Trajectory Tracking Control for Autonomous Parking Based on Adaptive Reduced-horizon Model Predictive Control

In this paper, an adaptive reduced-horizon model predictive control is proposed for autonomous parking trajectory tracking. Given the reference trajectory, the discrete linear time varying model is obtained by linearizing and discretizing along the reference trajectory point. Furthermore, the model is reformulated into a combined incremental form. Then, a standard quadratic programming problem is established, and the optimal control strategy is obtained by solving the problem online at every time instant. Meanwhiles, the prediction horizon will reduce adaptively by solving the constrained optimization problem, and it will minimize the computation time complexity of the MPC-based controller. The actual parking scenarios are co-simulated in Simulink and Carsim, which shows the effectiveness and feasibility of the proposed method.