Uncalibrated hand-eye coordination based HRI on humanoid robot

Abstract

In the field of human-robot interaction (HRI), the robot's hand-eye coordination ability is an important basis for the robot to achieve a natural interaction with human being. How to endow a humanoid robot with hand-eye coordination ability is the main content of this paper. In this paper, the issue is studied from the perspective of uncalibrated hand-eye coordination. Since humanoid robot has the nature of uncertainty and instability, the method of Kalman filter based online identification of image Jacobian matrix is applied, and an uncalibrated hand-eye coordination controller is designed for humanoid robot. Simulation shows the effectiveness of the method as well as its robustness to image observation noise. The HRI experiments on the humanoid robot NAO further illustrate the practical application value of the method.