Neural Network-Based Learning Impedance Control for a Robot

In this paper, a neural network-based learning approach for robot impedance control is presented to accomplish a contact task. Firstly, a discrete-time impedance control algorithm is obtained to control the contact task of the robot. Secondly, on-line learning algorithms based on a new evaluation function are developed for the neural networks which adjust the inertia, damping and stiffness parameters of the robot in order to adapt it to the unknown contact environment. Thirdly, experiments are carried out and the effecttiveness of the present approach is verified by pressing a spring using a 6 degrees of freedom robot. The adaptiveness, stability and flexibility of the present approach are also confirmed by the experimental ersults.