Adapting probabilistic roadmaps to handle uncertain maps

Randomized motion planning techniques are very good at solving high-dimensional motion planning problems. However, most planners assume complete knowledge of the environment, an assumption that can lead to collisions if there are errors in the world model due to uncertainty. We propose an extension of the probabilistic roadmap algorithm that computes motion plans that are robust to uncertain maps. We show that the adapted PRM generates less collision-prone trajectories with fewer samples than the standard method