Abstract
The rise in online interactions has introduced multiple challenges, including confusion during virtual meetings and fatigue associated with prolonged video conferencing. To address these issues, this study advocates using computer graphics (CG) avatars in dialogue systems that do not rely on camera feeds. Avatars have the potential to diminish gaze-related misunderstandings and reduce dissatisfaction arising from viewing oneself in video calls. Moreover, fatigue from continuous face-to-face interaction in multi-person conversations can be mitigated by avatar usage. Previous studies indicate that avatars' verbal and nonverbal communication play significant roles in facilitating social support and cooperation in online environments. In the field of robotics, nonverbal cues including movement and spatial positioning are crucial for improved communication. For instance, robotic movement can indicate shifts in attention and speaker preferences during human-robot interactions. However, the impact of movement in dialogue contexts, especially in comparison to a neutral stance lacks thorough evaluation. This study conducted a video-based experiment to assess the impact of a robot's approach movement on the perception of emotions and intentions in a multi-robot interaction scenario. The results suggest that approaching movement can enhance the expression of intentions and emotions, indicating desired turn-taking and conveying perceived impressions of a robot as positive or negative. Future research will focus on implementing these findings in real-time conversational experiments.