基于改进人工势场法的智能车辆避撞路径规划

Path planning for intelligent vehicle collision avoidance based on improved artificial potential field method

针对传统人工势场法存在道路边界势场不完善和局部最优问题,文章提出一种改进人工势场法的智能车辆避撞路径规划。引入道路势场函数来描述道路边界,设立虚拟目标点来摆脱局部最优,建立道路环境模型;为了根据周边环境和车辆状态进行实时规划,设计分层避撞路径规划控制器,将道路环境模型引入上层路径规划器的目标函数中,利用模型预测控制(model predictive control,MPC)的优化算法规划出局部避撞路径,再将路径信息输入到下层跟踪控制器进行跟踪。MATLAB/Simulink与CarSim联合仿真实验结果表明,该避撞路径规划对于静态障碍物和动态障碍物都可以规划出平滑无碰撞的路径,保证车辆行驶的稳定性和安全性。

In view of the problems of the traditional artificial potential field method such as imperfect road boundary potential field and local optimum,a path planning for intelligent vehicle collision avoidance based on improved artificial potential field method is proposed in this paper.The road potential field function is introduced to describe the road boundary,and the virtual target point is set up to get rid of the local optimum.In order to make real-time planning according to the surrounding environment and vehicle state,a hierarchical collision avoidance path planning controller is designed.The road environment model is introduced into the objective function of the upper path planning controller.The local collision avoidance path is planned by using the optimization algorithm of model predictive control(MPC),and then the path information is input to the lower tracking controller for tracking.The co-simulation experiments of MATLAB/Simulink and CarSim show that smooth collision-free paths can be planned for both static and dynamic obstacles to ensure the stability and safety of vehicles.