Journals Proceedings

Abstract

Sampling-based motion planning algorithms have been proven to work well with difficult planning tasks in a variety of problems. Recently, asymptotic optimal algorithms have been proposed to overcome the non-optimality inefficiency of these planners but with extra computational costs associated with the additional processing requirements. In this paper, new extensions of optimal sampling-based motion planning algorithms are presented which overcome this drawback by utilizing the Poisson-disk sampling distribution. The proposed planners replace the original uniform sampling with the Poisson-disk sampling by defining a sampling radius along with the neighborhood radius in the original optimal planners. The main advantage of the proposed planners is their ability to reach different levels of optimality with fewer sampling attempts which reduces the running time of the planner significantly. The proposed algorithms have shown to solve different motion planning tasks with considerably smaller set of samples. The simulation studies have been conducted and support the superiority claim of the proposed algorithms.