An optimal path tracking architecture for automated vehicle

Abstract

Nowadays, automated vehicle has attracted a lot of attention with the advantages of safety, comfort, and efficiency. This paper presents a path tracking architecture synthesizing a preview feedforward controller and an adaptive sliding mode feedback controller. First, the vehicle dynamics model and the geometric relationship between the target path and vehicle are described. Then, the feedforward controller is designed based on the multipoint preview, which could reduce the interference of road curvature and time delay of vehicle actuator. Subsequently, the feedback controller utilizing adaptive discrete sliding mode is proposed considering the robustness in different conditions. It has a few control parameters, fast convergence, and small stationary error of the direction and lateral distance. Eventually, real-time simulation results show that the automated vehicle could track the target path accurately under varying time delay, vehicle speed, and road adhesion. Furthermore, vehicle experiment results verify the effectiveness of path tracking.