Modeling of the Traffic Vehicle Lane Change Execution in Autonomous Driving Virtual Test

Abstract

In order to simulate plausible lane change behavior of traffic vehicles for autonomous driving virtual test, this paper proposes a lane change execution model based on vehicle dynamics. According to linear two-degree-of-freedom dynamics model, a mathematical of the lane-changing trajectory function is derived. To ensure vehicle lane change stability in different driving speed, considering that the maximum lateral acceleration can reflect the aggressiveness of lane change behavior, a trajectory function parameter setting method considering lane change characteristic is proposed. The lane change execution model is verified by the fitting effect on the real lane changing trajectories from NGSIM dataset. The result show that the lane change execution model has the similar expressing ability of the lane-changing trajectory with the geometric curve model, and even better fitting effect. The simulation of the model is implemented in PanoSim, which is an intelligent driving simulation platform. The simulation result show that the discretionary lane changing behavior of traffic vehicles can be realistically simulated by lane change execution model. Moreover, by adjusting the vehicle speed and maximum lateral acceleration can flexibly adjust the aggressiveness of the lane change behavior.