Statistical motion planning for mobile robots: paradigms for efficiency and optimality using a discrete model

Abstract

In recent papers we have proposed a new path planning approach for mobile robots, which employs statistical data in order to plan efficient paths in environments with moving obstacles. In this paper, a more systematic insight is offered using a simplified model with discrete space and time. We present complete algorithms yielding globally optimal solutions without requiring any heuristics. Regarding the reactive behaviour, benefits of statistical motion planning are specified in terms of reliability and expected time to reach the given goal positions.