Fuzzy adaptive model predictive control and sparse identification for unmanned vehicles

Vehicle operating conditions vary significantly across different road sections and speeds, yet traditional model control methods often lack the ability to adapt to these changing environments, resulting in prediction errors. This paper introduces an advanced motion control technique for unmanned vehicle trajectory tracking, leveraging an innovative approach to model predictive control. We develop a comprehensive dynamic model of the vehicle, incorporating relevant constraints, and propose an error trigger adaptive sparse identification (ETASI) method for efficient parameter identification under diverse operating conditions. Our novel fuzzy adaptive model predictive control (FAMPC) technique substantially enhances the adaptability of classical MPC controllers across a wide range of scenarios. Through rigorous analysis, we demonstrate the iterative feasibility and stability of the proposed system. Extensive numerical simulations and real-world vehicle experiments validate the efficacy of our approach. Results reveal that FAMPC achieves robust trajectory tracking across various operating conditions, significantly outperforming both classical MPC and adaptive MPC (AMPC) in terms of accuracy and adaptability. This research provides valuable insights for the future development of unmanned vehicle trajectory tracking control systems capable of handling complex, dynamic environments.