Speech-driven pupil response robot with high-arousal backchannel for active listening

In primitive communication where talkers are divided into speaker and listener, not only verbal messages but also nonverbal behaviors such as facial expressions, body movements, prosody, gaze, and pupil response are mutually synchronized. This synchrony among nonverbal behaviors creates a sense of unity and emotional empathy in embodied communication. Previous studies have reported that an active listening attitude by the listener induces feelings of acceptance and empathy in the speaker. Therefore, it is expected that speaker’s emotions and self-disclosure can be drawn out by designing an attentive listening attitude of a listener. This study developed a speech-driven pupil response robot that generates high-arousal backchannels based on both a nodding response model and an interaction-activated communication model. The robot expresses an attentive listening attitude by producing emotionally aroused backchannels according to the rhythm and activation level of the speaker’s speech. The effectiveness of the developed robot was demonstrated by sensory evaluations.