Research on Automatic Parking Path Planning Based on Bidirectional Hybrid A-Star Algorithm

An improved bidirectional hybrid <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$A^{*}$</tex> algorithm is proposed to address the issue of slow parking path search in narrow scenarios for vehicles using the hybrid <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$A^{*}$</tex> algorithm. Firstly, in the search process, both forward and reverse directions are used simultaneously, and RS curves are employed to connect the two paths, thereby improving search efficiency. Then, to address the problems of long search paths and poor comfort, the cost function of the algorithm is optimized. Finally, vertical and parallel parking scenarios were built in the MATLAB simulation platform. In both scenarios, the improved algorithm proposed in this paper has less search time and shorter search path compared to the hybrid <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$A^{*}$</tex> algorithm, demonstrating the effectiveness and superiority of the improved algorithm.