Planning and control system for an anthropomorphic grippers

This paper presents a complete planning and control strategy to the manipulation of objects and tools using an anthropomorphic gripper. To approximate, in an efficient way, the functions accomplished by a human hand, the controller was conceived taking into consideration: the gripper geometry, the gripper elastic characteristics, the object geometry and the intermittent contacts possibility. To take the gripper and object characteristics into consideration the screw theory was adopted. The screw theory allows the analysis in the three-dimensional Euclidean space of six-dimensional complex problems. The planning task is accomplished using an artificial neural network based system, suitable for real-time application.