Artificial emotion of face robot through learning in communicative interactions with human

摘要

It is pointed out that human-robot interface or human interface must be newly interpreted as intelligence of communicative interaction between robot and human, and is again a key issue for development of a new species of robot that is able to serve humans. We point out a concept of "active human interface (or media) " to be composed of three functions and essential to such new robots. When applying such machine to use for human service, psychological familiarity of robot form is another issue to ensure the psychological acceptance. We refer again to the existence of uncanny valley in familiarity to robot appearance or form. We have been developing a life-like face robot that has been a human media for realizing such intelligence of communicative interaction and technological issues and performance of face robot's recognition function of human emotion and actuation function of facial expressions are briefly summarized to understand the state-of-the arts about the face robot Mark II. Then basic studies for characterizing face robot behavior through communicative interactions with a human partner are described by showing two test results of imitation of facial expressions and personality creation in face robot response. We point out the importance of learning method employed by face robot in communicative interactions is also key issue. We can show that the most suitable learning method for human partner is OR type supervised reinforced one. Finally we discuss the value system in relation to unsupervised learning of face robot in communicative interaction to characterize its emotion or selection criteria or bias at taking a certain response facial expression to the partner's state of mind. At the end concluding remarks and future studies are briefly pointed out.