Calculation of grasping force to adapt mass variation of object considering human motion

This paper proposes a way to calculate a grasping force that could adapt to a variation of mass of a grasping object. In order to grasp an object, resultant force vector consisted of grasping force and friction force has to be within a friction cone. In conventional method to grasp an object, mass of a grasping object or acceleration caused by transportation have to be measured in order to calculate an optimal grasping force. However, it is difficult to measure these parameters quickly; therefore grasping an unknown object is difficult. The proposed method is a data-driven approach using bilateral control. While human grasps an object, the condition that a resultant force vector is within the friction cone, then the direction of the resultant force vector in the stored human motion are reproduced so as to control grasping force. By using the proposed method, it is able to grasp an object without any need to measure the mass of the object. Experiments are conducted in order to show the validity of the proposed method.