An improved RRT* algorithm for robot path planning based on path expansion heuristic sampling

Abstract

Rapidly-exploring Random Tree Star (RRT*) algorithm and its variants based on random sampling can provide a collision-free and asymptotic optimal solution for many path planning problems. However, many RRT* based variants have low sampling efficiency and slow convergence rate in the environment which consists of long corridors, due to a large number of iterations are required in sampling critical nodes. To overcome this problem, the paper proposes the Expanding Path RRT* (EP-RRT*) based on heuristic sampling in path expansion area. By combining the greedy heuristic of Rapidly exploring Random Tree (RRT)-Connect, EP-RRT* quickly explores the environment in order to find a feasible path, and then expands it to obtain the heuristic sampling area. It iteratively searches in the heuristic sampling area which also changes with the continuous optimization of the path, and finally obtains an optimal or suboptimal path connecting starting point and target point. Comparisons of EP-RRT* with RRT* and Informed RRT* in four simulation environments verify that EP-RRT* improves the node utilization, accelerates the convergence rate, and obtains a better path for the same number of iterations.