Method for behavior arbitration in autonomous mobile robots using emotion mechanisms

The conventional AI (artificial intelligence) technology has been criticized for its many drawbacks, one of which is brittleness under hostile dynamically changing environments. To overcome this problem, other approaches, such as behavior-based AI, new AI, emergent computation and animat approach, have been proposed and their usefulness has been confirmed. Since computational ability of mobile robots is limited in general, it is required that robots arbitrate output behaviors appropriately using obtained sensory data. On the other hand, in living organisms, there are two systems for coping with a dynamically changing environment: 1) the immune system and 2) the emotion system. Based on this, we present a new method for behavior arbitration in autonomous mobile robots by paying close attention to the emotion system in living organisms. We confirm the feasibility of our proposed method by applying it to an obstacle avoidance problem of a mobile robot as a practical example.