Intelligent Vehicle Lateral Control Method Based on Feedforward + Predictive LQR Algorithm

• DOI: 10.3390/act10090228
• Publication Date: 2021-09-09T16:20:37Z
• Authors: Tao Yang; Ziwen Bai; Zhiqiang Li; Nenglian Feng; Liqing Chen
• Keywords: Feed forward; CarSim; Model Predictive Control; Tracking error

Aiming at the problems of control stability intelligent vehicle lateral method, single test conditions, etc., a method with feedforward + predictive LQR is proposed, which can better adapt to problem tracking under complex working conditions. Firstly, dynamics error model built by using two degree freedom model, then controller, controller and are designed separately based on path system built. Secondly, YOLO-v3 algorithm, environment perception urban roads established, road information collected, equation fitted sent system. Finally, joint simulation carried out CarSim software Matlab/Simulink tested combined hardware in loop platform. The results hardware-in-loop show that transverse effectively improve accuracy distance course compared control, MPC premise track real time.