基于改进 RRT 算法和人工势场平滑的航迹规划

To address the shortcomings of low computational efficiency, high randomness, trajectory inflection points, lengthy paths, and poor safety of the combination of the rapid-exploring random tree (RRT) algorithm and conventional smoothing methods in solving unmanned aerial vehicle (UAV) path planning problems under complex environments, a trajectory planning method combining the improved rapid-exploring random trees-connect (RRT-Connect) algorithm with the artificial potential field (APF) smoothing method is proposed. Firstly, two random trees are simultaneously expanded from the start point and the target point, with a greedy strategy introduced to enhance search efficiency. Then, a target bias strategy is adopted to optimize the selection of random nodes. Additionally, an improved APF component is added to reduce the randomness of the search direction and accelerate the intersection of the two random trees to obtain a feasible trajectory avoiding obstacles. Finally, based on the initial trajectory, a smoothing optimization method using the APF is applied: the trajectory nodes are regarded as points with attractive potential field, which only acts gravitationally on the adjacent nodes, and the repulsive potential field is set up at each obstacle. Through this dual-action mechanism combining attraction and repulsion, trajectory nodes are adjusted and optimized to improve trajectory smoothness and safety. The feasibility and effectiveness of the proposed method in different scenarios are verified through simulation experiments.