基于可变预测时域及速度的车辆路径跟踪控制

非线性模型预测控制(NMPC)在车辆路径跟踪控制中的应用日益广泛,但目前的研究成果中尚未深入考虑预测时域和速度对车辆路径跟踪控制性能的影响。为此,分析了预测时域、速度与车辆路径跟踪控制性能之间的关系;采用三次多项式拟合获得了能够保证车辆路径跟踪横向误差小于0.1 m的最佳预测时域和参考速度的控制律;改进了用于车辆路径跟踪控制的NMPC控制器,且改进后的NMPC控制器的性能通过仿真进行了验证。仿真结果表明:改进后的NMPC控制器的横向误差在0.092 8 m以内,航向误差在0.072 4 rad以内。相比传统NMPC控制器,改进后的NMPC控制器将最大横向误差减小了4.267 1 m以上,将最大航向误差减小了0.392 7 rad以上,路径跟踪控制性能得到了较大幅度的提高。

Path Tracking Control of Vehicles Based on Variable Prediction Horizon and Velocity

NMPC was widely applied to the path tracking control of vehicles. However, the impacts of prediction horizon and velocity on the performances of path tracking control were not considered in reported researches. Thus, the relationship among prediction horizon, velocity and the performances of path tracking control was analyzed. Through the cubic polynomial fitting, the control laws of the optimal prediction horizon and reference velocity were obtained, which may guarantee the lateral error of path tracking less than 0.1 m. Then, the NMPC controller for path tracking control was improved, and the performances of the NMPC controller were vertified by simulation. The simulation results show that, for the improved NMPC controller, the lateral error is within 0.092 8 m and the heading error is within 0.072 4 rad. Compared with the traditional NMPC controller, the improved NMPC controller reduces the maximum lateral error by more than 4.267 1 m and reduces the maximum heading error by more than 0.392 7 rad, and the performances of path tracking control are improved.