With the gradual application of commercial Automated Valet Parking (AVP) systems in common parking slots, effective collision detection methods in narrow space have received considerable attention. This paper proposes a fast and accurate collision detection method for autonomous vehicle along the Reeds-Shepp curve in narrow space including various static obstacles. Taking advantage of Bounding Volume Hierarchy (BVH) and quadtree, a tree structure based on hierarchical inflated maps is established by decomposing vehicle into circles recursively, which is suitable for narrow space. To improve the tree search efficiency, an incremental strategy that uses boundary envelope circles to replace full coverage circles is proposed to simplify the tree. Besides, this paper theoretically derives the max sampling distance to accurately guarantee a continuous collision-free path which also minimizes the number of detections. Furthermore, a range of challenging simulation experiments show that the proposed method combined with the Optimal Rapidly exploring Random Tree (RRT*) planner effectively returns a safe and high-quality path. Most importantly, our work out-performs the widely used method in terms of detection accuracy and computation time.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A Fast and Accurate Collision Detection Method During Automated Valet Parking for Autonomous Vehicle in Narrow Space


    Contributors:
    Zheng, Han (author) / Ma, Xihan (author) / Kong, Weiwei (author)


    Publication date :

    2023-09-24


    Size :

    832013 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English