased creepage and localized wear—factors essential for understanding vehicle dynamics through turnouts and improving railway infrastructure design and maintenance."

Speakers: Asier Alonso, Björn Pålsson, Juan Giner-Navarro, Luis Baeza

3:30 PM → 5:00 PM Road 3: Powertrain and Driveline

Convener: Edoardo Sabbioni (PoliMi, Italy)

3:30 PM Self-excited torsional vibrations caused by clutch differentials in high-performance car drivelines

Paper ID: 111

Torque splitting strategies adopting clutch units are increasingly used in modern powertrains to improve vehicle dynamics, enabling the transition between two-wheel drive and four-wheel drive. A significant drawback, however, is represented by the possible occurrence of self-excited vibrations. Therefore, a nonlinear multibody model of a full driveline has been developed to analyze these issues, validating the results with experimental data. The non-stationary behaviour of the driveline is found to be strongly dependent on the nonlinear friction characteristics. Stability maps let identify negative friction gradients of the clutch torque splitter as responsible of unstable oscillations, while modal analysis revealed a dominant localization of some modes within the whole transmission.

Speakers: Alberto Bodini (Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, Italy), Alessandro De Felice (Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, Italy), Daniele Gualdi (Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, Italy), Guido Ricardo Guercioni (Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, Italy), Silvio Sorrentino (Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, Italy)

4:00 PM Deceleration Control of an Electro-Hydraulic Brake System Using Dual Fuzzy Logic Controllers

Paper ID: 126

This study presents a deceleration control strategy for electro-hydraulic brake (EHB) systems using dual fuzzy logic controllers (DFLC). One controller manages vehicle deceleration, while the other regulates wheel slip ratio. By synthesizing the outputs of both controllers, an optimal pressure change rate is determined and used to adjust the duty cycles of pulse-width modulation (PWM) signals. Unlike conventional high-frequency PWM, low-frequency PWM is employed to operate on-off solenoid valves, producing a stepwise pressure response that emulates the behavior of linear valves. The proposed algorithm dynamically detects insufficient road surface friction and adjusts reference targets for deceleration and slip ratio to prevent wheel lockup. Simulation results, based on actuator dynamics identified from a production EHB module, demonstrate that the system accurately tracks commanded deceleration under adequate friction conditions. Under low-friction scenarios, the controller limits deceleration to the maximum allowable by the road surface while effectively preventing wheel lockup.

Speakers: Bo-Chiuan Chen (National Taipei University of Technology, Taiwan), Hsin-Han Chiang (National Taipei University of Technology, Taiwan), Zhen-Yu Zhou (National Taipei University of Technology, Taiwan)

4:30 PM Dynamic Features of Electrified Heavy-duty Trucks Considering the Drive System

Paper ID: 280

The electrification of heavy-duty trucks is a highly effective approach to promoting the energy transition in the transportation sector, significantly reducing carbon emissions. However, the dynamic behavior of these trucks, especially their interaction with the pantograph-catenary system, is not yet fully understood. For the reliable operation of electrified heavy-duty trucks, ensuring stable current collection from the pantograph-catenary system is crucial. Thus, this paper develops a dynamic model of electrified heavy-duty trucks considering a dual-parallel pantograph-catenary system, focusing on the effect of the drive system on dynamic features. The results show that the magnitude of the torque of the drive system not only affects the inconsistency of the pantograph-catenary contact forces on the left and right sides, but also alters the mean value of the contact forces. This indicates that it is vital to study the influence of different excitations on the pantograph-catenary contact characteristics of heavy-duty truck vehicle systems.

Speakers: Gino D'Ovidio (University of L’Aquila, Italy), William Zhendong Liu (KTH Royal Institute of Technology, Swenden), Xian Chen (Kunming University of Science and Technology, China), Yan Xu (Kunming University of Science and Technology, China), Ziwei Zhou (Kunming University of Science and Technology, China)

3:30 PM → 5:00 PM Road 3: Suspension I

Convener: Xinjie Zhang (Jilin U, China)

3:30 PM A novel semi-active suspension control strategy: Sky-hinge control

Paper ID: 194

The theme of this paper is the design and analysis of a control strategy for semi-active suspensions in road vehicles. Existing closed-loop control strategies, such as Sky-hook control, typically require two sensors per wheel-side suspension, usually comprising either two accelerometers or a combination of one accelerometer and one stroke sensor. This paper introduces a novel yet straightforward algorithm that utilizes a single Inertial Measurement Unit (IMU) sensor for the entire vehicle to enhance suspension performance. The starting point of this work is the Sky-hook control, leading to the development of the Sky-hinge algorithm. The advantages of the Sky-hinge algorithm are discussed. The effectiveness of the proposed algorithm is validated through simulations using a full-vehicle model. Compared to passive suspensions and conventional semi-active algorithms, the proposed algorithm not only reduces costs but also enhances comfort and handling stability. Therefore, this algorithm has great application potential in practical applications.

Speakers: Mingxing Liu (Hefei University of Technology, China), Minyi Zheng (Hefei University of Technology, China), Nong Zhang (Tongji University, China), Weimin Zhong (Hefei University of Technology, China), Zhengfeng Yan (Hefei University of Technology, China)

4:00 PM Virtual Reference Feedforward Active Suspension Control with Half-Car Model for Ride Comfort

Paper ID: 222

This paper presents virtual reference feedforward control (VRFC) with half-car model and active suspension for ride comfort. In previous paper, VRFC was designed with a quarter-car model. For the reason, it cannot control the pitch motion of a sprung mass. In this paper, VRFC is designed with half-car model to control the vertical and pitch motions of a sprung mass. Feedback controller is designed with LQ static output feedback (SOF) control with half-car model. With the LQ SOF controller, VRFC is optimized with simula-tion-based optimization. VRFCs designed with half-car model are validated through simulation. From simulation results, it is shown that VRFC is effec-tive in controlling the vertical and pitch motions of a sprung mass.

Speakers: Jinwoo Kim (Seoul National University of Science and Technology, Republic of Korea), Seongjin Yim (Seoul National University of Science and Technology, Republic of Korea), TAESEOK JIN (Seoul National University of Science and Technology, Republic of Korea)

4:30 PM Research on Semi-Active Vibration Control Performance of Electromagnetic Suspension Based on Intelligent Tire Load Informatio

Paper ID: 262

With the development of intelligent tire technology, the ability to sense real-time load distribution offers new opportunities for optimizing suspension system control performance. This study explores a semi-active vibration control strategy for electromagnetic suspension based on intelligent tire load information. By using tire load data as input, the suspension system dynamically adjusts the damping force to optimize vehicle ride comfort and handling stability.

Speakers: Lin He (Hefei University of Technology, China), Pengfei Liu (Anhui Agricultural University, China), Xiaolong Zhang (Anhui Agricultural University, China), Yu Tang (Hefei University of Technology, China)

5:00 PM → 6:00 PM Shuttle bus to Welcome Reception Venue hotel

Shuttle buses leave from the Pick-up Point from 17:00 to 18:00; return buses from 20:00 to 21:00.
Please bring your welcome reception ticket.

6:30 PM → 8:30 PM Welcome Reception

The Welcome Reception will be held on the evening of the first day of the
symposium. All registered participants are invited to join us for an informal
gathering, providing an excellent opportunity to meet fellow attendees, network,
and enjoy light refreshments. The reception will take place at the main foyer of
the conference venue.

Note: There will be shuttle-bus service from the conference venue to the
banquet venue, please assemble at the pick-up point at 17:30.

Tue, August 19

8:30 AM → 9:30 AM State of the Art II

Conveners: Manfred PLOECHL, Oldrich Polach

8:30 AM Dynamics and Control of Articulated Passenger Vehicles on Roads

Speakers: Chaojie Zhang (Tongji University, China), Jun Wang (Tongji University, China), Qingwei Liu (University of Lincoln, UK), Tim Gordon (University of Lincoln, UK)

9:30 AM → 10:30 AM Plenary Presentation

Conveners: Manfred PLOECHL, Oldrich Polach

9:30 AM Scalable High-Speed Lateral Control for Single-Body and Articulated Autonomous Vehicles

Paper ID: 260

This paper presents a scalable lateral control framework designed to enable robust path tracking for both single-body and articulated autonomous vehicles under high-speed conditions. Traditional lateral controllers often struggle with oscillations, instability, and reduced accuracy when applied to high-speed maneuvers or complex vehicle configurations such as tractor-trailers. To address these challenges, a clothoid-based lateral controller is extended with several high-speed adaptations: 1) an integrated tangential check and Fréchet distance method for optimized lookahead and oscillation mitigation; 2) real-time trajectory segment classification for dynamic lookahead search space adjustment; and 3) a dual-adaptive, rate-controlled lookahead mechanism responsive to both cross-track error and proactive segment analysis. These enhancements enable dynamic tuning of the controller’s lookahead distance to improve stability and accuracy during transitions between straight and curved paths. To ensure applicability across different vehicle types, the framework incorporates a flexible tracking point selection scheme and a curvature-to-steering lookup table (LUT). This allows accurate tracking of key vehicle points, such as the tractor rear axle or the hitch point, without structural changes to the control logic. The proposed controller is validated in simulation using high-fidelity TruckSim® models across varied driving scenarios, including dual lane changes and winding road sections. Results demonstrate improved stability, reduced oscillations, enhanced tracking accuracy, effective lookahead adaptation, and maintained lateral acceleration within safe limits across a wide speed spectrum and vehicle configurations. This study highlights the potential of a unified lateral control architecture for diverse autonomous vehicle platforms and lays the groundwork for further development toward multi-trailer systems.

Speakers: Aashish Shaju (Virginia Tech, USA), Mehdi Ahmadian (Virginia Tech, USA), Steve Southward (Virginia Tech, USA)

10:00 AM Quantum computing in vehicle dynamics and an example of its application to locomotive traction control

Paper ID: 14

Quantum computing is an experimental technology in its early stages of development. It harnesses the power of quantum mechanics to tackle problems that are too time consuming for classical computers. It is projected that the quantum computers will be able to simulate complex physical systems, such as the dynamic behavior of an entire locomotive, with unprecedented comprehensiveness and computation speed. Considering that a locomotive is a complex system, and its traction control system and adhesion control algorithms implemented in a locomotive are key elements that still require investigation for ensuring optimal locomotive performance and safety during train operation, there is some space for further research and development. However, the adhesion control systems still face some problematic challenges in real-time optimization of their operation and prediction mechanisms due to the complexity of dynamic friction interactions between the locomotive wheels, rails, track structure, and environmental factors. The application of quantum computing may solve adhesion detection problem efficiently considering its ability to process vast amounts of data simultaneously to find an optimal dynamic performance characteristics of a locomotive. This paper describes an idea of the development of quantum computation framework for locomotive traction modelling explores the potential benefits of quantum computing in enhancing locomotive traction control.

Speakers: Colin Cole (CQUniversity, Centre for Railway Engineering, Australia), Dan Agustin (CQUniversity, Centre for Railway Engineering, Australia), Esteban Bernal (CQUniversity, Centre for Railway Engineering, Australia), Ingemar Persson (AB DEsolver, Optand 914, Sweden), Maksym Spiryagin (CQUniversity, Centre for Railway Engineering, Australia), Qing Wu (CQUniversity, Centre for Railway Engineering, Australia)

10:30 AM → 11:00 AM Coffee Break

11:00 AM → 12:30 PM Rail 4: Track System II

Convener: Xinwen Yang (Tongji U, China)

11:00 AM Integration of Floating Slab Track with Quasi-Zero Stiffness Mechanism and Dynamic Vibration Absorber for Low-Frequency Vibration Mitigation

Paper ID: 234

Vibration amplification at the natural frequency of steel spring floating slab track (FST) remains a significant concern. This study integrates quasi-zero stiffness mechanisms combined with dynamic vibration absorbers (QZSDVA) into FST, resulting in a novel formulation of vibration reduction track (QZSDVA-based FST) to effectively suppress low-frequency track vibrations. The analytical expression for the dynamic response of the two-degree-of-freedom QZSDVA system is derived and subsequently validated through numerical approaches. Building on this, a vehicle-QZSDVA-based FST dynamic model is established and optimal parameters for DVA are identified guided by H_∞ optimization. Further exploration highlights the capability of QZSDVA-based FST in significantly suppressing track vibrations. Results emphasize that optimally tuned parameters enable QZSDVA-based FST to reduce low-frequency vibrations effectively, even with a minimal DVA mass.

Speakers: Shengyang Zhu (Southwest Jiaotong University, China), Wanming Zhai (Southwest Jiaotong University, China), Yuhao Ren (Southwest Jiaotong University, China)

11:30 AM Suppressing localized-rail bending modal vibration by designing high-frequency vibration-absorbing fasteners

Paper ID: 74

Since the view that the localized rail 3rd-order bending mode (B3 mode) can cause high-order polygonization (mainly 18~23) of high-speed train wheels was put forward in 2017, many scholars have attempted to link the connection between the localized rail bending modes and wheel polygonization phenomenon and polygonal wheel passing frequency (PWP frequency). Since most of the reported studies did not consider the multi-component assembly characteristics of the fastener system, the phenomenon of localized rail bending modes could not be accurately reproduced by many theoretical models, which hindered scholars from exploring the solution to the polygonal wheel problem. To advance the research on this scientific issue, this paper first establishes a flexible track model considering the structural and parametric characteristics of fasteners, verifies the model by using vehicle tracking test data, and then investigates the influence of fastener parameter matching on the localized rail bending modes, and obtains the following conclusions: (1) There is nearly 1:1 mapping relationship between the localized rail bending modal frequency and PWP frequency, which supports that the localized rail bending mode is one of the causes of wheel polygonization; (2) The iron plate of the fastener system plays the role of dynamic vibration absorber (DVA), in the vehicle-rail coupled system, and the quality of the iron plate and the stiffness of the upper/lower rubber pads significantly influence the localized rail bending modal frequency. Finally, this paper proposes the design principle of a high-frequency vibration-absorbing fastener, which provides a feasible solution to mitigate the localized rail bending modal vibration and high-order wheel polygonization. Meanwhile, it points out that this measure may induce other high-frequency vibration problems, e.g., aggravating modal vibration above 800 Hz. Further, this paper proposes the concept of differentiated arrangement of fasteners, suggesting that different high-frequency vibration-absorbing fasteners be installed in different sections of the whole line to make the localized rail bending modal frequency of the whole line disordered, thus disrupting and further mitigating the development of the wheel polygonization.

Speakers: Dadi Li (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Huanyun Dai (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Yunguang Ye (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Zhecheng Tao (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

12:00 PM Prediction of Derailment Coefficient from Onboard and Ground Sensors using LSTM model

Paper ID: 253

This study investigates the use of machine learning techniques for predictive maintenance in railway systems. By employing Long Short-Term Memory (LSTM) neural network models, this research analyses the data collected from onboard sensors on Tokyo Metro trains for the prediction of derailment coefficient, a key indicator of derailment risk. The results demonstrated that LSTM models can effectively predict its trend. This research has the potential for real-time monitoring and early intervention strategies to enhance railway safety.

Speakers: Abhinav Srivastava (Graduate School of Engineering, The University of Tokyo, Japan), Kenji Ejiri (Institute of Industrial Science, The University of Tokyo, Japan), Yoshihiro SUDA (Institute of Industrial Science, The University of Tokyo, Japan)

11:00 AM → 12:30 PM Rail 4: Vehicle Dynamics IV

Convener: Oldrich Polach

11:00 AM An Investigation into Hunting Stability Considering the Creep Nonlinear Saturation Characteristics under Traction/braking conditions

Paper ID: 270

Railway vehicles are in traction and braking conditions for a large amount of time and previous studies have always regarded traction/braking force as an external input force that can hardly influence the stability of system, and ignored the influence of traction/braking force on creep nonlinear saturation characteristics. Therefore, it is necessary to take creep nonlinear saturation characteristics into consideration and study the variation of vehicle’s stability under traction and braking conditions. The equivalent creep coefficients have been proposed to quantify the nonlinearity of creep, together with the simulation method to obtain them. Then a dynamic system has been established and the nonlinear critical velocity is found to increase with the increase of creep saturation degree under traction/braking conditions.

Speakers: Jinsong Zhou (College of Transportation, Tongji University, China), Junhao Zhang (College of Transportation, Tongji University, China), Xiaoyu Li (College of Transportation, Tongji University, China), Zhanfei Zhang (College of Transportation, Tongji University, China)

11:30 AM Hardware-In-the-Loop test rig to assess the influence of secondary dampers on hunting motions of rail vehicles

Paper ID: 158

Nowadays several research efforts have being invested in the prevention of un-stable running condition in railway vehicles due to the continuous arise in the commercial speed. Indeed, the arising of car body hunting and bogie hunting phenomena is also fostered by the wider range of operating conditions that rail-way vehicles must face, both in terms of vehicle speed and wheel-rail profiles coupling. Previous works proposed to mitigate these unstable phenomena either by optimizing traditional passive suspension components or by introducing ad-vanced smart suspensions. Nevertheless, the experimental validation of these suspension components is still costly and time consuming. Therefore, this paper proposes a cost-effective Hardware-In-the-Loop methodology designed to test the influence of secondary suspension components on the dynamics of railway vehicles by replicating various operating scenarios. The proposed methodology is implemented to show its capability of simulating a wider range of operating con-ditions in a fast and economic way, supporting the design and optimization of new secondary suspension components for railway vehicles.

Speakers: Egidio Di Gialleonardo (Department of Mechanical Engineering, Politecnico di Milano, Italy), Francesco Ripamonti (Department of Mechanical Engineering, Politecnico di Milano, Italy), Gioele Isacchi (Department of Mechanical Engineering, Politecnico di Milano, Italy), Stefano Bruni (Politecnico di Milano, Italy)

12:00 PM Research on Diagnosis Method of Hunting Instability of High-speed Trains Based on Transfer Learning

Paper ID: 103

Hunting instability poses a critical safety risk in high-speed train operation, especially in its early stages when low amplitude, complex disturbances, and weak features often lead to oversight, potentially causing bogie instability alarms, carbody swaying, and shaking. This study proposes a method combining system dynamics modeling and domain-adversarial transfer learning to identify small hunting instability. A simulation model incorporating nonlinear wheel–rail interactions and track irregularities was developed on the SIMPACK platform to reproduce the stability-to-instability transition, and integrated with roller rig test data to construct, for the first time, a cross-domain binary classification dataset targeting small hunting instability. Furthermore, four types of features—time domain, frequency domain, wavelet, and HHT—were systematically evaluated, and a multi-channel convolutional fusion architecture was designed to enhance feature representation and transferability. Experimental results demonstrate that the proposed approach achieves an accuracy of 97.55% with combined frequency-domain and HHT features, confirming its practical effectiveness and engineering applicability in diagnosing small hunting instability.

Speakers: Jianfeng Sun (School of Engineering, Zhejiang Normal University, China), Jiayi Liang (School of Engineering, Zhejiang Normal University, China), Maoru Chi (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Siyu Liu (School of Engineering, Zhejiang Normal University, China), Weidong Jiao (School of Engineering, Zhejiang Normal University, China), Yonghua Jiang (School of Engineering, Zhejiang Normal University, China), Zhilin Dong (School of Engineering, Zhejiang Normal University, China)

11:00 AM → 12:30 PM Rail 4: Wheel-Rail Interaction IV

Convener: Zili Li (TU Delft, NE)

11:00 AM Dynamic vehicle-track interaction at swing nose crossings – validation of an MBS model and evaluation of instrumentation for condition monitoring

Paper ID: 278

This paper makes preliminary investigations into how well the condition of crossing panels in high-speed turnouts can be monitored using embedded sensors. To this end an advanced track model with flexible rails and sleepers is built in the multibody simulation software Simpack and then calibrated to measurement data from an instrumented high-speed turnout in operation in Spain. The model accounts for the non-linear support of the swing nose due to base plate contacts and locks. The model has been applied to make preliminary investigations into the feasibility of monitoring running surface damage on the swing nose using embedded sensors. The conclusions are that the sensor must be located close to the region of interest in order to detect changes in wheel-rail interaction forces. Further it’s preferable if the sensor is located on the wing rail over the sleeper as soft rail pads will isolate much of the higher frequency excitation. Finally, the non-linear baseplate contacts do introduce high frequency dynamics compared to regular rail fastenings. This makes it more difficult to perform inverse load identification from track excitations as performed previously for fixed crossings.

Speakers: Björn Pålsson (CHARMEC, Chalmers University of Technology, Sweden), Christian Bucher (Voestalpine Railway Systems GmbH, Zeltweg, Austria), Thomas Titze (Voestalpine Railway Systems GmbH, Zeltweg, Austria)

11:30 AM Design, Development, and Instrumentation of the KTH Scaled Roller Rig

Paper ID: 196

The Rail Vehicles research group at KTH is developing an in-house roller rig for educational and research purposes within rail vehicle dynamics. The design, de-velopment, and instrumentation of the KTH scaled roller rig are the focus of this paper. The KTH Roller Rig is designed with three modularity principles, namely dimensional scaling, railway condition, and operation scenarios. The roller rig has three possible operation modes: Tangent track without irregularities, curved track without irregularities, and tangent track with vertical irregularities. A single wheelset running on a single roller set has been commissioned; the upcoming second phase will entail a two-wheelset bogie, with the possibility of adding car-body mass, over two roller sets. A set of measurement instruments have been in-stalled. This set includes load cells for longitudinal and vertical suspension forc-es, LVDTs in longitudinal and lateral directions to measure wheelset movement relative to the frame, two incremental encoders to monitor wheelset and roller ro-tational speed, and a triaxial accelerometer on one of the axle boxes. A dynamic verification process has been carried out.

Speakers: Alireza Qazizadeh (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Alireza Qazizadeh (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Prapanpong Damsongsaeng (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden、), Prasidya Wikaranadhi (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Rocco Libero Giossi (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Rohan Kulkarni (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden)

12:00 PM Design and Calibration of a Twin-Disc Test Rig for Wear and Thermal Tests

Paper ID: 328

This paper presents the design and experimental calibration of a novel scaled twin-disc test bench developed at Politecnico di Torino for investigating wear and thermal phenomena in wheel-rail and tread braking operations. A key innovation of the test bench is the adoption of scaling rules to effectively correlate experimental data to the full-scale systems. A comprehensive calibration campaign was conducted to characterize the main bench subsystems. These include the wheel-rail contact load application system, the proportional valve controlling the pneumatic cylinder, and the pneumatic cylinder pushing brake shoes against the wheel. The calibration was successful, and linear characteristics were obtained for the main bench components. Preliminary tests run to estimate the wheel-shoe friction coefficient exhibited high repeatability, showing consistent friction values regardless of the rotational direction or side of brake application. Therefore, the robust design of the test bench makes it suitable for extensive experimental campaigns that will be performed in future activities.

Speakers: Antonio Gugliotta (Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino 10129, Italy), Matteo Magelli (Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino 10129, Italy), Nicol`o Zampieri (Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino 10129, Italy), Nicola Bosso (Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino 10129, Italy), Rosario Pagano (Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino 10129, Italy), Rosario Pagano (Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino 10129, Italy)

11:00 AM → 12:30 PM Road 4: Driver Assistance

Convener: Patrick Gruber (University of Surrey, UK)

11:00 AM HMI Based on Human Foresight Prediction for Advisory Take-over Timing

Paper ID: 73

When the SAE Level 2 driving automation system reaches the bound-ary of its Operational Design Domain (ODD), the driver must promptly take over the vehicle. Although the system can no longer operate safely, it can still provide advisory information via the Human-Machine-Interface (HMI). Therefore, a key challenge for HMI is to ensure that drivers can quickly and safely return to the driving loop. This study introduces a Pre-Timing parameter to evaluate the ap-propriate timing for issuing advisory assistance in response to various emergency situations, particularly in cut-in scenarios. A driving simulator experiment was conducted with twelve participants, testing five Pre-Timing groups with 1.5s in-tervals across four scenarios with varying gaps and relative velocities. A Gated Recurrent Units (GRU) model was trained with three types of inputs, vehicle dy-namics, driver operation states and gaze behavior, to predict optimal Pre-Timing. The result demonstrated the effects of Pre-Timing on subjective and dynamic metrics and concluded that Pre-Timing between 6 and 7.5 seconds was optimal for most cut-in scenarios. Training results were consistent with the statistical findings, showing that the predicted Pre-Timing was negatively correlated with both scenario urgency and gaze weight. This study provides a reference for future studies on the advisory HMI for take-over, and the prediction model incorporat-ing human foresight inputs.

Speakers: Kimihiko Nakano (Institute of Industrial Science, The University of Tokyo, Japan), Linyan Wang (Institute of Industrial Science, The University of Tokyo, Japan)

11:30 AM Pedestrian Crossing Intention Prediction Using a Multimodal Fusion Network

Paper ID: 124

Pedestrian crossing intention prediction is essential for the deployment of auton-omous vehicles (AVs) in urban environments. Ideal prediction provides AVs with critical environmental cues, thereby reducing the risk of pedestrian-related collisions. However, the prediction task is challenging due to the diverse nature of pedestrian behavior and its dependence on multiple contextual factors. This paper proposes a multimodal fusion network that leverages seven modality fea-tures from both visual and motion branches, aiming to effectively extract and in-tegrate complementary cues across different modalities. Specifically, motion and visual features are extracted from the raw inputs using multiple Transformer-based extraction modules. Depth-guided attention module leverages depth infor-mation to guide attention towards salient regions in another modality through comprehensive spatial feature interactions. To account for the varying importance of different modalities and frames, modality attention and temporal attention are designed to selectively emphasize informative modalities and effectively capture temporal dependencies. Extensive experiments on the JAAD dataset validate the effectiveness of the proposed network, achieving superior performance compared to the baseline methods.

Speakers: Steffen Müller (Chair of Automotive Engineering, Technische Universit¨at Berlin, Gustav-Meyer-Allee 25, Germany), Yuan-Zhe Li (Chair of Automotive Engineering, Technische Universit¨at Berlin, Gustav-Meyer-Allee 25, Germany)

12:00 PM Evaluating Interactions between Automated Vehicles and Cyclists using a coupled In-the-Loop Test Environment

Paper ID: 198

Testing and validation of automated driving systems in interactions with vulnerable road users, such as cyclists, are essential for ensuring safety, but often lack realism. We demonstrate and validate a Cyclist-in-the-Loop (CiL) test bench combined with a Vehicle-in-the-Loop (ViL) test bench via a virtual environment (VE). This setup enables safe, authentic, closed-loop interactions between a real automated vehicle and a human cyclist, with their actions dynamically influencing each other in real time. Validation experiments with an automated shuttle bus and a track-and-follow function are conducted with the combined test benches and on a proving ground. The results are compared by responses to cyclist gestures, trajectory adherence, and latency induced by the test setup. We discuss strengths and weaknesses of the proposed approach and determine the suitability of our setup for analyzing more complex interaction scenarios between automated vehicles and cyclists in the future.

Speakers: Clemens Groß (Department of Automotive Engineering, Technische Universit¨at Berlin, Gustav-Meyer-Allee 25, Germany), Lisa Marie Otto (Department of Automotive Engineering, Technische Universit¨at Berlin, Gustav-Meyer-Allee 25, Germany), Michael Kaiser (Department of Automotive Engineering, Technische Universit¨at Berlin, Gustav-Meyer-Allee 25, Germany), Steffen Müller (Department of Automotive Engineering, Technische Universit¨at Berlin, Gustav-Meyer-Allee 25, Germany)

11:00 AM → 12:30 PM Road 4: Vehicle Design

Convener: Giampiero Mastinu (PoliMi, Italy)

11:00 AM Investigation of a 2-DOF Rotational Energy Harvester with Stiffening and Softening Effects for Automotive Applications

Paper ID: 146

Harvesting energy from automobile tire rotation presents a viable approach for powering tire sensors, removing the dependency on external power. This research introduces a rotational bistable vibration energy harvester with two-degree-of-freedom (2-DOF), markedly expanding its operational bandwidth relative to typical single-degree-of-freedom (1-DOF) designs. The harvester comprises a standard piezoelectric cantilever beam, subject to centrifugal stiffening effects, and an inverted piezoelectric cantilever beam, experiencing centrifugal softening effects. Simulation outcomes indicate that integrating the centrifugal softening beam establishes a secondary operational frequency band. This adaptation enables the harvester to harness vibration energy across two distinct frequency ranges, yielding an operational bandwidth increase exceeding 112.5%. The proposed device functions effectively at rotational frequencies between 0 and 125 rad/s, equivalent to vehicle speeds under 120 km/h, encompassing typical road driving scenarios. The results of this study substantially enhance the performance of rotational bistable energy harvesters and offer critical insights for designing and developing advanced energy harvesting technologies suited to dynamic automotive tire ap-plications.

Speakers: Yunshun Zhang (The University of Tokyo, Japan), Yuyang Qian (Jiangsu University, China)

11:30 AM Research on the Identification Method for Tire Turn-Slip Coupling Conditions

Paper ID; 147

With the increasing application of advanced driver assistance systems (ADAS) in new energy vehicles, the number of driving conditions that must be simulated and verified during the vehicle research and development phase has also risen. As a result, there is a growing demand for tire models that include turn-slip expressions in vehicle dynamics models, particularly for scenarios such as automatic parking and automatic U-turns. Additionally, the shortening of the development cycle for new models and the introduction of virtual research and development technology routes have created new demands for tire suppliers' ability to verify the compatibility between tires and vehicles. The lack of test data for current tire turn-slip conditions, along with the absence of initial reference values for the model parameters expressed by tire turn-slip, presents an urgent challenge that must be addressed to meet these requirements. This paper, set against the background of virtual tire sample delivery, addresses the challenge that mainstream tire test benches lack the ability to test steady-state tire turn-slip conditions. It presents a method to obtain virtual test data for tire external characteristics under turn-slip coupling conditions, using tire finite element model simulations. Additionally, the paper explores a model parameter identification method for the Magic Formula tire model with a turn-slip expression. First, the finite element simulation methods for pure turn-slip, side-slip-turn-slip combined, and camber-side-slip-turn-slip combined conditions are introduced and computed. The virtual test data for the external characteristics of the tire under turn-slip coupling conditions are then obtained. Second, based on the magic formula tire model, and in the absence of initial reference values for the turn-slip model parameters, the particle swarm optimization algorithm is employed to determine and optimize the preliminary range of these parameters. This approach yields the initial optimal values for the turn-slip model parameters within the defined range, thereby enhancing the convergence of the model parameter identification process and improving the accuracy of the model identification. Third, based on the characteristics of the magic formula tire model in expressing turn-slip conditions, a comparison is made between the step-by-step parameter identification and the global parameter identification for both the tire side-slip expression and the turn-slip expression. Finally, based on the differences in friction characteristics between the tire side-slip test condition and the turn-slip condition, a comparison and identification of the turn-slip model parameters are performed using high-speed tire side-slip data and low-speed tire side-slip data. This research presents a method for obtaining data on the virtual external characteristics of tire turn-slip coupling conditions based on the finite element method, as well as a technique for determining the initial value of model parameters when the initial turn-slip model parameters are unavailable. The study demonstrates the effectiveness of PAC tire model identification using tire finite element simulation data. Additionally, it compares and discusses the influence of step-by-step versus global identification on the parameter identification of the tire turn-slip model, and examines how the expression of side-slip friction characteristics affects tire turn-slip identification. The results of this study offer an effective method for obtaining and identifying tire models with turn-slip expressions, benefiting both tire dynamics researchers and tire suppliers.

Speakers: Dang Lu (Jilin University, China), DeKUAN Liu (Jilin University, China), Haitao Min (Jilin University, China), Hengfeng Yin (Jilin University, China), Longfei Zhou (Jilin University, China)

12:00 PM Sensitivity of road vehicles dynamic handling behaviour to structural changes

Paper ID: 361

The Body-in-White (BiW) is the backbone for the structural integrity of road vehi-cles. The use of new materials and structural joints in the BiW requires to under-stand how the BiW construction affects the vehicle dynamic behaviour. To that purpose, suitable computational tools are required. The Finite Element (FE) meth-od allows to develop models to study the static structural deformation and vibra-tion characteristics of components, and to simulate vehicle crash scenarios, at the cost of oversimplifying the tire-road interaction and the suspension elements. Thus, FE is often used to study the BiW alone, for Noise-Vibration-Harshness (NVH), instead of vehicle handling and ride. Flexible Multibody (FMB) simula-tions allow considering the tire-road contact, the suspension systems, and to in-clude the structural flexibility of components, being suitable to study how BiW construction affects the vehicle dynamics. Most of the works on the matter consid-er simple tubular chassis structures, or, when they consider more complex BiW constructions, the FMB models resort to a simplification of the structure. Such works are not unanimous about the relevance of the BiW flexibility into the multi-body simulations of road vehicles or the BiW stiffness in the ride and handling behaviour of the vehicles. This work extends the understanding on the topic by incorporating a high-fidelity BiW model in the FMB model of a luxury sports car, and exploring the impact of BiW design details, such as adhesive joints, in vehicle dynamics.

