Goal-biased Bidirectional RRT based on Curve-smoothing

Abstract In this paper, a goal-biased bidirectional Rapidly-exploring Random Trees (RRTs) algorithm based on curve-smoothing is newly proposed. The main contribution of this work is that the two-parts of the rapidly-exploring random trees generated in the bidirectional search process are smoothly connected by Bezier curves, so that the whole path satisfies kinematic constraints. Comparative experimental results with the naive RRT algorithm are presented to demonstrate that the proposed algorithm can achieve superior performance in terms of higher success rate, shorter search time, shorter path length and fewer number of the search nodes. Finally, in order to simulate the motion of the robot in a real environment, we track the trajectory through a controller under the visual robot simulation platform V-Rep.