Path Planning of UAV Based on Improved Bidirectional F-RRT* Algorithm
Weijie Cai, Xiafu Peng, Xiaoli Zhang · 2024
Unmanned aerial vehicles (UAVs) face challenges in path planning within complex environments. Traditional RRT* algorithms exhibit drawbacks, including long initial path, slow convergence, and non-smooth paths. To address these issues, this paper proposes the Improved Bidirectional F-RRT* Algorithm. Utilizing informed and probability goal sampling, the algorithm enhances convergence speed by guiding the random tree towards the goal. Turn angle constraint is introduced to generate paths better suited for UAVs. Similar to F-RRT*, dichotomy is adopted for parent node creation. Using obstacle avoidance adjustment when collisions occur to reduce the time consumption of excessive sampling. The rewire process considers both the new node and its parent, expediting convergence. Integration with RRT*-Connect enhances overall efficiency. When dynamic obstacles arise, a local path correction strategy enables dynamic obstacle avoidance. Simulation results confirm the algorithm’s capability to swiftly find optimal paths and efficiently handle dynamic obstacles.