Industrial implementation of a multi-task redundancy resolution at velocity level for highly redundant mobile manipulators

Abstract:
Mobile manipulators with many degrees of freedom (DOFs) have increasingly become of interest to industrial and service robotics due to their mobility and remarkable versatility. The redundancy of these robots can be conveniently exploited to generate internal joint motions that reconfigure the structure according to given task specifications. The availability of many DOFs allows the robot to execute various tasks at a time. Therefore, an efficient redundancy resolution becomes crucial to achieve multiple tasks simultaneously and use the capabilities of the complete robot system. Tasks can be of different nature and defined in different spaces; relative priorities among tasks are strictly enforced which is appropriate for situations that cannot tolerate compromises as in industrial environments. Null-space based behavioral approaches have been deeply investigated at velocity level as it is the core of the proposed control strategy. We implemented an effective priority strategy approach using a closed-loop version of the inverse kinematics of the robot together with Damped Least-Squares operators on Jacobian matrices. Different tasks with different priorities have been specified for our robot mimicking a typical industrial manufacturing setting. Simulations and real experiments on our 12 DOF mobile manipulator have been conducted to validate the implemented technique.