Robust feedback predictive path tracking control for autonomous vehicles considering safe geometric constraints

Path tracking (PT) control of autonomous vehicles (AV) plays a significant role in ensuring the safety and accuracy when following the intended trajectories. Consequently, the PT control algorithm should consider two crucial factors: (1) safety in the presence of road boundaries and obstacles, (2) accuracy and robustness in the face of uncertainties. To tackle these challenges, a robust PT control strategy is presented for AV with bounded disturbances, incorporating geometric safety constraints. First, the PT problem is formulated based on the vehicle dynamics model and PT kinematics mechanism. Subsequently, a robust model predictive control (RMPC) framework is introduced, incorporating an ancillary feedback controller and a constraint-tightening approach to mitigate the effects of disturbance on PT operation. Additionally, multiple constraints are structured considering the geometric characteristics of AV, obstacles, and road boundaries, and are incorporated into the receding-horizon optimization control problem to ensure vehicle safety. The PT optimization control problem is converted into a quadratic programming problem for solving. Finally, the effectiveness of the presented scheme is validated under three different scenarios, through Simulink-CarSim co-simulation and field tests. The results reveal the superior performance of the proposed approach in terms of PT accuracy and robustness, without compromising vehicle safety.