Cooperative Visual Servoing System Nicol´asGarc´ ıa, M. Asunci´on Vicente, Carlos P´ erez, C´ esarFern´andez,and Rafael Puerto Dpto. Ingenier´ ıa, Miguel Hern´ andez University, Avda. Ferrocarril s/n, 03202 Elche (Alicante), Spain nicolas.garcia@umh.es http://lorca.umh.es Abstract. A simple and efficient control law that combines the infor- mation of two cameras in order to realize the positioning task of a robot end-effector is presented in this paper. One of these cameras is rigidly mounted on the robot end-effector(eye-in-hand configuration) the other one observes the robot within its work space(eye-to-hand configuration). The aim of this work is to take advantage of both in a cooperation scheme. The control architecture of this system is presented. Compari- son between simulations and experimental results with only one camera and with the two cameras in a cooperative way are shown. 1 Introduction Nowadays, the great majority of robot population operates in factories where the work environment is structured and previously well-known. The application of a robot to carry out a certain task depends, in a high percentage, on the previously knowledge about the work environment and object placement. This limitation is due to inherent lack of sensory capability in contemporary commercial industrial robots. It has been long recognized that sensor integration is fundamental to increase the versatility and application domain of robots. One of these sensor systems is Computer Vision. Computer vision is a useful robotic sensor since it mimics the human sense of vision and allows for non contact measurement of the work environment. In- dustrial robot controllers with fully integrated vision systems are now available from a large number of suppliers. In these systems, visual sensing and manip- ulation are typically integrated in an open-loop fashion, looking then moving. The precision of the resulting operation depends directly on the accuracy of the visual sensor and the robot end-effector. An alternative solution for the position and motion control of an industrial manipulator evolved in unstructured environments is to use the visual informa- tion in a feedback loop. This robot control strategy is called visual servo control or visual servoing. Visual servoing systems have recently received a growing interest, as the computational power of commercially available computers became compatible with real time visual feedback [9], [3].