A Linear Affective Space Model based on the Facial Expressions for Mascot-Type Robots

Abstract

Although many researchers have tried to find out the bases of emotion in human mind, there is still no conclusion about that question. However, it is necessary to find out or establish the bases of emotion since they could be adapted to control facial expressions of robots effectively. This paper introduces a new linear model to define a linear affective space, which is based on the six basic emotional expressions of the face of a mascot-type robot. We attempt to determine which affective space is suitable for the facial robot, and define a linear expressional space. A new linear model is built up regarding relations between affects and facial expressions. This paper demonstrates efficient control of various and continual emotional expressions with a low dimensional linear affective space. The results are displayed through simulator experiments and mascot-type facial robot system