Adaptive Trajectory Tracking of Heavy-Duty Unmanned Tracked Vehicles Based on a Data-Driven Time-Varying Linear Model

Trajectory tracking control plays a crucial role in ensuring the stable maneuverability of heavy-duty unmanned tracked vehicles (HDUTVs). It requires consideration of both the vehicle dynamic characteristics and external disturbances during operation. Accordingly, an adaptive trajectory tracking control strategy based on a data-driven time-varying linear model is proposed. First, multiple parallel neural networks are designed to learn the evolution of each vehicle state variable. Through training, a vehicle system parameter inference model is obtained, allowing the model to preserve both interpretability and time-varying characteristics. Then, based on the proposed vehicle model, a model predictive control-based trajectory tracking controller is designed to achieve accurate tracking performance. Finally, the proposed method is validated with real-world driving data and simulations. The results demonstrate that the proposed trajectory tracking method improves the trajectory tracking performance of the HDUTV, achieving a 41.13% reduction in lateral tracking error compared to the widely used method.