Speakers: Jorge Ambrosio (University of Lisbon, Portuga), Pedro Millan (University of Lisbon, Portuga)

11:00 AM → 12:30 PM Road 4: Yaw Moment Control

Convener: Daofei Li (Zhejiang U, China)

11:00 AM Research on Control of Pivot Steering for Multi-Axle Distributed Drive Electric Vehicles

Paper ID: 144

Multi-axle distributed drive electric vehicles, as an essential component in civilian transportation sector, offer advantages such as high power, heavy load capacity. As the driving scenarios for multi-axle vehicles become increasingly complex, higher demands are being placed on multi-axle vehicles maneuverability. Pivot steering, achieved by adjusting the wheel speeds on either side, is regarded as a significant method for enhancing vehicle maneuverability. However most pivot steering strategies are primarily applied to passenger vehicles, which are not sufficiently designed to address the specific characteristics of multi-axle vehicles, such as heavy duty and large moment of inertia. Therefore, it is essential to conduct in-depth research on pivot steering control strategies for multi-axle vehicles. In this paper, a dynamic model of a multi-axle distributed drive electric vehicle is established and analyzed. Considering the challenges posed by intense tire dynamics during the pivot steering process, a hierarchical control architecture is pro-posed, consisting of a model-free controller with online self-adaptive capability and a robust anti-slip regulation (ASR) controller. Finally, the effective-ness of the proposed control strategy is validated through co-simulation experiments.

Speakers: Bo Leng (Tongji University, China), Lu Xiong (Tongji University, China), Wei Han (Tongji University, China), Wenhai Piao (Tongji University, China), Zhengkai Zhan (Tongji University, China), ZongYu Lv (Tongji University, China)

11:30 AM On the analysis of vehicle understeer behavior with different torque vectoring strategies

Paper ID: 202

Electric vehicle with multiple motors allow Torque Vectoring, i.e. individual wheel torque control. So, a direct yaw moment may be generated with a purposely-induced left-right torque bias. Having such feature available, the engineer is left to decide on what basis such yaw moment should be defined. This paper provides insights on that, by analysing various scenarios in which the yaw moment is either constant or linearly correlated to other relevant vehicle parameters, alternatively steering angle, yaw rate, or sideslip angle.

Speakers: Basilio Lenzo (Universit`a degli Studi di Padova, Italy), Gabriele Fichera (Universit`a di Catania, Italy), Marco Adami (Universit`a degli Studi di Padova, Italy), Marco Cazzola (Universit`a degli Studi di Padova, Italy), Mariagrazia Tristano (Universit`a degli Studi di Padova, Italy), Mario Costantini (Universit`a degli Studi di Padova, Italy), Matteo Massaro (Universit`a degli Studi di Padova, Italy), Mauro Andriollo (Universit`a degli Studi di Padova, Italy), Ricardo de Castro (University of California at Merced, USA)

12:00 PM Cornering with driving torque: steady state cornering, donuts, drifting

Paper ID: 208

A single-track vehicle model with a variation of the brush tire model is presented that takes into account how the driving torque limits the available lateral friction of the tires. This allows us to investigate the effects of the driving torque on the different steady state maneuvers of the vehicle, which is not captured by simpler vehicle models. In addition, the vehicle model is used to design a linear feedback controller that can both initiate and stabilize drifting.

Speakers: Denes Takacs (Budapest University of Technology and Economics, Hungary), Gabor Orosz (Department of Mechanical Engineering, University of Michigan, USA), Illes Voros (Department of Mechanical Engineering, University of Michigan, USA)

12:30 PM → 1:30 PM Lunch

1:30 PM → 3:00 PM Poster Session (Rail)

Note: The poster session is scheduled for the afternoon of Tuesday, August 19, 2025, at the Sino-French Centre, Tongji University. Delegates can present their posters and engage with other participants. The posters will be open for viewing from Day One to Day Three. Your participation in selecting the Best Poster Award is encouraged.

1:30 PM A Feed-forward Spatial-delay Control Strategy for Active Steering of Trailing Bogies in Trams

Paper ID: 86

This paper presents the design and evaluation of a feed-forward controller for the active steering system of trailing bogies in railway vehicles. The controller is designed to enhance curving performance, reduce wheel and track wear, and mitigate noise pollution. It employs a feed-forward control approach that uses the angular position of the leading bogie and incorporates a distance-delay algorithm to account for the spatial separation between leading and trailing bogies. The controller's performance is validated through co-simulations and field tests.

Speaker: Xiaotao REN

1:30 PM A Hybrid-Driven Bogie Stress Monitoring Method Based on the Integration of TF-CNN-LSTM

Paper ID: 321

The structural health monitoring (SHM) of railway vehicle bogies presents considerable challenges due to the distributed nature of critical stress locations and limitations in sensor deployment, making it difficult to comprehensively capture the stress state at key structural points. To address this issue, this paper proposes a novel stress estimation method based on a Convolutional Neural Network (CNN). The method employs a CNN to directly learn the nonlinear mapping between axle box vibration signals and bogie structural stress. A two-dimensional input tensor is constructed from the vibration acceleration data and fed into the CNN for stress prediction. This data-driven design enhances the network’s ability to capture complex input–output relationships without relying on extensive sensor instrumentation. Using the bogie antenna beam as a case study, the proposed model effectively reconstructs the stress response from axle box vibrations. The estimated stress is further used for fatigue damage evaluation. Experimental results show strong agreement between the predicted and measured stress values, demonstrating the accuracy, generalization capability, and engineering applicability of the proposed CNN-based method for SHM of railway bogie components.

Speakers: Maoru Chi, XINGWEN WU, junzuo liu

1:30 PM A METHOD FOR IDENTIFYING WHEEL-RAIL LATERAL FORCES BASED ON BOGIE VIBRATION ACCELERATION

Paper ID: 59

The interaction between the wheel and rail is a fundamental characteristic of railway transportation, and the forces between them are key parameters in assessing the safety of railway operations. The most direct method for measuring wheel-rail forces is through the use of instrumented wheelsets, while due to its relatively intricate installation and maintenance, requirement for specialized calibration equipment as well, only a few vehicles are equipped with such devices. Therefore, Indirect method based on system parameters or other test data obtained through measurements, employing structural dynamics methods is developed in this study. Considering the derailment factor as a key safety indicator for evaluating vehicle operation, it is essential to separately identify the lateral forces on the left and right wheels, which is the focus of this paper. This represents the core research emphasis of this paper, as existing indirect methodologies to date have only been capable of identifying the composite lateral forces exerted on both wheels collectively. The algorithm developed within this method is employed to predict the wheel/rail forces acting upon a metro system, utilizing measured bogie accelerations. Meanwhile, the lateral wheel-rail forces gathered from an instrumented wheelset operating on the same track under identical conditions are leveraged to validate the efficacy of the proposed algorithm.

Speakers: jiawei huang, shuoqiao zhong, xiaozhen sheng, xin Zhou

1:30 PM A minimal and scalable architecture for Hardware-in-the-Loop simulations of railway signalling systems

Paper ID: 104

Speakers: Andrea Collina, Ivano La Paglia, Lorenzo Bernardini, Marco Barbaro

1:30 PM A neural network approach to the modelization of railway couplers

Paper ID: 217

The progressive lengthening of freight trains, adopted by railway operators to in-crease the quantity of goods transported poses challenges - especially for couplers, which must endure significant forces, thus impacting safety. Existing derailment risk simulations rely on complex coupler models to predict longitudinal forces act-ing on the trainset, but these often fail to capture nonlinear behaviors accurately. In this work, the possibility of using Artificial Neural Networks (ANNs) to model rail-way couplers is explored, with the aim of reducing computational complexity while maintaining accuracy. A Gated Recurrent Unit (GRU) neural architecture is selected as the most suitable choice. The ANN is trained using two sets of data, the first com-ing from the combination of Longitudinal Train Dynamic simulations and a rheolog-ical model currently present in literature, and the second coming from experimental measurements taken on an instrumented freight wagon. After training, the ANN-based coupler demonstrates improved efficiency over traditional models while maintaining high accuracy in reproducing coupler forces.

Speakers: Alessio Cascino, Enrico Meli, Francesco Mazzeo, Laurens Lanzillo, Leandro Nencioni, Stefano Melzi

1:30 PM A new fast and approximate method for solving tangential contact in simulations of rail traction vehicles

Paper ID: 31

The Chopaya (Chollet-Pascal-Ayasse) model is a creep method in which forces are evaluated by a function. The difficulty in this class of methods lies in considering creepage combinations including spin, since analytical solutions exist only in no-spin configurations. The Chopaya method is here extended to address traction/braking applications. The formulations of the original and extended methods are presented. The Chopaya method returns to the Shen-Hendrick-Elkins (SHE) method, with the difference that curve-fitting functions have been introduced to more accurately evaluate the creep forces in the presence of spin, taking FASTSIM as a reference. The extended Chopaya method is inspired by the modified FASTSIM method proposed by Spiryagin et al. Taking CONTACT as a reference, the Chopaya method is compared with the SHE method, Polach’s method and FASTSIM: Chopaya results are closest to those of FASTSIM. The ability of the extended Chopaya method to represent realistic traction curves is demonstrated in a set of locomotive measurements studied by Polach in his publication. It is hoped that this method will represent a new compromise in terms of computational time, accuracy and versatility.

Speakers: Honoré Yin, Michel Sebès

1:30 PM A numerical study on the impact of automatic train operation (ATO) on rail wear

Paper ID: 350

Automatic Train Operation (ATO) systems automate train acceleration, cruising, coasting, and braking to replicate and assist driver operations, improving operational efficiency. While ATO enhances driving performance and energy efficiency, the effects of its traction and braking control strategies on wheel–rail contact behaviour and consequent wear are insufficiently investigated. This study investigates the impact of ATO on rail wear by numerically simulating train-track interactions with and without ATO. A multi-body dynamics (MBD) train–track interaction model was established, which utilized wheel torques and rotational speeds measured from real-life trains operating with and without ATO on the same section of the Dutch railway. Dynamic wheel–rail contact forces and slip ratios were thus obtained and used as inputs of the CONTACT program to calculate the normal contact stresses and microslip distributions. Archard’s wear model was then applied to compute the peak rail wear. The results suggest that the ATO system decreased slip ratios and peak wear along the track section of concern. This work enhances understanding of the influence of ATO on wheel–rail contact behaviour and wear evolution. It promotes the proper implementation of ATO systems in railway networks.

Speakers: Chunyan He, Zhen Yang, Zili Li

1:30 PM A path-following control strategy for statically indeterminate virtual rail trains: a strategy based on minimum total lateral force

Paper ID: 224

This paper introduces the characteristics of a statically indeterminate vehicle and designs a path following control strategy for a certain type of four-module six-axles virtual rail train. The control algorithm is designed based on the planar dynamic model of the vehicle, and the cost function with the minimum total wheel lateral force is introduced to solve the system control quantity. The simulation test of SIMPACK virtual prototype shows that the control method can effectively reduce the max lateral force of circular curve passing by more than 40%. The control method can effectively improve the limit operating condition of the vehicle.

Speakers: LIHUI REN, Yuanjin Ji, huang youpei

1:30 PM A quick analysis approach for multiple pantographs operation

Paper ID: 151

To simulate the load of freight locomotives on the test line, multiple locomotives are typically utilized in coupling and reverse traction to smoothly carry out the traction brake test of freight locomotives. This brings about the problem of multi-ple pantographs operating concurrently. When multiple pantographs are operated, the dynamic performance of the pantograph is often disturbed by the operation of more than one pantograph due to the fluctuating effect of the contact line. When working with two pantographs, it is frequently possible to optimize the dynamics of the trailing pantograph by altering the interval, but the issue becomes more problematic as the number of pantographs grows. In this context, this paper at-tempts to propose a quick analysis approach for the contact state of pantograph and catenary during the operation of multiple pantographs (three pantographs). First, analyze the operation state of each pantograph at different intervals when three pantographs are operating at the same time, and then propose an approach to find a reasonable arrangement of pantographs when multiple pantographs are op-erating using the double pantograph intervals.

Speakers: Hongming Chen, Ning Zhou, WenWei Lu

1:30 PM A Rapid Simulation Method for Dynamic Interaction Analysis of Long-Distance Flexible Overhead Contact System-Pantograph

Paper ID: 141

To address the issue of low computational efficiency in the dynamic simulation of long-distance flexible overhead contact system-pantograph systems, this paper proposes a method called: a fast solution method for the interaction between the flexible overhead contact system and pantograph on long railway lines based on a coupled window approach (hereinafter referred to as: CW-FCP). This method utilizes both online and offline windows to simulate the entire flexible overhead contact system, with special treatment applied to the overlap section. The displacement and dropper force of the online window are replicated to the offline window, and forced vibrations are applied to the offline window using contact forces, completing the preprocessing before the offline window is converted to the online window. The effectiveness of the CW-FCP method was demonstrated through various means, including field test data, international simulation standards, and traditional methods. The results indicate that the CW-FCP method exhibits high computational accuracy and possesses significant advantages in calculating the dynamics of long-distance flexible overhead contact system-pantograph systems.

Speakers: Bingrong Miao, Guiming MEI, Shiyu Zhao, Weihua Zhang, zhong huang

1:30 PM A real-time wheel-rail force identification method using inertia parameters of the axlebox via a CNN-ATT model

Paper ID: 162

Wheel-rail force is one of the most valuable physical parameters in rail transit, and it is very important to evaluate the safety of vehicle operation and the durability of structure service. However, existing studies have shown that the real-time wheel-rail force identification requires multiple-types of data input, which poses a great challenge to its engineering application. To tackle this issue, this paper proposes a 1DCNN and multi head attention (CNN-ATT) model with high-accuracy and real-time performance for vertical and lateral wheel-rail force identification, which only uses the triaxial acceleration signal of the axlebox as input. Subsequently, a multibody dynamics model of` B-type urban railway vehicle is established and simulated under different situations to provide the wheel-rail force and acceleration data. The research results show that high-precision identification of wheel-rail forces, achieving an R-squared (R²) value exceeding 92% on the test set. Experimental results confirm that the proposed model achieves high accuracy and robustness in wheel-rail force identification, offering reliable technical support for real-time railway safety monitoring.

Speakers: Jianwei Yang, Jinhai Wang, Zhiqiang Zhang

1:30 PM An optimization strategy rail grinding for improving the dynamic performance of high-speed trains passing through turnouts

Paper ID: 284

High-speed trains often experience excessive carbody lateral acceleration (CLA) and swaying when traversing worn turnouts. Rail grinding is an effective solution, but current practices lack precise strategies. This study develops a Grinding Strategy Model (GSM) driven by hybrid intelligent algorithms to address this. Worn rail profiles are analyzed, and target profiles are optimized using GA-BP-NSGA III, a vehicle-turnout coupled dynamic model, and NURBS theory. The GSM, integrating NRBO, Transformer, BiGRU, and KELM, achieves 98% training accuracy and 97% testing accuracy. After application, rail profile deviations are controlled within 0.2 mm, resolving CLA issues and improving grinding efficiency and vehicle stability.

Speakers: Chao Chang, Yihang Yang

1:30 PM Analysis of modal coupling effect between hunting motion and carbody flexibility of high-speed train

Paper ID: 123

On the basic of typical 17-DOF lateral simplified model, this study develops a full-DOF linear analysis model considering the carbody flexibility of high-speed train. The aim is to analyze the coupling effect between hunting motion and car-body flexible modes, particularly focusing on carbody diamond mode closely re-lated to the carbody chattering phenomenon. The continuous modal tracking tech-nology is applied to describe and explain this mode-coupling instability phenom-enon. The mode veering phenomenon is observed when the hunting frequency approaches the natural frequency of carbody’s diamond mode, and a strong cou-pling effect between carbody’s diamond mode and hunting mode is clearly identi-fied. Furthermore, a covariance method is used to examine the influence of sus-pension parameters and carbody structure parameters, including modal frequency and modal damping, on vibration characteristics. The results indicate that a small-er frequency difference between the hunting frequency and the modal frequency of the carbody diamond mode, along with a smaller damping ratio of the hunting mode and carbody structure damping ratio, exacerbates the carbody chattering phenomenon.

Speakers: Heng Zhang, Liang Ling, Wanming Zhai

1:30 PM Analysis of Wheel-Rail Contact Stability under Variable Friction Coefficients

Paper ID: 70

Abstract. The stability of wheel-rail contact is critical for safe and efficient railway operations, especially when wheel-rail interactions exhibit unstable tendencies due to decreasing friction coefficients with sliding velocity. This study proposes a novel stability analysis model based on Hertzian contact theory, which integrates steady-state wheel responses on curves with radii ranging from 300 to 600 m. This model uses finite element analysis and frequency domain methods to identify dominant unstable modes, and assess the influence of key factors such as wheelset lateral displacement, contact angles, and friction coefficients on system stability. The results highlight the significant impact of specific friction coefficients and lateral displacements on the onset of self-excited vibrations, providing insights into potential noise reduction and vibration mitigation strategies in rail systems.

Speakers: Shaolong Lv, Shutong Liu, Xinwen Yang, Yingtao Qu

1:30 PM Augmented State Functionals: a proposal to improve the accuracy of Neural Networks by incorporating physical knowledge in the Data Models

Paper ID: 155

Traditionally, most of the models used to predict the behavior of dynamical systems have been based on physical laws. Over the last few years, data-based models have gained importance due to the fact that the amount of data registered by the systems has increased significantly. In the railway industry, the same trend has been followed and, currently, most railway vehicles register a large number of variables of their different systems while operating. However, the success of data-based models in the railway industry has been limited, since the results provided by them are usually not accurate enough, due to the specific characteristics of the field. On the other hand, physical models provide a deep and precise understanding of the systems, but sometimes lack of the required accuracy too. Given these considerations, combining physical models with available data knowledge appears to be a promising approach for developing models with increased accuracy. In this paper, a new type of neural network that allows directly incorporating the physical knowledge based on differential equations is defined. In particular, a new structure, called Augmented State Functional, is proposed. The equations that define this model are given and its training process is explained. In addition, the model is applied to two simple mechanical systems and some considerations about its application to general railway cases are given.

Speakers: Ane Oyarzabal, Asier Alonso, Nere Gil-Negrete

1:30 PM Comparison of train-single/double line track-subgrade spatial coupled dynamics models using Green’s function method

Paper ID: 289

The double-track subgrade and single-track subgrade structures are two typical forms of railway infrastructure. At present, studies on the dynamic responses of subgrade under double-track high-speed railways are quite limited due to the complexity of these railways. In this work, a comparative analysis is conducted, based on the frequency-domain Green’s function method (GFM), on the subgrade dynamic characteristics caused by a high-speed train moving on the double-track and single-track. First, the modeling strategy and numerical solution approach for the train-track-subgrade (TTS) system are introduced. Then, the distribution of typical subgrade dynamic responses along longitudinal, lateral and vertical directions is discussed, and the influence of random track irregularities on subgrade vibration distribution is further investigated.

Speakers: Liming Zhu, Yu Sun, sen zhang

1:30 PM Coupled dynamics characteristics of straddle-monorail train with steel track beam system

Paper ID: 254

Since the continuous steel track beam is applied in the straddle-monorail opera-tion line for the first time, research on the coupled dynamics characteristics of ve-hicle-bridge is essential to ensure the stability and safety of train operation. Con-sidering the motion relationship between the key components of the train and the refined structure of the 52+80+52 m continuous steel beam, a spatial coupled dy-namics model of the four-group straddle monorail train and the track beam bridge system was established. Firstly, the key parameters of the rubber tire were deter-mined by the static and dynamic stiffness and damping tests of the traveling wheels. Under the excitation of the measured track irregularity, the dynamic re-sponse and the first three modes of the bridge are in good agreement with the re-sults of the field test. The numerical results show that the vertical vibration of the continuous steel beam is obvious. The nonlinear stiffness of the traveling wheel and beam joints cause a great impact on the track beam and bogie. The coupled dynamics characteristics under a series of suspension parameters are calculated. When the stiffness of the traveling wheels is reduced from 1.44 MN/m to 0.72 MN/m, the wheel-track force and vertical acceleration of the bridge are reduced by 6.7% and 13.83%, respectively.

Speakers: Jiayuan Song, Kaiyun Wang, Qinglie He, Zhida Yang

1:30 PM Design and application of a new wheel profile for mitigating abnormal carbody swaying of a metro vehicle

Paper ID: 203

Abnormal carbody sway occurs in Chinese metro vehicles equipped with linear induction motors and standard LM wheel profiles when operating on straight and shallow-curved tracks. This significantly degrades passenger ride comfort. This phenomenon arises from the coupling of the excitation vibration of periodic track irregularities with a wavelength of 11–13 m, bogie hunting, and the upper sway mode of the carbody. Mitigating this issue requires suppressing resonance by controlling both the frequency of bogie hunting and the track-induced vibrations. Reducing wheel-rail equivalent conicity below 0.05 effectively lowers the hunting frequency (less than 1 Hz) and prevents it from aligning with the carbody’s rolling eigenfrequency of about 2 Hz. However, conventional low-conicity solutions (e.g., the LMA profile) compromise safety in sharp curves. To resolve this conflict, the novel LMX wheel profile was developed. Field tests demonstrate that the LMX profile simultaneously eliminates the abnormal sway and enhances curving performance. Specifically, the maximum lateral Sperling index of the carbody decreases from 4.10 to 2.36, and the average maximum decreases from 3.55 to 2.47. Additionally, the derailment coefficients for the leading wheelsets on 200-meter-radius curves decline from 0.78 to 0.60. Following its successful implementation on Guangzhou Metro Lines 4 and 5, future work will investigate the long-term evolution of wheel-rail wear

Speaker: Wei Li

1:30 PM Design and Evaluation of Lateral Dynamic Vibration Absorber for Speed Increased Motor Vehicles Oriented to Uncertain Influences of Wheel Spin Dissipation

Paper ID: 62

Speed increased motor vehicles must change to use the independent mode of fore and rear traction motors elastically hung on bogie frames, enhancing the lateral DVA vibration reduction effectiveness with ZF Sachs T70 dampers, and reasonably extending the rail grinding AHCP / C processing periods with the wheel AHC modified design. System safety is a warrant of structural reliability, thereby reasonably reducing the costs of operations and maintenances. In the previous stage of this research, applying the novel industrial digitalization methods based on highly-integrated dynamic systems, the self-adaptive improved MDO achievements were gained to better coordinate the following three conflicting technical goals: 1) Rational wheel-rail matching and carbody instability at lower conicity; 2) Constraint stiffness of wheelset positioning and detrimental wear in wheel-rail contact; 3) Uncertain influences of wheel spin dissipation on system safety and structure reliability. For both full service motor / trailer vehicles, the use of traction motors / inverters is taken as lateral DVA designs respectively, the rigid-flex coupling simulation analyses show that the maximum safe speed Vmax can reach 550 km/h under the wheel-rail matching condition of UK-P8/ CN60KG: λeN=0.155, by which some dynamic technical indicators can be accepted according to the requirement of (1.1 - 1.2) times the design speed. Considering the particularities of Chinese newly-built HSR practices, e.g. vibration alarming, fluttering / chattering phenomena, etc., the original design manufacturing of HSRS should respond actively the challenge of wear-vibration-fatigue.

Speakers: Jun FAN, Ming-wei PIAO, Peng DANG, Qi-bin WANG, Si-yang PIAO, Xi-bei CHEN

1:30 PM Disturbance Observer-based Sliding Mode Control of Active Yaw Damper for High-speed Rail Vehicles

Paper ID: 67

A disturbance observer-based sliding mode control (SMC) is developed for an active yaw damper (AYD) to improve the hunting stability of high-speed rail ve-hicles. Firstly, a lateral dynamic vehicle model is constructed, including the body roll, pitch and yaw, and the bogie pitch and yaw motions. Sliding surfaces are then designed for SMC, requiring only the states of the bogie yaw motion, but with the non-linear wheel/rail forces and unknown track irregularities estimated by the disturbance observer (DO). The control effectiveness of the AYD on the vehicle running stability is verified for various operating speeds, wheel/rail wear conditions and suspension arrangements by comparison with passive suspension and sky-hook control. Numerical experiments indicate that AYD can sufficiently extend the vehicle's adaptability to wheel/rail conicity and then go for a longer profiling cycle and higher running speed. Furthermore, AYD using SMC can re-alise radial steering control on the bogie in tight curves, then reduce the wheel/rail forces, then reduce the wear. It is concluded that multi-objective control on AYD can be sufficiently realised by the DO-based SMC.

Speakers: Huailong Shi, Jing Zeng, Jinying Guo, Shiqiao Tian

1:30 PM Dynamic loading of bulk cargo containers under impacts during rail transportation on 1520 gauge

Paper ID: 80

"Objective: Conducting a set of tests on the impact of a platform car with a container for the transportation of bulk cargo. Experimental establishment of the loading of container structures under the dynamic impact of bulk cargo, and development of recommendations for the calculation method.
Methods: Full-scale impact tests in the entire range of operating speeds permitted for the 1520 gauge.
Results: The dependencies of the loading of container elements were established, actual values of longitudinal accelerations of cars and containers were determined, with a distribution by height depending on the impact speed"

Speakers: Denis Danilenko, Dmitry Beyn, Ludmila Tsyganskaya, Nataly Tanicheva

1:30 PM Dynamic performances of DEM-based ballasted track subjected to polygonal wear of locomotive wheels

Paper ID: 195

Wheel polygonal wear is a focused issue of long-term concern in terms of its consistent harm and destructiveness upon wheel/rail matching relationship and train operating quality and safety. In this study, an innovative heavy-haul locomotive-ballasted track coupled dynamics model is developed, incorporating the flexibility of the wheel/rail system and a detailed ballast layer modeled using the Discrete Element Method (DEM). A simulation-based study upon the dynamic performances of ballast particles due to polygonal wear of locomotive wheels is conducted. Simulation results reveal that wheel polygonization considerably enhances the dynamic behavior of ballast particles, especially in terms of acceleration or velocity, which is affected far more than displacement. Compared to the non-polygonized case, polygonal wheel conditions amplify ballast acceleration and velocity amplitudes by factors of 20 and 8, respectively. The high-order polygonal wear can sharpen the medium- and high-frequency behaviors of the ballast vibration due to the deteriorated wheel/rail vibrations. This study could provide further guidance for the maintenance works of heavy-haul ballasted track with presence of wheel/rail injury and abnormal interactions.

Speakers: CAN SHI, Yunfan Yang

1:30 PM Effects of interphase short-circuit fault of permanent magnet traction motor on the dynamic performance of a direct-drive locomotive

Paper ID: 134

Permanent magnet direct-drive locomotives employ permanent magnet synchro-nous motors (PMSMs) as traction motors, where the electromagnetic torque gen-erated by the motors is directly transmitted to the wheelsets. As a common fault in the PMSMs, interphase short-circuit fault can induce significant ripple in elec-tromagnetic torque and result in abnormal vibrations within the direct-drive trac-tion system, thereby compromising the safety and reliability of the locomotive operation. In order to investigate the dynamic performance of a direct-drive loco-motive under permanent magnet traction motor interphase short-circuit faults, this paper develops an electromechanical coupling dynamic model of a direct-drive lo-comotive. The model considers the complete power transmission path within the drive system, including the electric motor control system and the mechanical sys-tem. Simulation results indicate that the ripple torque caused by interphase short-circuit faults can deteriorate the operation conditions of the locomotive system and exacerbate vibrations of the key components at frequencies related to fault frequency. The findings in this paper can provide theoretical guidance for the fault diagnosis and intelligent operation of direct-drive locomotives.

Speakers: Zaigang CHEN, Zirui YE, Ziwei ZHOU

1:30 PM Estimation of the local stiffness between derailed railway axles and containment structures for multibody post-derailment simulations

Paper ID: 229

Train derailment represents a critical challenge in railway safety, posing significant risks to human lives and infrastructure integrity. To mitigate its consequences, various derailment mitigation measures (DMMs), such as guard rails, kerbs and containment walls are implemented on high-risk zones of the railway line. However, the design of these devices remains complex and lacks standardized guidelines. Previous research has approached the design of containment devices using a combination of finite element analysis (FEA), for detailed structural modelling and multibody approach for the study of vehicle dynamics. Anyway, one of the challenging points is the definition of local contact-impact stiffness parameters. This study aims to address this limitation by proposing an enhanced methodology, based on FEA, for determining local contact stiffness between the derailed wheelset and DMMs, enabling more accurate and reliable multibody analysis.

Speakers: Edoardo Montini, Egidio Di Gialleonardo, Francesco Mazzeo, Matteo Santelia, Stefano Bruni

1:30 PM Evaluation Method for Wheel-Rail Relationship Based on Nonlinear Characteristics of Equivalent Conicity

Paper ID: 38

The nonlinear characteristics of equivalent conicity have a significant impact on the hunting stability of railway vehicles. In particular, its "negative slope" behav-ior can reduce the critical speed. To enhance the accuracy of using equivalent conicity for evaluating wheel–rail interaction, this study applied the wheelset mo-tion equation to calculate the equivalent conicity and investigates its nonlinear features. Two evaluation methods that account for the nonlinear behavior of equivalent conicity—the integration-based method and the differentiation-based method—are proposed as supplements to the conventional 3 mm equivalent conic-ity. These methods are validated using long-term monitoring data from high-speed electric multiple units (EMUs). The results demonstrate that the differentiation method, when combined with the 3 mm equivalent conicity, more accurately re-flects the actual wheel–rail contact conditions and offers valuable guidance for the operation and maintenance of EMUs.

Speakers: Huanyun Dai, Zhenhuan Yang

1:30 PM Experimental Investigation of the Influential Factors of Abnormal Vehicle Interior Noise in a Metro Line

Paper ID: 207

"In the past decades, as the most efficient urban transportation system, metro is fast developed in large cities in China. Due to the extremely limited space in urban area (usually more than 10 thousand people per square kilo meter), the route designs of metro lines are significantly constrained and adverse situations such as small radius curves and short transition zones are obliged to be introduced. Such situations often lead to the fast degradation of railway track and resulted in the increase of noise and vibration. With the continuously restricted environmental laws and regulations, the metro-induced noise and vibration have attracted increasingly more public attention and series of complaints. With the wide application of anti-vibration track, especially the steel-spring floating slab track, the environmental impacts of the train-induced ground vibration can be effectively controlled. However, the abnormal vehicle interior noise that resulted in an increasing number of complaints has no good solution yet. Therefore, investigate the key influential factors of abnormal vehicle interior noise and figure out effective remediation schemes are urgently needed.

Up till now, research on the vehicle interior noise, especially the connection between abnormal noise and track alignment and condition, are rather limited. Most of the studies are focused on the characteristics of the noise itself. Soeta et al.[1] pointed out that in the static state, the most important noise source of the vehicle is auxiliary equipment noise. Auxiliary equipment operation will produce vibration, which leads to unstable airflow around the equipment and generate noise. This type of noise is high in frequency and low in sound level. Thompson et al.[2] discovered that rolling noise reigns supreme at low-to-medium speeds with its frequency firmly nestled in 500~5000 Hz range. Zheng et al.[3] analysed the quality of in-vehicle noise in different zones and found that rail wave abrasion is prone to cause noise in the frequency bands of 63~125 Hz, 315~500 Hz and 630~1000 Hz.

As train speeds escalate, aerodynamic factors come to the fore. Jia et al. [4] illuminated the significance of internal aerodynamic noise within ventilation and air - conditioning systems. Lan et al.[5] carried out in-vehicle noise tests with a high-speed train as the research object and found that when the train runs in a tunnel at a higher speed, the in-vehicle noise mainly comes from the aerodynamic excitation on the surface of the vehicle body, and the main frequency range is 160~1000 Hz.

