Hierarchical roadmap based rapid path planning for high-DOF mobile manipulators in complex environments

The probabilistic roadmap method (PRM) has been widely used in the field of robot path planning and based on it, a great number of variants have been developed addressing for different purposes, e.g. the lazy PRM method, the dynamic roadmap (DRM) method, and the dynamic bridge builder method (DBB). In this paper, PRM is extended as well-a hierarchical roadmap based rapid path planner is presented, which combines DRM with DBB to achieve good planning performance in environments including moveable obstacles. Through observation, we find that how to determine precise location of a robot's end-effector is quite important for nearly all practical planning tasks. Moreover, the workspace that can be swept by its end-effector determines scope of action of a robot. In consideration of the safety factor, the velocity of planning should be limited. Accordingly, its scope of action must be limited during a certain time interval, as well. In this paper, a hierarchy sampling strategy oriented towards the end-effector position is employed. Initially only a fundamental global roadmap is constructed and visible. Then the scope of a manipulator, called manipulator space in this paper, is computed. Moveable obstacles lying inside the manipulation space are used to update the global roadmap and activate other two layers of nodes with dynamic bridge builder. The approach has been implemented and tested in simulation, and results show that it meets the real-time demand well.