Guided-TargetTree-Hybrid_A* Path Planning Algorithm for Vertical Parking

To enhance the safety of automatic parking in complex environments and reduce the time consumption of path searching, a vertical parking path planning method based on the optimization of Hybrid A* is proposed. Firstly, a costmap containing semantic information is generated to produce a Voronoi diagram. The starting point and the target point of the vehicle are projected onto the Voronoi map, and an A* search is used to create a global guiding line for the Hybrid A* algorithm. Simultaneously, a target tree is generated from the parking space to extend the target point for the Hybrid A*. The Hybrid A* is then optimized to search for a parking path that satisfies the vehicle's kinematics, and an SQP method is used for smoothing. Through real vehicle verification, this method meets the parking requirements in narrow spaces, demonstrating strong adaptability and feasibility.