Controlling Formations of Robots with Graph Theory

A number of techniques that allow autonomous multi-robot systems to be held within formation-like structures exist but they are limited by a high communication load, high energy usage and a lack of robustness. This research improves on state-of-the-art formation control schemes for leader-follower type multi-robot systems by employing mechanisms that enable groups of robots to move in two-dimensional formations without the need for inter robot communication. We also incorporate techniques that enable the robots to move back into formation in a precise manner when external interferences have caused the formation shape to deteriorate. The control system is derived through the use of graph theory and has been tested rigorously in a realistic simulator to prove its applicability to multi-robot control.