Human-Robot Interaction Using Markovian Emotional Model Based on Facial Recognition

Abstract

Interactive emotion communication (IEC) is a research on the communication between human and robot with emotional behaviors. In this paper we propose a method of IEC for emotion generation using Markovian emotional model (MEM) based on the difference of the characteristic quantity in the facial expression of human by self-organizing map (SOM). For example, it is thought that an angry person bursting into tears is caused with high possibility than the angry person bursting into laughter when the emotion of person changes. In other words the expression which a characteristic resembles is easy to change as emotions. In this study, the emotion transition probability is found by the questionnaire, and an interaction experiment with the proposed emotion transition is carried out using a communication robot.