Path Planning of Manipulator Based on Improved RRT* Algorithm

Abstract

The traditional RRT* algorithm lacks directionality in sampling, and often generates a large number of invalid nodes, resulting in low utilization of later nodes. A large number of calculations between nodes cause the time spent by the manipulator to plan a path greatly increased and the path is not smooth. For the shortcomings of the above RRT* algorithm, an improved RRT* algorithm is proposed in this paper. First, a random point is generated with variable probability bias. Second, the sampling method is improved and an adaptive step size strategy is introduced. Finally, for the initial path, a path point connection screening strategy is introduced, and the screened path points are optimized again using a cubic b-spline curve. The experimental results show that the improved RRT* algorithm can quickly find a collision-free path, the path is more optimized, and the planning time is greatly reduced.