Tube-Based Robust Model Predictive Control for Autonomous Vehicle with Complex Road Scenarios

Abstract

This study proposes a Tube-based Robust Model Predictive Control (Tube-RMPC) strategy for autonomous vehicle control to address model parameter uncertainties and variations in road–tire adhesion coefficients in complex road scenarios. More specifically, the proposed approach improves the representation of vehicle dynamic behavior by introducing a unified vehicle–tire modeling framework. To facilitate computational tractability and algorithmic implementation, the model is systematically linearized and discretized. Furthermore, the Tube-based Robust Model Predictive Control strategy is developed to improve adaptability to uncertainty in the road surface adhesion coefficient. The Tube-based Robust Model Predictive controller ensures robustness by establishing a robust invariant tube around the nominal trajectory, effectively mitigating road surface variations and enhancing stability. Finally, a co-simulation platform integrating CarSim and Simulink is employed to validate the proposed method’s effectiveness. The experimental results demonstrate that Tube-RMPC improves the path-tracking performance, reducing the maximum tracking error by up to 9.17% on an S-curve and 2.25% in a double lane change, while significantly lowering RMSE and enhancing yaw stability compared to MPC and PID.