In order to investigate the factors that caused the abnormal vehicle interior noise, the noise level of a metro line is measured and analysed. The sound pressure sensors are setup upward and 1.2 m above the floor, and the sampling frequency is set as 51.2 kHz to make sure that the main noise frequency bands are covered. To eliminate interference, the broadcast, air conditioner as well as the station announcement are turned off during the measurement.

There are totally 3 sound pressure sensors installed inside the measurement train, respectively one in the driving cab, one in the front passenger cab and one in the middle passenger cab. The train is operated according to the actual operation diagram and the sound pressure is recorded continuously. The travelling distance is calculated based on the operation diagram. By transferring the sound pressure signal into 1/3 octave band sound pressure level, the noise distribution along the track can be obtained. An example of vehicle interior noise distribution map shows that the sections of 22.7~22.8 km and 23.0~23.1 km have high interior noise and the dominant frequency bands are around 800 Hz. For the whole metro line, there are more than 30 intervals and roughly 50~100 sections with comparatively high interior noise will be screened up. By establishing the relationship between vehicle interior noise and the track structure parameters, the key influential factors of such abnormal noise can be found.

This study is of great guiding significance for the construction, operation and maintenance of metro lines. In case of high noise caused by the adverse situations, additional control measures such as track lubrication need to be taken into account in the construction stage; while for the abnormal noise introduced by the track degradation, effective maintenance should be carried out in a timely manner. By reducing the vehicle interior noise, the metro travelling comfort can be greatly improved, which will further contribute to the continuous promotion of the green development of urban rail transit.

Speakers: Long Wang, Xiangming Liu, Ying Chen, Yuxuan Wei

1:30 PM Experimental study of a scaled running gear based on independently rotating wheels with caster angles

Paper ID: 165

The mechanism of front wheel systems of automobiles and bicycles was utilized to make a breakthrough in the self-steering ability for the independently rotating wheel (IRW) system. An idea is proposed to recover the self-steering ability for the running gear composed of independently rotating wheels. The idea was put forward to achieve the self-steering ability for the running gear with four independently rotating wheels by using caster angles. Yaw moments were generated to assist the automatic steering of the novel running gear which is the second generation. Experiments and numerical simulations were carried out to demonstrate the self-steering behavior of the running gear. The results show that the novel running gear possesses the self-centering ability on straight tracks. It can restore to the track center when it is drift to one side of the straight track. The lateral shift of wheels can be damped out by the wheel-rail friction. The exponentially damped hunting motion is observed for the new running gear. It is helpful to mitigate the wheel-rail wear by automatically reducing the attack angle. The recorded accelerations of a scaled prototype validate the lateral oscillation phenomenon of the running gear. The experiment gives an obvious demonstration of the stable hunting phenomenon of the running gear.

Speakers: Chao Yang, Haotian Wang, Jianxin Zhou, Linxuan Wu, Ning Xu, Yuxian Liu, Zunsong Ren

1:30 PM Explicit Speed-Integrated LSTM Network for Non-stationary Bogie Vibration Representation and Beam Joint Detection under Varying Speed Conditions

Paper ID: 139

The joint of the monorail track beam is crucial for the strength design of the tires and framework, and it is necessary to specify requirements for the joint's connection shape. Existing fault detection methods have not explored the scientific structure that integrates speed signals into Long Short-Term Memory (LSTM) networks, leaving room for improvement under varying speed conditions. This paper proposes a novel explicit speed-integrated LSTM (SI-LSTM) model to enhance the representation accuracy of non-stationary vibration signals and improve gearbox fault detection capability. Comprehensive experiments are first conducted using a dataset of axle box vibration data obtained from the bogie on an actual railway line, which was collected through track irregularity tests. The superiority of the proposed method is validated, and the model is further applied to real lines (Wuhu,China), achieving the highest AUC of 0.9998 on the dataset and demonstrating the best performance in vibration representation accuracy.

Speakers: Lihui Ren, Xiaoguang Ma, Yuanjin Ji, Yuejian Chen

1:30 PM Impact assessment of high-speed train suspension parameters and track characteristics on smooth running by computer modeling method

Paper ID: 239

In this paper, the influence of the condition of the wheel and rail, the supereleva-tion and suspension damping value on the ride index of the high-speed train was estimated using computer simulation methods. Simulation results show that the combination of worn wheel and rail affects the vehicle ride index, superelevation value does not have a significant effect on the lateral ride index. The smooth ride of the train under study could be improved by regulation its suspension damping value.

Speakers: Alina Saidova, Anastasiia Semenova, Anna Komarova, Veronika Fedorova

1:30 PM Impact dynamic behaviors of the pantograph-catenary system in the overlap section at a speed of 400 km/h

Paper ID: 261

The impact phenomena and contact separation process within the overlap section of the catenary system deteriorate the current collection quality of the pantograph catenary system (PCS), induce structural damage, and pose a significant threat to the train’s operation safety, especially at the speed of 400 km/h or higher. The mathematical model of the PCS is established based on the finite element(FE) method and validated against the line test data. The impact dynamic behaviors of the PCS crossing through the overlap section are revealed. The optimal parame-ters in the sensitive transition section are proposed to reduce the impact and to improve the contact quality in the contact separation process at a speed of 400 km/h, consisting of the span length, the suspension height, and the dropper ar-rangement. The results benefit the structural design of the 400 km/h railway cate-nary in the overlap section.

Speakers: Guiming MEI, Jiangwen Wang, Weihua Zhang

1:30 PM Improving the Lateral Stability of High-Speed Trains using Elastic-Suspended Motors Based on Inerter Structure

Paper ID: 32

As a new type of structural control element, the damper container is widely used in structural vibration systems. However, its application in the elastic-suspended motor (SD) has been less studied. Therefore, this paper establish-es a model of the motor suspension with inerter (ISD) for high-speed trains, analyzing the frequency-varying characteristics of the ISD structure and its impact on the stability of the bogie. Secondly, considering the influence of wheel-rail contact on vehicle stability, a global sensitivity analysis of the main suspension parameters and optimization of the suspension parameters were carried out, followed by a comparative study of the dynamic responses of the two. As a dynamic vibration absorber (DVA) for the bogie system, the motor can improve the bogie's stability by sacrificing the motor's stabil-ity. The research results can provide a reference for the design of the elastic bogie suspension of high-speed train motors.

Speaker: Yayun Qi

1:30 PM Intelligent condition monitoring of wheel polygon using a deep learning model with novel metrics and loss functions

Paper ID: 95

Accurate and fast condition monitoring of wheel polygon is of great importance for uncovering its mechanism, predicting the evolution and planning maintenance. Traditional physical model-based monitoring methods rely on high-fidelity simulations and are poorly automated. The popular data-driven methods have good availability. However, they ignore the power of expert experience, resulting in limited accuracy in quantitative monitoring. Driven by big data and domain knowledge, this study presents a quantitative detection model for wheel polygon using deep learning algorithms. The roughness characteristics of the wheel polygon are thoroughly analyzed. Adaptive metrics and loss functions are then developed for each sample so that the monitoring model can obtain better global feature regression and local detail convergence capabilities. The effectiveness of the proposed method is verified using on-site measurement data from a commercial high-speed rail line.

Speakers: Gongquan Tao, Qinglin Xie, Zefeng Wen

1:30 PM Investigation of ballast degradation and fouling based on in-situ sampling

Paper ID: 69

During the ballast degradation process, fine particles resulting from particle abrasion and breakage, along with fine particles from the environment, contribute to ballast fouling. This fouling results in differential settlement and reduces both the drainage capacity and structural performance of the ballast bed. The issue of ballast fouling is significant due to the considerable effort and costs required for maintenance. To better understand the ballast degradation and fouling status, this research collected and sieved over 900 samples from active railway lines. The samples were obtained under various weather conditions, geographical settings, and types of transportation. The results comprehensively reflect the evolution of ballast fouling over its service life and provide a database for further validation using GPR (Ground Penetrating Radar) in rapid fouling inspections.

Speakers: Fei Yang, Jinzhao Liu, Weile Qiang, Wenli Jia

1:30 PM Investigation on Coupled Vibration Characteristics of the Metro Vehicle-Floating Slab Track System

Paper ID: 313

Vehicle component vibrations intensify when external excitation frequencies approach their natural frequencies, elevating the risk of vibration fatigue damage. To avoid resonance, vehicle component designs must be adjusted to shift their natural frequencies away from external excitation frequencies. A primary source of such excitation is vehicle-track coupling vibration. Identifying key frequencies in the vehicle-track coupled system is thus crucial for frequency matching during vehicle design. This study focuses on the vehicle-track coupling system involving metro vehicles and floating slab tracks, commonly used in urban rail transit. A model based on the rigid-flexible coupling method, considering the flexibility of both wheelsets and track structures, is proposed. Simulations reveal the frequency responses and modal characteristics of the system, identifying key frequencies for metro vehicle-floating slab track coupling vibration, which provides a foundation for vehicle frequency matching design.

Speakers: Bingbin Guo, Longjiang Shen, Xuancheng Yuan, Zhixiang Luo, Zhongcheng Jiang, Zuobin Zhou

1:30 PM Low-Cycle Fatigue Damage Assessment and Strategy for Some Stainless Coal Hopper Wagons Based on Rigid-Flex Coupling Simulation Technique

Paper ID: 112

The proposed rigid-flex coupling multiaxial fatigue damage assessment tool is necessarily used in the backward design of dynamic systems for the rail freight wagons with heavy axle-load, so as to accurately predict the occurrence location of weld low-cycle fatigue damage and correctly guide the decision-making direction in managing operation and maintenance. Unlike the case of carbon steel coal gondolas, some stainless steel coal hopper wagons adopt the monocoque structural design with curved sidewalls to enhance the structural rigidity of fully-loaded carbody. This substructure proxy model forms the dual-layer transition with tetrahedral elements of 50 / 150 mm using the shell extracting / nesting technique. Considering the anti-overturning capability of bolster suspensions, the improved processing technique of solid elements with boundary common nodes is implemented to make the operating load variations of upper side-bearings closer to the actual situation. Under the input excitation of local twisted track, the rigid-flex coupling simulation analyses show that significant local torsional elastic deformation will occur diagonally at the upper-corners. Considering the effects of relevant strong constraints, the predicted location of weld low-cycle fatigue damage is consistent with the experimental conclusion from full-wagon accelerated fatigue testrig. The foregoing fully-loaded carbody better solves the difficult problem in rigid-flex coupling simulations of rail freight wagons for transportation of bulk cargos.

Speakers: Chun-ge Nie, Jun-lin ZHANG, Ming-wei PIAO, Shi-ying JIN, Si-yang PIAO, Tian-ci HUANGFU, Xing-jia XU, Zong-yu HU

1:30 PM Modeling and Analysis of the Dynamic Interaction between Derailed Trains and Trackside Infrastructure

Paper ID: 94

Modeling and simulating the dynamic behavior of post-derailment trains presents significant challenges due to the complexity and unpredictability of contact interactions between the train and trackside guardrail. This study establishes a comprehensive dynamic interaction model between the train and guardrail to investigate post-derailment train behavior. Furthermore, a dynamic optimization method is applied to improve the accuracy of model parameter determination, ensuring reliable simulation results. The study evaluates the guardrail effectiveness in mitigating derailment consequences by analyzing key factors, including height and thickness, under various collision conditions (impact velocity and angle). A safety evaluation method is proposed to assess the guardrail performance based on impact forces, collision duration, lateral displacement, and wheel vertical rise. The results demonstrate that while guardrails can effectively contain post-derailment trains at low speeds and small impact angles, their protective capability diminishes at higher speeds and larger angle, where excessive lateral forces and wheel climb-over occur.

Speakers: Tao Chen, Zhao Tang, Zhiming Qu

1:30 PM Numerical Simulation-Based Study on the Dynamics of the Pantograph-Catenary System at Railway Turnouts

Paper ID: 56

This study investigates the dynamic interaction between the pantograph and catenary system as an electric train passes through the catenary line branch-ing section associated with a No. 42 turnout, which refers to the specially configured catenary structure designed to accommodate track divergence in high-speed railway turnout areas. Numerical simulations are conducted to evaluate the contact force and vertical displacement of the pantograph head under varying speeds and transition directions—specifically, transitions from the main line to the side line and vice versa. Results show that at speeds be-low 120 km/h, the contact force remains stable, with minimal differences be-tween leading and trailing pantographs, and between transition directions. However, as speed increases beyond 120 km/h, the standard deviation of the contact force becomes noticeably higher when transitioning from the side line to the main line, indicating greater dynamic instability. The maximum contact force exhibits a clear upward trend with speed, while the minimum contact force shows a non-monotonic behavior—first decreasing, then in-creasing—highlighting complex dynamic interactions. The findings also demonstrate that the use of auxiliary lines effectively smooths the transition process, mitigates impact loads, and enhances current collection stability, thereby supporting safe and reliable high-speed train operation through turn-out zones.

Speakers: Guiming MEI, Jiangwen Wang, Shiyu Zhao, Wenyan Qi, Zhong huang

1:30 PM Numerical Study on Wear Mechanism of Pantograph Contact Strip in Electrical Road System under Complex Truck Driving Behavior

Paper ID: 21

The heavy-duty truck plays an important role in road freight transportation sys-tem, and the electrical road system (ERS) is one of the most possible electrification solution of heavy-duty truck to reduce carbon emission and operation cost in long-distance travel. However, while the pantograph is one of the most important compo-nents in ERS, its contact strip’s wear mechanism is not fully investigated at present. In this work, the wear mechanism of ERS’s pantograph contact strip is studied. The ERS’s pantograph-catenary-truck-road interaction model and truck traction model are formulated based on multibody dynamic theory, finite element method, and ex-isting traction model, and the truck running at constant speed, acceleration, braking, and lane changing are considered. It is found that the truck running at constant speed has normal wear behavior, but the braking and lane changing can results in obvious abnormal wear behavior, which should be handled through driving behavior and pantograph structure optimization.

Speakers: Bastian Schick, Dietmar Göhlich, Park Sangyoung, William Zhendong Liu, Zhiwei Zhou, yan xu

1:30 PM Optimization of wheel profile according to wheel/rail friction temperature

Paper ID: 166

With the improving of adhesion utilization of heavy-haul locomotive, the friction power between wheel and rail increases significantly. The friction power would cause temperature rising in the contact area. The high temperature in wheel/rail contact area would decrease the strength of wheel material and cause wheel tread spalling. An analysis of BEM-FEM coupling procedure is presented and the key factors that influence wheel/rail friction temperature are studied. A method for wheel profile optimization is presented based on wheel/rail friction temperature and locomotive dynamic characteristics.

Speakers: Bo Peng, Longjiang Shen, Xiaobo Zhong, Xihong Chen

1:30 PM Performance Evaluation of Regenerative Dampers in High-Speed Trains Wheel Wear, Dynamics, and Energy Recovery

Paper ID: 176

The study provides a quantitative evaluation of hydraulic-electric regenera-tive dampers (PSRDs) applied to high-speed train suspensions, focusing on their effects on wheel wear, profile optimisation, and vertical dynamics of carbody and bogie. Using a co-simulation framework combining MATLAB and SIMPACK, the investigation analysed vehicle-track interaction under various operational speeds up to 350 km/h. Results reveal that PSRDs achieve a measurable 12% reduction in wheel wear and ensure consistent profile interaction, minimising maintenance requirements. Time-domain analysis indicates a 25% decrease in peak vertical accelerations, while fre-quency-domain analysis demonstrates significant vibration attenuation with-in critical ranges of 1–20 Hz for carbody and 20–60 Hz for bogie. Further-more, PSRDs generate an average power output of 21.72 W with an efficien-cy of 45.28%, contributing to onboard energy savings. The findings under-line the dual advantages of enhanced dynamic performance and energy re-covery, offering a pathway to sustainable advancements in high-speed rail systems.

Speakers: Adam Bevan, David Crosbee, Paul Allen, Ruichen Wang, Yating Han, Yunfan Yang

1:30 PM Physics-Guided Reinforcement Learning for Straddle Monorail Active Suspension with Image-Based Beam Joint Detection

Paper ID: 232

The joints of monorail track beams cause impacts on the vehicle body and framework, reducing ride comfort and leading to uneven tire stress. Unlike passive suspension systems with constant spring and damping coefficients, active suspension systems incorporate electronic actuators to dynamically control stiffness and damping variables. This paper proposes a Physics-Guided Deep Reinforcement Learning (DRL) method based on image recognition for real-time adjustment of the active suspension system in monorail vehicles. Specifically, an innovative control strategy combining image preview information and physical models is introduced to enhance the suspension system’s response to track joint impacts. Real-time image processing techniques are used to detect the presence and location of track joints, predicting potential impact events. This information is integrated with the suspension system’s dynamic model to adjust control strategies in real time. Additionally, prior knowledge of joint locations is incorporated into the DRL training process to improve adaptability to sudden impacts. For scenarios where only the presence and location of joints are available, a preview-based pre-adjustment control method is proposed, extending the reinforcement learning state space to include the predicted distance between the joint and the vehicle. The model is trained on stochastic road profiles conforming to ISO 8608 standards to optimize actuator control strategies.

Speakers: Jiakang Li, Jian Li, Kaiying Zhang, Lihui Ren, Yuanjin Ji

1:30 PM Physics-Informed Data-Driven Prediction of Abnormal Tram Wheel Wear

Paper ID: 110

Trams with independently rotating wheelsets are widely used in urban rail transit systems to accommodate sharp curves. However, this design can lead to abnormal wear on wheel flanges and backs, impacting operational effi-ciency and maintenance costs. This study presents a five-year field meas-urement analysis of identical articulated low-floor trams operating in two cit-ies with distinct track conditions. In City A, wheel wear was predominantly observed around the flange region, with significant reductions in flange thickness as mileage increased, while tread wear remained minimal. Con-versely, in City B, wear was distributed more evenly across the tread and flange root areas, but wheel back wear was significantly higher. To address these phenomena, a physics-informed data-driven wear prediction model was developed, incorporating the Archard wear model and adapting it to dif-ferent wheel positions. The model estimates future replacement needs and costs for non-compliant wheelsets over seasons and years, providing a robust reference for optimizing tram maintenance strategies and reducing operation-al disruptions

Speakers: Lihui Ren, Maozhenning Yang, Yuanjin Ji, Zhangxing Huang

1:30 PM Predicting Light Rail Vehicle Wheel Profile Evolution Under Severe and Mild Wear Conditions Using Numerical Simulations

Paper ID: 172

Urban rail systems operating in densely built environments often incorporate tight curves, which exacerbate wheel and rail wear, thereby increasing maintenance costs and reducing operational efficiency. To investigate the problem of accelerated wheel wear in Indonesian light rail systems, a numerical model of the railway vehicle–track interaction was developed using a multibody software package. This model integrated actual wheel–rail profiles and realistic operating conditions. Field observations revealed that the accumulation of metal debris from worn material was significantly more pronounced on tight curves than on those with larger radii. In response to this, the wear simulation was designed to differentiate between mild wear occurring on large-radius curves and severe wear on tighter curves. This differentiation was achieved using the Specht model, which is based on the Krause–Poll law and is capable of replicating both mild and severe wear regimes. The transition between wear regimes was defined by the power density of the friction force at the wheel–rail contact patch. The simulated wear results showed close agreement with field measurements, with differences in flange and tread wear limited to just 0.05 mm and 0.07 mm, respectively. These findings validate the accuracy of the model in predicting wear behaviour and support its use in further investigations into the root causes of excessive wear in light rail systems.

Speakers: Eki Putra, Fariz Kharisma, Yunendar Handoko

1:30 PM Rail Vehicle Collision Dynamics-driven Study on the Practical Application of Train Collision Derailment Prevention

Paper ID: 77

Derailment caused by train collision is a major threat to traffic safety. This study breaks through the tradition and innovatively incorporates the initial height and pitch angle of the train into the dynamic model, which is rarely considered in previous studies. The research optimizes the wheel-rail relationship solving module, and establishes a mathematical model for the key components such as coupler buffer device, energy absorption and anti-climbing device for the first time. These components are either inaccurate or neglected in previous studies. The development of a unique train collision calculation program further demonstrates its innovation. The accuracy and stability of the new developed dynamic model are verified by comparison with the finite element simulation analysis. Subsequently, the improved model system was used to analyze the influence of the initial vertical height difference and pitch angle of the train on the derailment behavior at different collision speeds. The research results are of great significance for improving the safety of train collision derailment, improving the relevant derailment standards and guiding the optimization of train design, and have made unique and valuable contributions to the field of train safety research.

Speakers: Haoxu Ding, Jingke Zhang, Tao Zhu, Wenyue Yuan, Xiaorui Wang

1:30 PM Reproduction of wheel polygonization and rail rutting corrugation on a full-scale roller rig

Paper ID: 75

"Wheel polygonization and rail rutting corrugation are increasing issues for rail-way systems with a high number of tight curves. The excitation sources and excited resonance frequencies of the coupled vehicle-track system that contribute to the formation of these quasi-periodic wear patterns are not yet fully understood. In this paper, a full-scale roller rig is used to investigate the initiation and growth of wheel polygonization and rail corrugation.
Two test campaigns of increasing complexity were conducted to systematically isolate and assess the influence of key parameters. The wheel polygonization and rail corrugation generated during the tests were found to correspond to the resonance frequencies of the rail-wheel, identified via impact hammer testing. These modes were excited by the initial surface roughness of the reprofiled wheels during test runs at a high angle of attack.
The results demonstrate that wheel polygonization and rail rutting corrugation can be systematically produced and analyzed on a full-scale roller rig under controlled conditions, supporting future efforts in modeling and the development of mitigation strategies."

Speakers: Felix Platzer, Klaus Six

1:30 PM Research on collision dynamics modeling and derailment behavior based on speed-sensitive coupler

Paper ID: 100

The derailment of a train is the main factor that causes casualties in severe collisions, so the in-depth study on the passive safety design of trains is of great significance in reducing the loss of collision accidents. In this paper, the viscoelastic characteristics of the coupler's mastic damper are analyzed, and it is equivalent to the system composed of variable stiffness spring and damping elements. Combined with the characteristics of slot laminar flow, thin hole flow, and the compressibility of the mastic material, the elastic force and damping force are calculated, respectively, and the corresponding mathematical model is established. The reliability of the model is verified by comparison with the trolley test. Secondly, based on the theory of train collision dynamics, a speed-dependent coupler model is introduced to modify the collision dynamics equation further. Finally, the influence of various parameters of the head coupler on the maximum wheelset lifting is analyzed. The research results provide theoretical support for optimizing the passive safety design of trains and have important guiding significance in reducing derailment risks under different collision speeds.

Speakers: Binlin Wang, Haoxu Ding, Jingke Zhang, Tao Zhu, Tianyi Lv, Wenyue Yuan, Xiaorui Wang

1:30 PM Research on dynamic parameters matching of pantograph and overhead conductor rail system at speed of 200km/h

Paper ID: 133

Overhead conductor rail (OCR) have been widely used and proven effective in metro lines operating at speeds below 160 km/h. However, existing designs perform poorly at higher speeds (200 km/h and above). This paper focuses on the dynamic parameters matching issue of pantograph - overhead conductor rail system (POCRs) at a speed of 200 km/h. Key parameters for POCRs dynamic simulation model were obtained through experimental methods. Model validation was achieved by comparing modal calculation values with experimental results. The influence of factors such as suspension stiffness, span length, overlap section design, and pantograph selection on the dynamic characteristics of the POCRs were analyzed. This research provides support for the design of POCRs operating at 200 km/h and above.

Speakers: Congjian Jiang, Guiming MEI, Jiangwen Wang, Jun Li, Mingguo Chen, Shiyu Zhao, Yanlin Zhang, Zhenbao Chen, Zhong Huang

1:30 PM Research on Evaluation Method for Conformal Wear of High-Speed Train Wheel-Rail Surface

Paper ID: 330

The vehicles operated by China's high-speed rail are fixed EMUs, which causes conform wear in wheels and rails. To maintain wheels and rails, wheels are reprofiled based on the equivalent conicity, which is calculated with standard rail profile. Rails are ground into optimized profiles, which are also calculated using standard wheel profiles. The existing evaluation method of wheel and rail profiles does not consider the characteristic of conform wear of wheels and rails and both of them require complete profiles. A research is carried out on the wheel-rail concentrated wear evolution and two indices are created to describe those regional wear via the degree of concavity Aw for the wheels and the degree of collapse Tr for the rails. The algorithm is proven to be reliable with less measuring points on profiles and two equipments for calculating these two indices are developed.

Speakers: Gang SHEN, Guangneng Fan, Hongxuan Chen

1:30 PM Research on Feature of Hunting Stability of EMU Based on Wheel Tread Hollow Worn and Modal Vibration Decomposition Method

Paper ID: 205

With the continuous increase in operating mileage and speed, the wheel wear of high-speed EMUs is gradually increasing, which in turn has a certain impact on vehicle stability. To analyze the influence of the development of hollow worn treads on hunting stability, the relationship between wheel wear and the hunting motion of high-speed EMUs is examined. Firstly, this paper proposes parameters that can accurately fit the distribution of hollow wear. Then, a dynamic model similar to the real vehicle is established and verified through experiments on high-speed trains. Modal tracking of the vehicle system is then utilized to reveal the relationship between low equivalent conicity and carbody hunting, as well as the relationship between high equivalent conicity and bogie hunting. Finally, a detection method for carbody hunting based on sample entropy and LS-SVM is proposed with an accuracy rate of 97%. The research results can provide reference for the maintenance and operation of rail vehicles.

Speaker: Xiaoyu Li

1:30 PM Research on modal optimization method and dynamic performance improvement of rack vehicle

Paper ID: 116

The stability and riding comfort of rack vehicle deteriorate due to track excitation such as wheel-rail irregularities and gear-rack meshing impact. The modal characteristics of the vehicle are important in the distribution of system responses and energy transfer. This paper proposed a Modal Optimization Strategy (MOS) based on the identification and analysis of critical vehicle modes. Subsequently, an adaptive frequency band weight and penalty function was introduced to design an objective function incorporating modal contribution and frequency response function. This function effectively ac-counted for the relationship between external excitation and vehicle modal properties. Finally, a Multi-Objective Whale Optimization Algorithm (MOWOA) was employed to optimize the suspension system. The results demonstrate that the optimized suspension system effectively reduces vibration response amplitudes and significantly improves the dynamic performance of the rack vehicle. Frequency sweep analysis indicates that the riding comfort is enhanced within the operating speed range. The finding provides valuable theoretical guidance and engineering application potential for the design and performance enhancement of rack vehicles.

Speaker: haitao zhang

1:30 PM Safety Assessment Based on Machine Learning Methods of the Heavy-Haul Trains during Emergency Braking Process under Different Line Conditions

Paper ID: 138

With the increase in heavy-haul train formations and speeds, issues related to train dynamics have become increasingly prominent, particularly concerning safety and efficiency during braking. Therefore, studying the dynamic responses and safety performance of heavy-haul trains under braking conditions is crucial. Traditional longitudinal train dynamic (LTD) models cannot directly derive dynamic performance indicators, while three-dimensional train dynamic models are complex and computationally expensive. To address these limitations, this paper develops a hybrid one-dimensional-three-dimensional train dynamic model that considers the effects of the braking system and line conditions, achieving a balance between detailed simulation at critical segments and simplified calculations at relatively non-critical parts. In addition, machine learning techniques are employed to establish a surrogate model for rapid and accurate assessment of the dynamic indicators of heavy-haul trains. Combined simulation data with machine learning algorithms—least squares support vector machine (LSSVM), back propagation (BP) neural network, and long short-term memory (LSTM) network—safety indicators during the emergency braking process, such as maximum wheel unloading rate, derailment coefficient, wheelset lateral force, and coupler compression force, are evaluated. Results indicate that the assessment accuracy of the BP and LSTM models exceeds that of the LSSVM model, with the LSTM model demonstrating the highest accuracy.

Speakers: Mingtao Zhang, Xinyuan Qu, Yukun Wang, Zhiwei Wang, kaizhong liu, weihua zhang

1:30 PM Sensor fusion for detection of rail bending at 320km/h

Paper ID: 335

The research project described in the document aims to develop an automatic system for detecting rail deformations caused by high-speed gravel impacts, a phenomenon known as rail bending. Currently, these defects are manually detected, often after track geometry problems arise. The proposed system utilizes data fusion between image analysis and the analysis of vertical accelerations from axle boxes. A neural network is trained to identify ballast impacts on rails from images and distinguish these impacts from other similar forms. Simultaneously, a frequency analysis of the accelerations confirms the presence of bent rails. These two combined methods offer early detection of bent rails, facilitating quicker corrective interventions, minimizing the impact on traffic regularity, and reducing costs. This system will increase network availability by enabling automatic monitoring onboard trains without needing to block tracks for manual inspections.

Speakers: Marc LEDET, Olivier VO VAN, Solenne THEVENET, Stéphane NEVEU

1:30 PM Study on contact force identification method considering vibrations of stable arm on pantograph

Paper ID: 58

The measurement of contact force has long been a central topic in pantograph–catenary research. As a key coupling variable between the pantograph and the catenary, the contact force contains rich information that can be used to evaluate current collection quality, the wear condition of the contact strip or contact wire, and to diagnose various fault states. In this study, the pantograph head is treated as an independent input–system–output subsystem, with the moving contact force and stable arm vibration as inputs. Based on the modal superposition method, both forward and inverse models of the pantograph head dynamics are established, forming the foundation for the contact force identification process. Furthermore, a novel L1-regularization approach based on cosine and wavelet dictionary matching is proposed to solve the inverse problem and estimate the con-tact force. Identification results demonstrate that the proposed method offers strong robustness and stability, even under noise interference. In addition, it is shown that considering the vibration of the balancing arm is essential for improving the accuracy of moving contact force identification.

Speakers: Haifei Wei, Ning Zhou, Weihua Zhang, Xingshuai Zhi

1:30 PM Study on the countermeasures for preventing coupler separation of middle locomotives in 20,000-tonne heavy-haul trains

Paper ID: 327

Coupler separation mainly occurs in combined heavy-haul trains, especially between locomotives and vehicles in the middle of them. This article aims to propose effective prevention measures to suppress the eventual occurrence of coupler separation when there is a tendency for separation between the connected couplers. Firstly, a three-dimensional dynamic simulation model of train short formation was established, which the contact characteristics of the coupler heads under real geometric profiles were considered in detail. Then, through numerical simulation, the effect of the free pitch angle of the locomotive coupler was discussed. Besides, the effect of the locomotive's secondary suspension system on coupler separation are explored, including the vertical stiffness of the secondary springs and the free clearances of the secondary vertical stoppers. The simulation results indicate that the free pitch angle of the locomotive coupler, as well as the vertical stiffness of the locomotive's secondary springs and the free clearance of the secondary vertical stoppers, all exert a certain degree of influence on coupler separation. When the connected couplers exhibit a tendency to separate, reducing the free pitch angle of the locomotive coupler, the vertical stiffness of the secondary springs, and the free clearance of the secondary vertical stoppers all have a positive effect on inhibiting coupler separation.

Speakers: Jian Wu, Kaiyun Wang, Liang Ling, Shiqian Chen

1:30 PM Study on Vibration Damping of Traction Transformers under High-Speed Trains Based on Acoustic Black Hole

Paper ID: 121

"With the widespread adoption of Electric Multiple Unit (EMU) technology, an increasing number of diverse types of devices of different masses such as transformers and electric power converters, which are suspended beneath the vehicle chassis have significantly deteriorated the vibration levels of rail vehicles, posing challenges to operational safety and service life. This paper presents a vibration damping solution for the traction transformer beneath a high-speed train, based on the concept of acoustic black hole (ABH). Using a specific high-speed train model as a case study, the dimensions of an acoustic black hole vibration absorber suitable for the traction transformer were designed based on engineering requirements. Finite element analysis was then conducted to validate the effectiveness and feasibility of the damping functionality.
Building on this foundation, the influence of damping layer parameters and layout strategies on vibration absorption performance was systematically analyzed through structural, layout, and topology optimization. A novel absorber topology was proposed, offering a balance between lightweight design and vibration damping performance. Finally, an application study was conducted on the traction transformer of a particular EMU model, confirming the excellent vibration damping performance across the full frequency range."

