Relative Positioning of Multiple Robots Based on Panoramic Vision

This paper focuses on the problem of multi-robot localization based on vision systems. Between two localization methods, relative and absolute localization, the relative localization method is chosen to obtain robot position information. The paper uses a panoramic vision system to capture the color ring image information on the robot, and carries out feature point categorization based on an improved K-means clustering algorithm. As the pixel point values are not the actual target body distance values, it is necessary to obtain robot coordinate information by converting between the color ring pixel values and the real distance values. The paper concludes with a pilot-follower formation experiment to verify the effectiveness and the paper concludes with a pilot-follower formation experiment to verify the effectiveness and feasibility of this paper’s positioning method.