Calibration and Analysis of Heterogeneous Car-Following Behaviors Based on a Naturalistic Trajectory Dataset on Highways

Modeling car-following behaviors is critical in traffic flow simulation and analysis. However, the investigation of car-following behaviors was limited in the small sample size of vehicle trajectories and different vehicle types. To overcome such shortcomings, this research calibrates and analyzes car-following behaviors based on a naturalistic trajectory dataset recorded in German highways. We divide vehicles into four types and calibrate an IDM car-following model for each vehicle based on a genetic algorithm. The calibration results-based IDM model shows heterogeneous and interesting parameters, which undoubtedly reveals heterogeneous car-following behaviors among different vehicle types on highways. A final simulation case study validates the performance of the vehicle type-based car-following model and, further pointed out potential applications of the research, for instance, models more realistic traffic simulations and provides parameters for adaptive cruise control systems for a different type of vehicles.