ABSTRACT: Calibration of traditional car-following models usually applies to drivers’ average parameters. This method cannot address the difference between individual driving behaviors. The purpose of this study is to analyze the consistency of drivers and the heterogeneity across drivers during car-following. A total of 2,580 car-following events were extracted from the Shanghai Naturalistic Driving Study. The Intelligent Driver Model (IDM) for car-following was found to have more stable performance among different drivers and handling highly dynamic situations. IDM as used to calibrate each event for all drivers. The calibrated IDM parameters include desired speed, time gap, max acceleration, comfortable deceleration, beta, and the gap at standstill. To reflect the effect of diverse traffic conditions, car following periods were categorized into high-speed groups (HSG) and low-speed groups (LSG). Two-way Analysis of Variance was selected to evaluate intra and inter driver variability. The findings of this study show that 1) variability of parameter estimates within drivers greatly depend on the traffic conditions (low-speed vs high-speed). Driver behavior is highly influenced by traffic conditions rather than individual choices. 2) The following gap is closely associated with the acceleration and comfortable deceleration. 3) The only significant supported by ANOVA to distinguish between different driving characteristic is time gap All the additional factors were insignificant. Hence, it shows difficulty to support heterogeneity of drivers in the analysis.
Xuesong Wang, Dingming Qin, Cristhian Lizarazo Jimenez, Andrew Tarko. Consistency Analysis of Drivers Car Following Behaviors. Transportation Research Board 99th Annual Meeting, Washington D.C., USA, 2020. 1.12-16.