An asymptotically-optimal sampling-based algorithm for Bi-directional motion planning

Abstract

Bi-directional search is a widely used strategy to increase the success and\nconvergence rates of sampling-based motion planning algorithms. Yet, few\nresults are available that merge both bi-directional search and asymptotic\noptimality into existing optimal planners, such as PRM*, RRT*, and FMT*. The\nobjective of this paper is to fill this gap. Specifically, this paper presents\na bi-directional, sampling-based, asymptotically-optimal algorithm named\nBi-directional FMT* (BFMT*) that extends the Fast Marching Tree (FMT*)\nalgorithm to bi-directional search while preserving its key properties, chiefly\nlazy search and asymptotic optimality through convergence in probability. BFMT*\nperforms a two-source, lazy dynamic programming recursion over a set of\nrandomly-drawn samples, correspondingly generating two search trees: one in\ncost-to-come space from the initial configuration and another in cost-to-go\nspace from the goal configuration. Numerical experiments illustrate the\nadvantages of BFMT* over its unidirectional counterpart, as well as a number of\nother state-of-the-art planners.\n