Description
Despite the high quality of available traffic data, there is a significant gap in the understanding of heavy vehicle deceleration and the frequency of braking events involving these large vehicles. This paper presents a practical and scalable approach based on the use of smartphones to collect braking event data. Dedicated Android applications were developed to collect essential data, including accelerations, angular velocities (gyroscope), and GPS information. To optimize file sizes, acceleration data, sampled at a rate of 10 Hz, was only recorded when a threshold of 2 m/s² was exceeded. A parallel strategy was used for the gyroscope data. The geographic data, sampled at a rate of 1 Hz, provided invaluable insight into the location and causes of braking events. To complete the analysis, classification algorithms were used to estimate braking frequency rates on different road categories. Over approximately nine months, 22 heavy-duty vehicles of different types were equipped with smartphones, and their operation was monitored along the Swiss road infrastructure. The comprehensive dataset of collected braking events provides an in-depth understanding of braking behavior as well as a good basis to improve the braking forces estimation. Information such as deceleration profiles and braking rates (braking events per km driven) were therefore used to improve the braking force model developed in research project AGB 2011/003. The preliminary simulation results based on the updated model are presented in this paper.
