Lateral Trajectory Tracking Control of 4WIS Intelligent Vehicles Considering Time-Varying Cornering Stiffness Estimation

The uncertainty and simplification in the modeling process, along with unknown disturbances from the environment, are the primary factors affecting the trajectory tracking accuracy of autonomous vehicles. In response, this paper proposes a hierarchical control strategy for lateral trajectory tracking in four-wheel independent steering (4WIS) intelligent vehicles considering time-varying cornering stiffness estimation. At the upper level, using the deep deterministic policy gradient (DDPG) algorithm addresses the time-varying and nonlinear issues of cornering stiffness in the predictive model, achieving higher accuracy in lateral force estimation and reducing the impact of model mismatch. Subsequently, an adaptive model predictive controller (AMPC) is designed to minimize tracking errors. At the lower level, tracking the desired steering angle using adaptive integral terminal sliding mode control (AITSMC) enhances the robustness of the steering motor against environmental disturbances. Finally, based on a combined simulation platform, the feasibility and superiority of the proposed trajectory tracking control scheme are verified in scenarios of double lane change and serpentine turning.