A quaternion-based regulator for cooperative manipulators

The problem of controlling a system of two cooperative manipulators tightly grasping a common object is considered. Based on a suitable formulation of the cooperative task space, a novel regulator is devised which is based on two individual regulation schemes for each manipulator. Internal forces at the held object are cancelled via kineto-static filtering of the control action, while force regulation is achieved through force feedback. In order to overcome the problem of representation singularities, end-effector orientation errors are computed through quaternions. The proposed regulation strategy is experimentally tested on a laboratory setup consisting of two industrial manipulators with open control architecture.