Path Planning Based on Improved RRT and Bezier Curve Fusion Algorithm

To address the issue of excessive redundant nodes in the Rapidly-exploring Random Tree (RRT) algorithm, this paper proposes several measures for improvement. These measures encompass dynamically adjusting the sampling step size, end-point convergence radius, goal-oriented approach, node rejection mechanism, and curve optimization. The paper mainly introduces the improved principles. The simulation results show that the improved RRT algorithm improves the search efficiency and reduces the path length for unmanned surface vehicles (USVs). Moreover, the path is smoothed by using Bezier curve.