面向动态障碍物场景的自主代客泊车路径规划  

Autonomous valet parking path planning for dynamic obstacle scene

作　　者：张家旭 王志伟 郭崇[1] 赵健[1] Zhang Jiaxu;Wang Zhiwei;Guo Chong;Zhao Jian(State Key Laboratory of Automotive Simulation and Control,Jilin University,Changchun 130022,China;Intelligent Network R&D Institute,China FAW Group Co.,Ltd.,Changchun 130011,China)

机构地区：[1]吉林大学汽车仿真与控制国家重点实验室,长春130022 [2]中国第一汽车集团有限公司智能网联研发院,长春130011

出　　处：《东南大学学报（自然科学版）》2022年第2期369-376,共8页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金面上资助项目(51775235)。

摘　　要：针对存在动态障碍物场景的自主代客泊车路径规划问题,提出一种基于D;算法和动态窗口法的自主代客泊车路径规划方法.首先,利用栅格扫描算法快速准确地构建自主代客泊车场景的静态环境地图,并采用Dijkstra算法实时更新动态障碍物影响的局部静态环境地图信息.随后,利用D;算法将自主代客泊车场景静态环境地图转化为静态路径场,得到原始全局自主代客泊车路径,同时借助静态路径场重新规划动态障碍物影响的全局自主代客泊车路径.最后,采用动态窗口法和圆弧-直线组合方式规划最优的局部自主代客泊车路径,并在规划过程中融合几何边界碰撞检测方法和栅格空间覆盖枚举方法,快速剔除不可行局部自主代客泊车路径.利用仿真验证了所提出方法的可行性和有效性,结果表明,在动态障碍物存在的自主代客泊车仿真场景中,该方法能够安全无碰撞地引导汽车进入目标垂直泊车位.Aiming at the problem of autonomous valet parking path planning in the presence of dynamic obstacle,a novel path planning method for autonomous valet parking system is proposed based on D;algorithm and dynamic window method.First,the raster scanned algorithm is used to quickly and accurately construct the static environment map of the autonomous valet parking scene,and the Dijkstra algorithm is used to update the local static environment map affected by dynamic obstacle in real time.Then,the D;algorithm is used to transform the static environment map of the autonomous valet parking scene into a static path field in order to obtain the original global autonomous valet parking path.Meanwhile,the global autonomous valet parking path affected by dynamic obstacle is re-planned with the help of the static path field.Finally,the dynamic window method and arc-line combination method are used to plan the optimal local autonomous valet parking path,and the geometric boundary collision detection method and raster space coverage enumeration method are integrated to quickly eliminate the infeasible local autonomous valet parking path.The feasibility and effectiveness of the proposed method for autonomous valet parking system are verified in VC++6.0 environment.The results show that the proposed method can safely guide the vehicle into the target vertical parking space in the autonomous valet parking scene with dynamic obstacle.

关 键 词：栅格扫描算法 DIJKSTRA算法 自主代客泊车 路径规划 D 算法 动态窗口法 

分 类 号：U461.1[机械工程—车辆工程]