Lateral Path Tracking Control Based on Tube MPC

To solve the problem of external disturbances encountered in the lateral path tracking control process of autonomous vehicles, this paper proposes a TMPC algorithm based on the nonlinear vehicle dynamics model. The core of this method lies in tightening the constraint boundaries and increasing errors feedback compensation to ensure lateral tracking performance and driving stability. Additionally, the algorithm's real-time performance is further improved by using zonotopes instead of polytopes. Finally, the effectiveness of the method is verified through simulation experiments. The results demonstrate that compared to traditional MPC, TMPC reduces the maximum value and mean square value of lateral tracking errors by 80.1% and 86.6%, respectively. It also reduces the maximum value and mean square value of heading tracking errors by 37.2% and 33.8%, respectively.