DF-RRT*: A Dynamic-Fusion RRT* Algorithm for UAV Path Planning in Dynamic Environments
Ziqi Guo, Hai-Ning Zhang · 2025
Efficient path planning in dynamic, three-dimensional environments remains a major challenge for unmanned aerial vehicles (UAVs). To address the limitations of the traditional RRT* algorithm in terms of sampling randomness, poor convergence, and lack of adaptability to dynamic obstacles, this paper proposes DF-RRT*, a Dynamic-Fusion RRT* algorithm. DF-RRT* integrates three strategies: a local potential field-guided sampling strategy to reduce unnecessary exploration, a dynamic goal-biased strategy to enhance convergence, and a local re-planning mechanism to adaptively update paths in real time. Simulation experiments in 3D environments demonstrate that DF-RRT* significantly outperforms baseline algorithms (RRT*, G-RRT*, and APFRRT*) in terms of execution time, initial path quality, and final path length. The proposed method also improves path smoothness and robustness through B-spline optimization, offering a practical and efficient solution for UAV navigation in complex dynamic scenarios.