A proposal of sensor fusion in cooperative system

The main problems in visual servo system are long sampling period and extent of visual field. For sampling period of camera used in general is longer than joint servo period, hidden oscillation is controversial. Also, if a object is out of visual field, motion control is impossible. This paper discusses such problems, taking assist mobile robot system for application. An assist mobile robot which mounts a camera supports a working mobile robot with various tasks including delivery and receipt of baggage and removal of obstacles. At the first step of this study, an assist mobile robot is controlled to track a working mobile robot. Here, It doesn´t matter whether transmission facility between an assist mobile robot and a working mobile robot is available. This study proposes virtual image plane by sensor fusion of vision and encoder. Virtual image plane is developed by estimated feature. And estimation and correcting of image feature on virtual image plane are realized. In conclusion, motion control with disturbance observer on virtual image plane is designed. And the validity of proposed method is shown in simulation results.