Research And Implementation Of Path Planning Method In Automatic Parking Scene

In recent years,with the rapid rise of domestic automobiles,the total number of motor vehicles in the country has been increasing year by year.This not only exacerbates the problem of traffic congestion but also leads to the phenomenon of parking difficulties.In addition,some parking space designs are not reasonable,making parking even more challenging.To address this problem,automatic parking technology has emerged.Automatic parking is an assisted driving technology that enables automatic search for parking spaces and completion of parking,providing convenience to users while improving parking safety.In order to meet the needs of path planning in parking scenarios,two main approaches have been designed: an improved Hybrid A* algorithm for parking path planning and a parking path planning method based on a hierarchical node tree.The main works are as follows:(1)In order to design an heuristic cost that adheres to the optimal design principles in the Hybrid A* algorithm,the Basic Theta* algorithm is used to calculate the actual path cost between nodes and the goal node.This cost is then used as the heuristic cost for nodes in the Hybrid A* algorithm.To address the issue of detection errors in the visibility check method of the Basic Theta* algorithm,adjustments are made to the detection strategy of the existing method to avoid the generation of detection errors,thereby improving the rationality of the path.The improved Basic Theta* algorithm is used to traverse the planning space,obtaining a cost map composed of the actual path costs between all nodes in the map and the goal node.This facilitates the acquisition of heuristic costs in the Hybrid A* algorithm.(2)To address the issue of the Hybrid A* algorithm’s inability to meet the path planning requirements in parking scenarios,strategies are improved.Firstly,the characteristics of nodes are enhanced by adding steering angle and gear information to consider the limitations of the vehicle’s steering system and transmission system.Secondly,an improved Basic Theta*algorithm is utilized to estimate the heuristic cost between nodes and the goal node,guiding the search direction towards the goal node.Furthermore,penalty terms for steering angle and gear shifts are added to the cumulative path cost function to improve the comfort of the path.Lastly,the Reeds-Shepp curve is employed to achieve fast search and precise connections between nodes.These strategy improvements ensure that the Hybrid A* algorithm is capable of handling path planning tasks in parking scenarios,effectively guaranteeing path quality,planning accuracy,and planning efficiency.(3)To address the issues of low efficiency and insufficient generalization ability in the improved Hybrid A* algorithm,a parking path planning method based on a hierarchical node tree is designed.Firstly,different sampling strategies are designed based on the differences in node types.Samples are taken in a hierarchical order to construct a node tree with a hierarchical structure and interconnected nodes.Secondly,corresponding search spaces are built for different parking scenarios to prune the node tree.Thirdly,connection paths between the initial node and the node tree are established using straight or circular units.Backtracking from the initial node generates alternative paths that include key nodes.Finally,interpolation is performed on the key nodes in the alternative paths to obtain discrete parking paths.The optimal path is selected by designing a cost evaluation function for the alternative paths.This method efficiently accomplishes planning tasks in various parking scenarios.Finally,a parking path planning system for parking lot environment was built using the hierarchical node tree-based parking path planning method as the algorithm kernel and the Qt visualization framework.By simulating the parking lot environment,the efficiency and scene adaptability of the hierarchical node tree-based parking path planning method were verified,which has enlightening and reference significance for the path planning of vehicles in parking scenarios.