Rapidly-Exploring Random Tree Algorithm Based on Environmental Fuzzy Controller

This paper presents an improved unmanned aerial vehicle (UAV) path planning algorithm that combines the advantages of fuzzy control and the artificial potential field (APF) algorithm. This hybrid approach addresses the limitations of the rapidly-exploring random tree star (RRT*) algorithm, which suffers from poor search path quality and low search efficiency. We introduce a goal-biased strategy that divides the sampling process of RRT* into two stages: exploration and convergence. A dual-input single-output fuzzy controller dynamically adjusts the goal-bias based on obstacle changes and the distance to the target. The paper also proposes a directed node generation approach to enhance the algorithm's adaptability and purposefulness during the exploration stage. The APF algorithm is integrated to guide the random generation of the new node, improving the overall efficiency. The results of simulation experiments conducted in two environments show that the proposed algorithm outperforms other algorithms in terms of higher path quality, fewer redundant nodes, and faster search speeds.