A Bidirectional-RRT*-Connect-Assisted RRT*-Smart for a path planning algorithm

Abstract

Although Rapidly Exploring Random Tree Star (RRT*) has been considered to be able to achieve convergence to an optimal solution, this method has a slow convergence speed and requires an infinite amount of time to produce a truly optimal solution. For this reason, RRT*-Smart which includes path optimization and intelligent sampling processes was introduced. Although the addition of these methods can directly complete infinite-duration RRT* searches, they will work once the initial path obtained with RRT* is available. The effectiveness of reducing the optimality time is determined by the initial path formed. If this path is not close to optimal, the path optimization and intelligent sampling process will take a long time, and vice-versa. For this reason, RRT*-Connect, which has the advantage of searching from two directions, is proposed in this study. The goal is to replace the RRT* algorithm to produce a more optimal initial formed path. Based on this approach, this method will be named Connect-RRT*-Smart. Several methods, such as RRT, RRT*, RRT*-Connect, and RRT*-Smart, are compared to see their performance in producing the feasible path. Regarding this comparative result, the proposed method shows better performance in terms of convergence speed and path optimality.