An eye-in-hand stereo visual servoing for tracking and catching moving objects

Abstract

In this paper an image-based eye-in-hand stereo visual servoing system is proposed to perform a task of tracking and catching a moving object in real time. The presented method utilizes the conventional image based visual servoing algorithm for a stereo vision system mounted on the end-effector of 6 degrees of freedom (DOF) robot. In order to estimate the object motion trajectory in two dimensional image planes and predict the future positions, a Kalman filter with a linear model of the moving object and also an extended Kalman filter (EKF) with a non-linear model are employed. The procedure of tracking and catching a moving object is simulated and implemented based on a 6 DOF DENSO 6242G robot. Both simulation and experimental results are presented to verify the effectiveness of the proposed methods. The comparison with the case of a monocular image-based system is carried out.