--------------------------------------------
i n v e s t i g a t o r s :
Keith Evan Green
Architecture/ECE, PI
Ian D. Walker
ECE/Robotics, PI

Johnell Brooks
Human Factors Psychology, Co-PI

--------------------------------------------
A W E - f u n d i n g :

U.S. National Science Foundation # IIS-1116075

NSF logo

--------------------------------------------
t e a m :
Stan Healy
Healthcare Administration,
GHS
Tony Threatt
Ph.D. Student, Architecture
Jessica Merino
M.S. Student, ECE

Paul Yanik
Ph.D. Student, ECE

--------------------------------------------
Kinect GestureGesture captured by MS Kinect

 

Nonverbal Human-Robot Communication

We have been developing an appropriate and effective nonverbal communication (NVC) platform for robots, communicating with people; a mode of communication that dignifies what it is to be human by not competing with us, nor imposing on our social-emotional-cognitive constitution. We have initially developed a working prototype integrated wtih the our Assistive Robotic Table ("ART") with support from NSF's Smart Health & Wellbeing Program.

Our non-verbal communication (NVC) is conveyed by the familiar means of Audio-Visual Communication (AVC): low-cost lighting (colors, patterns) and sounds. The NVC we designed is based on an understanding of cognitive, perceptual processes of non-verbal communication in humans. The NVC, in turn, affords a communicative dialogue (i.e. acknowledging requests, or providing requested feedback) that conveys the purpose of accomplishing tasks. Our employment of learning algorithms offers both user and robot the capacity to interrupt, query, and correct the dialogue, and conveys in the robot some semblance of emotional information (e.g. urgency, respect, frustration) at a level that is not disconcerting to the user in a way that a user might misconstrue as human.

NVC loop
The Human-Robot Communication Between our ART [Assistive Robotic Table – IIS#-1116075] and both Post-Stroke Patients and their Clinicians.

Jessica gestring ART
Gesturing ART, and receiving a reply in lights and sounds.

NVC feedback loopSystem flow diagram for the gesture learning interface.

Publications:
Yanik, P.M., Merino, J., Threatt, A.L., Manganelli, J., Brooks, J.O., Green, K.E. and Walker, I.D. “A Gesture Learning Interface for Simulated Robot Path Shaping with a Human Teacher.IEEE Transactions on Human Machine Systems, 44(1): 41–54, 2014.

Yanik, P. M., Threatt, A.L., Merino, J., Manganelli, J., Brooks, J. O., Green, K. E. and Walker, I. D. “A Method for Lifelong Gesture Learning Based on Growing Neural Gas." Proceedings of HCI International 2014 (Crete, Greece, July 21-27, 2014). In M. Kurosu, Ed., Human-Computer Interaction, Part II, HCII 2014, LNCS 8511, Springer International Publishing Switzerland 2014, pp. 191–202.

Threatt, A. L., Green, K. E., Brooks, J. O., Merino, J. and Walker, I. D. Design and Evaluation of a Nonverbal Communication Platform for Human-Robot Interaction. Proceedings of HCI International 2013 (Las Vegas, July 21-26, 2013). In N. Streitz and C. Stephanidis, Ed.s, Springer Lecture Notes in Computer Science (LNCS) 8082 (25), Springer, Heidelberg, 2013, 505–513.

Yanik, P., Manganelli, J., Merino, Threatt, T., Brooks, J. O., Green, K. E. and Walker, I. D. Use of Kinect Depth Data and Growing Neural Gas for Gesture Based Robot Control. Proceedings of PervaSense2012, the 4th International Workshop for Situation Recognition and Medical Data Analysis in Pervasive Health Environments. May 21, San Diego, California, pp. 283-290.