Trajectory planning for autonomous vehicles based on improved hybrid A*

Abstract

Trajectory planning plays an important role in autonomous driving technology. This paper presents an improved Hybrid A* algorithm by considering vehicle dynamics in planning. First, a vehicle dynamic model with three degrees of freedom and a nonlinear UniTire tyre model are developed to describe the vehicle dynamics. Then, in order to generate curvature-continuous trajectory, the method of expanding nodes based on the rotation speed of steering wheel is introduced. Then, heuristic functions are designed for trajectory searching. In addition, the feasibility of the planning result is verified in the driver-vehicle-road closed loop system. Simulation results show the proposed trajectory planning method performs well under a variety of working conditions and the planned trajectory is trackable even at extreme conditions, e.g., high speed and low-adhesion road surfaces. Specifically, on the icy road, the planned results are good and the maximum lateral tracking error of this trajectory is limited to 0.24 m.