Abstract: Aiming at the autonomous valet parking path planning problems for the wire-four-wheel steering vehicles， an autonomous valet  path planning method with adaptive ability to dynamic obstacles was proposed.  Firstly， the grid map which might describe the environmental information of parking lot was used to transform the autonomous valet parking path planning problem in continuous state space into that in discrete state space， and the isotropic circular structure elements were used to expand the grid map to keep a safe distance between the planned autonomous valet parking path and the obstacles.  Then， the D* Lite algorithm was used to plan the global autonomous valet parking path with adaptive ability to dynamic obstacles， as well the dynamic window method and the arc-line combination method were used to plan the local autonomous valet parking path， which met the kinematic and mechanical constraints of the wire-four-wheel steering vehicles and might guide the vehicles to enter the oblique parking space safely.  Finally， the feasibility of the proposed method was verified by simulation， and the results show that the proposed method may plan a feasible path to guide the vehicles into the target oblique parking space safely in the presence of dynamic obstacles. 

Key words: vehicle engineering, grip map, autonomous valet parking, path planning, dynamic window method

摘要： 针对线控四轮转向汽车自主代客泊车路径规划问题，提出了一种对动态障碍物具有自适应能力的自主代客泊车路径规划方法。首先，采用栅格地图描述停车场环境信息，将连续状态空间中的自主代客泊车路径规划问题转化为离散状态空间中的自主代客泊车路径规划问题，并采用各向同性的圆形结构元素对栅格地图进行膨胀处理，使规划的自主代客泊车路径与障碍物之间保持一定的安全距离。随后，采用D* Lite算法规划出对动态障碍物具有自适应能力的全局自主代客泊车路径，并采用动态窗口法和圆弧-直线组合方式沿着全局路径规划出满足线控四轮转向汽车运动学和机械约束的局部自主代客泊车路径，引导汽车安全无碰撞地进入斜向泊车位。最后，通过仿真验证所提方法的可行性，结果表明：所提出的方法可以在动态障碍物存在的自主代客泊车场景中规划出安全无碰撞的可行路径，引导汽车进入目标斜向泊车位。

关键词: 汽车工程, 栅格地图, 自主代客泊车, 路径规划, 动态窗口法