Obstacle avoidance path planning of robotic arms based on an improved RRT algorithm
Yutao Niu, Wenjie Chen, Xiantao Sun, Xiaolong Cui · 2025
Aiming at the shortcomings of fast expansion random tree (RRT) algorithm such as slow convergence speed, long search time, low sampling efficiency and many redundant points, an improved RRT algorithm is proposed. Firstly, a two-step node expansion method based on target bias is proposed. Under the guidance of gravity of target point and random expansion nodes, the randomness is reduced while the probability is kept complete, and accelerates the expansion speed through the two-step expansion strategy to improve the convergence speed of the algorithm. The method of node rejection is used to avoid repeated search of the same area and avoid the generation of useless nodes. Finally, the redundant nodes are removed by node cutting and the path length is shortened. The simulation results show that the number of sampling nodes and time cost of the improved RRT algorithm are much smaller than that of the standard RRT algorithm and the traditional RRT* algorithm, and the path length of the improved RRT algorithm is close to that of the traditional RRT* algorithm, which verifies the superiority and feasibility of the algorithm.