Robot formations using local sensing and limited range vision

This paper addresses the problem of developing intelligent multi-robot team capable of collaborating to achieve team objectives. The proposed work focuses on robot formations and how robot team-mates could face many complex situations. In order to do this a new vision technique is proposed. There is neither global information nor a predefined team leader available. A behavior based controller is proposed and implemented to give the robot teammates the ability to cooperate in constructing different robot formations. A simulated team of robots based on the well known robot simulator Webots is constructed to validate the proposed approaches. Different robot formations´ experiments are conducted. The experiments show also how to the robot team could solve many difficult tasks such as large swerving angles or passage through narrow corridors.