Fingers of human allow an object to be lifted using adequate grasping force and without slippage, even when the weight and friction coefficient of the object are unknown. Grasping force is controlled by detecting complex response of the skin using tactile receptors incorporated in the finger tissue. In this study, we propose a method for controlling the grasping force when objects are grasped by artificial elastic fingers in which many sensors are incorporated. First the relationship between the stick area and internal strain distribution of the finger is calculated in detail by using FE (finite element) analysis. The quantitative relationship between the stick area in the contact surface and the velocity of internal strain distribution is obtained. From these, we propose a method for controlling the grasping force by decreasing the increasing ratio of the tangential force when the stick area is decreasing. Finally, the grasping force is controlled by using the actual elastic finger made of silicorn rubber in which strain gages are incorporated. It is confirmed that objects can be grasped using adequate grasping force without complete slippage even when the weight and the friction coefficient of the objects are unknown.
17th RSJ Best Paper Award in 2003.