AI-genererad bild på en robot
Photo: AI-genererad bild

The Importance of Gaze in Human-Robot Communication


Human communication extends beyond speech – gaze reveals much about our intentions and engagement. In a new thesis in computational linguistics at the University of Gothenburg, eye movements are investigated as a communicative phenomenon in human interaction, aiming to enhance dialogue with robots.

The more human-like behavior a robot/avatar displays, the more positively people perceive the interaction.

Gaze is a crucial nonverbal signal indicating attention and providing insight into others’ intentions and future actions. By observing different eye movements alongside speech, a new thesis explores the significance of various gaze patterns humans employ in conversation.

A study examines several interactions between individuals during a tasting session to study the relationship between laughter, speech, and accompanying eye movements.

"It demonstrated that laughter serves different functions linked to specific eye movement patterns. This underscores the importance of laughter and eye contact in understanding and coordinating multisensory communication, which is vital knowledge when designing human-like dialogue systems," says Vidya Somashekarappa, doctor in computational linguistics at the University of Gothenburg.

Forskaren framför en robot
Vidya Somashekarappa and the robot Furhat.

Promising Development Opportunities

The thesis also proposes a novel approach to estimate gaze using a neural network architecture considering the dynamics of real eye movement patterns during natural human interaction. This framework aims to facilitate responsive and intuitive interaction, enabling robots to communicate with humans through natural and multisensory dialogue.

The framework combines gaze prediction and prediction of the gaze target to generate a heat map. Evaluation shows superior results compared to previous methods, indicating promising opportunities for the development of robot interactions.

Vidya Somashekarappa.
Vidya Somashekarappa.
Photo: Monica Havström

Lastly, the thesis examines the significance of different eye movement patterns from a robot in interaction with humans.

"The results suggest that adapting the robot's eye movements based on patterns in human interaction positively influences user perceptions. It enhances engagement and the feeling of communicating with something human-like," says Vidya Somashekarappa.

The thesis "Look on my thesis, ye mighty: Gaze Interaction and Social Robotics" was defended in a dissertation on April 25.

Vidya Somashekarappa, email: