Three-dimensional path planning of manipulator based on improved RRT algorithm

摘要

This paper provides an improved algorithm which is proposed for the deviating optimal solution, non-target and path non-smooth problem of the original Rapidly-exploring Random Trees (RRT). The improved algorithm is based on the new parent node selection method to ensure that the random tree extends the asymptotic optimality of the new node. On this basis, the heuristic value function is introduced according to the Euclidean shortest distance thought, which improves the planning efficiency of the algorithm. In addition, the method of regression analysis is adopted to improve the search ability of the algorithm and solve the local minimum problem caused by the heuristic value function. The Bézier curve fitting method is then used to generate a smooth path, which further reduces the jitter during the 6-DOF manipulator motion. The simulation experiment of improved RRT algorithm is established in MATLAB platform, and the effectiveness and rationality of the algorithm are verified.