A probabilistic roadmap based planning algorithm for wheeled-tip robots manipulating polygonal objects

摘要

This work extends the probabilistic roadmap (PRM) framework by efficiently incorporating the force closure (FC) criterion into motion planning. A new FC verification approach is proposed that eliminates the need for testing FC in the intermediate configurations of the local planner. The algorithm is tested for special kinds of planar robots that have two wheeled-tip fingers. The wheels are actuated and can roll along the object's edges. We propose a general planning algorithm for this new mechanism without any limitations on the fingers structure and the task. The algorithm is implemented and the simulation results in ADAMS verify the validity of our algorithm.