“…Our focus in this paper is on robots with uncertainty in their motion and state estimation; we do not consider uncertainty in sensing of obstacle locations (e.g., [11]) or grasping (e.g., [20]). Extensive prior work has investigated motion planning under uncertainty for mobile robots that can be approximated as points or spheres in the workspace, e.g., [26,17,18,27,2,21,19,4,8]. For point or spherical robots, computing an estimate of the probability of collision with obstacles can be done in the workspace since the geometry of the C-obstacles is low dimensional and can be directly computed.…”