Taylor

As robots enter human spaces and begin to work proximately with people, it is important that they understand human social interaction. They must be able to perceive human social signals and understand how to adapt to groups. The goal of our work is to design robot perception algorithms that allow robots to understand human group dynamics via social cues, and understand how to behave collaboratively in groups. In this paper, we discuss the current state-of-the-art of two fields that have contributed methods to achieve this goal, social signal processing and computer vision. We describe recent advances in these fields, as well as some of the challenges faced when adapting them to mobile robots.