Speakers: Dao Gong, Jinsong Zhou, Jiwei WU

1:30 PM Tensor Singular Spectrum Analysis for rail corrugation detection by vibration and noise signals

Paper ID: 57

Short-wave irregularity of rail, such as corrugation, is not only likely to excite high-frequency vibrations in the vehicle-track system because of wheel-rail contact fatigue, but also may cause the failure of key track components, which makes it urgent to pursue a comprehensive inspection and evaluation system to detect short-wave defects besides the existing diagnostic method by axle box acceleration. In this study, we apply tensor singular spectrum analysis (TSSA) to detect rail corrugation by combining vibration with acoustic signals. The 3-order tensor is first constructed by phase space reconstruction, and then tensor singular value decomposition is utilized to investigate the low-rank intrinsic features. Finally, the effective component of rail corrugation is obtained by reconstruction. Results show that the exact frequency band of rail corrugation can be clearly collected from both vibration and noise signals without any contamination. Compared with traditional multivariate decomposition approaches, like EEMD and MVMD, TSSA has proved the superiority of feature extraction. The construction of tensor retains the original structure of multisource signals to the maximum extent. In the future research, we will continue to optimize the model parameters so as to detect different kinds of short-wave diseases. The combined diagnosis method based on vibration and noise signals will improve the evaluation of the rail short-wave regularity status system.

Speakers: Jinzhao Liu, XUEGENG MAO, Zezhou Liu, Zhehao Huang

1:30 PM The Influence of EHA Active Control Anti-yaw Damper on the Stability of High-Speed Trains

Paper ID: 37

The hunting stability problem is the most typical dynamic problem in the operation of high-speed trains. During the long-term operation of high-speed trains, due to the evolution of the equivalent conicity of wheel profile, high-speed trains are prone to hunting stability problems, which typically include bogie hunting and carbody hunting. Traditional anti-yaw dampers have certain limitations in suppressing hunting instability. This paper proposes an anti-yaw damper based on EHA active control. The actuator can control the relative displacement between the bogie and the carbody, thereby enhancing the stability of high-speed trains. An actuator system model was established, and a fractional-order PI controller was used to control the displacement. Using SIMPACK and MATLAB/Simulink software, a dynamic model of the active EHA anti-yaw damper in high-speed train was established. By analyzing dynamic indexes such as wheelset lateral displacement, and critical speed, the effect of the EHA active control anti-yaw damper on the stability of high-speed trains was evaluated, and the suppression effect of the active anti-yaw damper on typical bogie hunting and carbody hunting was analyzed. The relevant research can provide suggestions for further improving the stability of high-speed trains.

Speakers: Huanyun Dai, Yayun Qi

1:30 PM The mechanics of Shimmy vibration of monorail vehicle

Paper ID: 6

This study investigates the nonlinear vibration characteristics of a monorail vehicle equipped with tyres. A dynamic model of the monorail vehicle's bogie was first established using the magic tyre model. Numerical simulations identified a limit cycle in the system, manifesting as shimmy vibration in the bogie's yaw, coupled with roll motion. The trajectory equation of the limit cycle was derived, and its occurrence conditions were analyzed.

Speaker: Shiqiao Tian

1:30 PM The third generation of self-steering running gears with independently rotating wheels: experiment and simulation

Paper ID: 44

Toe angle was integrated into an independently rotating wheel (IRW) running gear to enhance the self-steering performance. The original idea was the utilization of toe angles which derive from the mechanism of wheel systems of automobiles. Experiments and simulations were carried out to demonstrate the self-steering performance of the running gear. The results demonstrate that the toe angle has an effect on the self-centering capability. When the toe angle is less than 0.2°, the IRW running gear fails to achieve self-centering capability. The IRW running gear with toe angles not less than 0.2°can automatically return to its initial lateral position after passing through the excitation zone. The toe angle configuration effectively restores the self-centering capability of IRW running gear. The wear of IRW is related to the value of the toe angle. The experiment of a scale model is carried out to validate the self-steering phenomenon of the new running gear with toe angles.

Speakers: Chao Yang, Haotian Wang, Jianxin Zhou, Yuanjie Hu, Zunsong Ren

1:30 PM Thermomechanical Fatigue Life Evaluation of Train Brake Discs: Integrating Slope Profiles and Wheel-Rail Dynamic Response

Paper ID: 283

This study investigates the thermomechanical behavior and fatigue life of train brake discs considering rail vehicle dynamics and various track slopes. The working history (the relationship among speed, running dis-tance and time) of a railway vehicle on a continuous slope track is calculat-ed through a self-developed program, and an efficient indirect dynamic-thermodynamic coupled method is proposed. Based on the thermomechani-cal modelling and its simulation results, the thermal fatigue life of the brake disc is predicted, and the effects of vehicle dynamic response on the disc's thermodynamic response and fatigue life are discussed. The findings reveal that braking pressure exhibits significant fluctuations at the initial stage of disc-pad contact, which diminish as speed decreases. Large slopes induce greater vertical vibration of the brake components, with the fluctuation am-plitude of the friction radius reducing as speed decreases. The track slope influences braking duration, peak temperature, and stress levels, while the vehicle dynamic response alters the temperature and stress field distribu-tions, shifting the peak stress location. Introducing dynamic response into the thermomechanical analysis leads to increased peak temperature and stress, significantly reducing the brake disc's predicted thermal fatigue life. These results highlight the necessity of considering vehicle dynamic re-sponse in future brake disc design and evaluation.

Speakers: Jiacheng Shen, Jianyong Zuo, Yu Pan

1:30 PM Trajectory Optimization of 30,000-Ton Large-Scale Heavy-Haul Trainsets with Train-to-Train Communication

Paper ID: 24

A major challenge limiting the industrial deployment of large-scale heavy-haul trainsets such as those weighing 30,000-ton is ensuring stable maneu-verability under complex railway lines and communication conditions. This paper studies the differences in communication range and maneuvering sta-bility between combined train and train platoon in the context of existing train-to-train (T2T) communication systems. The goal is to achieve both re-liable communication and stable maneuvering in challenging environments. To this end, a general optimization model for maneuvering heavy-haul train-sets with varying configurations was developed, integrating longitudinal train dynamic and communication factors. A convexity-preserving strategy was employed to approximate real-world three-dimensional railway lines, re-sulting in a globally efficient and stable solution. The proposed model and method were validated through engineering case studies. The results indicate that currently widely deployed 400 or 800 MHz radio stations can meet the low-capacity T2T communication requirements for heavy-haul trainsets, and that train platoons are more effective than combined trains for ensuring sta-ble maneuvering of 30,000-ton trainsets.

Speakers: Shengyang Zhu, Wanming Zhai, Xiangping Wang

1:30 PM Vibration Fatigue Damage Assessment of Motor-hanging Frames Using Rigid-flex Coupling Simulation Technique for Service Motor Vehicles

Paper ID: 61

Structural reliability cannot be guaranteed without system security. As the vehicle speed promotes and the equivalent conicity increases, the enhancement of wheel spin dissipation is irreversible, which results in the intermittent or continuous chattering phenomena of traction motors and forcing the bogies to lower the yaw phase margin. Considering the design default of primary hunting in the technical prototype of German ICE3 serial bogies, the main reason for the cracking of motor hanging frames found on site can be proved to be the stronger dynamic interaction on Coupling Interface of Interest, while the residual tensile stress is only one of the accidental factors. The influencing relationship of strong correlation can thus be conditionally formulated between the spin dissipation of wheels and the strain energy of motor hanging frames, which breaks the applicability conditions of fatigue equivalent Acceleration Spectral Density for evaluating the vibration safety of on-board electrical equipments. Applying the proposed rigid-flex coupling multiaxial fatigue assessment tool, the investigation results for cracking causes of motor hanging frames gained thereby are consistent with the layered tearing conclusion of metallographic analyses. To overcome the yaw redundancy constraint existing in prototype design and associated sensitive impacts on the weld fatigue damage of motor hanging frames, the speed increased motor bogies are necessary to be changed to use the independent mode of fore and rear traction motors elastically hung on bogies, in which the fatigue life of rubber nodes depends on the geometric nonlinear influences of worn wheel-rail contact, satisfying the requirement for the 3A repair period of 165×10^4 km.

Speakers: Cheng-zhuo SUN, Jun FAN, Ming-wei PIAO, Peng DANG, Qi-bin WANG, Si-yang PIAO

1:30 PM Wheel-Flat Identification Study by means of Experimental Tests in a Railway Closed Circuit

Paper ID: 197

"Early detection of wheel-flats is essential for railway safety and cost-effective maintenance, as undetected defects accelerate the degradation of rolling stock and infrastructure. This study explores wheel-flat identification using synthetic time-domain features—Root Mean Square, Crest Factor, and Wheel Flat Index—computed from vertical axle-box acceleration signals. A full-scale experimental ampaign was carried out on a dedicated closed railway circuit, enabling controlled and repeatable testing. Key operational factors—defect size, number of flats, axle load, speed (up to 65 km/hour), and track layout—were tested at different levels to correlate each condition with specific parameter values. Results confirm the features’ sensitivity to defect severity and their robustness across various conditions, especially on tangent tracks. Controlled testing revealed that wheel-flat visibility depends on the wheel–rail contact surface, underlining the importance of careful sampling for reliable onboard
monitoring."

Speakers: Arianna Cavallo, Daniele Regazzi, Francesco Castelli-Dezza, Gisella Tomasini, Henkeled Hoxha, Steven Cervello

1:30 PM Wheel-rail Force and State Estimation Method for Active Steering Bogie based on Distributed Actuation-Sensing Configuration

Paper ID: 9

Wheel-rail force and state estimation constitute a pivotal and intricate issue within the realm of rail vehicle dynamics. To address this challenge, a method for measuring wheel-rail force and estimating bogie state employing a dual-observer architecture within a distributed actuation-sensing configura-tion, predicated on the classical structure of an active steering bogie, is proposed. Initially, the force and moment equilibrium equations of the bogie are formulated, facilitating the realization of measur-ing wheel-rail longitudinal force and wheelset lateral force based on actuator displacement and force feedback. Subsequent, a design of a Kalman filter-based bogie state estimation scheme, proposition of a method for real-time calculation of pure roll line position, and introduction of measurement re-sults pertaining to wheel-rail force. These elements collectively constitute a bogie state observer char-acterized by a dual-observer structure, enabling the estimation of the nonlinear bogie state. Findings indicate that the incorporation of wheel-rail force mitigates the state estimation error stemming from the nonlinearity of the wheel-rail contact relationship.

Speakers: Huailong Shi, Lai Wei, Shiqiao Tian

1:30 PM YOLOv8-Pose for Wheel-Rail Lateral Displacement Detection of the Small Radius Curve

Paper ID: 143

Recent advances in computer vision have promoted digital transformation of rail transit maintenance, providing a novel solution for online monitoring of wheel-rail interaction. This paper proposes an end-to-end framework to achieve real-time detection of wheel-rail relative motion, using the YOLOv8-Pose-based feature-point identification network. Validated through field full-scale testing, the trained model attains a relative error of 2.78% and an inference speed of 3ms per frame, outperforming traditional computer vision algorithm in both accuracy and speed. The lightweight variant achieves 86.9% and 95.7% accuracy under 3- and 4-pixel tolerance, respectively. Across repeated trials, the predicted relative wheel-rail displacement of 2.625mm aligns closely with the measured and post-processing result, 2.7mm, confirming the method’s precision, robustness, and potential engineering application prospects.

Speakers: Xinwen Yang, Zhiang Sun

1:30 PM → 3:00 PM Poster Session (Road)

1:30 PM A MAC-SAC Controller for the Improvement of the Path Tracking Performance of Multi-Actuators' Vehicle

Paper ID: 170

"Path tracking performance is essential for the autonomous vehicles. Deep reinforcement learning (DRL) has been utilized to deal with it for the improvement of model uncertainties and computing efficiency. However, most of DRL controllers ignore the character of the chassis’s multi-actuators,
and directly acting by the acceleration and steering, which can’t activate the cooperation ability of chassis actuators and mismatches with actual signal interface. Therefore, we proposed a novel MAC-SAC (Multiple Actuators Coordination Soft Actor-Critic) algorithm for the vehicle’s chassis equipped with multiple actuators to improve the path tracking performance, vehicle stability and robustness, utilizing the interaction data and the coordination of the actuators. To improve the policy fluctuation caused by multiple high dimension policy permutation, a staged optimization method is proposed. Finally, a MPC imitation learning method is introduced by the action probability sample strategy to lead a way to enhance the learning ability of the network and avoid the divergence during the training progress. The simulation experiments based the software CarSim demonstrate that our algorithm can reduce the vehicle’s tracking error with a little longitudinal velocity decline and improve the stability and robustness to different adhesion roads, compared with the classical DRL control algorithm."

Speakers: Chen Tang, Lu Xiong, yu zhou

1:30 PM A NOVEL VARIABLE SUSPENSION FORCE APPROACH FOR VEHICLE ROLL-RESISTANT APPLICATIONS

Paper ID: 302

"To enhance the vehicle safety a simulation investigation was conducted to analyze the dynamic stability of commercial vehicle. A vehicle rollover is a hazardous accident that destroys roll stability, especially for higher centre of gravity vehicle like SUV. A standard mechanical stability measures repeatedly fails in high-speed dynamic driving scenario due to its passive damping characteristics. This investigation focuses the need for dynamically variable damping characteristics to improvise the roll stability. The main objective is to enhance vehicle safety in terms of roll stability by minimizing the roll angle in dynamic condition. A 14 degrees of freedom (DOF) vehicle model was developed in MATLAB/Simulink software. The proposed variable damping system (VDS) is integrated with the developed vehicle model and the dynamical behaviour of model is investigated. In the MATLAB/Simulink tool, the system was simulated under double lane change (DLC) driving condition according to ISO standard 3888-2.
A VDS was created using MATLAB/Simulink tool as represented in the Figure 1. It consists of double acting actuator fixed between the wheel axle and suspension mounts on both front and rear axle. The double acting actuator of left side is directly connected to right side actuator through hydraulic control valve (HCV). The compression chamber (CC) of left side actuator is connected to A port of left side HCV. The rebound chamber (RC) of left side actuator is connected to B port of left side HCV. The T and P port of left side HCV is directly connected to T and P port of right side HCV respectively. The A and B port of right side HCV is connected to CC and RC of right side actuator. This connection is similar for both front axle and rear axle. The HCV consists of 4 ways and 3 positions. Based on the input signal the position on all HCV’s will change and the fluid flows through the pipe which produces a damping force on all the actuators. The proposed VDS has the advantage of requiring minimal energy to generate damping force, reducing lateral load transfer and thus maintains vehicle roll stability.
To assess the effectiveness of the proposed damping system, the key parameters of roll angle and normal force were observed. The roll stability measures were done by the load transfer ratio (LTR) formulae: LTR=(F_zl-F_zr)/(F_zl+F_zr ) , where F_zl denotes normal force on left side wheels, F_zr denotes normal force on right side wheels. The safe margin of LTR is between -1 and 1. If it reaches the maximum 1 or -1 denotes the vehicle wheels are not contact with the road that leads to rollover within short time. Figure 2 shows that VDS integration reduces the LTR value by up to 30%, ensuring better vehicle safety during the double lane change test.
The proposed system significantly enhances the roll stability compared to a vehicle without VDS. With this system it is observed that it additionally enhances the lateral stability and pitch stability of vehicle. Future work will focus on validating the system with dSPACE MicroAutobox III hardware-in-loop testing using adaptive control algorithms, as well as improving ride comfort along with roll stability control."

Speakers: Sathishkumar Palanisamy, Vigneshwaran P

1:30 PM Active car suspensions with multiple actuation: a method for simultaneous camber and toe control

Paper ID: 109

Active suspensions, designed according to great variety of strategies, are in-creasingly adopted in high-performance cars to improve vehicle dynamics. In the process of structure first - control next, a method has been developed for a first insight into simultaneous camber and toe control in active suspen-sions with actuation. The method, based on multibody models of suspension mechanisms and on maps of the controlled parameters (camber, toe) with re-spect to positions of the actuators, has been applied in straight running ahead and steady-state stationary cornering.

Speakers: Alberto Bodini, Alessandro De Felice, Daniele Gualdi, Marco Malagò, Silvio Sorrentino

1:30 PM Analytical Insights into Drifting of Bicycle Mode Rear-Wheel-Drive Vehicle

Paper ID: 177

"In vehicle dynamics, drifting represents a distinct state of motion characterized by
significant sideslip angles and controlled instability, requiring deep understanding of
non-linear tire-vehicle interactions for effective control. Current approaches to analyz-
ing vehicle drifting rely either on computationally intensive numerical simulations that
offer limited theoretical insight, or on oversimplified linear models that fail to capture
essential non-linear behaviors. A particular challenge lies in representing combined tire
slip conditions, yaw moment generation, and load transfer effects while maintaining
mathematical tractability. The identification of steady-state drifting conditions has
remained especially difficult due to the inherent instability of these states.
This study addresses the need for an analytically tractable yet physically accurate
method to analyze steady-state drifting dynamics. Here we show that our novel Sim-
ple Modified Elliptical Method (SMEM) enables analytical derivation of steady-state
drifting conditions while maintaining physical accuracy, through a tire model that in-
corporates saturation drop-off characteristics. The mathematical foundation reduces
the vehicle’s equations of motion to a set of non-linear coupled algebraic equations,
capturing critical aspects of longitudinal and lateral dynamics through closed-form
solutions. A hybrid numerical approach combining Genetic Algorithm with Newton-
Raphson techniques validates these analytical solutions by systematically exploring
the vehicle’s state space of longitudinal velocity, lateral velocity, and yaw rate.
Our analysis reveals that the SMEM significantly improves upon the Basic Ellipti-
cal Method (BEM) by incorporating a saturation drop-off feature, enabling accurate
prediction of tire behavior beyond the saturation point - a critical region for drifting
dynamics. As shown in Figure 1, the complete spectrum of equilibrium points for both
steering-direction and counter-steering drifting conditions demonstrates the existence
of feasible steady-state solutions across the operational envelope. These results iden-

(a) Steering-direction drifting (b) Counter-steering drifting
Figure 1.: Equilibrium points mapping for steering-direction and counter-steering drifting conditions, showing
the complete feasible region of steady-state solutions in the vehicle’s state space. The analysis reveals distinct
regions of stability characteristics for different steering strategies, providing insights into the fundamental
mechanics of vehicle drifting.
tify specific conditions where rear wheel slip dominates the drifting behavior, providing
fundamental insights into the mechanics of steady-state drifting. The analysis demon-
strates that equilibrium points exist for any given steering angle, though these states
exhibit inherent instability, with the maintenance of yaw rate emerging as a critical
factor in sustained drifting.
The SMEM’s innovative design, incorporating carefully tuned parameters to rep-
resent tire characteristics near and beyond saturation, establishes a robust analytical
framework for investigating drifting phenomena. This formulation achieves a previ-
ously unattained balance between physical accuracy and mathematical tractability,
enabling detailed analysis of the underlying mechanisms governing vehicle drifting
dynamics.
This analytical framework advances beyond existing simplified tire models while
maintaining computational efficiency, offering new perspectives on tire-vehicle inter-
actions during aggressive maneuvers. The methodology’s capacity to capture post-
saturation tire dynamics through analytical expressions provides a rigorous founda-
tion for theoretical analysis and practical applications. These contributions establish
a systematic approach to understanding and controlling drifting behavior, with impli-
cations for advanced vehicle dynamics research and the development of robust control
strategies."

Speakers: Ahmad Salahuddin Mohd Harithuddin,, Javad Zolfaghari, Nuraini Abdul Aziz, RAJA MOHD KAMIL RAJA AHMAD, Reza Jazar

1:30 PM Autonomous Vehicle Localization Using Environmental Magnetic Field with Ferromagnetic Objects on the Road

Paper ID: 306

Localization is one of the key technical factors for autonomous driving. Existing methods include GNSS (Global Navigation Satellite System) or methods using visual sensors such as cameras or LiDARs (Light Detection and Ranging). However, there are conditions where their accuracy can degrade, even when multiple methods are utilized. This paper proposes using the environmental magnetic field (EMF) for localization. EMF is a combination of the natural geomagnetic field and the magnetic field induced by man-made objects. The proposed method was validated in simulation and experiment using a test vehicle equipped with magnetic sensors. The results showed that the proposed method can accurately localize the vehicle with errors of less than 0.5 m.

Speakers: Hirotaka Mukumoto, Kazuo Urakawa, Keisuke Shimono, Kyoya Ishii, Takayuki Ando, Yoshihiro SUDA

1:30 PM Bifurcation Analysis of a Bicycle Mode Rear-Wheel-Drive Vehicle with Three State Variables and Two Control Parameters

Paper ID: 185

Understanding bifurcation behaviour is essential in vehicle dynamics, as it enables
the identification and control of critical transitions where a system’s behavior changes
drastically, often leading to instability. This knowledge plays a fundamental role in ve-
hicle handling, stability, safety, and control system design. Traditional vehicle modeling
often employs simplified two-state variable approaches, which overlook the complex
interaction between longitudinal velocity, lateral velocity, and yaw rate, particularly
in other nonlinear regimes such as those encountered in vehicles with rear-wheel-
drive configurations. This paper addresses these limitations by employing a three-state
model to analyze the interplay between these coupled dynamics, focusing explicitly on
bifurcation phenomena. The model is applied to study the dynamics of a bicycle mode
rear-wheel-drive vehicle under varying rear-wheel longitudinal slip ratio (0–0.4) and
front-wheel steering angle (0–0.5 radians). These parameters are chosen based on their
influence on stability and their direct coupling with tire forces dynamics, particularly
in nonlinear scenarios involving longitudinal and lateral forces saturation, as well as
combined-tire forces effects.
A three-state dynamic model is formulated to investigate the longitudinal velocity
(vx), lateral velocity (vy), and yaw rate (r), encapsulating the core dynamics of the
vehicle under a bicycle mode rear-wheel-drive system. The model incorporates nonlin-
ear tire-ground interactions, specifically using a combined-slip variation of Modified
Elliptical Method to capture coupled lateral and longitudinal forces with considering
yaw and neglecting of roll and pitch motion. Albeit these nonlinear tire forces are
crucial for analyzing the stability characteristics under varying rear longitudinal slip
ratios and steering inputs.
The purpose of this study is to investigate equilibrium points, stability transitions,
and corresponding bifurcation phenomena under the defined parameter ranges. To
achieve this, a hybrid numerical approach combining a Genetic Algorithm (GA) and
the Newton-Raphson method is used. The Genetic Algorithm highlights its utility in
performing a global search of the parameter space to identify initial approximations of
equilibrium points, particularly in complex and highly nonlinear regions. The refined
Newton-Raphson method is then applied for precise localization of equilibrium points,
leveraging the initial conditions obtained from the GA. Combining the strengths of
both methods allows for the robust identification of equilibrium points in systems
where traditional methods often fail due to sensitivity to initial guesses and divergence
in nonlinear domains.
Once equilibrium points are identified, the stability of these points is analyzed
through eigenvalue decomposition of the system’s Jacobian matrix. The study fo-
cuses on identifying two primary bifurcation types: saddle-node bifurcations and Hopf
bifurcations. Saddle-node bifurcations occur when two equilibrium points merge and
annihilate, typically representing a sudden loss of stability as a vehicle moves beyond
its handling limits (e.g., at higher rear-slip or steering inputs). Hopf bifurcations occur
when a stable equilibrium loses its stability through an oscillatory route, generating
periodic behavior (limit cycles), which may manifest as lateral or yaw oscillations in
the vehicle’s dynamics. These bifurcations reveal critical transitions and thresholds in
the slip-steering parameter space, including the potential for codimension-two bifur-
cations.
In this study, analysis is conducted by using numerical tools such as MATLAB to
systematically trace bifurcation loci in the parameter space. This analysis involves
iteratively perturbing the rear-slip ratio and front steering angle while simultaneously
analyzing the Jacobian matrix’s eigenvalue spectra. Continuation methods also allow
for tracking the evolution of bifurcation points across the three-dimensional parameter
space. This is particularly significant, as the results not only uncover critical bifurcation
transitions but also allow for the mapping of stability regions in a more robust and
comprehensive way than traditional bifurcation diagrams.
The results of the study are presented using three-dimensional parameter visualiza-
tions, showcasing bifurcation loci and stability by plotting along three axes of state
variables (longitudinal, lateral velocity and yaw rate), considering variation of rear lon-
gitudinal slip ratio and front wheel steering angle, enabling a detailed understanding
of stability domains. This visualization effectively determines the regions of stabil-
ity (where equilibria are locally asymptotically stable) and regions where instability
emerges due to saddle-node or Hopf bifurcations. Saddle-node bifurcations are ob-
served as sharp creases or folds in the parameter space, representing the boundaries
where abrupt destabilization occurs. These bifurcations are strongly associated with
increasing in rear longitudinal slip ratio, where nonlinear force saturation effects dom-
inate the dynamics. On the other hand, Hopf bifurcations are identified as curved or
loop-like boundaries in the parameter space, marking the onset of oscillatory insta-
bilities and the generation of limit cycles. These periodic oscillations are indicative
of lateral instabilities such as fishtailing, understeer, or yaw oscillations, which are
particularly detrimental in high-performance or emergency maneuvering scenarios.
The study finds that the interplay between rear-wheel longitudinal slip and front-
wheel steering contributes strongly to the observed bifurcation behavior. In particular,
rear-wheel slip variations cause significant changes in the lateral and longitudinal force
balance between the tires, while steering angle sharpens the coupling effects by shifting
the load distribution across the front and rear wheels. These effects create rich bifur-
cation structures, highlighting a strong dependence on nonlinear tire dynamics and
force saturation effects. At higher slip ratios, the combined-slip tire model indicates
lateral force saturation, leading to critical destabilization events through saddle-node
bifurcations.
In addition to numerical insights, the findings have important implications for ve-
hicle control and design. The detailed mapping of bifurcations and stability regions
provides engineers with critical insights into operating limits under extreme slip and
steering conditions. For example, the ability to identify Hopf loci can inform controller
designs in advanced driver assistance systems (ADAS) or autonomous vehicles, ensur-
ing that oscillatory instabilities are predicted and mitigated effectively. Similarly, the
delineation of saddle-node bifurcation boundaries can guide vehicle chassis and sus-
pension designs to ensure stability in high-slip scenarios, such as aggressive cornering
or high-speed emergencies.
This analysis builds a strong foundation for understanding and predicting stabil-
ity trends in three-state vehicle models, with applications extending to autonomous
driving systems, high-performance driving, and vehicle handling optimization. By in-
tegrating a robust numerical framework with advanced nonlinear tire modeling and
bifurcation analysis, this work significantly enhances the ability to model, visualize,
and manage stability in rear-wheel-drive vehicle configurations. It ultimately bridges
critical gaps in the existing literature by providing a comprehensive bifurcation analy-
sis in three-dimensional parameter spaces, with an emphasis on rear longitudinal slip
and steering interactions. The visualized stability boundaries and critical transitions
offer a practical tool-set for vehicle engineers and researchers to optimize stability,
safety, and driver performance under realistic and aggressive conditions."

Speakers: Ahmad Salahuddin Mohd Harithuddin, Javad Zolfaghari, Nuraini Abdul Aziz, Raja Mohd Kamil Raja Ahmad, Reza Jazar

1:30 PM Design of High-speed Autonomous Road Vehicles with Active Aerodynamic Control

Paper ID: 34

For autonomous road vehicles, the two main design criteria are: 1) safety, and 2) transportation efficiency. In terms of control actuation schemes, typi-cal active vehicle safety systems (AVSSs), such as anti-slip regulation (i.e., ABS and traction control), active steering, differential braking, and anti-roll control, have been proposed and designed to increase the safety of road ve-hicles. To improve transportation efficiency, the vehicle’s forward speed should be increased. Unfortunately, there is a trade-off between the two de-sign criteria: the higher the forward speed at which the vehicle travels to im-prove transportation efficiency, the higher the safety risk the vehicle takes. Conventional AVSSs are designed within the limitations of tire/road interac-tions. Under severe operating scenarios, e.g., high-speed cornering, high lon-gitudinal/lateral accelerations, the AVSSs are hindered by saturations of tire/road forces and moments. However, little effort has been made to in-crease the safety of autonomous vehicles under high-speed operations. This study tackles the safety problem of autonomous vehicles under high-speed operations. A design of an active aerodynamic control (AAC) scheme is pro-posed for high-speed autonomous road vehicles. The proposed design is evaluated using numerical simulations. The results derived from this study may provide a guideline for the design of autonomous road vehicles with AAC.

Speakers: Chunyu Mao, Martin Agelin-Chaab, Yuping He

1:30 PM Designs of Autonomous Multi-trailer Articulated Heavy Vehicles with Active Trailer Steering

Paper ID: 271

Multi-trailer Articulated Heavy Vehicles (MTAHVs) are cost-effective for freight transportation since the wide application of MTAHVs can significantly reduce greenhouse gas emissions, improve fuel economy, and decrease traffic congestion. It is reported that 94% of fatal road accidents are caused due to drivers’ mistakes [1]. Autonomous driving provides a promising solution for increasing road safety by removing human drivers from the control loop of road vehicles [2]. To increase the safe operation of freight transportation, in recent years, a few studies have been conducted to explore the autonomous driving techniques for articulated heavy vehicles (AHVs) [3]. It is shown that these autonomous driving techniques provide an effective means for increasing the safety of AHVs through automatedly controlling the speed and direction of the leading vehicle units, i.e., tractors [1]. AHVs and, especially, MTAHVs exhibit a unique dynamics of rearward amplification, showing that during high-speed evasive maneuvers, the trailing units exhibit larger lateral motions than the leading units [4]. Unfortunately, these existing autonomous driving techniques for AHVs ignore the active motion control for trailing units [3]. To fill the research gap, this paper proposes an innovative autonomous driving technique for MTAHVs, which integrates the automated driving control for the leading vehicle unit with the active motion control of the trailing units. A B-train double configuration has been adopted as the subject vehicle, and the required vehicle models have been generated [5]. The proposed control system is assessed by co-simulations using TruckSim and MATLAB/Simulink. In the co-simulations, the nonlinear TruckSim model is employed as the virtual MTAHV plant, the linear model based integrated controller is designed in MATLAB/Simulink, by means of the S-function as the interface, the integrated controller is combined with the virtual vehicle plant and the numerical simulations can be conducted. In this study, vehicle model validations are performed through open-loop simulations, whereas the control systems are evaluated in closed-loop simulations to examine their applicability. Active trailer steering systems are proposed to control the lateral motions of trailing units of the MTAHV. The main contribution of this study include: 1) An integrated control system is designed to regulate the autonomous driving system with ATS using a model predictive control (MPC) technique; 2) A model-based predictive motion-planning modular is introduced. The MPC-based control method automates the steering of the tractor's front axle, while the ATS is to control the lateral motions of trailers. The sophisticated trajectory planner utilizes a model-based predictive approach to customize the trajectory, enhancing lateral stability during high-speed evasive maneuvers. The simulation results demonstrate that the autonomous driving system combined with ATS efficiently regulates the MTAHV to achieve the appropriate performance. Numerical simulations show the effectiveness of the proposed autonomous MTAHV with ATS.

Speakers: Haoxiang Lang, Md Omar Faruq Joney, Yuping He

1:30 PM Development of a Vehicle Behavior Prediction Algorithm Using GAN and High-Precision Map Fusion

Paper ID: 26

In this study, we propose a novel method for predicting the behavior of target vehicles in environments equipped with high-precision maps, while requiring on-ly a limited amount of data. By leveraging vector sequences, we construct sub-graphs and apply an attention mechanism to accurately model the interactions be-tween high-precision map information and target vehicles. Furthermore, a Gener-ative Adversarial Network (GAN) is employed to enhance data efficiency. Dur-ing the data extraction phase, the primary features of the target vehicle are effec-tively encoded using vector sequences, and multilayer Graph Neural Networks (GNNs) are utilized to construct subgraphs that facilitate deep representation learning of local data characteristics. These subgraphs are then integrated into a global graph that captures global interactions, enabling more accurate trajectory prediction. Experimental results demonstrate that, despite requiring significantly less training data than previous models, the proposed approach achieves perfor-mance comparable to state-of-the-art methods, highlighting its effectiveness and efficiency in vehicle behavior prediction.

