A Fast Trajectory Tracking Control Design for Autonomous Driving

Abstract

This paper presents a fast trajectory tracking control strategy based on the Linear-Quadratic Regulator (LQR) and Receding Horizon Control (RHC) for real-time autonomous driving. A path deviation parameter is introduced and integrated into the LQR. This parameter can play a role in controlling the vehicle to avoid collisions or change the motion lane. Under the framework of RHC, the LQR must be solved online repeatedly. The computational complexity of the closed-form solution to the LQR was the motivation behind obtaining a computationally fast approximate solution aiming to make the RHC fast and real-time. Simulation experiments where a vehicle tracks a curvilinear trajectory in a two-lane road are carried out to illustrate the performance of the presented fast LQR RHC controller.