A fuzzy intimacy space model to develop human-robot affective relationship

Abstract

This paper presents a novel method to establish human-robot affective relationship for long-term interaction. At first, in order for a robot to estimate human intimacy, a fuzzy space model to formally represent human intimacy is proposed. Proxemic, tactile and oculesic behavioral features which are dominantly used for intimacy exchange in human-human communication are analyzed for developing a 3 dimensional intimacy space between a human and a robot. Fuzzy Min- Max neural network is utilized to develop a general intimacy space from the training data obtained during human intimacy expressions. In second, criteria for designing a robot intimacy are provided according to the types of social robots. Our affective relationship model is being developed to make a robot differently respond to different users with a variety of emotions and therefore increase the believability of a robot.