Robust Control for Autonomous Vehicle Lateral Dynamics: A Comparative Study of Gain-Schedule LPV and Non-linear MPC

The increasing demand for autonomous vehicles necessitates advanced control strategies, particularly for managing complex driving scenarios like high-speed turns. In this study, we introduce a novel adaptive control approach using Gain-Scheduled Linear Parameter Varying (LPV) control and Non-linear Model Predictive Control (MPC) for controlling autonomous vehicle lateral dynamics. Our technical contribution lies in the development and application of these control strategies to a non-linear dynamic vehicle model, particularly during the challenging Fishhook Maneuver, which assesses rollover propensity. The validation of these approaches encompasses extensive testing under varied road conditions and wind effects. The results demonstrate LPV's enhanced robustness and superior performance compared to MPC, suggesting its greater suitability for complex, real-world driving scenarios. This research offers significant insights into optimizing control systems for autonomous vehicles, emphasizing safety and adaptability in dynamic environments.