Speakers: Xianjun Li, Yonghwan Jeong

1:30 PM Development of Semi-Active Rear Axle Suspension for Pick-up Truck to Improve Ride Comfort

Paper ID: 25

This paper presents a semi-active suspension system for a pickup truck aimed at improving ride comfort in scenarios involving changes in loaded cargo. To en-sure economic feasibility, a variable damper was installed on the rear axle. A magnetorheological damper, offering a faster damping force response compared to conventional variable dampers, was employed. Consequently, a feedback con-trol strategy based on the vertical speed of the vehicle body and stroke rate was designed and compared with a full semi-active suspension system utilizing the same control methods. Accelerometers mounted on the rear left and rear right of the vehicle body and wheels were used to estimate the vehicle state. For practical implementation in an embedded system, bandwidth filters were incorporated into the state estimator design. The feedback controller was developed using a half-car model with a passive damper on the front axle and a variable damper on the rear axle. The proposed algorithm was evaluated through simulation studies conduct-ed with TruckSim and MATLAB/Simulink. Results demonstrated that the pro-posed semi-active suspension system provided improved ride comfort compared to a passive suspension and achieved performance comparable to that of a full semi-active suspension system.

Speakers: DongHyun Kim, Yonghwan Jeong

1:30 PM Driver Aggressiveness Estimation Based on Historical Driving Data and Car-Following Models

Paper ID: 132

Connected vehicles can provide historical trajectory and velocity data of sur-rounding traffic participants during maneuvers such as lane changes and car-following. This information can be used to analyze the driving styles of sur-rounding drivers and conduct further quantitative assessments, which essen-tially correspond to the evaluation of driving aggressiveness. In this study, a nonlinear estimation framework is proposed to infer driver aggressiveness based on historical driving data and a car-following model incorporating in-dividual behavioral characteristics. Firstly, the driver’s aggressiveness is es-timated by solving a nonlinear optimization problem using historical data. The calibrated model is then used to compute the residual error sequence, which is further modeled using Gaussian Process Regression to capture pre-diction uncertainty. Finally, the estimation results are validated using real-world vehicle trajectory data from the NGSIM dataset. The proposed method demonstrates promising performance for short-term prediction of driver ag-gressiveness and car-following behavior through iterative state updates.

Speakers: Huangsong Chen, Jingliang Zou, Liang Shao

1:30 PM Framework study for cooperative system to define ODD for cooperative automated driving with infrastructure

Paper ID: 308

Cooperative automated driving system with infrastructure has garnered signifi-cant attention as a potential avenue for the expeditious social deployment of Lev-el4 automated driving. Despite the demonstration of cooperative automated driv-ing systems with infrastructure in limited scenarios, there is a need for guidelines that evaluate the design of the entire cooperative automated driving system with infrastructure. Such proposals are essential for implementing the system in a wide range of regions and routes. Once the impact of the integration of diverse func-tions and the variability in the performance of each function within a cooperative system is fully definite, it will be possible to devise a cooperative system that should be implemented and to set the Operation Design Domain (ODD) for co-operative automated driving. This study presents the methodology along with the case example in a scenario where an automated bus at the real intersection crosses a cyclist when turning right from the left-hand lane.

Speakers: Hiroshi Yoshitake, Keisuke Shimono, Motoki Shino, Shunta Horisawa, Wataru Kugimiya, Yoshihiro SUDA

1:30 PM Human-Centered Chassis Dynamics Control for Distributed Drive Electric Vehicles with Experimental Evaluation

Paper ID: 51

Most chassis dynamics control (CDC) strategies focus on improving vehicle’s objective performance, such as minimizing tracking errors, while neglecting the driver's subjective driving requirements. This may cause a mismatch between the vehicle motion response and the driver's expectations, ultimately affecting vehicle dynamic performance. To address this issue, this study proposes a human-centered chassis dynamics control (HCCDC) strategy. Unlike traditional vehicle parameter-based references, this study calibrates reference variables based on the closed-loop driver feedback, thus incorporating driving preferences into the control reference. A hierarchical predictive control strategy is then designed to track the human-centered control reference, ensuring the enhancement of both vehicle handling and driving stability. Real-world experiments, including objective handling tests and subjective evaluation tests, are conducted to assess the proposed control strategy. Compared to the controller with the parameter-based reference, the objective performance metrics of HCCDC in objective handling tests are significantly improved. Moreover, the subjective evaluation scores of HCCDC in subjective evaluation tests are also higher.

Speakers: Hanghang Liu, Haobo Sun, Hong Chen, Lin Zhang

1:30 PM Impact of Time Delay on Dynamic Response of an Independent Steering System in a Corner Module of Electric Vehicles

Paper ID: 297

This paper investigates the impact of the time delay on the dynamic response of an independent steering system in a corner module of electric vehicles. The corner module integrates the independent driving, braking, steering, and suspension system. Unlike conventional steer-by-wire systems, the independent steering system in the corner module eliminates mechanical linkages between wheels at the same axle. While this enhances flexibility, it also increases susceptibility to time delay effects, especially under high-frequency or rapid steering operations. A detailed dynamic model incorporating a tire model and time-delay elements is developed to analyze its influence on system stability. Frequency-domain analysis and step-response simulations reveal that even slight time delays can amplify steering response lag, introduce oscillations, and degrade overall system performance. Additionally, a parameter sensitivity study highlights how steering motor and planetary reducer delays further exacerbate instability. These findings underscore the necessity of accounting for time delay in the design and validation of independent steering systems to ensure safe and reliable operation.

Speakers: Cheng Wang, Jun Wang, NIng Zhang, Zihong Li, hangyu lu

1:30 PM Investigating the effects of lateral load transfer on the steady-state cornering of a vehicle

Paper ID: 55

This study examines strategies to improve the handling of high centre-of-mass vehicles using a validated full-vehicle dynamics model. A constant radius test is used to evaluate the steady-state cornering and results clearly show that changes in suspension settings and ride height significantly influence the vehicle’s understeer behaviour. The tyre’s cornering coefficient also plays a role, specifically at lower lateral accelerations, but this diminishes at high lateral acceleration where lateral load transfer is the dominating factor. Reducing ride height is a particularly effective means of maintaining a neutral steer characteristic across a wide range of lateral accelerations because ride height is directly coupled to the lateral load transfer during cornering. The promising results should be further investigated for implementation in real-world vehicles to improve SUV handling.

Speakers: Bongani Zulu, Herman A. Hamersma, P. Schalk Els

1:30 PM Neural Network Vehicle Model for Real-Time Motion Control of Autonomous Vehicles

Paper ID: 20

This study presents a data-physics hybrid control framework termed neural network-aided fast iterative model predictive control (NN-FIMPC) for autonomous vehicle motion control. A neural network-based vehicle dynamics model is embedded to enhance the accuracy of nonlinear modeling under diverse driving conditions while maintaining real-time feasibility. NN-FIMPC introduces the modeling error between the neural network and a linear physical model as an explicit disturbance within the linear MPC structure. By sequentially solving two low-complexity linear MPC problems, the framework achieves improved control accuracy without violating onboard computational constraints. Extensive co-simulation experiments validate that NN-FIMPC delivers superior trajectory tracking precision and robustness against external disturbances while significantly reducing computational time compared to conventional linear and nonlinear MPC approaches.

Speakers: Bingzhao Gao, Dele Meng, Hong Chen, Hongqing Chu, Jinlong Hong

1:30 PM On Lateral Stability Control of Four-Wheel Drive Electric Vehicles Considering the Influence of the secondary yaw torque

Paper ID: 12

Existing stability control systems enhance handling performance by utilizing direct yaw moment control (DYC) through asymmetric longitudinal torque distribution. However, these systems often neglect the influence of secondary yaw torque (SYT), leading to deviations between the applied wheel torques and the required yaw moment—particularly under extreme conditions—which degrades control effectiveness. To address this limitation, this study proposes a hierarchical lateral stability control system that accounts for SYT effects. Co-simulations in Matlab/CarSim, using the double lane change (DLC) test, demonstrate that the proposed system significantly improves vehicle handling and stability on low- and medium-friction roads compared to conventional SYT-neglected approaches.

Speaker: Chih-Keng Chen

1:30 PM Optimization and Simulation Analysis of the Output Force Complementary Characteristics of Composite Electromagnetic Shock Absorber

Paper ID: 334

With the growing demands for vehicle performance, traditional active elec-tromagnetic shock absorbers encounter problems in force output, response speed, and energy consumption. This study introduces a composite design of an electromagnetic shock absorber, integrating a frameless motor and a rota-ry magnetorheological damper. Through theoretical analysis, model con-struction, and simulations, the objective is to extend the force output range to ±8000 N, and reduce energy consumption by 20%. The research results provide a theoretical basis for optimizing the shock absorber and its practical applications.

Speakers: Donghong Ning, Jing Cao, Pengfei Liu, Xiaolong Zhang

1:30 PM Real-time control strategy for extended-range electric vehicles considering cabin thermal comfort in high-temperature environments

Paper ID: 179

In order to ensure the cabin thermal comfort of Extended Range Electric Vehicles (EREVs) in low-temperature environments, the air conditioning system (AC) plays a crucial role in providing the necessary cooling capacity. However, the en-ergy consumption related to cabin cooling is enormous, affecting the overall en-ergy consumption of EREVs. To address these challenges, we propose a collabo-rative adaptive equivalent fuel consumption minimization strategy (A-ECMS) for EREVs, which includes specific cooling requirements for the cabin. In order to transform the fuel economy problem of EREV into a convex optimization prob-lem, we use polynomial fitting and variable transformation to change the differen-tial state equation. The solution to this problem was obtained using a quadratic programming based analysis method, and the Lagrange multiplier for prediction was obtained. Subsequently, we proposed a collaborative A-ECMS aimed at real-time optimization of EREV control. The simulation results show that the strategy proposed in this paper can always maintain SOC close to the target. In terms of fuel economy, the difference between the proposed A-ECMS and the globally optimized EMS based on dynamic programming is less than 2%, indicating better thermal management performance. The strategy proposed in this article has good performance in terms of fuel efficiency, thermal management performance, and real-time performance.

Speakers: Mingyao Yao, Nong Zhang, Sunan Hu, Zhengfeng Yan, Zhu Bo

1:30 PM Redundant Steering Control of a Novel Electro-hydraulic Coupled

Paper ID: 78

To improve the emergency steering performance of the electro-hydraulic coupling steering system in the event of a single subsystem failure, a new type of electro-hydraulic coupling steering system is proposed, and a coordinated control strategy including redundant steering functions is designed. The system and the control strategy are validated with simulation. The results show that the control strategy proposed in this paper reduces the mean square error of path tracking when a single power assist subsystem fails compared to the uncontrolled mode by 57.27% at 50km/h and 40.22% at 70km/h. This verified the assistance performance and redundant safety performance of the control strategy in the event of a single assistance subsystem failure

Speakers: Baohua Wang, shuang wang, shubo zhang, zhao yong

1:30 PM RTK-IMU Fusion-Based Dynamic Trajectory Generation and Coordinated Steering Control for Virtual Track Trains

Paper ID: 294

Accurate trajectory generation and precise path following control are essential for the efficient operation of Virtual Track Trains (VTTs). Poor coordination of mul-ti-wheel steering can cause excessive loads on tires and articulations, leading to uneven wear and shortened component lifespan especially for over-actuated sys-tems with statically indeterminate lateral force distributions. While theoretical frameworks for VTT control have been widely studied, experimental validation—especially under dynamic conditions—remains limited. This paper presents a new trajectory generation method utilizing RTK-IMU (Real-Time Kinematic and Iner-tial Measurement Unit) fusion, integrating real-time kinematic and inertial sensor data to enhance positioning accuracy and robustness. The resulting high-precision trajectory is used as a reference for tracking points of the VTT. Additionally, an Extended Ackermann geometry control method and a coordinated steering control method are developed to achieve path following and alleviate abnormal force distributions. The proposed approach is evaluated through both simulations and field experiments. Results demonstrate significant improvements in path following performance and force coordination, offering theoretical and practical foundations for advancing VTT control systems.

Speakers: Han Leng, Lihui Ren, Maozhenning Yang, Wei Huang, Yuanjin Ji, jun chen

1:30 PM Safety of Road Vehicles Against Crosswinds in Thunderstorm Downburst Conditions

Paper ID: 108

The safety of road vehicles can be compromised by crosswinds, particularly during severe thunderstorm conditions. Nevertheless, most studies on vehicle stability typically consider only generic synoptic wind conditions. In this work, a methodology to assess the safety of road vehicles under downburst winds is presented. The workflow includes defining the wind velocity field during the downburst, calculating the aerodynamic loads, and evaluating the dynamic response of the vehicle-driver system. Wind tunnel tests have been performed to obtain the aerodynamic coefficients. Co-Simulation with Simulink and VI-CarRealTime has been used to analyse the behaviour of the vehicle. To include the effect of human driver a driver controller has been defined and validated with the results obtained from the driving simulator. The results demonstrated that a high-sided lorry represents a serious safety risk when exposed to the intense and rapidly varying winds generated during a downburst event.

Speakers: Carlos Esteban Araya Reyes, Gisella Tomasini

1:30 PM The Opposite Lateral Force_A Key Factor in the Regular Variation of Load Influence on Tyre Cornering Stiffness

Paper ID: 339

The regular variation of tyre cornering stiffness influenced by load plays a crucial role in governing vehicle linear motion. Based on numerous test results, it has been widely acknowledged that cornering stiffness (absolute value) increases initially and then decreases with the increasing load. However, rare research is fond to explain how it happens. In this paper, this phenomenon is explained by proposing the concept of ‘opposite lateral force’. Using a finite element (FE) model, the distribution of free-rolling and sideslip lateral forces on the contact patch is analyzed. The equivalence of local carcass camber (LCC) between sideslip and camber is utilized to characterize the changes in opposite lateral force and pure-sideslip-induced lateral force with varying loads. Calculating the cornering stiffness based on the lateral force at a 1-degree slip angle, the sum of the opposite lateral force caused part K_(y_O) and pure-sideslip-induced lateral force caused part K_(yα_pure) reveals the regular variation of cornering stiffness with load. The opposite lateral force F_(y_O) is validated a key factor in the regular variation of load influence on cornering stiffness. Based on this theory, a semi-empirical model of cornering stiffness with changes in load is proposed.

Speakers: Dang Lu, Haidong Wu, Hengfeng Yin, Yandong Zhang, Yanru Suo, Yuhang Sun

3:00 PM → 3:30 PM Coffee Break

3:30 PM → 5:00 PM Rail 5: Pantograph-Catenary Systems II

Convener: Yongming Yao (China Academy of Railway Sciences, China)

3:30 PM Dynamic Performance Analysis of the Pantograph-Catenary System on Long Suspension Bridges Considering Rail Irregularities

Paper ID: 13

To meet the demands of transport development in mountainous regions, extensive efforts have been devoted to the ongoing construction of long suspension railway bridges. In the context of electrified railways, the crucial responsibility of trans-ferring electric current to the train falls upon the pantograph-catenary system. The installation of the railway catenary on the suspension bridge deck intertwines the dynamics of suspension bridges with trains, thereby directly influencing the dy-namic performance of the pantograph-catenary system and potentially degrading the current collection quality. An imperative aspect of ensuring operational safety is the thorough evaluation of the impact of train-rail-bridge interaction on the pan-tograph-catenary interaction performance. In this paper, we investigate the impact of rail irregularities as a disturbance source on the dynamic performance of the pantograph-catenary system on long suspension bridges. Dynamic models for the train-rail-bridge and pantograph-catenary systems are developed and coupled through spatial transformations. Rail irregularities are introduced into the coupled model using power spectral density (PSD) functions. The influence mechanism of rail irregularities on pantograph-catenary contact force is determined based on spectral analysis methods. Additionally, the contribution of rail irregularities at different levels to the dispersion of contact forces is analyzed.

Speakers: Xufan Wang (Southwest Jiaotong University, China), Yang Song (Southwest Jiaotong University, China), Zhigang Liu (Southwest Jiaotong University, China)

4:00 PM Extrapolating Railway Catenary Dynamics to Unobserved Operational Conditions Using Long Expressive Memory Networks

Paper ID: 348

Speakers: Alfredo Nunez Vicencio (Section of Railway Engineering, Delft University of Technology, The Netherlands), Hongrui Wang (Section of Railway Engineering, Delft University of Technology, The Netherlands), Rolf Dollevoet (Section of Railway Engineering, Delft University of Technology, The Netherlands), Taniya Kapoor (Section of Railway Engineering, Delft University of Technology, The Netherlands)

3:30 PM → 5:00 PM Rail 5: Vehicle Control I

Convener: Maksym Spiryagin (CQU, Austrilia)

3:30 PM Semi-Active Secondary Suspension with Model Predictive Control to Improve Ride Comfort

Paper ID: 236

Speaker: Alexander Posseckert (German Aerospace Center, Institute of System Dynamics and Control, Germany)

4:00 PM Observer based Control for Railway Vehicles with Independently Rotating Driven Wheels

Paper ID: 40

"Independently rotating driven wheels offer significant performance benefits, but achieving stable and effective control requires accurate estimation of lateral displacement. This paper presents a simple, tunable two-step observer that addresses this challenge, integrated with a cascaded PI-PD control system for enhanced performance. Additionally, we examine the distinct stability conditions of the front and rear wheel carriers and demonstrate how these differences affect both observer and controller design. The proposed approach is validated on a 1:5 prototype developed by the German Aerospace Center (DLR), showcasing its effectiveness in real-world applications."

Speaker: Tobias Posielek (Department of Vehicle system dynamics and control, German Aerospace Center (DLR), Germany)

4:30 PM Dynamic Trajectory Control System (DTCS): A Co-simulation Approach for Wear Reduction in Trailing Railway Bogies

Paper ID: 18

"This abstract presents a novel active steering system, the Dynamic Trajectory Control System (DTCS), aimed at minimizing wheel and rail wear in trailing railway bogies. The system utilises a co-simulation approach, integrating a detailed multibody vehicle model in SIMPACK with an advanced controller developed in Simulink. Field testing on a private track at Verkehrsbetriebe Zurich (VBZ) validates the simulation results and demonstrates the system's efficacy in real-world scenarios.
Building on our prior work, this research expands the scope of active steering technology to encompass trailing bogies and axles, a domain traditionally dominated by passive steering mechanisms. The limitations of passive steering, particularly the inherent wear and tear caused by fixed wheelset configurations, necessitate the exploration of active solutions for enhanced efficiency and sustainability."

Speakers: Sidrah lari (Traila AG, Switzerland), Simone Urbinati (Traila AG, Switzerland), Xiaotao Ren (Traila AG, Switzerland)

3:30 PM → 5:00 PM Rail 5: Vehicle Dynamics V

Convener: Qing Wu (CQU, Austrilia)

3:30 PM Numerical Simulation Model of Pneumatic Braking System for Heavy-haul Combination Trains

Paper ID: 68

The pneumatic braking system (PBS) is a critical component of heavy-haul trains, and its characteristics directly influence the operational safety of the train. To accurately simulate the characteristics of the PBS during the braking process, a PBS numerical simulation model is established in this paper. This model accounts for the airflow characteristics within the brake pipe and the motion states of the mechanical components inside the distribution valve. The brake pipe is modeled as a one-dimensional, isothermal, and isentropic pipe, with boundary conditions derived separately for the leading locomotive, central locomotives, tail locomotive, and pipe tee, and numerically solved using the finite difference method. The motion of the mechanical components inside the distribution valve is analyzed using the quasi-static method. Pressure changes in the reservoir and brake cylinder are simulated by the mass flow. The effectiveness of the proposed model is validated through field test data and bench test data. This heavy-haul combination train PBS model can be applied to combination trains of any configuration.

Speakers: Kaizhong Liu (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Weihua Zhang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Yao Luo (School of Mechanics and Aerospace Engineering, Southwest Jiaotong University, China), Zhiwei Wang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

4:00 PM Influence of Triggering Conditions for Cyclic Air Brake on the Safe Operation of Heavy-haul Combined Trains

Paper ID: 96

The cyclic air brake mode is adopted for speed regulation of heavy-haul combined trains under continuous long downhill slopes and is triggered by specific conditions. Exploring appropriate triggering conditions for brak-ing/releasing is crucial for reducing the longitudinal impulse and ensuring the safety of train operation. In this paper, a precise longitudinal dynamics model is established, and the experimental test data is utilized to verify the effectiveness of the model. Then, the principle of longitudinal impulse dur-ing the braking and releasing process is analyzed. Furthermore, different triggering conditions such as electric braking force, initial braking speed, ini-tial releasing speed, and gradient during braking/releasing have been studied for their impact on safe operation. Simulation results can provide a model and safety constraint guidelines for cyclic air brake strategies and the optimi-zation of longitudinal impulse performance.

Speakers: Pengfei Sun (School of Eletrical Engineering, Southwest Jiaotong University, China), Qingyuan Wang (School of Eletrical Engineering, Southwest Jiaotong University, China), Wei Mi (School of Eletrical Engineering, Southwest Jiaotong University, China), Xiaoyu Hu (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Xiaoyun Feng (School of Eletrical Engineering, Southwest Jiaotong University, China)

4:30 PM Analysis of Braking Force Distribution Ratio on the Dynamic Characteristics of Rack Vehicles under Large Slope Braking Conditions

Paper ID: 135

Rack vehicles are widely used in mountain rail transit due to their strong climbing ability. However, previous studies have mostly focused on the dynamic characteristics of rack vehicles during upward movement. Few studies have explored the dynamic characteristics of rack vehicles during the braking process.

Speakers: Guojun Yang (Southwest Jiaotong University, China), Zaigang Chen (Southwest Jiaotong University, China)

3:30 PM → 5:00 PM Rail 5: Vehicle Dynamics VI

Convener: Weihua Zhang (Southwest Jiaotong U, China)

3:30 PM Research on biased loading of helical gear in high-speed vehicle with inner axle-box bogie

Paper ID: 228

Curve line is the weak link in the railway lines. There are complex interaction relationships among the passing velocity of the vehicle and the planar direction, the radius, and the superelevation of the curve line, which leads to different wheel-rail dynamic interactions on both sides. When the high-speed vehicle passes through the curve line, the transfer of the wheel-rail dynamic interactions on both sides will induce the non-uniform load distribution on the tooth surface of the helical gear. To explore the load distribution characteristics on the tooth surface of helical gear transmission systems under different curve line conditions, the high-speed vehicle with inner axle-box is taken as the research object. The typical operating condition of the vehicle at a speed of 200 km/h is taken as an example, and the irregularity spectrum of ballastless track in China's high-speed railway is taken as the external excitation conditions encountered by the vehicle during actual operation. Here, the first gear drive system is used as the analysis object. The results indicate that there is obvious non-uniform load distribution phenomenon on the helical gear tooth during the vehicle passing through the curve line.

Speakers: Jieyu Ning (Southwest Jiaotong University,China), Zaigang Chen (Southwest Jiaotong University,China)

4:00 PM Software integration and experimental testing of a hybrid powertrain for an Unmanned Railway Vehicle

Paper ID: 160

Railway vehicles have historically represented one of the most promising transportation systems, but infrastructure failures (especially from intentional threats) can lead to severe consequences. To enhance surveillance on high-speed lines, this paper presents the development of an Unmanned Railway Vehicle (URV) equipped with vision systems for anomaly detection. The need for compactness led to the exclusion of overhead line power, designing the powertrain with a series-type hybrid architecture. This choice combines batteries, traction motors and a Range Extender (internal combustion engine with generator), aligning with sustainability goals. Key development activities included high and low voltage systems, pneumatic braking system and control algorithms. The core of the URV is the Vehicle Control Unit (VCU), managing all vehicle operations. A Software-in-the-Loop (SiL) test bench was used to simulate dynamics and verify control laws. Experimental validation on a roller test bench confirmed vehicle performance and software accuracy.

Speakers: Davide Tarsitano (Politecnico di Milano, Department of Mechanical Engineering, Italy), Michele Vignati (Politecnico di Milano, Department of Mechanical Engineering, Italy), Nicola Debattisti (Politecnico di Milano, Department of Mechanical Engineering, Italy)

4:30 PM Safety of High-speed EMUs Under the Braking Condition

Paper ID: 301

"When the High-speed EMUs encounters unexpected situations such as multiple natural disasters and signal faults during operation, when the low-adhesion third media such as oil stain, rain and snow exist, the adhesion control through creep rate is still limited for the improvement of braking ability. Some vehi-cles may adopt the mode of anti-skid removed, namely locking the wheelset to improve the braking ability to a certain extent. However, it will cause wheel tread or rail surface scratch, resulting in reducing the service duration. Besides, the wheel-rail contact relation will be also changed after the tread wear, which will impact dynamic responses or even running safety of the train.
In this study, the wear model of the wheel tread in the state of anti-skid is es-tablished by numerical method, and the wheel-rail contact relation is also sim-ulated. The dynamic simulation model of High-speed EMU-track system is es-tablished for the purpose of calculating the dynamic responses of the vehicle, and the running safety of the vehicles under different running conditions are also evaluated. This study provides a theoretical basis and data support for the exploration of the running safety boundary of the new generation of high-speed EMUs."

Speakers: Honglin Zu (Locomotive and Car Research Institute, China Academy of Railway Sciences, China), Jingcheng Wen (Locomotive and Car Research Institute, China Academy of Railway Sciences, China), Qing Wei (Locomotive and Car Research Institute, China Academy of Railway Sciences, China)

3:30 PM → 5:00 PM Road 5: Heavy Vehicles I

Convener: Yuping He (Ontario Tech U, Canada)

3:30 PM Loading Effects on Low-speed Motions of a Tractor-Trailer Vehicle

Paper ID: 227

The kinematic model is often selected as the control model for path-following control of reversing articulated vehicles. To some extent, the feedback component in control is designed to account for variations and uncertainties between models and actual vehicles. However, existing research also suggests that derivation can limit the feasible range for controllers. According to truck drivers, the presence of bogies and load differences are two factors that can significantly change the low-speed behaviors of articulated trucks. Therefore, a preliminary study is conducted to explore the effects of load positions and bogies on the low-speed motions of a tractor-trailer vehicle. These effects are evaluated as deviations from a kinematic model that simplifies bogies with single axles at their geometric center. Based on tests and simulations with selected maneuvers, the offsets between the zero sideslip axle positions and the kinematic model’s geometric axle positions may reach 20 % of the tractor's geometrical wheelbase or the trailer’s coupling length. The offsets are affected by vehicle velocities, load conditions, axle configurations, driving directions, and steering directions.

Speakers: Bengt Jacobson, Daniel Poveda Pi, Fredrik Bruzelius, Fredrik von Corswant, Tommi Saarikoski, Zhaohui Ge

4:00 PM An Anti-Rollover Control Strategy of Multi-Axle Vehicle Using Hybrid Data-Mechanism Driven Model

Paper ID: 72

Multi-axle vehicle are prone to significant rollover risks under extreme driving conditions due to pronounced load transfer, strong nonlinearity, and multi-degree-of-freedom coupling characteristics. To enhance vehicle dynamics modeling accuracy and rollover prevention performance, this paper proposes an anti-rollover control strategy for Multi-axle vehicle using a hybrid data-mechanism driven model. The proposed strategy consists of a hybrid driven model layer and a rollover control layer. In the model layer, a Dropout-LSTM based neural network architecture is developed to compensate for modeling errors caused by nonlinear features in the dynamics model. In the control layer, an improved rollover index is constructed based on the hybrid driven model and an anti-rollover controller is designed using a linear quadratic regulator (LQR). The effectiveness of the proposed anti-rollover control strategy is verified through co-simulation using MATLAB & TruckSim. The results demonstrate that the RMSE of roll angle and roll rate is reduced by 93.51% and 65.21% under the Fishhook maneuver, respectively and the RMSE improvements reach 95.31% and 74.66% under the steering angle step input maneuver. The strategy significantly enhances the rollover stability performance of the multi-axle vehicle.

Speakers: Bo Leng (Tongji University, China), Lu Xiong (Tongji University, China), Quan Zhou (Tongji University, China), Wei Han (Tongji University, China), Wenhai Piao (Tongji University, China), Zongyu Lv (Tongji University, China)

4:30 PM Development and Simulation of Multi-Axle Active Steering for Articulated Heavy Vehicles Using Driver-in-the-Loop Systems

Paper ID: 316

Articulated heavy vehicles (AHVs) operating under high-speed and heavy-load conditions often suffer from poor lateral stability and reduced handling performance, posing significant safety concerns. To address these issues, this study proposes a Multi-Axle Steering (MAS) control strategy that simultaneously actuates the rear axle of the tractor and the front axle of the trailer. The control architecture is developed based on Linear Time-Varying Model Predictive Control (LTV-MPC), enabling real-time coordination of steering angles under dynamic constraints. To account for human uncertainties, a Driver-in-the-Loop (DIL) system is introduced to capture actual driver steering inputs, allowing the investigation of their impact on the performance of MAS. On this basis, a Lyapunov-based stability compensator is designed to suppress disturbances caused by irregular driver operations, thus enhancing system robustness. The proposed strategy is validated on a co-simulation platform integrating TruckSim and Simulink. Simulation results under standard lane-change scenarios demonstrate that the MAS controller significantly improves lateral stability compared to baseline configurations, especially in the presence of human-induced uncertainties. This work highlights the importance of integrating real driver behavior into the control loop for the development of more reliable and adaptable active steering systems in AHVs.

Speakers: Wei Gao (School of Automotive Engineering, Hubei University of Automotive Technology, China), WenQian Hu (School of Automotive Engineering, Hubei University of Automotive Technology, China), Yuping He (Department of Automotive and Mechatronics Engineering, University of Ontario Institute of Technology, Oshawa, Canada), Zhaowen Deng (School of Intelligent Connected Vehicle, Hubei University of Automotive Technology, China)

3:30 PM → 5:00 PM Road 5: State/Param Estimation I

Convener: Basilio Lenzo (University of Padova, Padua, Italy)

3:30 PM IMU-based Road Elevation Estimation in the Presence of Road Slopes

Paper ID: 122

"This paper presents an observer design for estimating road elevations at all tire contact patches using only an IMU sensor when a vehicle is driving on a hill. IMU-based road elevation estimation is attractive for road profile measurements due to its cost-effectiveness and flexibility compared to direct-contact-based methods. However, designing an observer based solely on an IMU is challenging due to the observability of the system in the observer. Many researchers have addressed this issue by mounting additional sensors on the tires or suspensions, which increases costs, or by neglecting the tire model in the observer, which sac-rifices accuracy. The concept of virtual measurement, which assumes that the excitations caused by road elevation changes longer than the wheelbase have a minimal effect on the IMU sensor signals, solves the observability problem with-out the need for additional sensors or model simplifications. However, this ap-proach is only valid for flat roads.
The proposed observer enhances the virtual measurement-based road eleva-tion observer to be applicable on road slopes without requiring additional sen-sors or unphysical assumptions. The key idea is that longitudinal road slopes induce pitch between the front and rear tires, while lateral road slopes induce roll between the left and right tires. Therefore, the virtual measurements should be designed to represent the road elevations at the current time step based on the base tire contact patch. However, the observer estimates road elevations with a one-step delay. To address this time-delay issue, an 8-degree-of-freedom vehicle model is first introduced to represent road elevations at all tire contact patches through pitch, roll, and vertical motions. The measurement model coordinates are then calibrated to the global coordinate system to express the IMU signals as functions of the road slope. Next, the virtual measurement model is integrated with the estimated road elevations from the base tire contact patch, and the one-step delay is compensated using the kinematic chains between the estimated height changes of the sprung mass and unsprung masses. Finally, the model is completed by integrating a sensor bias model to compensate for errors resulting from sensor installation. The designed model is synthesized within the frame-work of the unknown input Kalman filter, allowing the observer to simultaneous-ly output road elevations at all tire contact patches and the road slope.
Fig. 1 shows the estimated results of road inclination and road elevations at all tire contact patches, simultaneously estimated using the proposed observer, in three driving cases on a 300 m road with slopes: the first driving involves a com-pact-sized sedan with an average speed of 56 km/h; the second is the same car with an average of 62 km/h; and the third involves a compact-sized SUV with an average of 57km/h. The ground truth is derived from signals directly measured from the ground using an inclinometer and laser sensors. The estimated results consistently have similar patterns to the ground truth for any vehicle type and speed ranges. The root mean square errors are 0.125 degrees for road slopes and 0.000876 m for road elevations, respectively. The high repeatability of the results demonstrates that the proposed observer can accurately estimate road elevations using only an IMU, even in the presence of road slopes, regardless of vehicle types and dynamics. Therefore, the proposed observer offers a cost-effective alternative to direct-contact-based road profile measurement methods."

