Micro-Vehicle-Based Automatic Parking Path Planning

The path planning problem of the whole process, from given point in the parking lot to the parking slot, is considered in this paper. A hierarchical path planning framework is proposed. Firstly, the map of the parking lot is transformed into directed graph and A* algorithm is used to complete global path planning. Secondly, based on the global path planning, Rapidly-exploring Random Tree-Fixed Nodes (RRT*FN) is applied for local path planning. After that, the local path is simplified by pruning process based on maximum steering curvature, and B-Spline curve based method is introduced to smooth the path. Thirdly, at the end of path, double cubic Bezier curve is used to generate feasible path into the parking slot. At last, automatic parking experiment is carried out on the micro-vehicle-based hardware simulation system, and the experimental result verifies the feasibility and effectiveness of the proposed method.