UAV Path Planning Based on GeoSOT Grid and JPS3D Optimized Algorithm

Abstract. With the rapid advancement of drone technology, the demands for accuracy and real-time performance in air route planning within modern battlefield environments have significantly increased. In high-precision three-dimensional battlefield modeling, a surge in the number of targets, along with complex obstacle distributions and dynamic threat factors, greatly raises the complexity of path computation.Especially in large-scale three-dimensional complex environments, traditional A* algorithms suffer from efficiency losses due to a lack of effective pruning mechanisms, which leads to excessive redundant node expansions. To address this challenge, this paper implements a three-dimensional Jump Point Search algorithm based on the GeoSOT global discrete grid system. Leveraging the spatial partitioning characteristics of GeoSOT encoding, we propose an efficient path planning algorithm called GPJ3D-RP (GeoSOT Parallelized JPS3D for Rapid Pathfinding). This approach decomposes the global high-precision path planning task into multiple low-resolution subregions, enabling parallel processing of local path searches within each sub-block and ultimately integrating the results into a complete path. Through simulation experiments involving both fighter aircraft route planning and UAV path planning at two different scales, the GPJ3D-RP algorithm demonstrates significant improvements in search speed compared to traditional A* and basic JPS-3D algorithms, making it better suited for real-time path planning requirements in dynamic and complex battlefield environments.