Speakers: Dongin Kim (Changwon National University, South Korea), Hosik Choi (Changwon National University, South Korea), Juhui Gim (Changwon National University, South Korea)

4:00 PM A UKF-Informed Transformer Network for Tire Road Friction Coefficient Estimation

Paper ID: 127

Accurate estimation of the road-tire friction coefficient is critical for ensuring vehicle safety and performance. This paper proposes a hybrid approach that combines the unscented Kalman filter (UKF) with a Transformer neural network (TNN) to estimate the road friction coefficient under varying driving conditions. The hybrid model is trained using an end-to-end framework, optimizing both the Transformer Network and UKF estimator simultaneously. Simulation and real vehicle experiments are conducted to compare the proposed method with model-based UKF and data-driven GRU approaches, demonstrating substantial improvements in estimation accuracy and validating the effectiveness of the proposed method.

Speakers: Runzhe Yu (Shanghai Jiao Tong University, China), Sheng Zhao (Shanghai Jiao Tong University, China), Xiaodong Wu (Shanghai Jiao Tong University, China), Zhicheng Fu (Shanghai Jiao Tong University, China), Zhuo Kong (China National Heavy Duty Truck Group Co., Ltd, China)

4:30 PM Parameter Estimation for a Bi-articulated Autonomous-Rail Rapid Transit Vehicle

Paper ID: 357

Multi-articulated passenger vehicles are gaining popularity due to their efficiency, safety, and environmental benefits for urban public transport. Accurate parameter estimation is essential for the optimal performance and safety of these vehicles. This paper presents a study on parameter estimation (the effective rolling radius, total vehicle mass, and center of gravity height) for a bi-articulated 6-axle Autonomous-Rail Rapid Transit (ART).

Speakers: Jun Wang (Tongji University, China), Ruijun Gao (Tongji University, China), Sheng Zhou (Hunan CCRC Intelligent Transport Technology Co., Ltd, China), Timothy Gordon (Timothy Gordon, University of Lincoln, UK), Yukun Jia (Tongji University, China)

Wed, August 20

8:30 AM → 9:30 AM State of the Art III

Conveners: Changsun Ahn, Simon IWNICKI

8:30 AM Train–Track Coupled Dynamics Problems in Heavy-Haul Rail Transportation

Speakers: Liang Ling (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Sebastian Stichel (School of Engineering Sciences, KTH Royal Institute of Technology, Sweden), Wanming Zhai (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

9:30 AM → 10:30 AM Plenary Presentation

Conveners: Changsun Ahn, Simon IWNICKI

9:30 AM Assessing the Results of Simplified Non-Hertzian Contact Algorithms

Paper ID: 256

Many methods have been presented for the modelling of wheel/rail contact forces that often work by simplifying the full contact model in one way or another. Promising results are presented for each method. However, these results are restricted typically to isolated cases, and there are limited results on the comparative performance of different methods. We combine and extend two benchmarking approaches that have been presented previously, viz. the Manchester contact benchmark that focused on the embedding of wheel/rail contact in multi-body dynamics simulation, and a statistical approach of assessing tangential creep forces for Hertzian circumstances. Extensions are made for the evaluation of local quantities like contact patch shape, contact stresses, and distribution of frictional power. Results are presented comparing Kpec and Analyn for the normal problem, and comparing Fastrip and Fastsim, with parabolical and elliptical traction bounds, for the tangential problem. Test cases are identified where the original Fastrip method works out poorly, and a new algorithm ‘modified Fastrip’ is proposed that outperforms the existing Fastrip and Fastsim approaches.

Speaker: Edwin Vollebregt

10:00 AM The effect of tyre coupling on the raceline of road vehicles

Paper ID: 85

"Minimum-lap-time simulation (MLTS) has been applied over the years to analyse various aspects of road vehicles. Such problems have been addressed using different methodologies and assumptions. An approach proposed in the literature employs GG diagrams, which represent the maximum performance that can be extracted from a vehicle.
Tyres play a crucial role in vehicle performance. The coupling between longitudinal and lateral tyre forces has often been modelled using the well-known friction ellipse approach. Nevertheless, tyres can exhibit combined characteristics that deviate from an exact ellipse model.
In this work, the effect of tyre coupling on the MLTS of road vehicles is considered. A free-trajectory quasi-steady-state approach is employed to solve the MLTS. The problem is investigated considering different shapes of the GG diagram, which are associated to different tyre coupling models.
The results suggest that different tyre coupling, and thus different GG diagram shapes, call for different optimal racelines to extract the maximum performance of the vehicle. Indeed, combined areas of the GG diagram can be leveraged to obtain faster manoeuvring time."

Speakers: Matteo Massaro (Department of Industrial Engineering, University of Padova, Italy), Roberto Lot (Department of Industrial Engineering, University of Padova, Italy), Stefano Lovato (Department of Industrial Engineering, University of Padova, Italy)

10:30 AM → 11:00 AM Coffee Break

11:00 AM → 12:30 PM Rail 6: Optimization, Maintenance and Monitoring II

Convener: Björn Pålsson (Chalmers University of Technology, Sweden)

11:00 AM In-Motion Monitoring of Lateral Track Stability using Doppler LiDAR Systems

Paper ID: 269

This study introduces a non-contact, in-motion system for assessing railroad track stability using Doppler LiDAR velocity sensors. By measuring rail and tie velocities, the system enables the early detection of less-stable track segments. The system, mounted onboard on a Hyrail vehicle, accurately collects data on rail lateral velocity and vertical velocities of rails and ties, which are then processed through a multistep filtering process and a deep learning-based model to detect potential track vibration anomalies that may indicate low lateral stability. The analysis integrates multiple field data obtained onboard a Hyrail truck on a revenue-service track. Six LiDAR sensors, oriented vertically and laterally, are used to measure rail and tie velocities with high precision. The velocity data is used to assess the tie and track vibrations, through a multi-step filtering process that eliminates measurements insignificant to track vertical and lateral vibrations. The processed data is then assessed using Variational Autoencoders (VAEs), a type of Deep Learning algorithm, to encode data into a structured latent space. Then, Reconstruction Error (RE) method is used to identify track segments with excessive vibrations that can indicate conditions that are more susceptible to buckling. The identified section can be visually inspected by the maintenance of way engineers to monitor less-stable track that can lead to thermal buckling.

Speakers: Carvel Holton (Center for Vehicle Systems and Safety (CVeSS), Virginia Tech, USA), Mehdi Ahmadian (Center for Vehicle Systems and Safety (CVeSS), Virginia Tech, USA), Robert Wilson (U. S. Department of Transportation, Federal Railroad Administration (FRA), Track Division, USA), S. Morteza Mirzaei (Center for Vehicle Systems and Safety (CVeSS), Virginia Tech, USA)

11:30 AM The influence of tread modification on the wheel profile comprehensive wear of high-speed trains: characteristic and optimization

Paper ID: 340

Tread modification is an emerging wheel profile maintenance technique. Compared to wheel re-profiling, it performs incremental corrections during vehicle operation, extending service life while reducing maintenance costs. However, suboptimal operational strategies may negatively impact its effectiveness. Therefore, refining optimization methodologies for tread modification strategies is crucial. Addressing this, based on the wheel profile comprehensive wear (WPCW) prediction model incorporating tread modification, this study reveals the initial desynchronization phenomenon (IDP) and analysis its mechanism. Subsequently, a novel adaptive optimization strategy is proposed, overcoming the limitations of conventional fixed-parameter approaches. Results demonstrate that the strategy fosters more intelligent tread modification implementation, effectively balancing vehicle dynamic performance with wheel maintenance economic efficiency.

Speakers: Maoru Chi (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Shulin Liang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Wubin Cai (School of Mechanical Engineering, Southwest Jiaotong University, China), Yabo Zhou (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Yuchen Xie (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

12:00 PM TrainPulse: A Novel Approach to Predicting Squat Formation Using Forward-Facing Cameras

Paper ID: 310

"Early detection of squat defects in railway tracks is essential for proactive maintenance and safe train operations. Current methods rely on high-resolution railhead images and ultrasound data, but these require specialized equipment that is not widely available. This research introduces TrainPulse, a novel approach that uses standard low-resolution forward-facing cameras to predict squat initiation by analyzing train movement patterns.

Our initial work leverages deep neural networks trained on high-resolution images from Network Rail’s Automated Intelligent Video Review (AIVR) tool to automatically detect large squats. Additionally, we are developing machine learning models that integrate AIVR data with subsurface anomaly detection from the Rail Defect Management System (RDMS). While effective, these methods require costly, specialized sensors.

To address this limitation, TrainPulse provides a more accessible alternative by utilizing forward-facing cameras and train speed data. Although these cameras do not directly capture rail defects, they can detect subtle irregularities in train motion caused by track issues. The key research question is whether TrainPulse can predict squat formation before defects become large and visible.

To test this, we develop a machine-learning model that correlates train movement patterns with known squat defects from AIVR and RDMS data. We validate our approach through retrospective analysis—once a squat is detected in AIVR, we examine past forward-facing camera footage using TrainPulse to determine if early indicators of the defect were present.

This research highlights the potential of low-cost, widely available sensors for track defect detection, enabling earlier interventions and reducing the risk of rail failures. Beyond squat detection, TrainPulse can be adapted to identify other track issues, such as rail breaks, providing a scalable solution for railway safety and maintenance optimization."

Speakers: Artur Gower (University of Sheffield, UK), Matthew Jones (University of Sheffield, UK), Roger Lewis (University of Sheffield, UK), Tomlinson Kate (University of Sheffield, UK)

11:00 AM → 12:30 PM Rail 6: Track System III

Convener: Yoshihiro SUDA (Tokyo University of Technology, Japan)

11:00 AM Accurate measurement and modelling of load- and frequency-dependent properties of an embedded rail structure for vibration transmission analysis

Paper ID: 259

Accurate measurements and modelling of the dynamic properties of the con-tinuous rail support of an Embedded Rail Structure (ERS), provided by the elastomeric material where the rail is embedded, are presented. Laboratory tests to determine the dynamic properties of the rail support were carried out with a high-performance dynamic force-measuring platform. The dynamic properties are determined for different static preloads and dynamic loading frequencies. The experimental measurements exhibit strong dependencies on the applied load and frequency. A load-dependent three-element viscoelastic model is proposed and identified, and good agreement is achieved between the experimental and reproduced results. The identified viscoelastic model is integrated into a train-track interaction model for time-domain simulations. The force transmitted to the substructure is analysed based on the simulation outcomes, and the results are compared with those obtained by using a con-stant-value model of the rail support. Significant differences between the re-sults are observed in several frequency ranges, demonstrating the importance of accurately modelling the load- and frequency-dependent dynamic proper-ties of the rail support for an ERS.

Speakers: Andrea Collina (Department of Mechanical Engineering, Politecnico di Milano, Italy), Egidio Di Gialleonardo (Department of Mechanical Engineering, Politecnico di Milano, Italy), Qianqian Li (Institute of Sound and Vibration Research, University of Southampton, UK), Roberto Corradi (Department of Mechanical Engineering, Politecnico di Milano, Italy)

11:30 AM Low-Frequency Vibration Isolation in Metro Floating Slab Tracks Using Chiral Phononic Crystals

Paper ID: 142

Coupled compression-torsion chiral phononic crystals offer an advanced solution for low-frequency vibration isolation by exploiting inertial amplification and cre-ating broad band gaps that outperform conventional phononic designs. This study presents a chiral phononic crystal track for railway applications, which enhances low-frequency vibration suppression through the coupling of orthogonal polari-zation modes within the chiral unit cell. Mechanical modeling and simulations identify the key mechanisms behind the improved performance. The track system is integrated into a coupled vehicle-floating slab track (FST)-tunnel model, and its performance is compared with that of traditional steel spring FST systems. The results show that the chiral phononic crystal track significantly reduces vibrations in the 200 Hz frequency range, achieving a maximum insertion loss of 7.19 dB for the slab and 5.69 dB for the tunnel. Further analysis of rail and slab displace-ments, along with the wheel load reduction, indicates that the proposed track sys-tem does not compromise operational safety. This work provides a promising ap-proach to vibration control in urban rail systems and contributes valuable insights for future developments in this field.

Speakers: Shuai Qu (Tongji University, China), Wanming Zhai (Southwest Jiaotong University, State Key Laboratory of Rail Transit Vehicle System, China), Wei Ding (Southwest Jiaotong University, School of Mechanical Engineering, China), Xinwen Yang (Tongji University, China)

11:00 AM → 12:30 PM Rail 6: Vehicle Dynamics VII

Convener: Sebastian Stichel (KTH Royal Institute of Technology, Sweden)

11:00 AM Optimization of Dynamic and Wear Performance of Railway Vehicle Using Traction/Braking Torque

Paper ID: 113

The traction and braking torques applied to wheelsets significantly influence the wheel–rail creep forces, thereby affecting the dynamic behavior and wear performance of railway vehicles. The steady-state creep forces at the wheel–rail contact interface alter the gradient of dynamic creep forces, leading to reduced creep capability—meaning that a given increase in creepage results in a smaller increase in creep force. In curved track segments, sufficient creep forces are essential for steering and derailment prevention, whereas in straight segments at high speeds, excessive creep forces may induce unstable self-excited oscillations, known as hunting instability. In this study, a two-vehicle model is developed to investigate the potential for optimizing dynamic performance and wear behavior by redistributing traction and braking torques under both curved and straight track conditions. The proposed strategy effectively reduces the total wear number in curves and maintains hunting stability on straight tracks at high speed.

Speakers: Jinsong Zhou (Tongji University, Shanghai 201804, China), Zhanfei Zhang (Tongji University, Shanghai 201804, China)

11:30 AM Quantitative Detection of OOR Roughness Level of High-Speed Train Wheels Using On-site Measured Datasets of Axle Box Acceleration

Paper ID: 249

Wheel polygonal wear of high-speed trains is a common problem, resulting in some operational issues. Timely quantita-tive detection of wheel polygonal wear has become an urgent issue that needs to be addressed, which can be achieved by using axle box acceleration (ABA). The features of wheel out-of-roundness (OOR) and ABA are first analyzed, which provides a guiding for the construction of ABA datasets. Second, stationarity tests with different thresholds are per-formed on ABA signals to construct datasets with different qualities. Third, a lightweight deep learning (DL) model combining convolutional network (CNN) and attention mechanism (AM), namely AMCNN-Net, is designed to detect the roughness level of wheel OOR. Finally, the influence of the qualities of datasets on the detection accuracy is ana-lyzed, and the detection performance of the model proposed is compared to those published in the literature. The results show that the AMCNN-Net can achieve quantitative identification of the roughness level of wheel OOR under complex operating conditions.

Speakers: Cheng Zhou (Technical Center, CRRC Qingdao Sifang Co., Ltd, China), Gongquan Tao (State Key Laboratory of Rail Transit System, Southwest Jiaotong University, China), Qinglin Xie (State Key Laboratory of Rail Transit System, Southwest Jiaotong University, China), Tao Huang (State Key Laboratory of Rail Transit System, Southwest Jiaotong University, China), Wentian Xu (State Key Laboratory of Rail Transit System, Southwest Jiaotong University, China), Zefeng Wen (State Key Laboratory of Rail Transit System, Southwest Jiaotong University, China), Zhaowen Hou (Technical Center, CRRC Zhuzhou Institute Co., Ltd., China)

12:00 PM Dynamic study and analysis of a novel energy harvesting inerter damper on rail vehicle suspension

Paper ID: 318

This paper presents the influence of a novel energy harvester inerter damper on railway vehicle dynamics. The proposed inerter damper is composed of mechanical elements that can translate the bidirectional motion of suspen-sion into a unidirectional rotation of an electromagnetic generator. In such a configuration, the system could be not only to harvest the mechanical vi-bration energy from the suspension, but also works like shock absorber that mitigates the suspension vibration and improve the vehicle dynamics. A nonlinear railway vehicle model integrating the proposed energy harvester inerter damper at the primary suspension is developed in MATLAB, and simulations with analysis are presented in the paper. Results show that compared with a vehicle with traditional shock absorbers, railway vehicles with the proposed electromagnetic inerter damper could reduce accelera-tions by 5~12.3% when the equivalent inertia mass is properly designed according to the vehicle parameters, which will improve the ride comfort of the railway cars.

Speakers: Jianyong Zuo (Tongji University, China), Peng He (Tongji University, China), Wei Liao (Tongji University, China), Yu Pan (Tongji University, China)

11:00 AM → 12:30 PM Road 6: Motion Comfort II

Convener: Georgios Papaioannou (TU Delft, Netherlands)

11:00 AM A Motion Sickness-Aware Path Planning Algorithm for Autonomous Vehicles

Paper ID: 15

This study proposes a path planning algorithm to mitigate motion sickness in autonomous vehicles by integrating the 6DOF subjective vertical conflict model into a reinforcement learning (RL) framework. A head motion estima-tion model estimates passenger head movement from vehicle dynamics us-ing transfer functions to calculate motion sickness incidence (MSI). The RL agent optimizes path and speed profiles while directly considering MSI. Val-idation in CarSim showed significant MSI reduction compared to bench-marks without compromising travel efficiency. These results demonstrate that directly considering MSI in path planning is an effective approach to enhancing passenger comfort in autonomous vehicles.

Speakers: Changsun Ahn (Pusan National University, South Korea), Seongjae Mun (Pusan National University, South Korea)

11:30 AM Reinforcement Learning for Enhanced Ride Comfort: Vision-Based Speed Control over Road Obstacles

Paper ID: 346

This paper presents an integrated vision-control approach to optimise ride comfort while traversing a road obstacle using Reinforcement Learning. The learned speed controller plans the longitudinal motion of the vehicle based on an estimation of the road bump's location and dimensions using a single front-facing camera. The ride comfort is defined by an objective function that simultaneously ensures both vertical and longitudinal comfort. Exposing the Reinforcement Learning controller to noisy road bump estimations during training achieves robust control performance.

Speakers: Florian Jaumann (TU Wien, Institute of Mechanics and Mechatronics, Vienna, Austria), Florian Klinger (TU Wien, Institute of Mechanics and Mechatronics, Vienna, Austria), Gasser Elazab (TU Berlin, Computer Vision & Remote Sensing Laboratory, Germany), Johannes Edelmann (TU Wien, Institute of Mechanics and Mechatronics, Vienna, Austria), Manfred Ploechl (TU Wien, Institute of Mechanics and Mechatronics, Vienna, Austria), Michael Unterreiner (CARIAD SE, Vehicle Energy, Motion & Body, Wolfsburg, Germany), Olaf Hellwich (TU Berlin, Computer Vision & Remote Sensing Laboratory, Germany), Tobias Schuster (TU Wien, Institute of Mechanics and Mechatronics, Vienna, Austria), Torben Gräber (CARIAD SE, Vehicle Energy, Motion & Body, Wolfsburg, Germany)

11:00 AM → 12:30 PM Road 6: Stability and Control

Convener: Steffen Müller (TU Berlin)

11:00 AM NMPC-Based Stability Control without Direct Sideslip Angle Regulation

Paper ID: 336

This paper presents a new Electronic Stability Control strategy that does not rely on sideslip angle beta, neither for stability assessment nor as a control variable. While sideslip angle has been extensively employed to enhance vehicle stability, its measurement is unfeasible in standard vehicles, and estimation errors can significantly degrade control effectiveness. To address this, a nonlinear Model Predictive Control based strategy is developed, assessing stability using only measurable or robustly estimatable signals: yaw rate, lateral acceleration, and sideslip rate. The proposed strategy optimises the derivative of the corrective yaw moment, subject to physical constraints, including lateral load transfer. The controller is validated in simulation using a high-fidelity IPG CarMaker model and compared to a beta-based baseline strategy. Results from a Sine with Dwell manoeuvre show improved stability, reduced yaw rate and sideslip excursions, and higher robustness to beta estimation error. Driver-in-the-loop experiments conducted on a moving-based driving simulator further confirm real-time feasibility and effectiveness, particularly in safety-critical scenarios such as double lane change at high speed.

Speakers: Alberto Bertipaglia, Barys Shyrokau, Basilio Lenzo, Giovanni Righetti

11:30 AM Integrated vs. Uncoordinated Control for a High Centre of Mass Vehicle

Paper ID: 322

High Centre-of-Mass (CoM) vehicles are inherently susceptible to rollover, posing a significant safety risk. This study investigates whether the increased complexity of an integrated chassis control system provides a quantifiable performance benefit over traditional uncoordinated control strategies for a high CoM vehicle. A high-fidelity ADAMS model of a Land Rover Defender, modified with a semi-active suspension system, active rear-wheel steering, and differential braking with realistic actuator delays and response times, serves as the simulation platform. An integrated controller, based on a single Model Predictive Control formulation, is compared against an uncoordinated approach combining individual controllers for each subsystem. Performance is evaluated during a severe Double Lane Change manoeuvre at various speeds on both smooth and uneven road surfaces. The results demonstrate that the integrated controller significantly improves vehicle handling and stability, evidenced by reduced roll and side-slip angles, particularly on smooth roads. Furthermore, it achieves these enhancements with less control effort, effectively mitigating the subsystem conflicts apparent in the uncoordinated strategy. While the integrated controller's execution time is slightly higher (12 ms) than the combined time for the uncoordinated controllers (9 ms), both approaches operate well within real-time constraints. The findings confirm that an integrated control strategy offers substantial advantages, justifying its implementation for enhancing the safety of high CoM vehicles.

Speakers: Andries Peenze, Schalk Els

12:00 PM Modeling on Online Reconfigurable Dynamic Behavior of Multi-Axle Distributed Drive Vehicles

Paper ID: 332

Given the flexible and diverse configurations of multi-axle distributed drive vehicles, which need to adapt to various complex driving conditions, this paper explores the modeling of online reconfigurable dynamic behavior for such vehicles. First, an online reconfigurable vehicle dynamic model is established. The actuator configuration matrix is constructed to change the vehicle's actuator configuration, and the callback function is established to update the actuator configuration matrix. The callback function will be called as required to reconfigure the actuator configuration matrix online during the operation of the vehicle. Subsequently, the online reconfigurable dynamic model is validated and its potential applications are discussed via co-simulation with Matlab and Trucksim. Simulation results indicate that the model proposed in this paper is feasible for online reconfiguration of vehicle models and is also effective for addressing various driving conditions or actuator failure scenarios, providing a theoretical and model foundation for the dynamic control and performance optimization of multi-axle distributed drive vehicles in complex dynamic behaviors.

Speakers: Cong Fang, Konghui Guo, Luoqi Zhang, Xinjie Zhang, Ye Zhuang

11:00 AM → 12:30 PM Road 6: Suspension II

Convener: Davide Tavernini (University of Surrey, UK)

11:00 AM Road-Roughness-Adaptive Vehicle Vibration Control via Controllable Suspension Systems Based on Standard-Guided Reinforcement Learning

Paper ID: 125

With the electrification and intelligence of vehicles driving the chassis toward drive-by-wire evolution, traditional passive suspension technology is shifting toward controllable height, stiffness, and damping. The vibration control effect of controllable suspensions relies on adjusting damping or control forces in a continuous action space. However, the actuation of four-wheel controllable suspensions exhibits coupling characteristics on vehicle posture, and current vertical vibration control rarely considers full-vehicle models. Reward function design for learning-based methods lacks a solid foundation, and reinforcement learning (RL) suffers from poor adaptability to different road roughness due to identical training and testing datasets. In response, this paper makes key contributions: using a seven-degree-of-freedom full-vehicle model as the RL control plant, calculating control forces for each suspension considering actuator capacity; proposing a Standard-Guided Reinforcement Learning (SGRL) method for reward function design based on professional standards; and expanding the Markov Decision Process (MDP) state space to include different road roughness classes, enhancing adaptability to diverse road conditions.Results show that the expanded state space and SGRL method achieved 42.12% and 27.56% reductions in the weighted root mean square (RMS) value of acceleration compared to the baseline TD3 algorithm, respectively. Their synergistic effect further improved performance, yielding reductions of 57.05% and 45.93% in the weighted RMS acceleration value compared to the baseline TD3 and LQR algorithms.

Speakers: GuiRong Zhuo (School of Automotive Studies, Tongji University, China), Lu Xiong (School of Automotive Studies, Tongji University, China), Sizai Zhou (School of Automotive Studies, Tongji University, China), Wei Han (School of Automotive Studies, Tongji University, China)

11:30 AM Influence and Countermeasure of Vehicle Electrification on Ride Comfort

Paper ID: 317

Changes due to vehicle electrification, such as sprung mass inertia specifications, power plant suspension specifications, and powertrain rotational inertia and stiffness, influence vehicle dynamics. In this paper, the effects of vehicle electrification on various phenomena of vehicle dynamics, such as body attitude due to operational inputs, primary ride, and secondary ride, are studied, and a suspension design method, including a suspension control system, is proposed.

Speakers: MASAKI YAMAMOTO (Kanagawa Institute of Technology, Japan), MASATO ABE (Kanagawa Institute of Technology, Japan), Makoto Yamakado (Kanagawa Institute of Technology, Japan), SHINGO KOUMURA (Toyota Motor Corporation, Japan), TSUYOSHI YOSHIMI (Toyota Motor Corporation, Japan)

12:00 PM An optimal control algorithm for a novel vehicle semi-active suspension with dual-valve controlled CDC dampers

Paper ID: 333

This paper introduces a novel dual-valve electronically controlled Continuous Damping Control (CDC) semi-active suspension system. A nonlinear half-car model incorporating factors such as tire lift-off, suspension bump stops, and damper characteristics is established. A practical road surface identification algorithm is proposed to evaluate road conditions in real-time. Control strategies for the dual-valve CDC damper system are developed using a multi-objective particle swarm algorithm, with the results stored in a lookup table for real-time application. Experimental and simulation studies demonstrate that the dual-valve CDC dampers exhibit superior damping flexibility and performance compared to traditional single-valve systems under various time delays. A modified NSGA-II genetic algorithm is employed to optimize suspension conditions. Compared with passive, adaptive, and single-valve CDC dampers, the dual-valve damper significantly enhances overall vehicle suspension damping performance. Thus, the results provide valuable insights for the design of CDC dampers and the implementation of advanced semi-active suspension control algorithms.

Speakers: Fan Yu (Shanghai Jiaotong University, China), Jiannan Luo (Shanghai University, China), Shilong Xue (Shanghai University, China), Zhen Huang (Shanghai Linden Chassis-tech Co.Ltd, China)

12:30 PM → 1:30 PM Lunch

1:30 PM → 5:30 PM Technical Tours

The IAVSD 2025 conference will feature an engaging technical tour, offering participants insights into the latest advancements in vehicle system dynamics.
The tour will include visits to leading research facilities, industry partners, and demonstration sites relevant to the conference themes.

1:30 PM Rail

The Shanghai Metro is one of the world’s top metro systems, under operation of Shanghai Shentong Metro Group. The Shanghai Metro is the second longest metro system in the world by route length, at 826 Km (513 miles), second largest metro system in the world by number of stations, with 508 stations across 20 lines.
The high-speed rail (HSR) network in China is the world’s longest and most extensively used – with a total length of 45,000 km (28,000 miles) by the end of 2023. The HSR network encompasses newly built rail lines with a design speed of 200–380 km/h (120–240 mph). Shanghai High-speed Train Depot is a supporting project for Beijing - hanghai, Shanghai - Nanjing and Shanghai -Hangzhou intercity high-speed railway. It is located at the south side of Shanghai Nanxiang Marshalling Station, covering an area of about 2,000 acres. There are 32 maintenance lines, 6 temporary maintenance lines, 6 wheel profiling
lines, 156 storage lines and 15 car wash lines. It mainly undertakes the maintenance and operation tasks of the assigned EMUs from the first to the fifth level. It is the High-speed Train Depot with the largest number of assigned EMUs and the most complete vehicle models in China.

1:30 PM Road

Yangshan Port is the world’s largest automated container terminal, showcasing
a high degree of automation and effciency in cargo handling, with operations
largely controlled remotely. It’s the world’s largest automated container terminal
in terms of scale and size. The port is a leader in automated container handling,
using ”5G+L4” self-driving trucks and other technologies to improve effciency
and cargo transportation capacity. Operators can control the automated
machinery from a central location, even up to 100 kilometers away, thanks to
the advanced fiber-optic network. Container handling, loading, and unloading
are done with minimal human intervention. In 2024, the annual throughput
exceeded 26 million TEUs, surpassing the design throughput capacity of 15.6
million TEUs across its four phases and accounting for more than half of
Shanghai Port’s world-record total of 50 million TEUs. [TEU = twenty-foot
container equivalent units]

1:30 PM Road and Rail

Shanghai Ground Vehicle Wind Tunnel Center, located inside the Jiading campus of Tongji University, is the first wind tunnel for whole vehicles in China. It contains an aero dynamic wind tunnel, a thermal dynamic wind tunnel, a digital wind tunnel and a wind tunnel for models.

The Laborotory of Intelligent and Connected Driving aims HIL Testing System for Intelligent and Connected Driving. Its core architecture comprises one main vehicle simulator (test vehicle) and three traffc vehicle simulators. The algorithm under test can be deployed on the main vehicle, which is also equipped with a 6-DoF motion platform to deliver realistic driving sensations from all six directions. The three traffc vehicles can be either human-driven or controlled by autonomous driving algorithms, creating interactive scenarios for the main vehicle under test. All four vehicles operate within a unified virtual environment.

College of Transportation of Tongji University established a comprehensive Test Center& Platform for Rail & Maglev, consisting of an 800m test rail line, a test train, and a measuring system for education, training, research, and equipment testing. This platform supports core technology R&D for rail transportation and serves as a public testing hub for collaboration with institutes, universities, and industries. IRT aims to become a national test and certification center for rail transit vehicles and equipment, with plans to extend the line to 2100m, achieve a maximum speed of 120 km/h, and connect to Shanghai Metro Line 11.

Tongji University’s National Intelligent Connected Vehicle (ICV) Testing and Evaluation Base is the first national-level intelligent connected vehicle testing demonstration area authorized by the Ministry of Industry and Information Technology (MIIT) of China and located within a university. The base is strategically situated around Shanghai Jiading International Automobile City, adjacent to Jiangsu, Zhejiang, and Anhui provinces. It focuses on research, innovation, and testing services related to intelligent connected vehicles and has established comprehensive testing facilities and multifunctional sites.

The 8-DoF Driving Simulator is able to simulate 80%∼90% of realistic driving scenarios, including situations with mountain roads, acceleration, turning with small radius, etc. Load and vehicle parameters can be adjusted, and full data log can be captured for analysis and research.

6:30 PM → 9:30 PM River Cruise

Note: For attendees who will make their own way to the cruise, please assemble
at the meeting point by 18:30 at the latest. The cruise will depart at 18:45
sharp, and latecomers may miss the boat.
For those who will depart from the conference venue, please assemble at the
pick-up point at 17:30.

Thu, August 21

8:30 AM → 9:30 AM State of the Art IV

Conveners: Giampiero Mastinu, Stefano Bruni

8:30 AM Post Impact Motion Control of Road Vehicles

Speakers: Derong Yang (Volvo Car Corporation, Department of Vehicle Energy and Motion Control, G¨oteborg, Sweden), Tim Gordon (School of Engineering & Physical Sciences, University of Lincoln, Lincoln, United Kindom), Ulrich Sander (Safety Centre, Volvo Car Group, Gothenburg, Sweden)

9:30 AM → 10:30 AM Plenary Presentation

Conveners: Giampiero Mastinu, Stefano Bruni

9:30 AM Limit Cycles in Road Vehicle-and-Driver Dynamics: A Force-Based Instability Detection Approach

Paper ID: 258

The aim of the paper is to show that the dynamic behavior of road vehicles and drivers in evasive maneuvers is described by saddle-type limit cycles. The paper further explores how such limit cycles can be exploited for early detection of loss-of-control scenarios. To this end, a Degree of Stability (DoS) index is developed. The index is computed by using Floquet theory applied to the limit cycle, which is represented on the Force-Moment diagram. This representation enables straightforward scaling of the cycle with respect to the tyre-road friction coefficient, which can be estimated using force sensors. Several evasive maneuvers were performed at the dynamic driving simulator of the Politecnico di Milano to evaluate the effectiveness of the DoS index. The vehicle model used was a 14 DoF system incorporating all major nonlinearities and without any controls such as ABS, ESP TCS. The DoS index proved effective in detecting unstable motion and quantifying its severity.

