EvBots - The Design and Construction of a Mobile Robot Colony for Conducting Evolutionary Robotic Experiments.

Abstract

This paper describes a method for building small, inexpensive, autonomous mobile robot systems in order to study into robot colonies. We describe how to build EvBot mobile robot colonies that can be used to navigate through mazes of varying complexity, display intelligent control, produce evolutionary computing algorithms, and use low-bandwidth distributed RF networks to communicate. Experimental results show that EvBots can navigate through a real-world maze using an evolutionary algorithm based on complex fuzzy neural paradigms. These algorithms were first developed in an equivalent simulated maze and transferred to the EvBot operating in the real world. On-going experiments are being conducted using shared learning between individual EvBots in a colony, as well as shared learning in complex robot colony task cooperation.