Overview

Drawing on our previous work in hands-off assistive human-robot interaction for stroke rehabilitation, this project seeks to understand the role of embodiment in such supervisory interactions.

Project Details

Unlike most other research into rehabiliation robots, this project seeks to understand how robots can be used in a safe, non-contact way to encourage patients through their rehabilitation. Since our robots don't physically contact people, however, we need to ask: what benefit do we gain from using a physical robot, rather than something like an on-screen computer agent? We are therefore looking at ways that the robot's embodiment might affect people's behaviors in physical therapy tasks. We are examining two aspects of embodiment: proxemics and engagement.

Proxemics is a measure of space around a person (Hall 1966). There are four distances, each with a "close" and a "far" phase:

1. intimate: involves physical contact; uncomfortable if strangers enter
2. personal: "arm's-length" away
3. social: distance assumed when conducting business
4. public: greater than 4 meters away

Since intimate distance would risk collision between the robot and the patient, and public distance is generally beyond the robot's sensor range, we are comparing the personal and the social distances.

We view engagement as a measure of how closely the robot followed a person's movements. Since people tend to prefer and be more persuaded by agents that mimic their movements (Bailenson 2005), we expect that a robot that moves back and forth as a person exercises will be more encouraging to a person than a robot that is primarily stationary.

Pictures

Publications

• Rachel Gockley and Maja J. Mataric´. Encouraging Physical Therapy Compliance with an Hands-Off Mobile Robot. To appear in Proceedings of Human-Robot Interaction 2006.

References

1. E. T. Hall. The Hidden Dimension. Doubleday, New York, 1966.
2. J. N. Bailenson and N. Yee. Digital chameleons: Automatic assimilation of nonverbal gestures in immersive virtual environments. Psychological Science, 16(10):814-819, Oct 2005.

Support

This work is supported by the National Science Foundation via a Graduate Research Fellowship and grants #IIS-032914 and #IIS-0121426, and is due in large part to the support of Jodi Forlizzi and Reid Simmons of Carnegie Mellon University.

The USC resources were supported by the Okawa Foundation and the USC Provost's Center for Interdisciplinary Research.

Contact

Rachel Gockley
Emily Mower