Robot Dynamic Path Planning Based on the Fusion of Jump Point Search Algorithm and Dynamic Window Approach

Abstract

To address the problem of slow route computation caused by excessive exploration of unnecessary nodes in traditional A* algorithm combined with the dynamic window approach, as well as the challenge of effectively leveraging global planning data for local navigation, this paper proposes an integrated method that combines the jump point search (JPS) algorithm with the dynamic window approach (DWA). First, the JPS algorithm is utilized in a pre-defined map environment for global route generation. Next, a three-stage spline optimization is applied to improve the continuity and flow of the computed path. Finally, the optimized path nodes are selected and used as intermediate waypoints for the dynamic window approach to facilitate efficient local navigation. During the motion, these nodes guide the dynamic window approach algorithm in stages, aligning it more closely with the global planning path and allowing real-time avoidance of dynamic and static obstacles in the environment. Experimental results demonstrate that the fusion algorithm exhibits superior path planning capabilities, achieving global optimality and effective obstacle avoidance.