Dynamic behavior arbitration of autonomous mobile robots using immune networks

Conventional artificial intelligence systems have been criticized for their brittleness under dynamically changing environments. Therefore, in recent years, much attention has been focused on reactive planning systems (e.g. behavior-based AI, emergent computation, and so on). In this paper, we propose a new inference/consensus-making system inspired by immune systems in living organisms, and we apply our proposed method to the behavior arbitration of an autonomous mobile robot as a practical example. Furthermore, we try to evolve affinities among antibodies using genetic operators. To confirm the validity of our method, we carry out some simulations.