Abstract:

In order to improve the universality of the traditional visual servoing control methods based on image fea- tures and to reduce the complexity of the marking, extracting and matching of image features, a novel plane visual servoing control method is proposed based on artificial immunity and direct image feedback. In this method, first, the translation and rotation errors between the reference and feedback images are obtained by employing an artificial immune algorithm. Then, the position and orientation increments transformed from the image errors via a visual controller are transmitted to the robot controller, which implements the closed-loop visual servoing control without marking and extracting any image features. Experimental results indicate that the proposed method makes the robot accurately track objectives with an error under one pixel, showing its effectiveness and correctness

Key words: artificial immune algorithm, direct image feedback, robot, visual servoing