An improved search direction-optimized jump point search path-planning algorithm for autonomous vehicles

Traditional jump point search algorithms in path planning often suffer from issues such as increased search blindness, higher memory usage, and redundant nodes in the output path due to the exploration of numerous irrelevant expansion nodes. To address these issues, an improved search direction-optimized jump point search algorithm that leverages optimized goal-point search directions and safe node updates is proposed. First, by introducing priority sorting for search directions, the possible movement directions are reordered to prioritize expanding nodes toward the goal, thereby accelerating the jump point search process. Subsequently, a redundant node elimination strategy removes unnecessary path points by checking direct collision-free connectivity between nodes, thereby shortening the path length. Finally, the trajectory’s continuity and smoothness are further improved by using a cubic quasi-uniform B-spline curve to smooth the path. Experimental results reveal that, compared to the classic jump point search method, the proposed search direction-optimized jump point search technique diminishes search time by 88.73% in intricate settings and by 72.18% compared to the A* algorithm. The number of expanded nodes is reduced by 75.78% compared to jump point search and 98.58% compared to A*. In more complex simulation environments, ROS2 simulation vehicle experiments indicate that the search time increases by 9.16% compared to A*, while improving by 83.00% compared to jump point search. These findings indicate that the study’s search direction-optimized jump point search algorithm provides better path-planning performance, higher efficiency, and stronger trajectory optimization capabilities compared to the traditional jump point search algorithm.