基于MPC的智能车辆路径规划与跟踪控制

Intelligent Vehicle Path Planning and Tracking Control based on MPC

针对在障碍物环境下的避障路径动态规划效果较差,以及在面对复杂工况和曲率较大的路况时,跟踪控制的效果仍然不理想等问题,本文以智能车辆为研究对象,提出了一种模型预测控制(MPC)结合人工势场(APF)算法的路径规划跟踪系统。将改进的势场模型函数引入到MPC的目标函数和约束中,设计了基于MPC和APF的避障路径动态规划器。。运用模糊控制对MPC的车辆横向路径跟踪控制器的权重系数进行优化。仿真结果表明:在干燥路面下,与MPC控制器相比,模糊MPC路径跟踪控制器的最大横向偏差减少19.14%。在湿润路面下,模糊MPC控制器最大横向偏差减少0.55 m。基于MATLAB/Simulink与Carsim软件搭建避障路径规划与跟踪控制联合仿真模型,选择动态障碍物不同速度进行障碍物路径动态规划及跟踪控制仿真试验。实验结果表明:跟踪规划路径过程中的最大横向偏差约为0.170 m,说明规划的避障路径能够安全有效地避开障碍物。

The dynamic planning effect of obstacle avoidance paths in obstacle environments was poor and the tracking control effect was still not ideal when facing complex working conditions and high curvature road conditions.Taking intelligent vehicles as the research object,a path planning and tracking system was proposed by combining model predictive control(MPC)algorithm with artificial potential field(AMF)algorithm.The improved potential field model function was intrduced into the objective function and constraints of MPC.The dynamic obstacle avoidance path planner based on MPC and improved APF was designed.Fuzzy control was used to optimize the weight coefficients in the MPC path tracking controller.The simulation results show that the maximum lateral deviation of the fuzzy MPC path tracking controller is reduced by 19.14%compared with the MPC controller on dry road.The maximum lateral deviation of the fuzzy MPC controller is reduced by 0.55 m on wet road.The joint simulation model of obstacle avoidance path planning and tracking control is built based on MATLAB/Simulink and Carsim software.Dynamic obstacle path planning and tracking control simulation experiments are conducted under different speeds of dynamic obstacles.The experimental results show that the maximum lateral deviation in the process of tracking the planned path is about 0.170 m,which indicates that the planned obstacle avoidance path can avoid obstacles safely and effectively.