Speakers: Fabio Della Rossa (Dep. of Electronics, Information Technology and Bioengineering, Politecnico diMilano, Piazza Leonardo da Vinci 32, Italy), Giampiero Mastinu (Dep. of Mechanical Engineering, Politecnico di Milano, Piazza Leonardo da Vinci32, Italy), Giorgio Previati (Dep. of Mechanical Engineering, Politecnico di Milano, Piazza Leonardo da Vinci32, Italy), Massimiliano Gobbi (Dep. of Mechanical Engineering, Politecnico di Milano, Piazza Leonardo da Vinci32, Italy), Samuele Giacintucci (Dep. of Mechanical Engineering, Politecnico di Milano, Piazza Leonardo da Vinci32, Italy)

10:00 AM Active Wheelset Steering Based on Suspension Deflection Control Considering Non-Compensated Lateral Acceleration

Paper ID: 241

Active wheelset steering can improve curving performance and can be real-ised through various strategies, e.g., radial and perfect steering strategies. One of the control strategies for perfect steering is a perfect angle of attack aiming at compensating lateral forces due to the non-compensated lateral acceleration (NLA). Radial steering control usually aims at forcing the wheelset into a radial position in curves. However, the angle of attack is difficult to obtain. The longitudinal suspension deflection, which can be measured relatively easy, is often used as a feedback signal. This study analyses a control strategy based on wheelset suspension deflection by combining the control strategy based on radial position and perfect yaw an-gle to cope with NLA. Two control approaches for NLA compensation are investigated. The first method uses a constant value of lateral creep coeffi-cient, and the second one utilises varying lateral creep coefficient. Co-simulation is used in performance evaluations with 24 running cases at var-ious NLA. The control methods for NLA compensation result in a wear number reduction of up to 49% compared to the radial steering. The NLA compensation method with varying creep coefficient mostly provides even lower wear number than the one with a constant value. Later, the effect of secondary suspension on steering performance is investigated as the per-formance of NLA compensation approaches in a tight curve is lower than with a pure radial steering control strategy. Softer secondary yaw stiffness thus provides lower wear number in almost all cases compared to radial steering.

Speakers: Carlos Casanueva (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Prapanpong Damsongsaeng (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Rickard Persson (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden), Sebastian Stichel (Department of Engineering Mechanics, KTH Royal Institute of Technology, Sweden)

10:30 AM → 11:00 AM Coffee Break

11:00 AM → 12:30 PM Rail 7: Track System IV

Convener: Liang Ling (Southwest Jiaotong U, China)

11:00 AM Development of a Test Zone for High Speed Stability Tests of Freight Cars

Paper ID: 353

In 2023 MxV Rail began work to create a test zone suitable for performing high speed stability testing of North American freight railcars. The test zone was to be located on the west tangent of the High-Speed Loop (HSL) on MxV Rail's newly built test site. The test zone was designed to meet the requirements of AAR Chapter 111,2 and 49 CFR 213.3333, with the intent that it would produce high speed stability test results that were consistent with tests conducted at Federal Railroad Administration Transportation Technology Center (TTC). Tests on the as-built track showed that a loaded and empty car were more stable than expected. Vehicle dynamics simulations were used to investigate potential modifications to the test zone. The primary variable examined as part of this study was rail profile. The simulation results showed that a 2mm overlap between a worn rail shape and the new rail profile produced the desired performance in the model. This rails of the test zone were ground to match the desired shape. Final tests on the test zone produced results within expectations.

Speakers: Michael Craft (MxV Rail, Pueblo CO 81001, USA), Nicholas Wilson (MxV Rail (Retired), Colorado Springs, CO 80903, USA), Russell Walker (MxV Rail, Pueblo CO 81001, USA), Yuqing Zheng (MxV Rail, Pueblo CO 81001, USA)

11:30 AM Study on detecting of the early signs of derailment using OCSVM with experiment data

Paper ID: 298

This study investigates the feasibility of using roll and pitch angular velocities to early detect flange climb derailments on tight curves by One-Class Support Vec-tor Machine (OCSVM). To find early signs of the derailment and develop its de-tection algorism, we needed to obtain data for the derailment. We conducted an experiment using a bogie with intentionally increased derailment coefficients. Flange climb derailments occurred twice in this experiment. We observed data which was measured by gyro sensors during the experiment and found that roll and pitch angular velocities of the bogie may indicate early signs of flange climb derailments. The early signs were identified by frequency analysis of the angular velocities using the continuous wavelet transform (CWT). Detection of the early signs was attempted by OCSVM using the wavelet coefficients. The results indicate that this approach can detect the early signs of derailment and may contribute to the development of early detection systems for the derailment.

Speakers: Jaehyeon Park (Akebono Brake Industry Co.,Ltd., Japan), Keiji Fujimoto (West Japan Railway Company, Japan), Keisuke Shimono (Institute of Industrial Science, The University of Tokyo, Japan), Kenichi Kitano (West Japan Railway Company, Japan), Kenji Ejiri (Institute of Industrial Science, The University of Tokyo, Japan), Minoru Mashiko (Akebono Brake Industry Co.,Ltd., Japan), Shih-Pin Lin (Institute of Industrial Science, The University of Tokyo, Japan), Shogo Matsuo (Graduate School of Engineering, The University of Tokyo, Japan), Takashi Arita (West Japan Railway Company, Japan), Tetsuya Kawanabe (Akebono Brake Industry Co.,Ltd., Japan), Yoshihiro SUDA (Institute of Industrial Science, The University of Tokyo, Japan)

12:00 PM Identification of the lateral track irregularities from the behavior of the train's axles on a surrogate railway dynamics model

Paper ID: 47

"Irregularities in railway tracks are critical factors that must be routinely monitored to maintain comfortable, safe, and punctual train operations. Additionally, understanding axle-rail contact forces and track irregularities is vital for analyzing the dynamic behavior and mechanical degradation of both trains and tracks. This data serves as a key input for various computational models, including those addressing rail and wheel damage, track settlement, and wheel wear.

Currently, specialized trains are used to measure and record track geometry, requiring dedicated personnel and equipment. Installing accelerometers on commercial trains and using the recorded accelerations to assess track geometry could enable more frequent and cost-effective monitoring, improving the organization and optimization of maintenance operations.

Recent studies [1-4] have explored the reconstruction of vertical track irregularities using on-board accelerometers. This involves double integration of the measured accelerations, allowing the vertical track offset of each rail to be estimated from the vertical position of the corresponding wheel. However, this method cannot be directly applied to lateral track irregularities. This is because lateral movement of the train's axles is subject to a clearance (or play) in the wheel-rail contact, enabling the axles to shift laterally between the rails. Consequently, while lateral axle position is strongly influenced by lateral irregularities, it cannot be directly inferred from them as it can for vertical irregularities. The axle's lateral dynamics between the tracks are complex, depending heavily on the train's mechanical properties and speed. Additionally, the train-track system exhibits significant non-linearity and is influenced by various uncertainties, both internal (e.g., mass, suspension characteristics) and external (e.g., wind, rain), which further complicates the analysis.

In this work, the Klingel formulas [5], which describe the natural oscillations of axles on rails, are utilized and adapted to develop a surrogate model for railway dynamics. Initially, a linear surrogate model is constructed, incorporating the train carbody and track geometry, including certain irregularities. A stability analysis is conducted to assure the model's validity. Lateral offset irregularities are identified using linear estimators, and the sensitivity to uncertain parameters is analyzed. Subsequently, system non-linearities are introduced incrementally. Their influence on the train dynamics is discussed and their influence on the reconstruction of lateral offset irregularities is systematically examined."

Speakers: Christine Funfschilling (SNCF, Direction Technologies, Innovation et Projets Groupe 1 to 9 Avenue Fran¸cois Mitterrand, 93210 Saint-Denis, France), Denis Duhamel (Laboratoire Navier, Ecole nationale des ponts et chauss´ees 6 and 8 Avenue ´ Blaise-Pascal, 77455 Champs-sur-Marne, France), Guilaume Perrin (COSYS, Universit´e Gustave Eiffel, 5 Boulevard Descartes, 77420 Champs-sur-Marne, France), Malek Chihaoui (Laboratoire Navier, Ecole nationale des ponts et chauss´ees 6 and 8 Avenue ´ Blaise-Pascal, 77455 Champs-sur-Marne, France)

11:00 AM → 12:30 PM Rail 7: Vehicle Control II

Convener: Hitoshi Tsunashima (Nihon University, Japan)

11:00 AM Robustness of Active Control Strategies for Railway Wheelset

Paper ID: 60

This paper presents a fundamental analysis on the control design for active wheelset control and studies the dependency of different control strategies on condition changes in vehicle speed and wheel-rail interface. Changes in vehicle speed are an important factor for consideration as it closely influence the dynamic properties of railway wheelset and it is not clear an active con-troller tuned at one particular speed would also best suit-ed for other speeds, whilst the robustness again variations at the wheel-rail interface is also a key design requirement. The primary focus of the study will be concerned with the stability control. Three different active control strategies (active yaw damping, active lateral damping and sky-hook stiffness) are investigated and assessed against the variations in vehicle speed and wheel-rail interactions.

Speakers: Hong Li (School of Engineering, Manchester Metropolitan University, UK), TX Mei (School of Science, Engineering and Environment, University of Salford, UK)

11:30 AM Application of an advanced railway vehicle motion simulator for the assessment of passenger comfort in curves and transitions

Paper ID: 220

Increasing linespeeds can enhance railway capacity and reduce journey times. However, this may lead to higher passenger acceleration, negatively affecting perceived ride comfort, especially in curves and transitions. The study investigates passenger perceptions on curves when exceptional design geometry limits are applied alongside varying levels of track quality. Passen-ger comfort was assessed using a combination of vehicle dynamics simula-tions and the THOMoS motion simulator, with ride comfort quantified through the PCT and PDE indices. The findings suggest that linespeeds can be increased without the need to re-design or re-lay the track, offering a cost-effective approach to improving journey times on specific routes.

Speakers: Adam Bevan (Institute of Railway Research, University of Huddersfield, UK), David Crosbee (Institute of Railway Research, University of Huddersfield, UK), Mark Burstow (Network Rail, UK), Paul Allen (Institute of Railway Research, University of Huddersfield, UK), Ruichen Wang (School of Mechanical Engineering, Shijiazhuang Tiedao University, China)

11:00 AM → 12:30 PM Rail 7: Vehicle Design

Convener: Lai Wei (Southwest Jiaotong U, China)

11:00 AM Investigating the role of bogie architecture in tramcar curve squeal

Paper ID: 107

This article uses numerical simulations to investigate the impact of bogie architecture on curve squeal in a modern articulated tramcar. The study presents the influence of potential design solutions aimed at reducing the propensity for curve squeal in a tram with Independently Rotating Wheels (IRW). First, the influence of the primary suspension stiffness and the vehicle wheelbase is analyzed on curves with different radii. The performance of the IRW tram is then compared with that of a Solid Axle (SA) configuration. The potential benefits of a steering axle solution in mitigating the phenomenon are also discussed. The analysis shows that, although adjusting the primary suspension stiffness does not effectively reduce squeal in an IRW configuration, it is still possible to optimize vehicle behavior in tight curves by slightly reducing the wheelbase. Furthermore, results indicate that a SA configuration may be more effective in mitigating the phenomenon on very tight curves. Numerical simulations also suggest that even partial axle steering is sufficient to effectively reduce wheel squeal in tight curves.

Speakers: Egidio Di Gialleonardo (Politecnico di Milano, Department of Mechanical Engineering, Italy), Federico Castellini (Politecnico di Milano, Department of Mechanical Engineering, Italy), Leonardo Faccini (Politecnico di Milano, Department of Mechanical Engineering, Italy), Roberto Corradi (Politecnico di Milano, Department of Mechanical Engineering, Italy), Stefano Alfi (Politecnico di Milano, Department of Mechanical Engineering, Italy)

11:30 AM Analysis of electromechanical coupled resonance mechanism of permanent magnet direct-drive inboard bearing bogie

Paper ID: 169

The permanent magnet direct-drive (PMDD) inboard bearing bogie is a novel lightweight bogie, where the motor’s output torque directly drives the wheelset via a hollow transmission shaft. However, the harmonic torque frequency of the motors may couple with the torsional modal frequency of the hollow shaft transmission system; therefore, it is essential to analyze the electrome-chanical coupled resonance mechanism. In this study, an electromechanical coupled dynamic model of the PMDD inboard bearing bogie was established, and the electromechanical coupled resonance mechanism of the bogie was further revealed. In addition, the control method for resonance is discussed. The results indicate that the inverter switches to square wave modulation at high speeds, resulting in an increase in the amplitude of the 6th harmonic torque. Subsequently, this harmonic torque excites the torsional mode of the hollow shaft transmission system, leading to electromechanical coupling resonance. Modal matching is crucial for mitigating electromechanical coupling resonance.

Speakers: Chen Yang (Southwest Jiaotong University, China), Guangtong Ma (Southwest Jiaotong University, China), Guanzhou Ren (Southwest Jiaotong University, China), Maoru Chi (Southwest Jiaotong University, China), Shulin Liang (Southwest Jiaotong University, China), Wubin Cai (Southwest Jiaotong University, China), XINGWEN WU (Southwest Jiaotong University, China)

12:00 PM Research on vibration and noise control of end-wall structure of high-speed EMU under aerodynamic excitation

Paper ID: 118

Abnormal vibration of the end-wall structure in high-speed trains can ad-versely affect ride quality. To address this issue, this study systematically in-vestigates the mechanism of end-wall abnormal vibration through aerody-namic simulations, line vibration and noise tests, and end-wall modal analy-sis. The findings reveal that the coupling between the 40 Hz aerodynamic excitation induced by the semi-enclosed windshield structure and the natural frequency of the end-wall structure is the root cause of abnormal vibration and noise. Based on this analysis, three vibration and noise reduction strate-gies are proposed: installation of aerodynamic deflectors, enclosing the up-per section of the external windshield (full-enclosed windshield), and usage of particle dampers. The experiments research demonstrates that all three so-lutions effectively reduce the root mean square acceleration and noise of the end wall at various speed levels. Among these, the full-enclosed windshield achieves optimal performance by reducing the power spectral density of aerodynamic loads, minimizing end-wall vibration frequency shifts, lowering fatigue stress, and extending structural lifespan. Meanwhile, particle dampers exhibit exceptional noise reduction performance specifically in the 40 Hz frequency range within the train cabin. This study provides both theoretical and practical solutions for mitigating vibration and noise issues in high-speed train end-wall structures.

Speakers: Heng Zhang (State Key Laboratory of Traction Power, Southwest Jiaotong University, China), Kaiyun Wang (State Key Laboratory of Traction Power, Southwest Jiaotong University, China), Lei Zhang (CRRC Changchun Railway Vehicle Co., Ltd., Ltd., China), Xin Ding (CRRC Changchun Railway Vehicle Co., Ltd., Ltd., China), Yizheng Yu (CRRC Changchun Railway Vehicle Co., Ltd., Ltd., China)

11:00 AM → 12:30 PM Road 7: Collision Avoidance

Convener: Gabor Orosz (UMICH, US)

11:00 AM Multi-Modal Model Predictive Path Integral Control for Collision Avoidance

Paper ID: 226

This paper proposes a novel approach to motion planning and decision-making for automated vehicles, using a multi-modal Model Predictive Path Integral control algorithm. The method samples with Sobol sequences around the prior input and incorporates analytical solutions for collision avoidance. By leveraging multiple modes, the multi-modal control algorithm explores diverse trajectories, such as manoeuvring around obstacles or stopping safely before them, mitigating the risk of sub-optimal solutions. A non-linear single-track vehicle model with a Fiala tyre serves as the prediction model, and tyre force constraints within the friction circle are enforced to ensure vehicle stability during evasive manoeuvres. The optimised steering angle and longitudinal acceleration are computed to generate a collision-free trajectory and to control the vehicle. In a high-fidelity simulation environment, we demonstrate that the proposed algorithm can successfully avoid obstacles, keeping the vehicle stable while driving a double lane change manoeuvre on high and low-friction road surfaces and occlusion scenarios with moving obstacles, outperforming a standard Model Predictive Path Integral approach.

Speakers: Alberto Bertipaglia (Delft University of Technology, The Netherlands), Barys Shyrokau (Delft University of Technology, The Netherlands), Dariu Gavrila (Delft University of Technology, The Netherlands)

11:30 AM Collision avoidance analysis of an articulated heavy vehicle in CARLA

Paper ID: 246

This paper presents the development and validation of an articulated heavy vehicle, tractor-semitrailer model in the CARLA simulation environment. The model addresses limitations in previous work related to the vehicle geometry, mass, powertrain, and steering curve to match a real-world vehicle. Validation is conducted using both low-speed experimental measurements and higher-speed data generated from a high-fidelity vehicle dynamics model. Results demonstrate good agreement in yaw dynamics, with a slight underestimation of offtracking. The model is further used for collision avoidance analysis using the two-dimensional time-to-collision measure developed in the literature. Code and implementation details are shared publicly to support reproducibility and future research.

Speakers: Abhijeet Behera (The Swedish National Road and Transport Research Institute, Sweden), Erik Frisk (Link¨oping University, Sweden), Sogol Kharrazi (The Swedish National Road and Transport Research Institute, Sweden)

12:00 PM An Autonomous Collision Avoidance Function for Articulated Commercial Vehicles Based on An Improved Artificial Potential Field

Paper ID: 299

The design of active safety system for the articulated commercial vehicles (ACVs) presents significant challenges due to the articulated configuration. This study investigates autonomous collision avoidance function for ACVs. The off-tacking phenomenon is analyzed to determine the safe space for avoiding collision with obstacle. An improved artificial potential field method is proposed for path planning, where potential functions are used to model the traffic scenario. The safe space is applied in the design of obstacle repulsive potential field. A model predictive controller, based on the dynamics model, is designed to track the planned path. Co-simulation results demonstrate that the proposed path planning and tracking methods enable safe collision avoidance for ACVs, while achieving a root mean square tracking error of 0.096m.

Speakers: Cheng Wang (School of Engineering and Physical Sciences, Heriot-Watt University, UK), Jun Wang (School of Mechanical Engineering, Southeast University, China), NIng Zhang (School of Mechanical Engineering, Southeast University, China), Qianchen Zhang (School of Mechanical Engineering, Southeast University, China)

11:00 AM → 12:30 PM Road 7: Heavy Vehicles II

Convener: Mehdi Ahmadian (Virginia Tech US)

11:00 AM Lateral Performance Enhancement in A-Double Vehicle Combinations Through Active Dolly Steering

Paper ID: 156

This study investigates the performance enhancement of long combination vehicles (LCVs) through the use of a steerable dolly unit. The focus is on improving lateral dynamics with particular attention to lateral off-tracking, swept path width, and rearward amplification. To enhance maneuverability and stability, these performance indicators are incorporated into scenario-specific cost functions within the proposed optimal control framework. An A-double combination vehicle, equipped with a dolly with steerable axles, is used to evaluate the approach. Simulation results demonstrate significant reduction in lateral off-tracking, swept path width, and rearward amplification across relevant driving scenarios, indicating the effectiveness of the proposed control strategies for dolly steering to improve the lateral performance of LCVs.

Speakers: Kartik Shingade, Maliheh Sadeghi Kati, Mats Jonasson, Shreekara Ramesh, Umur Erdinc

11:30 AM Trailer Reversing Made Easy with Steer-by-Wire

Paper ID: 245

"Reversing a trailer is a complex and challenging task, even for experienced drivers, due to the non-minimum-phase property of car-trailer kinematics and the inherently unstable nature of trailer reversal. These challenges require drivers to steer in the opposite direction to initiate a turn and continuously stabilize the trailer's motion to maintain the desired angle. This paper explores the potential of steer-by-wire technology to simplify trailer reversal. Steer-by-wire systems replace the mechanical connection between the steering wheel and road wheels with an electrical connection, enabling new control strategies.

In this study, the steering wheel is repurposed to directly control the angle between the car and the trailer, decoupling the driver’s input from the complexities of car-trailer kinematics. This approach allows drivers to focus solely on maneuvering the trailer without compensating for unstable dynamics. The system integrates steer-by-wire technology, sensors, and software to create an intuitive and effective trailer backup assist function.

Building on prior work in automatic trailer reversal control, this paper introduces a simplified algorithm and demonstrates its implementation on a new vehicle platform. Experimental results highlight the system's capability to enhance safety and usability, making trailer reversal accessible and stress-free for a wider range of drivers."

Speakers: Alice Finne (Sentient, Sweden), Eric Dahl (Sentient, Sweden), Erik Johansson (Sentient, Sweden), Fredrik Bruzelius (Chalmers University of Technology, Sweden), Hongchao Yu (Volvo Cars, China), Jakob Roempke (Volvo Cars, Swenden), Matthijs Klomp (Volvo Cars, Swenden)

12:00 PM Learning Efficiency Maps of Electric Trailers for Power Loss Minimization

Paper ID: 352

Optimizing energy consumption in over-actuated electric vehicles is essential, and this is often achieved by power loss minimization. Electric heavy vehicle combinations, consisting of an electric tractor or electric truck and one or more electric trailers, are even more over-actuated due to multiple driven units, and their energy consumption can be optimized in a similar way. A key challenge in such combinations is that the efficiency maps of the trailers may not be available to the tractor's controller. This study proposes a method for learning the trailers' efficiency maps and performing control allocation based on power loss minimization. The results show that it is possible to accurately learn the efficiency maps of the trailers while still preserving energy efficiency.

Speakers: Esteban Gelso (Volvo Group tractors Technology, Sweden), Junyan Tian (Chalmers University of Technology, Sweden), Maliheh Sadeghi Kati (Chalmers University of Technology, Sweden), Osama Al Sheikh Ali (Chalmers University of Technology, Sweden), Umur Erdinc (Chalmers University of Technology, Sweden)

11:00 AM → 12:30 PM Road 7: Suspension III

Convener: Jiannan Luo (Shanghai U, China)

11:00 AM A Tuning Framework for Vehicle Suspension Design Accounting for Pitch and Roll Motions

Paper ID: 192

This paper proposes a tuning for adjusting the suspension stiffness of a multipurpose agricultural vehicle (MPAV) equipped with a hydropneumatic suspension system. These vehicles are very versatile and can adapt to varying operating conditions by doubling their fully loaded weight and traveling both on roads and off-road. A 7-degree-of-freedom (DOF) model is developed to accurately capture pitch and roll dynamics alongside heave motion, with road irregularities modeled according to ISO 8608:2016 standards. By adjusting the air volume in the nitrogen chamber of the suspension actuators, the equivalent vertical stiffness can be varied to optimize key comfort indices: Road Holding (RH), Suspension Working Space (WS), and Discomfort Index (DI). A numerical framework is presented to identify optimal front and rear suspension stiffness values for varying load cases, ensuring that vehicle natural frequencies and damping ratios remain within acceptable ranges.

Speakers: Edoardo Sabbioni (Department of Mechanical Engineering, Politecnico di Milano, Italy), Mattia Belloni (Department of Mechanical Engineering, Politecnico di Milano, Italy), Michele Vignati (Department of Mechanical Engineering, Politecnico di Milano, Italy)

11:30 AM The Degradation of Suspension Damping: A Laboratory and Simulative Investigation

Paper ID: 212

The effects of degraded shock absorbers on the safety-critical driving dynamics of passenger cars have not yet been fully researched. This paper presents a validated suspension model capable of simulating the vertical dynamics of passenger cars with degraded shock absorbers using the phase-shift method. For this purpose, several identical front axle shock absorbers of a test vehicle were brought into different degraded states, measured on a test rig and a semi-physical phenomenological model of the degraded twin-tube shock absorber was parameterized. Test rig experiments were conducted using the phase-shift method with a test vehicle and different shock absorber variations. It was demonstrated that the correct representation of the non-linear dynamic properties of the stationary tire is necessary for the simulation of test rig experiments using the phase-shift method. Furthermore, it was found that the oil level of a shock absorber is strongly non-linearly related to the occurring wheel load fluctuations.

Speakers: Günther Prokop (University of Technology Dresden, Germany), Tobias Schramm (University of Technology Dresden, Germany)

12:00 PM A Review of System Design and Methodology of Controlled Suspension for Road Vehicle

Paper ID: 325

This paper aims to provide insights into key technologies and recent advancements in controllable suspension systems, which are crucial in improving vehicle performance. It reviews the core technologies within variable damping, variable stiffness, Inertial elements, variable suspension geometry, and coordination control of controllable suspension. Operational principles of active suspension components are introduced, and priority research directions and hot issues on structure and control design in existing approaches are reviewed. This state-of-the-art review discusses the latest design methodologies and their application in controllable suspension systems, which synthesize the design of suspension mechanical components, vehicle coordinated control, full life-cycle management and the coupling and coordination among subsystems from a system-level perspective and points the way toward the future of controllable suspension design development and vehicle dynamics control research.

Speakers: Gengrui Jin (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Konghui Guo (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Luhang Wang (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Mehdi Ahmadian (Center for Vehicle Systems and Safety, Virginia Tech, USA), Xinjie Zhang (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Yang Liu (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China), Ye Zhuang (National Key Laboratory of Automotive Chassis Integration and Bionics, Jilin University, China)

12:30 PM → 1:30 PM Lunch

1:30 PM → 3:00 PM Rail 8: Track System V

Convener: Akira Matsumoto (Nihon University, Japan)

1:30 PM Stability Analysis of Continuous Welded Rails on Long-Span High-Speed Railway Bridges under Temperature Gradient Effects

Paper ID: 175

The dynamic behavior of vehicles on long-span bridges, influenced by thermally induced long-wavelength irregularities in Continuous Welded Rails (CWR), is critical for ensuring operational safety and structural efficiency. This study inves-tigates the steady-state deformation of CWR under thermal effects and its subse-quent impact on the dynamic response of the vehicle-bridge coupled system. An integrated rigid-flexible coupling finite element model of the train-track-bridge system was developed using ANSYS and Universal Mechanism (UM) for multi-physics co-simulation. The geometric deformation characteristics of the track-bridge system under varying thermal conditions were quantified using ANSYS, and thermally induced track irregularities were incorporated into the UM dynamic model. Parametric simulations were conducted to elucidate how thermal effects influence the dynamic response mechanism of the vehicle-cable-stayed bridge coupled system. Key findings include: (1) CWR at beam-end regions is highly sensitive to uniform seasonal temperature variations; (2) Asynchronous thermal evolution between bridge components and running rails generates short-wavelength irregularities, significantly increasing wheel unloading rates; (3) Hor-izontal thermal gradients caused by oblique solar radiation induce curvature dis-tortion in mid-span CWR segments, which may pose a derailment hazard.

Speakers: Jijun wei (Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongii University, China), Shuai Qu (Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongii University, China), Xinwen Yang (Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongii University, China)

2:00 PM Origin of train-induced vibrations at insulated joint: contributions of track irregularity and stiffness variations

Paper ID: 279

Insulated rail joints (IRJs) are a major source of vibration nuisance along railway lines. This study investigates the excitation mechanisms at an IRJ and their contributions to train-induced vibrations through comprehensive field measurements, focusing on vibration generation and transmission near the source, i.e., at the rail and sleeper. Two types of measurements were conducted before and after the replacement of a degraded IRJ: falling weight (FW) impact tests and train pass-by measurements. Three excitation mechanisms caused by wheel passing, track irregularity and track supporting stiffness were identified. The results show that wheel passing frequencies dominate the vibration response near the source. After replacing the IRJ, improved track geometry significantly reduces vibrations at both the rail and sleeper. Furthermore, increased support stiffness reduces the transmission of vibrations from the rail to sleeper. However, whether these reductions near the source translate to lower vibration levels in the far field remains a question. The presented findings serve as a first step towards characterizing vibration transmission and wave propagation across the full rail-sleeper-ballast-embankment-subsoil chain, ultimately informing more effective mitigation strategies.

Speakers: Chen Shen (Secton of Railway Engineering, Delft University of Technology, The Netherlands), You Wu (Secton of Railway Engineering, Delft University of Technology, The Netherlands), Zili Li (Secton of Railway Engineering, Delft University of Technology, The Netherlands)

2:30 PM Research on short-wave impact of turnouts based on empirical mode decomposition and reconstruction of axle box acceleration

Paper ID: 204

The operational states of high-speed turnouts directly affect their service performance. Excessive short-wave irregularity may lead to deteriorated ride quality and abnormal wheel-rail impacts. As most vehicle wheel-rail impact vibration signals are non-stationary and nonlinear, this study proposes the Hilbert-Huang transform (HHT) to analyze train vibrations in turnouts. Complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is applied to decompose the axlebox vibration data.The processed data undergoes detrending through correlation coefficient analysis and denoising through an improved wavelet threshold method before signal reconstruction, effectively suppressing noise interference while preserving critical impact features. Based on this approach, the study further reveals the unique frequency characteristics of impact vibrations in turnout point rails,welded joints and insulated joints. The analysis specifically evaluates the variation of impact intensity over the 300 to 380 km/h speed range for these critical components, establishing an effective method for analyzing turnout impact mechanisms and providing theoretical support for condition monitoring and maintenance of high-speed turnouts.

Speakers: Hangyuan Qin (Infrastructure Inspection Research Institute, China Academy of Railway Sciences Corporation Limited, China), Tian Yang (Infrastructure Inspection Research Institute, China Academy of Railway Sciences Corporation Limited, China)

1:30 PM → 3:00 PM Rail 8: Vehicle Dynamics VIII

Convener: TX Mei (University of Salford, UK)

1:30 PM Theoretical investigation on coupler buckling instability of high-speed trains based on collision dynamics

Paper ID: 105

With the continuous improvement of the train speed and the deterioration of the train service environment, the coupler is prone to local plastic deformation leading to buckling instability, which induces collision accidents such as derailment and vehicle overturning. Consequently, the establishment of a train collision dynamics model considering coupler buckling instability can provide theoretical guidance for the train passive safety. To begin with, in accordance with the coupler refined finite element model with material failure, the deformation characteristics of the coupler instability process were examined, and the simplified mechanical model of the coupler was constructed. On this foundation, the theoretical investigation of the crush tube and drawbar stability was carried out, the obtained theoretical models were coupled by the connecting structures, the prediction model of the coupler instability process was achieved, and the accuracy of the model was verified by comparing with the finite element results. Subsequently, the model was incorporated into the train three-dimensional collision dynamics model by coordinate transformation, while the nonlinear factors of the dynamic differential equation were decoupled to acquire an efficient solution algorithm. Furthermore, on the basis of the formed train collision dynamics method considering the coupler instability, the dynamic behavior changes and connections of the body and coupler under various collision speed levels were explored.

Speakers: Binlin Wang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Haoxu Ding (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Jingke Zhang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Shoune Xiao (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Tao Zhu (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

2:00 PM A multi-layer progressive analysis method for collision energy flow in rail trains

Paper ID: 64

The huge impact kinetic energy cannot be quickly dissipated by the energy-absorbing structure and transferred to the other vehicle through the car body structure, which will cause structural damage and threaten the lives of the occupants. This study proposes a multi-layer progressive analysis method of energy flow during train collisions, considering the characteristics of train. In this method, the train collision system is divided into conversion, dissipation, and transfer layers to analyze the energy conversion, dissipation and transfer characteristics. The decay rate of the collision energy along the direction of train operation reaches 79%. The proposed method helps to understand the train collision energy flow law and provides theoretical support for the train crashworthiness design in the future.

Speakers: Jingke Zhang (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Tao Zhu (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

2:30 PM Adaptability Study of a Semi-active Anti-rolling Device for Active Tilting Trains

Paper ID: 128

The Active Tilting System (ATS) enables tilting trains to operate at higher speeds on low-speed rail lines. However, conventional anti-rolling device may interfere with the ATS, leading to increased energy consumption and diminished tilting performance. To enhance the energy efficiency and fault tolerance of the air-spring-based ATS, this paper proposes an improved Semi-active Anti-rolling Device (SAD). Firstly, the parametric model of the SAD is developed in AMESim, based on the mathematical model. Secondly, the model accuracy and SAD’s roll stiffness characteristics are validated through bench testing and simulation, respectively. Finally, the impact of the SAD on the dynamics of a high-speed active tilting train is compared with that of the Anti-roll Bar (ARB) through multi-scenario adaptability simula-tions. The SAD enables the removal or reduction of additional roll stiffness typically provided by the ARB during ATS actuation. Furthermore, the SAD provides multi-level, non-constant roll stiffness, ensuring stable train opera-tion even in the event of ATS and/or SAD failures. The proposed SAD of-fers a feasible anti-roll scheme, reducing energy consumption and improving line adaptability of high-speed active tilting trains.

