Chinmaya Mishra defends thesis on 17 April

Link to live-stream

With the rapid advancements in artificial intelligence and robotics technologies, social robots are poised to have greater integration in society. These robots are designed specifically to conduct human-like interactions. So, understanding and replicating essential non-verbal cues, such as facial expressions and gaze, are essential for enhancing the effectiveness, human-likeness, and acceptance of these robotic systems. Social robots are already being employed in a variety of domains, including healthcare, education, and assistive roles, where their capacity to convey and interpret human emotions and intentions can significantly impact the quality of interactions. Modeling non-verbal behaviors on these robots would make them more capable of providing a richer user experience. For example, a social robot designed to provide companionship to the elderly could express happiness when the user is cheerful and sadness when the user is upset, enhancing emotional connection, or look directly at the user when speaking, creating a more personalized and attentive interaction.

This research investigates methods for making human-robot interactions (HRI) more seamless and human-like by modeling non-verbal behaviors on social robots and is centered on two key areas: developing architectures to model the eye gaze and emotional behaviors of social robots and evaluating the human perception and influence of these behaviors during HRI.