The kinematic roadmap: A motion planning based global approach for inversekinematics of redundant robots

This paper proposes a novel and global approach to solving the point-to-poi nt inverse kinematics problem for highly redundant manipulators. Given an i nitial configuration of the robot, the problem is to find a reachable (path -connected) configuration (and a feasible path to it) that corresponds to a desired position and orientation of the end-effector, Our approach is insp ired by recent motion planning research and explicitly takes into account c onstraints due to joint limits, self-collisions, and static obstacles in th e environment. Central to our approach is the novel notion of kinematic roa dmap for a manipulator. The kinematic roadmap captures the connectivity of the connected component (that contains the initial configuration) of the fr ee configuration space of the manipulator in a finite graph like structure. The point-to-point inverse kinematics problem is then solved using this ro admap, We provide completeness results for our algorithm, Our current imple mentation, based on Ariadne's Clew Algorithm [5], [32], is composed of two sub-algorithms: EXPLORE, an appealingly simple algorithm that builds the ki nematic roadmap by placing landmarks in the free configuration space; and S EARCH, a local planner, that uses this roadmap to reach the desired end-eff ector configuration. Our implementation of SEARCH is an efficient closed fo rm solution, albeit local, to inverse kinematics that exploits the serial k inematic structure of serial manipulator arms. Initial experiments with a s even-DOF manipulator have been extremely successful.