A Control Strategy of Path Tracking for Unmanned Mining Truck Based on Model Predictive Control and Considering Steering Delay

To achieve accurate and stable path tracking for unmanned mining trucks in the face of changing paths and response delays in steering, this study raised a lateral control strategy for unmanned mining trucks based on MPC and considering steering delay response characteristics. Under the basis of deriving the state space equation from the commonly used two degrees of freedom truck dynamics model, this method introduces the dynamic relationship between steering angle issuance and actual response to form an augmented form of state vector to overcome the control instability caused by steering response delay. Then, based on the MPC method, a constrained objective function is constructed to solve for the optimal control law. In response to the problem of inaccurate selection of prediction and control time domains, this article proposes an adaptive selection method for prediction and control time horizon based on a modified particle swarm optimization (MPSO) algorithm, which obtains the optimal prediction and control time horizon that meet the preset training road conditions in this study, preventing the problem of control accuracy and control oscillation hard to balance caused by the horizon being too small or too large, thereby improving the control effect. Finally, the tracking performance of this algorithm was compared with pure tracking algorithms using a truck dynamics model bench simulation developed independently based on Simulink and a mining truck real-vehicle verification. The algorithm demonstrated good path tracking performance.