Our research draws on a wide toolkit: social robots, physiological sensors, and AI driven software and hardware. Combining these tools lets us study human-robot interaction with a precision that observation alone cannot offer, capturing not just what people say about a robot but how their bodies and behavior respond to it in the moment.
Social robots as research instruments
Social robots serve as the interactive centerpiece of many studies, letting us observe real-time human responses to humanlike agents in controlled but naturalistic settings. We work with a range of platforms, from robots designed for simple, focused interactions to more expressive systems capable of nuanced gesture and speech. This range lets us test how specific design choices, such as a robot’s voice, facial expressiveness, or movement style, shift the way people perceive and trust it.
Physiological sensing
Physiological sensors, including heart rate monitors, skin conductance sensors, and eye-tracking equipment, capture the subconscious reactions that surveys and interviews cannot reach. People often cannot articulate, or may not even be aware of, their own discomfort or ease around a humanlike agent. Physiological data lets us see those responses directly, giving us a layer of evidence that complements what participants tell us in their own words.
AI-driven adaptability
AI-driven platforms allow us to adapt robot behavior dynamically during a study, testing how subtle changes in tone, responsiveness, or conversational style shift human trust and comfort in real time. Rather than testing a single fixed version of a robot’s behavior, we can explore a range of variations within the same study, giving us more precise insight into which specific design elements drive which specific human responses.
Keeping pace with the field
As new sensing technologies and AI models emerge, our toolkit evolves with them. We regularly evaluate new hardware and software platforms for research use, and we prioritize tools that reflect the technologies people are actually likely to encounter in workplaces, homes, and public life, not just what happens to be available in a lab. This keeps our findings grounded in current and near-future reality rather than outdated assumptions about what robots and AI systems can do.