基于非线性模型预测控制的拖挂车系统泊车轨迹规划方法

The tractor-trailer is a highly coupled nonlinear system with underactuated and nonholonomic constraints, which has the characteristics of high state dimension, complex constraints and poor internal stability, resulting in a long solution process. Therefore, a parking trajectory planning method for tractor-trailer system is designed based on nonlinear model predictive control (NMPC), which constructs an optimization problem by integrating the multivariate constraints of the system, and plans a set of collision free parking trajectories. In order to accelerate the solution of the optimization problem, firstly, random sampling based on the RRT*-RS curve is carried out in the high dimensional state space of the trailer system, and an approximate optimal trajectory is found as a reference solution for the warm start of the optimization problem with the RS curve meeting the characteristics of reverse driving. Then, in the process of solving the optimization problem, the Proximal Averaged Newton-type method for Optimal Control (PANOC) is used to save the solution time and improve the real-time performance of the system. Finally, simulation experiments of reversing parking, side parking, multi-obstacle and other scenarios are carried out. The simulation results show that the proposed method can realize the fast trajectory planning of the tractor-trailer system in different scenarios. Compared with the traditional interior point method and sequential quadratic programming method, the solution time has increased by more than 40%.