Variable Sampling Domain and Map Compression Based on Greedy RRT Algorithm for Robot Path Planning

Abstract

The traditional rapidly-exploring random tree (RRT) algorithm and the improved algorithms use a fixed sampling domain in each iteration. This way of sampling will result in much repeated exploration. Also, a fixed step causes the inefficiency of expansion in open areas and the incremental algorithm consumes much time in collision detection. In order to solve these problems, this paper proposes an improved RRT algorithm based on a variable sampling domain and a map compression algorithm along with a strategy of greedy growing (VSDGC-RRT). A variable sampling domain and the greedy strategy increase the utilization of the sample node and reduce the number of iterations. The map compression algorithm reduces time consumption in collision detection. These improvements can make the tree expand more rapidly to the target. The simulation results indicate that the improved algorithm can effectively reduce the quantity of iterations and take less time while retaining the probabilistic integrity of the traditional RRT.