Communication learning for cooperation among autonomous robots

Communication among robots is necessary for autonomous robots to execute tasks cooperatively. Previous researches related to cooperative behavior among robots have largely depended on a predefined action or communication system.. However, it is difficult to set a fixed system., and more difficult for the agents to adapt to novel situations or tasks in a dynamic environment. Against these conventional fixed systems, new systems such as an emergence scheme have been proposed. In these systems, agents are required to understand the environment through experiences interacting with the environment, and adapt to it. We describe a method for robots to learn appropriate actions including communication to adapt to the environment without giving a fixed behavior or communication system precedency