Behavior selection system based on emotional variations

Abstract

Designing robots with cognition and consciousness resembling that of human or animal has become an important application of intelligent autonomous robot, in order to achieve a more effective human-robot interaction. On the grounds of dynamic system, the primary structure of a behavioral-emotional model of the robots consists of three processes: recognition process, cognition process and behavioral-emotional expression process. Therefore, in this paper, we propose a behavioral-emotional selection model for a conscious behavior robot (Conbe-I) based on the Self-Organizing Map (SOM) learning and the discrete stochastic state-space mathematical model (Markov model) that mainly consider the issues of an autonomous action selection corresponds to the emotional state transition and the emotional expression of an autonomous robot. Finally, the experimental results show the efficiency of the proposed system.