This paper presents a visual servoing method using the stereo camera to control the manipulator pose (position and orientation) with respect to an object pose (position and orientation). To represent the object pose with respect to cameras mounted at the end of manipulator, we use the virtual object points. Based on the translation and the rotation operation applied to a predefined template points, the virtual object points are generated. The translation operation is determined from the single feature point that represents the estimation of the object position. The rotation operation is determined from the virtual angle calculated from the difference of the object area and the reference model area. The control signal to move the manipulator is derived from the error signal between the virtual object points and its desired values. Experiments conducted on a 3-links serial planar manipulator demonstrated the performance of the proposed method.
ASJC Scopus subject areas