評価関数形ニューラルネットワークによる3本指のは持制御に関する研究

概要

論文の詳細を見る
Many grasping control methods for multifingered endeffector, such as the compliance control method or hybrid position/force control method, have been studied. There still remain some problems in the subtle control of finger motion as in human fingers. In this paper, we describe the grasping motion of a 3-fingered endeffector with each finger having 4 degrees of freedom. In order to derive the dynamics of the system, we formulated the equation of motion for 3 fingers grasping a stiff object with their tips. Then the neural network controller, which has a criterion function learning algorithm, is applied for the 3-fingered endeffector motion control. Simulation results show the effectiveness of this control method for grasping, and finger motion is characterized by a properly designed criterion function in the NN control.
一般社団法人日本機械学会の論文
1993-09-25