Cascaded MPC with Kinematic-Dynamic Dual-ESO for Trajectory Tracking of Tracked Vehicles

To address trajectory tracking deviations in tracked vehicles caused by significant unmodeled kinematic slip and dynamic terrain resistance uncertainties, this paper proposes a novel cascaded Model Predictive Control (MPC) architecture integrated with a Kinematic-Dynamic Dual Extended State Observer (Dual-ESO). Specifically, the inner-loop dynamic ESO estimates physical torque resistances, while the outer-loop kinematic ESO reconstructs time-varying geometric slips for active feedforward compensation. The proposed framework is comprehensively validated under significant compound disturbance scenarios, including continuous dynamic slips, step lateral skids, sudden yaw drifts, and abrupt terrain resistance surges. Simulation results demonstrate that the constructed Dual-ESO rapidly and accurately reconstructs the multi-source disturbances. Consequently, compared to the conventional cascaded kinematic MPC and Active Disturbance Rejection Control (ADRC) framework, the proposed architecture yields superior tracking precision, reducing both the Integral Time Absolute Error (ITAE) and the Root Mean Square Error (RMSE) along both the X and Y axes by over 73%.