Safety research on stabilization of autonomous vehicles based on improved-LQR control

Abstract

Mediating the divergent interest of vehicle stability and strengthened path tracking performance when aiming at the design of a path tracking controller for autonomous vehicles is a challenging issue. Accordingly, this paper proposes an improved-LQR (linear quadratic regulator) control applied using an improved path planning algorithm. A feedforward and feedback LQR control is constructed by applying the path optimization solution method, which is a different traditional polynomial trajectory fitting method, and then solving the path planning information and the control input parameter in real time to make the tracking error as convergent as possible. To verify the superiority of the improved-LQR, this study compares the proposed controller and model predictive control by the traditional path solving method on a closed-loop test road using Carsim/Simulink. The comparative results show the efficiency, accurate tracking, vehicle stability, and reliability of the proposed controller.