Our work is divided into several areas, all of which are actively growing.
How can a robot walk naturally with a person? This subject is the topic of Rachel Gockley's dissertation research. Rachel's dissertation work is developing a robot called Companion, armed with algorithms that sense and navigate the physical and social setting.
Receiving help from a robot. Receiving help from a robot. This subject is the topic of Cristen Torrey's dissertation research. When face to face, a help giver has contextual information that can be used to tailor communication. Current research on robotic helpers seeks to leverage these benefits of physically embodied interaction without acknowledging that receiving help is not always a positive experience. Help recipients may feel scrutinized, or they may feel the helper has taken control of their activity. Cristen's dissertation research examines both positive and negative byproducts of receiving help, testing whether help from a robot is perceived as evaluative or controlling.
How can intelligent objects in the home sense the patters of inhabitants and infer when to help? The Sensechair was a project by Jodi Forlizzi, Carl DiSalvo, Bilge Mutlu, and Francine Gemperle that explored this question.
Elders and products What can we learn through the principled study of elders and products they use? These projects attempt to understand opportunities for assistive technologies for elders and their families and caregivers.
Designing gaze behavior for humanlike robots … was Bilge Mutlu's dissertation topic. The project involved observations of the way people use gaze in their communications with others, and then modeling those patterns to create eye movements by a robot to help guide the flow of a conversation. The work was developed in collaboration with researchers from Japan's Osaka University and from ATR Intelligent Robotics and Communication Laboratory.
People's mental models of robots is a long-standing debate in HRI. Do people assume robots that look like people can respond as people do? Do people respond to interactive robots as though they were people? Our work shows how people's mental models and behavior vary with what robots look like, what they say, and even where they were built. We also show how mental models influence how people interact with robots.
Robots are being deployed in hospitals not just in surgeries, but also as autonomous delivery vehicles. In this role, they have to interact with staff and patients.
Snackbot is a mobile, semi-autonomous robot …built by an interdisciplinary team at Carnegie Mellon University. Snackbot has two jobs. One job is to serve as a research platform for projects in robotics, design, and behavioral science. We welcome new partners or sponsors for this work. Snackbot's other job is to serve snacks.