An underactuated gripper based on car differentials for self-adaptive grasping with passive disturbance rejection

We introduce an underactuated differential-based robot gripper able to perform self-adaptive grasping with passive disturbance rejection. The gripper utilises three car differential systems to achieve self-adaptiveness with a single actuator: a base differential for distributing power from motor to fingers, and two independent finger differentials for control-ling the proximal and distal joints. Linear and torsional springs are cleverly added to these differentials to allow the return of the fingers and the gripper-object system to equilibrium, thus enabling the gripper rejecting unexpected external disturbance forces applied to the fingers after securing a grasp. Moreover, the differentials allow the gripper to perform not only self-adaptive power grasps but also precision grasps, provide it with a large force transmission efficiency, and facilitate the prediction of grasping position. We analyse the static model of the introduced differential system and evaluate the gripper design via four sets of experiments. Numerical and empirical results clearly demonstrate the viability of the proposed grasper