Connecting with emotionally competent robots

Today’s robots can lift heavy objects, navigate busy streets, and perform complex routines. But connecting with humans on a social and emotional level is still a major challenge. Bahram Salamat Ravandi is a PhD student in Cognitive Science. His work tackles this gap by developing a framework that explains how engagement between humans and robots works as a dynamic process. The dissertation introduces a deep learning model that can detect how engaged a person is in real time. This means a robot could notice when someone loses focus, becomes frustrated, or needs encouragement and adjust its behavior accordingly.

"We’re entering a new era where robots don’t just follow commands but act as partners that can truly engage with us, emotionally, mentally, and socially. This direction opens enormous possibilities," says Bahram Salamat Ravandi.

Experiments conducted as part of the research reveal important insights. One is that emotional feedback from a robot can make people feel more socially connected to it. Task-focused feedback, on the other hand, boosts performance and accuracy. Understanding when to use each type of feedback could help designers create robots that both feel supportive and help people perform better, especially in demanding settings such as therapy or education.

The dissertation also points to broader possibilities. Imagine a tutor robot that adapts to a student’s motivation, a healthcare assistant that reacts with empathy, or a car system that helps drivers stay alert without being intrusive. These technologies are not meant to replace human connection, the research emphasises, but to strengthen it.

Bahram Salamat Ravandi will defend his dissertation on Thursday 11 December 2025.

Text: Agnes Ekstrand