Development of behavior and emotion models for Conbe-I using a Self-Organizing Map learning

We are attempting to develop a robot with a “consciousness” resembling that of a human or an animal in order to enhance an intellectual activity of the robot. In our previous study, we have constructed a model of mechanism of consciousness and behaviors based on the relationship between the action strategies and consciousness levels, this software architecture is called Consciousness-Based Architecture (CBA) module. In addition, the robot should select the behavior and express the emotion by itself, which depends on its surrounding environment. Therefore, in this paper, we have examined the intrinsic motivation and the external situations of robot are used to analyze and classify the behavior and the emotion by a Self-Organizing Map (SOM) that is trained using unsupervised learning, and we attempt to describe the integration of a behavior map and an emotion map into a CBA module. The conscious behavior robot (Conbe-I) has been implemented to show the effectiveness of the proposed system.