Behaviour-based Multi-Robot Formations using Computer Vision

L. Payá, A. Gil, O. Reinoso, M. Ballesta, and R. Ñeco (Spain)


Tracking, multi-robot systems, formations, behaviours, control schema and fuzzy logic


Multi-robot systems is a very attractive field of research in robotics, due to the fact that a team of robots is able to accomplish certain tasks that a single robot cannot carry out on his own. This paper presents a behaviour-based method for multi-robot formations using computer vision. A team of robots must move around the environment, keeping a relative position between them and avoiding the collision with possible obstacles. Computer vision has been used for the localization of the robots using the images captured by the camera they carry on their top. Different methods have been tested to obtain the leader robot pose, developing finally a very stable method to calculate distance and orientation. The final control has been built as a composition of several basic behaviours that calculate an action model using the information of the environment. These actions are composed using fuzzy logic to obtain a high-level behaviour. The algorithms developed have been tested over a team of WiFiBot mobile robots, with good results.

Important Links:

Go Back