基于预定义几何集的斜列式泊车路径规划方法

Oblique Parking Path Planning Method Based on Predefined Geometric Set

目前预定义几何集方法主要用于平行车位和垂直车位2种类型的泊车路径规划,对于未涵盖的停车位种类而言则存在无法规划泊车路径的问题,使得自主泊车系统对于能够泊入的车位种类存在局限性。为了设计斜列式泊车方式用以提高自主泊车系统中车位种类的覆盖率,提出一种基于预定义几何集的斜列式泊车路径规划方法。首先,提出了斜列式C字形泊车路径规划方法,利用几何学设计一种圆弧相切斜线的路径,并通过多项式建模表征任意角度停车位下的斜列式C字形泊车路径;其次,通过分析泊车路径与停车位中可能发生碰撞的关键点之间的距离,构建了泊车路径碰撞约束模型,以此分析不同规格停车位下约束条件的参数,从而获得规划路径安全区域;然后,结合拓扑地图信息与斜列式C字形泊车路径规划的碰撞约束模型,获得合适的泊车路径起始点范围,继而生成与全局路径无缝衔接的安全泊车路径;最后,通过MATLAB/Simulink仿真验证此方法的有效性,并通过实车试验证明实用性。结果表明:所提方法针对任意规格的斜列式停车位,均能有效地规划出无碰撞条件下的斜列式泊车路径;同时,车辆能够良好地跟踪所规划的路径,实现精准泊入停车位,证明所提方法能够提高自主泊车系统中停车位种类的覆盖率,并大幅度提高自主泊车系统的实用性。

Currently, most automated valet parking systems adopt a predefined geometric set for parking path planning within vertical and parallel parking. The system cannot plan a parking path without covered parking slots, so it has limitations because of the lack of a variety of parking slots. The purpose of this study is to improve the variety of parking slots for a system by designing a path-planning algorithm for oblique parking based on a predefined geometric set. In this study, a “C-like oblique parking path-planning” algorithm was developed;it was designed as a slash tangent to an arc based on geometry, and the oblique parking planning path was characterized through modeling as a polynomial. Second, the constructed parking-path crash constraint model was used to analyze all possibilities of distance constraints between the planning path and the crash point based on geometry to obtain the critical parameter and safety zone. Third, the parking path was generated in a suitable area and spliced to the global planning path by obtaining a suitable interval for the starting point of the parking path based on the information from the topological map and constraints of the C-like oblique parking path planning. Finally, the effectiveness was demonstrated by using MATLAB/Simulink simulation, and the feasibility was verified based on an actual vehicle experiment. The results indicate that this algorithm can effectively plan a crash-free oblique parking path for different types of oblique parking slots, and the results of actual vehicle experiments show that vehicles can be parked precisely in slots. For the automated valet parking system, using a predefined geometric set with the C-like oblique parking path planning can cover most types of parking slots without high calculation costs. Therefore, this algorithm can significantly improve the practicality of automated valet parking systems.