Robust Image-Based Visual Servoing System using a Redundant Architecture

The reliability and robustness of image-based visual servoing systemshave recently received a growing interest. In order to address this issue, a redundantand cooperative 2D visual servoing system based on the information provided bytwo cameras in eye-in-hand/eye-to-hand configurations to control the 6 dof of anindustrial robot manipulator (Mitsubishi PA-10) is proposed. Its control law hasbeen defined to assure that the whole system is stable if each subsystem is stableand to allow avoiding typical problems of image-based visual servoing systems liketask singularities, features extraction errors, disappearence of image features, localminima, etc. Experimental results are presented which demonstrate the stability,performance and robustness of the system proposed.