I. Introduction
The canonical motion planning problem involves computing a valid path for a robot to move from a given start to a given goal state while respecting a set of physical constraints [1]–[3]. This problem is motivated by an ever-growing number of practical applications such as autonomous exploration, search-and-rescue, robotic surgery, and warehouse management just to name a few. As robots become more mobile, articulated and dexterous, it is important to find not only a feasible plan for the robot, but also one that optimizes one or more criteria for a given high-level task. The quality metric is problem-specific, and this paper will consider optimizing path length. The method described, however, is applicable to many different metrics like smoothness or obstacle clearance.