Henny Admoni, a human-robot interaction researcher at Carnegie Mellon’s Robotics Institute, explains how robots can use and interpret nonverbal cues to collaborate with people.
Admoni argues that robots intended for everyday settings must be designed with people in mind. Human-robot interaction combines robotics with insights from psychology and cognitive science to understand how people perceive, communicate, and act.
In a study with CMU’s HERB robot, Admoni explored whether head direction could suggest where a person should sort an object after receiving it. Participants often missed the cue during smooth handovers. A one-second pause before release drew more attention to the robot and increased compliance with its suggestions on ambiguous items.
Admoni describes shared autonomy for wheelchair-mounted robot arms, which can reduce the effort required to control a six-degree-of-freedom arm with a joystick or other limited input device. Eye tracking showed that users mostly watched the robot’s hand, with occasional glances toward targets. Changes in gaze also revealed when users were struggling or when the arm blocked their view, suggesting moments when the robot could help or move aside.
Admoni’s group also studies posture, gestures, and facial expressions to estimate when restaurant customers need attention and can be interrupted. For social navigation, torso orientation can help a robot anticipate where someone will be a few seconds later. She closes by arguing that effective collaboration requires combining robotics with knowledge of human behavior.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
Robert Scoble of Microsoft’s Channel 9 visits Carnegie Mellon’s Robotics Institute with students and researchers to explore medical robots, educational robotics, manipulation research, and Segway soccer.
