Self-organizing multi-robot system based on personality evolution

Abstract

Coordination of the multi-robot system is an important problem to system efficiency. Traditional methods based on both planning and learning have limited adaptability in complex environments for the needs of the top-bottom supervisor message. Self-organization is a very effective bottom-up design approach, but most previous discussions have focused on the concept level. Some concrete realization methods still need to be found. It can be seen that the same worth orientation and decision tendency of each robot is one of the reasons for the resulting conflicts in the usual multi-robot system. So it is necessary to introduce a "personality". We represent personality with the evaluation function and give a personality evolution mode. With a different decision made by a different personality, the self-organizing coordination of a multi-robot system is realized. Simulations show the validity of the idea.