This project, developing a new generation of social robots for studies in human-robot interaction, aims at creating robots and modular robotics kits to be used to study robot social behavior, mutual understanding in human-robot communication, and the impact of assistants in group dynamics. The research involves understanding and testing theories of how these three aspects of human-robot interaction independently and interactively affect collaborative work and investigating how assistive robots can be designed best to aid people in domains such as health and aging. Designed to motivate and test theories, the work studies . Fundamental laboratory research on behavioral characteristics of robots performing social tasks, especially as these characteristics reflect likeness; . Controlled experiments and field studies of interpersonal communication and development of mutual understanding between robot and human, and . Robots in work groups. Aiming to better understand the societal impact of robots, a multidisciplinary team (robotics, computer science, social psychology, engineering, organization science, design) draws from and builds on basic research in cognitive and social psychology, and on recent research in robotics, computer graphics, and design. The team seeks a foundation for understanding and designing collaborative work with robots in critical environments like mines, hospitals, households with elderly or disabled residents, in challenging scientific settings and in situations in which the robot is remote.
Broader Impacts: This development contributes to public and student awareness of the human side of robotics. A life-size robot will be taken to school for educational purposes.
With advances in computing technology and artificial intelligence, autonomous robots are becoming viable in such critical domains as search and rescue, military battle, mine and bomb detection, scientific exploration, law enforcement, and hospital care. Robotic assistants ranging from museum guides to forestry scouts are being developed to interact with people "in person" or remotely, as agents that collaborate with the work team. The main goal of this project was to create a human-robot interaction laboratory for understanding and improving human-robot interaction. In the lab, we aimed to design and develop embodied robots for HRI research, including the design and development of a lifesize robot that could interact over time, learn people’s preferences, and offer appropriate levels of help. The main outcome of the effort was to create a collaborative environment for HRI research and, within it, the Snackbot robot--a lifesize robot that was able to deliver snacks and operate in a real office environment. There were many challenges to designing this robot. The robot had to navigate without running into people or objects. It had to understand people, even in noisy settings. It had to remember what snacks it was supposed to deliver and to whom, and adapt to changes in people's preferences (e.g., they preferred candy to an apple). In experiments conducted over several years, with the participation of students, we learned how a robot should act so as to engage people, how the robot could guess from their actions whether people wanted to be sociable or not, and how to obtain people's cooperation in helping the robot do its tasks.
