Master-slave mapping and slave base placement optimization for intuitive and kinematically robust direct teleoperation

This paper presents a technique to map two dissimilar master-slave devices with similar, human-like workspaces for telemanipulation. The mapping focuses on achieving full geometric pose correspondence between the operator and the slave device while avoiding reaching into areas close to singularities and joint limits. The approach followed can be divided in 4 steps, namely: master workspace offset and scaling adjustment to ensure that the operator always controls the manipulator within reachable areas; optimized mounting of the slave end-effector to eliminate manipulator wrist singularities; redundancy mapping using the arm angle criteria for geometric pose correspondence; slave manipulator base placement optimization to place the task-space area in the zone of highest manipulability. A setup using the EXARM exoskeleton and a virtual model of the Kuka 7-DOF Lightweight Robot is used to show that the proposed method generates geometric correspondence and places the operation within areas far from joint limits and singularities for the human arm workspace, thus ensuring robust tracking of human operators in real-time for arbitrary trajectories inside the task-space.