Abstract
We present a cooperative formation control method for multiple mobile robots based on visual information. Each robot needs recognize other robots for cooperative tasks. In order to reduce false recognition of the robots, we propose the detection of the robots by fusing visual information. In this paper, each robot estimates the center of gravity of tracked robots from the image data captured by equipped camera based on a particle filter. Then, the robot determines the velocity command so as to track other robot using the image-based controller and the estimated image positions. Experimental results using real three robots demonstrate the effectiveness of the proposed method.
