Simulation and Experimental Study on Trajectory Tracking of Tracked Unmanned Vehicle on Nonstructured Roads in the Field

Abstract

An improved model predictive control (MPC) trajectory tracking controller is proposed for the automatic driving of tracked unmanned vehicle (TUV) on the nonstructured roads in postdisaster field, and experiments and debugging are carried out in real environments. The TUV trajectory tracking controller based on the MPC algorithm is designed according to the kinematic model of the TUV. Aiming at the model uncertainty problem caused by the vehicle body sinking and track slipping during the traveling process of the TUV, a driving wheel rotation speed correction controller is proposed. The controller can effectively suppress external interference through experimental data fitting, thereby improving tracking performance, especially during curve tracking. Adaptive Kalman Filtering technology is introduced to improve the vehicle position accuracy. The model and parameters are optimized through model simulation and debugging of real vehicle experiments in the field roads. When compared with the nonlinear model predictive control (NMPC) algorithm, the improved MPC controller demonstrates significant reductions in trajectory tracking deviations. Specifically, for the three different road conditions, the maximum positional deviation is reduced by 41.21% on average, the average positional deviation is reduced by 42.95%, the maximum heading angle deviation is reduced by 27.84% on average, and the average heading angle deviation is reduced by 19.39%. These results clearly indicate that the improved MPC controller proposed in this paper outperforms the NMPC algorithm in terms of trajectory tracking effectiveness.