Motion Planning for Multilink Robots by Implicit Configuration-Space Tiling

We study the problem of motion-planning for free-flying multilink robots and develop a sampling-based algorithm that is specifically tailored for the task. Our approach exploits the fact that the set of configurations for which the robot is self-collision free is independent of the obstacles or of the exact placement of the robot. This allows for decoupling between costly self-collision checks on the one hand, which we do off-line (and can even be stored permanently on the robot's controller), and collision with obstacles on the other hand, which we compute in the query phase. Our algorithm suggests more flexibility than the prevailing paradigm in which a precomputed roadmap depends both on the robot and on the scenario at hand. We demonstrate the effectiveness of our approach on open and closed-chain multilink robots, where in some settings our algorithm is more than fifty times faster than commonly used, as well as state-of-the-art solutions.