Improved RRT Algorithm Based on Adaptive Hyperspherical Sampling Space and Local Motion Policy for Robot Path Planning

The Rapidly-exploring Random Tree (RRT) algorithm is widely used in robot path planning, but the slow convergence rate and poor quality of planned paths have always been its problems. To overcome these limitations, this paper proposes an RRT algorithm based on the improvement of the sampling space and local motion planner. Firstly, an adaptive hyperspherical sampling space is proposed, of which the radius is adjusted according to the minimum distance of the search tree nodes from the goal, which can effectively reduce the number of low-quality random sampling states; secondly, a local motion planner combining optimal motion policy and random motion policy is proposed, which improves the smoothness of the paths and accelerates the convergence. Simulations are conducted with a seven-degree-of-freedom space manipulator, and the results show that compared with the RRT*algorithm, the average search time of the improved RRT algorithm is reduced by 34%, the average path length is shortened by 2.5%, and the average path smoothness is improved by 59%, which verifies the effectiveness and practicality of the improved RRT algorithm proposed in this paper.