Classification of Human Grasping Posture Suitable for Catching and Its Application to Robotic Hand

Abstract

Although humans and robots are capable of catching a ball, it is difficult to catch a ball with a myoelectric prosthetic hand, which is a man-machine cooperative system. The purpose of this study is to observe the catching motion of humans, identify the suitable hand posture for the catching a ball, and apply it to the posture of the robotic hand. As a result of an experiment in which human subjects caught a free-falling ball, the catching postures were classified into five types. It was concluded that the catching posture with the highest catching success rate was the one in which the ball was touched distally to the palm, and that the oppositeness of the thumb and little finger was important. As a result of applying the identified catching posture to the robotic hand, the success rate of the catching a ball with the robotic hand increased dramatically. These results suggest that the user can perform the catching motion by using robotic hand for powered prosthetic hand with opposable thumb and little finger.