Diffeomorphism‐based trajectory tracking control for autonomous vehicles with time‐varying uncertainty

Abstract Trajectory tracking is one of the most essential tasks in autonomous driving. The driving environment and vehicle dynamics involve numerous complex uncertainties. For four‐wheel independent drive vehicles powered by in‐wheel motors, the torque output capability of each in‐wheel motor is constrained by its rotational speed, resulting in limited control inputs. To address this challenge, a robust control scheme for nonlinear vehicle systems is presented. This scheme aims to achieve satisfactory trajectory tracking performance in complex driving environments while maintaining in‐wheel motor torque output within an expected range. Initially, diffeomorphism theory is employed to handle bounded input constraints in vehicle speed control, addressing the mismatch between control inputs and the operational constraints of in‐wheel motors. A robust control approach based on an input diffeomorphism scheme is developed for longitudinal dynamics. Subsequently, the trajectory tracking problem is elegantly reformulated as an approximate constraint‐following control problem by converting trajectory tracking tasks into equality constraints. A constraint‐following‐based robust control methodology is proposed for vehicle lateral dynamics to address trajectory tracking challenges. Finally, CarSim‐Simulink co‐simulations verify that the proposed strategy achieves accurate trajectory tracking while maintaining in‐wheel motor torque within the maximum execution capacity, despite nonlinear and time‐varying uncertainties.