Speakers: Liang Luo (Hefei University of Technology, China), Mingxing Liu (Hefei University of Technology, China), Minyi Zheng (Hefei University of Technology, China), Nong Zhang (Tongji University, China), Weimin Zhong (Hefei University of Technology, China), Zhengfeng Yan (Hefei University of Technology, China)

1:30 PM → 3:00 PM Rail 8: Wheel Wear

Convener: Dao Gong (Tongji U, China)

1:30 PM Prediction and optimization of wheel asymmetrical wear under traction/ braking conditions

Paper ID: 33

Due to terrain and urban construction constraints, tracks often include curves and rail vehicles need to apply traction/braking torque to pass through such tracks. The traction/braking adversely affect the curve passing performance of the rail vehicle, causing more severe wheel wear problems such as wheel flange wear. In this paper, based on the scenario of rail vehicle passing through the curved track under traction/braking conditions, a dynamics simulation model is established in SIMPACK and the wheel wear profile of 100,000 km is estimated by using the Archard method. The optimization strategy of applying a larger traction/braking torque to the rear wheel under the appropriate total torque requirement can effectively reduce the wheel flange wear. This strategy can provide a reference for the study of wheel asymmetrical wear and wheel flange wear on curved track.

Speakers: Jinsong Zhou (College of Transportation, Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongji University, China), Yiyang Song (College of Transportation, Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongji University, China), Zegen Wang (College of Transportation, Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongji University, China), Zhanfei Zhang (College of Transportation, Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety, Tongji University, China)

2:00 PM The mechanism of high-order wheel polygonal wear of two types of metro trains on the same line

Paper ID: 300

Trains from two manufacturers with different bogie structures run on a certain metro line in China. Both types of trains exhibit significant wheel polygonal wear, which leads to considera-ble vibration and noise problems, thereby substantially diminishing ride comfort. To investigate the mechanism of wheel polygon wear, a large number of field tests were carried out on the two trains, including the measurements of wheel out-of-roundness, vehicle vibration, track fre-quency response, and track condition. The results show that the dominant orders of wheel po-lygonal wear for trains A and B were 12–15 and 14–16, respectively. Through comprehensive analysis, it was determined that the primary contributor to wheel polygonal wear is the first-order bending vibration of the wheelset, with the P2 resonance of the wheel-rail system identi-fied as a secondary factor. Notably, the first-order bending vibration frequency of the wheelset for train B is higher than that for train A, resulting in higher-order polygonal wear. Additional-ly, rail weld irregularities, rail corrugation on small radius curves, and the residual wheel polyg-onal wear after re-profiling were the potential sources of excitation for the natural vibrations of wheelsets.

Speakers: Gongquan Tao (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Pengfei Liu (State Key Laboratory of Mechanical Behavior and System Safety of Traffic Engineering Structures, Shijiazhuang Tiedao University, China), Wei Li (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Xiaoxuan Yang (State Key Laboratory of Mechanical Behavior and System Safety of Traffic Engineering Structures, Shijiazhuang Tiedao University, China), Xuesong Jin (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China), Zefeng Wen (State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, China)

2:30 PM A Method for Suppressing the Development of Wheel Polygonal Wear Considering Electromechanical Coupling in Traction Drive Systems

Paper ID: 36

Wheel polygonal wear (WPW) is commonly observed in heavy-haul locomo-tives, posing significant threats to operational safety and drastically reducing wheel service life. Most existing methods for suppressing WPW focus pri-marily on mechanical vibrations, overlooking the electromechanical coupling (EC) interactions of traction drive systems. As a result, these methods fail to comprehensively and fundamentally curb the progression of WPW. To ef-fectively mitigate this drawback, this study develops an improved field-oriented control (IFOC) method based on current compensation, considering the EC characteristics of traction drive systems. Subsequently, Subsequently, a WPW evolution model was developed, incorporating the EC model with IFOC and the wheel wear function, to evaluate the effectiveness of the pro-posed strategy in suppressing additional harmonics and the progression of WPW. The findings demonstrate that the presented method effectively sup-presses the harmonic currents and torques in traction drive systems caused by WPW, thereby significantly slowing down the evolution of WPW. This study provides a new perspective for the alleviation of WPW and holds promising potential for engineering applications.

Speakers: Bing Lu (School of Electrical Engineering, Southwest Jiaotong University, China), Zhigang Liu (School of Electrical Engineering, Southwest Jiaotong University, China)

1:30 PM → 3:00 PM Road 8: Heavy Vehicles III

Convener: Umur Erdinc (Chalmers Tech, Sweden)

1:30 PM Investigating the matrix regularities governing the state-space equations of lateral dynamics in long combination vehicles

Paper ID: 157

"Dynamic modeling plays a critical role in the stability control of Long Combination Vehicles (LCVs). However, accurately capturing the complex lateral dynamics of LCVs remains a significant challenge due to the presence of multiple articulations between vehicle units. In particular, solving for the articulation forces is often computationally intensive and analytically complex.
This paper presents a structured approach to derive the state-space matrices governing the lateral dynamic model of LCVs directly from their design parameters. Two generalized methods for formulating articulation forces are proposed: the Calculating from Front to Rear (CFTR) method and the Calculating from Rear to Front (CRTF) method. By integrating one of these general force expressions into the system’s differential equations, we identify consistent regularities in the resulting state-space matrices.
The proposed framework simplifies model construction and enhances its scalability to multi-unit vehicle systems. The findings offer valuable insights for model-based controller design, system optimization, and linear stability analysis of LCVs."

Speakers: Bo Cheng (School of Vehicle and Mobility, Tsinghua University, China), Heqian Wang (School of Vehicle and Mobility, Tsinghua University, China), SHUO WANG (School of Vehicle and Mobility, Tsinghua University, China), Shixian Huang (School of Vehicle and Mobility, Tsinghua University, China), Wenjun Wang (School of Vehicle and Mobility, Tsinghua University, China)

2:00 PM Mechanism information guided data-driven discrepancy modeling of articulated vehicles

Paper ID: 250

This paper proposes a dynamic modeling approach for articulated vehicles, integrating mechanistic models with data-driven methods to overcome challenges arising from the system's complexity, strong coupling, and high nonlinearity. Traditional mechanistic models are limited in accuracy due to the system's complexity and cognitive constraints, while data-driven approaches often lack interpretability and generalization. The proposed method utilizes data-driven techniques to correct errors in mechanistic models, compensating for discrepancies with real-world data, while ensuring interpretability by incorporating physical constraints. Experimental validation using simulation and real vehicle data demonstrates that this hybrid approach effectively improves model accuracy without sacrificing interpretability, offering a promising solution for dynamic modeling in articulated vehicles. In addition, this paper illustrates the improvement of model accuracy on control effectiveness through control experiments in a simulation environment.

Speakers: Chao Liang (Department of Control Science and Engineering, Tongji University, China), Jun Wang (Department of Control Science and Engineering, Tongji University, China)

2:30 PM Dual-Mode Hydraulic Interconnected Articulation System for Virtual Rail Train

Paper ID: 264

The Virtual Rail Train (VRT), a multi-section articulated vehicle, suffers from yaw instability during high-speed operation. To address this, a Hydraulic Interconnected Articulation System (HIAS) is proposed to provide dynamic articulation stiffness, which conventional dampers cannot achieve. A planar dynamic model of a three-car, six-axle VRT is developed and linearized to reveal inherent instability and the critical role of stiffness in yaw stabilization. A dual-mode hydraulic articulation system is designed: in stiffness mode, it resists yaw motion at high speeds; in complementary mode, it reduces output torque for improved curve negotiation. The system’s flow-pressure behavior is analyzed, and a coupled mechanical–hydraulic model is built. Stability is verified via complex eigenvalue and frequency response analysis. A co-simulation framework combining AMESim, Simulink, and Simpack validates the system under varied operating conditions. Results show that the VRT maintains yaw stability across a wide speed range, and the system ensures curve-passing performance with output forces below 3.5 kN.

Speakers: Jun Chen (College of Transportation, Tongji University, China), Lihui Ren (College of Transportation, Tongji University, China), Yuanjin Ji (College of Transportation, Tongji University, China)

1:30 PM → 3:00 PM Road 8: Motion Control

Convener: Johannes Edelmann (TUWien, Austria)

1:30 PM On the minimum-time-to-drift manoeuvre of cars

Paper ID: 87

Drifting is an unstable steady-state cornering motion with large rear tire slip angles, that are generally combined with counter-steering. Skilled drivers typically take advantage of drifting during corners in low grip conditions and/or sharp turns. Controlled drifting can also have practical implications in road safety, in particular in emergency scenarios such as accident avoidance. The typical manoeuvre consists of 3 phases: preparation, where the vehicle reaches a normal cornering motion, transient, where the driver applies control actions to drive the vehicle into the drifting condition, and drift, where the driver applies control actions to maintain such unstable condition. In this work, the minimum-time optimization of the transient phase is considered under an optimal-control framework and compared against experimental findings. Such transient is the least investigated phase, since most of the time the focus is on the characteristics and stabilisation of the sustained drifting condition. Nevertheless, this phase is especially relevant if the final goal is an autonomous system that takes advantage of drifting in relevant scenarios. The results of the investigation suggest that the human driver triggers the drift with a minimum-time-to-drift strategy. Ultimately this unveils a method to objectively design controllers that trigger drift on demand.

Speakers: Basilio Lenzo (University of Padova, Italy), Giovanni Righetti (University of Padova, Italy), Matteo Massaro (University of Padova, Italy), Roberto Lot (University of Padova, Italy), Stefano Lovato (University of Padova, Italy)

2:00 PM Towards Agile Autonomy: An Optimal Planning and Control framework for Agile Autonomous Vehicles

Paper ID: 120

Agile autonomous vehicles are breaking through conservative stable boundaries, enabling a broader spectrum of maneuvers to enhance agility. To extend current research, this study proposes a hierarchical planning and control framework within the OCP paradigm for executing swift and aggressive maneuvers. The optimal path planning for agile maneuvers is formulated based on motion agility quantification and the feasibility analysis of active drifting. Leveraging the optimized trajectory and state references, a hybrid MPC controller is developed, employing a piecewise affine (PWA) approximation strategy to strategically simplify the coupled nonlinear dynamics. Extensive simulations and experimental road tests validate the feasibility of the proposed agile path planning and demonstrate the precision and robustness of the predictive motion control.

Speakers: Caixin Zhang (Beijing Institute of Technology, China), Jianping Gao (Henan University of Science and Technology, China), Jinfa Hu (Beijing Institute of Technology, China), Xiaokai Chen (Beijing Institute of Technology, China), Yongyuan Liang (Beijing Institute of Technology, China)

2:30 PM HL4IT - Interaction of automated vehicles and human drivers during take-over

Paper ID: 267

This study examines human-machine interaction during Take-Over (TO) manoeuvres in Level 4 (L4) Autonomous Vehicles (AVs). It is part of the project “Human interaction with Level 4 AVs in an Italian environment - HL4IT,” which aims to define the Operational Design Domains (ODDs) representative of the Italian context. A preliminary experiment was conducted with eight participants at the Dynamic Driving simulator of Politecnico di Milano. Drivers operated a L4 AV in an urban scenario, incorporating a real roundabout located in Milan, Italy. Due to ongoing roadworks, the L4 AV exited its ODD, prompting to the driver a Take-Over Request (TOR). The grip force was recorded during the simulation using a proprietary Instrumented Steering Wheel (ISW). Furthermore, drivers filled out the Differential Emotion Scale (DES) psychological questionnaire before and after driving. Grip force mean and standard deviation increased during the TO manoeuvre, indicating raised stress levels. However, DES scores showed that, although stressful, the TO manoeuvre did not elicit drivers' negative emotions.

Speakers: Andrea Fossati (Universit`a Vita-Salute San Raffaele, Italy), Andrea Galbiati (Universit`a Vita-Salute San Raffaele, Italy), Antonella Somma (Universit`a Vita-Salute San Raffaele, Italy), Edoardo Sabbioni (Politecnico di Milano, Italy), Giorgio Previati (Politecnico di Milano, Italy), Leda Mastinu (Universit`a Vita-Salute San Raffaele, Italy), Linda Boscaro (Universit`a Vita-Salute San Raffaele, Italy), Lorenzo Uccello (Politecnico di Milano, Italy), Luca Subitoni (Politecnico di Milano, Italy), Marco Ponti (Politecnico di Milano, Italy), Maria Gabriella Signorini (Politecnico di Milano, Italy), Massimiliano Gobbi (Politecnico di Milano, Italy), Veronica De Guglielmo (Politecnico di Milano, Italy)

3:00 PM → 3:30 PM Coffee Break

3:30 PM → 5:00 PM Rail 9: Vehicle Dynamics IX

Convener: Yu Pan (Tongji U, China)

3:30 PM Wheel Profile Optimization: A Hybrid Approach Combining Dynamic Simulations, Evolutionary Algorithms, and Wear Modeling

Paper ID: 324

The growing demands of modern rail transport generated by increased axle loads and operational speeds have intensified wheel-rail interaction damage such as wear, rolling contact fatigue (RCF), and track geometry degradation. These interrelated damage mechanisms form a self-reinforcing cycle that can compromise safety, increase maintenance costs, and disrupt operations. A key strategy to control these damages is the optimisation of wheel and rail profiles to improve contact conditions. Over the years, a variety of optimisation techniques have emerged, ranging from target-based geometric function methods to advanced evolutionary algorithms. While genetic algorithms and other stochastic algorithms have shown promising results in handling the complex, multi-objective nature of wheel-rail profile design, the influence of algorithm selection on optimisation performance remains underexplored. This study investigates the comparative performance of two prominent stochastic optimisation algorithms, Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Multi-Objective Particle Swarm Optimisation (MOPSO), in optimising the wheel profile of a Swedish passenger train.

Speakers: Carlos Casanueva (KTH Royal Institute of Technology, Sweden), Elham Khorazmad (KTH Royal Institute of Technology, Sweden), Saeed Hossein-Nia (National Research Council (NRC), Canada), Sebastian Stichel (KTH Royal Institute of Technology, Sweden)

4:00 PM A fault diagnosis method for axlebox bearings of high-speed trains based on multi-source vibration signal fusion and transfer learning

Paper ID: 223

Axle-box bearing is a key component of high-speed trains. To address the issue of limited diagnostic performance of single-source signals and the poor cross-case performance of deep learning models, a fault diagnosis method based on the fusion of multi-source vibration signals and transfer learning is proposed in this paper. Firstly, to enhance the feature representation capability of multi-source vibration signals, a multi-channel parallel convolutional neural network based on continuous wavelet transform (CWT-MPCNN) is constructed for feature extraction and fusion of the axle-box bearing. Secondly, to solve the small-sample learning problem in cross-condition scenarios, transfer learning is introduced into the fault diagnosis process, resulting in a CWT-MPCNN based on transfer learning. Finally, the effectiveness of the proposed method is verified by using the fault datasets of axle-box bearing of high-speed train from Southwest Jiaotong University. Experimental results show that the proposed method provides more reliable and accurate diagnosis compared to other methods.

Speakers: Dongli Song (Southwest Jiaotong University, China), Xiao Xu (Southwest Jiaotong University, China), Zejun Zheng (Southwest Jiaotong University, China), Zifan Wang (Southwest Jiaotong University, China)

4:30 PM An investigation of wear and RCF damage mechanism on switchback turnouts of an urban intercity railway line through train-turnout interaction analysis

Paper ID: 359

"Railway crossings are essential components in the railway track to provide trains to pass from one track to another, yet also vulnerable elements due to the gap between wing rail and nose rail. The wheel-rail interaction in the railway crossings is complicated, which makes it more difficult to maintain good condition than the ordinary track[1]-[3].
The Hangzhou-Haining intercity railway, which started operation in 2021, has been plagued by severe wear and rolling contact fatigue (RCF) problems on its switchback turnouts. Compared to other metro lines with higher usage density, such significant damage should not have occurred on Hangzhou-Haining intercity railway, especially after merely three years of operation. Multiple turnout sections on this line, such as those at Linping South High - speed Railway Station, Zhejiang University International Campus Station, have experienced severe wear and RCF damage. Typical types of damage include the wear of switch rails, spalling on curved guard rails and corrugation on stock rail. Current measures to address these problems are mainly focused on enhanced maintenance, such as increasing the frequency of grinding, but the root causes of the damage remain unclear.

Figure 1 Spalling on the curved guard rail
Figure 2 Corragation on the stock rail
This research aims to determine the underlying root causes of such abnormal turnout damages through field measurements and numerical simulation, to identify the key factors which might lead to such damage and propose practical measures. The investigation is structured into three main parts. Firstly, field test of the dynamic interaction of turnouts when trains pass through will be performed. By measuring from the tip of the switch rail to the crossing nose, the vertical and lateral accelerations of the rails are measured simultaneously, providing basic data for subsequent research. Secondly, the impact of different working conditions on the dynamic response characteristics of turnouts will be studied. A vehicle - turnout multibody-system (MBS) dynamics analysis model will be established, by which the key influencing factors of vehicle - turnout interaction will be identified. The MBS model will be validated by the in-situ monitoring results of the turnouts. Finally, optimization schemes will be developed based on the above-mentioned analysis. These schemes may involve adjusting the traveling speed of passing trains, modifying the stiffness of the track structure, or combining both factors. Ultimately, if feasible, the optimization measures will be implemented and verified through field tests.

Figure 3 monitoring points on turnout for wheel-rail interaction
This investigation not only offers in - depth understanding of the root causes specific to the Hangzhou - Haining intercity railway but also presents a methodology for analyzing similar problems. This methodology can be applied to metro lines with comparable traveling speeds and track conditions, thereby contributing to extending the service life of railway turnouts."

Speakers: Lizuo Xin (Zhejiang Rail Transit Operation Management Croup CO., LTD. China), Xiangming Liu (China Academy of Railway Sciences Corporation Limited, Beijing, China)

3:30 PM → 5:00 PM Rail 9: Vehicle Dynamics X

Convener: Yuanjin Ji (Tongji U, China)

3:30 PM Research on the mass parameters identification of the complete carbody of railway vehicle based on the frequency response function mass line method

Paper ID: 82

The inertial parameters of the complete carbody of railway vehicle are the basic parameters for dynamic analysis and calculation. Since the complete carbody of railway vehicle is a symmetric structure with large aspect ratio, large mass and volume, it is difficult to establish a complete CAD model ac-curately, so the accurate identification of its inertial parameters has always been a difficult problem in engineering. Firstly, the article takes the complete carbody of railway vehicle as the research object, the mass and the gravity position in the horizontal plane are accurately measured by using four-point weighing; Secondly, a specially designed and customized low stiffness air spring is used as the support, and the modal test is carried out by using a force hammer and a shaker respectively, and the mass line method of the frequency response function is applied to identify the inertial parameters of the carbody; Considering the engineering application, the results are finally compared with those based on a six-degree-of-freedom excitation platform to analyze the reasons for the differences in the mass line method. The re-sults show that it is feasible to identify the gravity position and inertial pa-rameters of the complete carbody of railway vehicle by the mass line method of the modal test.

Speaker: HUIWEN PANG (CRRC Changchun Germany RailTech GmbH, Germany)

4:00 PM Co-simulation of lateral dynamic response of a high-speed EMU under unsteady aerodynamic loads using panel aerodynamics

Paper ID: 39

As the speed of high-speed trains continue to increase, the resulting aerodynamic forces are expected to have an increasing influence on its running behavior. Though researchers have investigated the impact of unsteady aerodynamic loading on the open-air running dynamics of high-speed trains in recent years, these studies employed a one-sided coupling approach, where aerodynamic forces are computed in isolation and applied as known loads to the multibody model of the train. This neglects aerodynamic forces induced by the vehicle’s unsteady body motion, the impact of which is investigated in this paper via the rapid co-simulation of the running dynamics of a Chinese high-speed EMU with panel aerodynamics. The results demonstrate that while the two-way interaction tends to suppress the amplitude of existing lateral vibrations on the head car, it tends to amplify existing lateral vibrations of the tail car, thereby influencing the ride comfort of the vehicle. For operation on a tangent track with track irregularities, the coupled simulation resulted in a 5% decrease in the lateral component comfort index of the head car and 3-4% increase in that of the tail car, as evaluated per EN12299.

Speakers: Gang Chen (CRRC Changchun Germany RailTech GmbH, Germany), Guozhen Jing (CRRC Changchun Germany RailTech GmbH, Germany), Hauke Schmidt (CRRC Changchun Germany RailTech GmbH, Germany)

4:30 PM Modelling of High Frequency Vibration of High-speed Railway Bogie Frame

Paper ID: 304

The high frequency vibration of bogie frame arising from the wheel/rail inter-action has become a main concern for railway operators due to its highly ad-verse influences on the structural integrity. This paper thus investigated on the methodologies for modelling of high frequency vibration of a high-speed rail-way bogie frame. To eliminate the influences caused by uncertainties of wheel/rail contact and track model as well as track irregularities, a random vi-bration model of bogie frame, neglecting both the wheel/rail contact model and track, was developed to reproduce the axle box acceleration based on the measurement obtained from the field tests. The key factors affecting the mod-elling accuracy were discussed. The results showed that the proposed random vibration model can effectively reproduce the vibration level of axle box based on the given vibration spectrum. The hydraulic damper was identified as a key component affecting the high frequency vibration of bogie. Therefore, the in-fluences of nonlinearity of hydraulic damper on the high frequency of bogie frame were further studied through laboratory tests, which further resulted in a neural network-based model for hydraulic damper. The nonlinearity of hydrau-lic damper was studied through a vibration shaker up to 150 Hz. The experi-mental results indicated that the vertical hydraulic damper expressed strong nonlinearity with the variation of excitation frequency, and a relatively high force transmissibility was observed when the resonance of hydraulic damper occurred near 60 Hz.

Speakers: Liangcheng Dai (Southwest Jiaotong University, China), Maoru Chi (Southwest Jiaotong University, China), Ningrui Yang (Southwest Jiaotong University, China), Shulin Liang (Southwest Jiaotong University, China), Wei Wang (Southwest Jiaotong University, China), Xingwen Wu (Southwest Jiaotong University, China), Yanyuan Ye (Southwest Jiaotong University, China), Yunhua Huang (Southwest Jiaotong University, China), Zefeng Wen (Southwest Jiaotong University, China), Zhenxian Zhang (CRRC Qingdao Sifang Co., Ltd., China)

3:30 PM → 4:30 PM Road 9: Motion Comfort III

Convener: Georgios Papaioannou (TU Delft, Netherlands)

3:30 PM Exploring the influence of motion sickness frequency sensitivity in trajectory planning for autonomous driving

Paper ID: 76

Motion sickness (MS) is a key challenge for the acceptance and comfort of autonomous vehicles (AVs), particularly as occupants are not driving and may engage in tasks like reading. This study investigates how personalised frequency weighting filters, which reflect individual sensitivity to MS, influence trajectory planning in AVs. To model individual variability in MS, second-order transfer functions are used to modify an approximated group-level filter based on ISO 2631 by adjusting its natural frequency and damping ratio. Trajectory planning aimed at minimising motion sickness is formulated as an optimal control problem (OCP) that incorporates different filters to evaluate the impact of filter parameter variations on planned vehicle speed, lateral position relative to the road centreline, and cornering behaviour. Simulations are conducted on a closed test track under a predefined travel time using different lateral filter settings. Results show that even small variations in filter parameters (MS sensitivity) notably affect the planned speed and path. These effects are also visible in planned vehicle accelerations and dynamic behaviour, as illustrated by G-G diagrams. The findings highlight the need to consider individual differences in MS sensitivity rather than relying solely on group-level filters that may be insufficient for delivering comfortable autonomous driving. To conclude, this work introduces a novel approach to personalised trajectory planning while also encouraging a rethinking of vehicle dynamics and control systems design based on individual MS variability.

Speakers: Ilhan Yunus (Volvo Cars Corporation, Gothenburg, Sweden), Jenny Jerrelind (University of Padova, Via Venezia 1, 35131, Italy), Lars Drugge (University of Padova, Via Venezia 1, 35131, Italy), Matteo Massaro (KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden), Stefano Lovato (KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden)

4:00 PM Exploring motion comfort and time optimal trajectory planning in autonomous vehicles through over-actuation

Paper ID: 265

The rapid advancement of autonomous vehicle technology has opened new opportunities for enhancing passenger comfort through trajectory planning and over-actuation. In the present study, the influence of these technologies on motion comfort is examined by formulating the trajectory planning task in an optimal control framework based on a double-track vehicle model that incorporates load-transfer effects and torque-vectoring actuation. The comfort–time trade-off is quantified by sweeping the comfort coefficient, and torque-vectoring and non-torque-vectoring configurations are compared in a representative country-road scenario. Results show that (i) comfort-oriented settings produce noticeably smoother trajectories, whereas time-oriented settings rely on aggressive driving styles to reduce travel duration; (ii) increasing the comfort coefficient consistently lowers the discomfort level while prolonging travel time; and (iii) torque vectoring simultaneously improves motion comfort and shortens travel time during highly dynamic manoeuvres, while its influence on comfort is negligible at low speed. These findings illustrate that over-actuation can reconcile passenger comfort with journey time objectives.

Speakers: Annika Stensson Trigell (KTH Royal Institute of Technology, Stockholm, Sweden), Derong Yang (Volvo Car Group, SE-405 31 Gothenburg, Sweden), Ilhan Yunus (KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden), Jenny Jerrelind (KTH Royal Institute of Technology, Stockholm, Sweden), Lars Drugge (KTH Royal Institute of Technology, Stockholm, Sweden), Mikael Nybacka (KTH Royal Institute of Technology, Stockholm, Sweden), Wenliang Zhang (KTH Royal Institute of Technology, Stockholm, Sweden)

3:30 PM → 4:30 PM Road 9: State/Param Estimation II

Convener: Matteo Massaro (University of Padova, Italy)

3:30 PM State Estimation of Articulated Passenger Vehicles in GPS-denied Environment

Paper ID: 356

Road bounded Articulated Passenger Vehicles (APVs) are emerging solutions for populating urban transportation systems. These vehicles are characterized by their flexibility, efficiency, and cost-effectiveness. The articulation coupling and over-length introduce challenges in terms of path tracking as well as stability control, while reliable and accurate state estimation is essential for both tasks. Though GPS (often in RTK configuration) is widely adopted, it is susceptible to signal loss in urban canyons, tunnels, and other GPS-denied environments, hindering safety and stability. To address this issue, we propose a state estimation algorithm that leverages onboard sensors to estimate states of the APVs, bi-articulated with 6 steerable axles.

Speakers: Jun Wang (Tongji University, China), Ruijun Gao (Tongji University, China), Sheng Zhou (Hunan CCRC Intelligent Transport Technology Co., Ltd, China), Timothy Gordon (University of Lincoln, UK), Yukun Jia (Tongji University, China)

4:00 PM State estimation and characterization of longitudinal vehicle dynamics at low speeds using experimental data

Paper ID: 292

This study investigates low-speed vehicle dynamics by combining experimental IMU (Muse 221e) data with a minimal model. Improving passenger comfort during low-speed driving—especially during start and stop maneuvers—requires a deeper understanding of vehicle dynamics, which is often overlooked. To address this, a real vehicle test has been conducted on an inclined road, where body acceleration and rotational wheel speed were recorded using IMU sensors. The measurements were processed using signal processing, followed by parameter estimation with a transient method to estimate wheel-suspension longitudinal stiffness and damping. The estimated suspension parameters were integrated into a minimal vehicle model to compute the total wheel torque. A Kalman filter was applied to this setup to enable state estimation and phase-space analysis, providing insight into vehicle dynamics and ride comfort during low-speed driving. This study analyzes jerk during the low-speed start and stop phases, highlighting moments of high discomfort. In addition, phase portraits provide insight into the dynamic response of the system, revealing how relative motion evolves as the vehicle approaches standstill. These findings contribute to the development of control strategies that aim to improve comfort in longitudinal vehicle movement.

Speakers: Mats Jonasson (Chalmers University of Technology, Sweden), Petri T. Piiroinen (Chalmers University of Technology, Sweden), Samira Deylaghian (Chalmers University of Technology, Sweden)

4:30 PM → 5:00 PM Road Discussion

5:00 PM → 5:40 PM Shuttle bus to Banquet hotel

Note: Shuttle bus will be provided from the conference venue to the banquet hotel.

6:30 PM → 9:00 PM Banquet

The Symposium Banquet is a highlight of the social program, offering a
memorable evening of fine dining and entertainment. Join us for a delightful
meal and celebrate the spirit of collaboration and innovation in vehicle system
dynamics.

Note: There will be shuttle-bus service from the conference venue to the
banquet venue, please assemble at the pick-up point at 17:30.

Fri, August 22

8:30 AM → 9:30 AM State of the Art V

Conveners: Mehdi Ahmadian, Sebastian Stichel

8:30 AM Recent Developments on PantographOverhead Line Interaction

Speakers: Alan Facchinetti (Politecnico di Milano, Italy), Giuseppe Bucca (Politecnico di Milano, ltaly), Joao Pombo (University of Huddersfield, UK), Santiago Gregori (Universitat Politècnica de València, Spain), Stefano Bruni (Politecnico di Milano, ltaly)

9:30 AM → 10:30 AM Plenary Presentation

Conveners: Mehdi Ahmadian, Sebastian Stichel

9:30 AM Study on the Ablation Characteristics of Contact Wire under Different Pantograph-Catenary Dynamic Separation Behaviors

Paper ID: 10

The frequent occurrence of offline arcs significantly affects the service life of the pantograph-catenary (PC) system. Compared to previous studies, this paper primarily focuses on the ablation characteristics of the contact wire caused by transient arcing during a single complete dynamic separation process of the PC system. First, based on a nonlinear finite element model of the catenary and a reduced three-degree-of-freedom mass model of the pantograph, the dynamic equations of the PC system were established. Subsequently, by accounting for heat loss due to material evaporation, an ablation model for transient arcing effects on the contact wire is developed and the validity is verified using experimental data. Next, the transient arcing temperature characteristics during a single complete dynamic separation behavior were obtained through simulation calculations. Finally, the ablation characteristics of the contact wire under different dynamic separation behaviors are revealed. The results indicate that the ablation area of the contact wire does not increase with the separation time, but is also related to the maximum separation distance of the PC system.

Speakers: Hao Liu (Southwest Jiaotong University, China), Jien Chen (Southwest Jiaotong University, China), Long Chen (Southwest Jiaotong University, China), Xiaokang Wang (Southwest Jiaotong University, China), Zhigang Liu (Southwest Jiaotong University, China)

10:00 AM A Unified Co-pilot Architecture for Integrated Dynamics Control of X-by-Wire Chassis

Paper ID: 326

The future of driving is likely to involve multiple driving modes, ranging from manual driving to full autonomy, and will require high levels of technology support, potentially including full X-by-Wire chassis technology and different levels of actuator capabilities, such as independently driven wheels and actively steered front and rear axles. Control integration can lead to highly capable driver support and vehicle automation, provided core methodologies are suitably chosen and rigorously implemented. And in place of multiple separately designed functions, a rigorously designed `unified copilot' should be developed, in this paper we propose an example of a unified co-pilot architecture for integrated dynamic control of a modern X-by-Wire chassis, and test its abilities via simulation and track testing.

Speakers: Jun Wang (Tongji University, China), Qingwei Liu (SAIC Z-ONE Technology Co., Ltd, China), Timothy Gordon (University of Lincoln, UK)

10:30 AM → 11:00 AM Coffee Break

11:00 AM → 12:30 PM Closing Ceremony

Note: Theres no parallel sessions on Friday. Delegates are free to leave after the closing ceremony. Lunch-box will be provided for all participants, while you can also enjoy lunch at the canteen on campus.

12:30 PM → 1:30 PM Lunch