Teleoperation of a vision-based mobile robot under office automation floors

This paper discusses human interface between a human operator and a vision-based mobile robot used under office automation floors. First, we explain the hardware of tele-operation system of the mobile robot, and the display system of the measured information of the robot. There are many supporting poles in the under floor, and the robot should estimate the self-location according to the layout of the surrounding poles. We apply a K-means algorithm and a steady-state genetic algorithm (SSGA) for extracting poles from an image. Next, we apply the proposed method for the control of a mobile robot used under an office automation floor. Furthermore, we propose an attention mechanism to the human operator. Experimental results show the effectiveness of the proposed method.