Efficient Path Planning for Automated Valet Parking: Integrating Hybrid A* Search with Geometric Curves

Abstract

An automated valet parking system, a critical element of autonomous driving technology, greatly enhances parking efficiency and reduces driver stress. In the realm of path-planning in tight spaces, traditional Hybrid A* algorithms often produce numerous invalid expansion nodes, leading to increased computational load and decreased spatial utilization efficiency. This study introduces a new path-planning approach that combines the Hybrid A* algorithm with geometric curves to overcome these challenges. Initially, the Hybrid A* algorithm conducts global path-planning to ensure the viability of the route from the parking entrance to the desired spot. Subsequently, a genetic algorithm optimizes the driving path locally, improving safety and efficiency. Finally, Bézier and clothoid curves are utilized to smooth the path, further enhancing planning efficiency and driving safety. Experimental results show that the proposed approach surpasses existing methods in terms of planning time efficiency and path quality.