Study on dynamically reconfigurable robotic system (13th report, optimization of group robotic system behaviour in dynamic environment)

This paper deals with the problem of optimizing the behaviour of a group robotic system, such as the Cellular Robotic System, in dynamic environments. The Cellular Robotic System(CEBOT) is an autonomous distributed robotic system composed of a large number of robotic units called 'cells'. Since the CEBOT is a group robotic system, cooperation and group organization are required for task execution. In this paper, we propose the concept of the 'self-recognition' of the decision-making in the group behaviour. To make a decision, each autonomous robot constructs an inner model based on the recognition of environment and the robots' state. If each robot behaves based on the inner model, which is constructed without consideration of the group behaviour in the group system, deadlock problem will occur. Therefore, each robot is required to consider the influence of self-action on other robots and group system. The concept of the 'self-recognition' is derived from consideration of this influence. We propose the learning and adaptation strategy for group behaviour based on the concept of 'self-recognition' and show its effectiveness through the simulation results.