Robot hand manipulation by evolutionary programming

Abstract

We propose a searching method of the grasping and manipulating an object in the 3D space by using a four-fingered robot hand, which is applicable to the real robot hand. To manipulate an object, it needs to teach contact points and forces of fingertip, but it is difficult to optimize each parameter for natural grasping. This paper shows the manipulation using evolutionary programming (EP) to optimize the motion of the fingers. In this case, EP determines not only when and where the fingers are moved, but also which finger is operated to the next contact point as the grasping attitude is changed. Moreover, the coding of individual is unique, which makes it easy to optimize the parameter. The result examines the effective in computer simulation in this paper but it would be applicable to a real robot hand.