The Application of RRT Fast Search Tree Algorithm in Route Planning and Obstacle Avoidance of Unmanned Vehicle

The author proposes an improved bidirectional RRT-based local path planning algorithm for unmanned boats, with the goal of quickly planning optimal local obstacle avoidance paths for high-speed sailing unmanned boats. In order to address the problem of strong randomness in traditional bidirectional RRT algorithms and frequent twists and turns in the planned path, corner constraints are applied to each newly extended node. In response to the problem of traditional bidirectional RRT algorithm where two trees often cannot be smoothly connected, angle constraints and distance constraints are applied at the active connection points. A new dynamic step-size strategy was also proposed. The research results show that, in the case of no dynamic step-size strategy, the average convergence time of the improved two-way RRT algorithm after removing the maximum and minimum values is 24.375 ms, which is 14.706% more than that of the algorithm with dynamic step-size strategy. It is verified that the dynamic step-size strategy can accelerate the Search tree to jump out of the area around obstacles, thus reducing the convergence time of the algorithm. Research has shown that the improved algorithm has better path planning quality and shorter convergence time.