基于Voronoi的新型垂直自动泊车路径规划启发函数

A Novel Perpendicular Parking Path Planning Heuristic Function Based on Generalized Voronoi Graph

即使对经验丰富的司机来说,停车也可能是一项艰巨的任务,尤其是在停车位很窄的老城区车位。考虑到Reeds-Shepp曲线和Dubins曲线的局限性,在满足启发函数设计原则的情况下不考虑车辆运动学约束,采用曼哈顿距离并引入广义Voronoi图设计Hybrid A*启发函数,并进行仿真分析。仿真结果表明,在设计的启发函数下规划的路径较为笔直和简洁,未出现不必要的倒车和不必要的转弯,较Reeds-Shepp曲线和Dubins曲线作为启发函数规划的路径有较大的改善。

Even for experienced drivers,parking can be a challenging task,especially in large cities where parking spaces are often narrow.Considering the limitations of the Reeds Shepp curve and Dubins curve,without considering vehicle kinematic constraints while meeting the design principles of the heuristic function,the Manhattan distance and the introduction of a generalized Voronoi diagram are used to design the Hybrid A*heuristic function,and simulation analysis is conducted.The simulation results show that the path planned under the designed heuristic function is relatively straight and concise,without unnecessary reverse and turning,which is significantly improved compared to the path planned using the Reeds Shepp curve and Dubins curve as heuristic functions.