Autonomous vision-based pose control of mobile robots with tele-supervision

In this research, a feedback control strategy based on tele-supervisory autonomous vision based pose control is proposed. In this system, client-server architecture is adopted. Based on the approach, client user can directly command the mobile robot, equipped with binocular vision on the server side, to accomplish pose control task with precision. Specifically, according to the binocular images observed by the mobile robot and transmitted from the server, client user can determine the desired pose of the mobile robot with respect to a particular target. Moreover, a visual servo control law synthesized on the server is able to autonomously drive the mobile robot to the desired pose with precision. The mobile robot together with the server is able to perform autonomous visual servoing tasks, while task commands are issued from client user based on observed binocular images via the Internet. Under the proposed tele-supervisory architecture, client user not only can directly manipulate the mobile robot via the Internet, but also can command and continue supervising the mobile robot to autonomously accomplish the desired tasks with precision. Therefore, precision control can be achieved without considering the propagation delay in the Internet. The proposed tele-supervisory visual servoing system is successfully validated through simulations and experiments. It can be further applied to vision-based security systems and tele-operation in hazardous environment.