Note: Most Internet Explorer 8 users encounter issues playing the presentation videos. Please update your browser or use a different one if available.

NAO-Base: A Multimedia Database to Support Socially-Assistive Robotics for Individuals with Autism Spectrum Disorder

Friday, 3 May 2013: 09:00-13:00
Banquet Hall (Kursaal Centre)
10:00
J. M. Vernon1, J. Kumar1, C. R. Crowell1, M. Villano1, K. G. Wier2, K. Tang1, J. Zona3, D. Portenier3 and J. J. Diehl4, (1)University of Notre Dame, Notre Dame, IN, (2)Logan Center, South Bend, IN, (3)Barber National Institute, Erie, PA, (4)Center for Children and Families, University of Notre Dame, South Bend, IN
Background:   Socially-assistive robotics is a growing field that shows particular promise for use in therapy for children with autism spectrum disorder (ASD).  Children with ASD:  (a) show relative strengths in understanding the physical (object-related) world and relative weaknesses in understanding the social (interpersonal) world, and (b) seem to be more interested in treatment when it involves electronic or robotic components.  At this point, social situations are so complex that the current state of robot technology and programming does not support the types of sophisticated social interaction necessary to have a robot act on its own in a seemingly autonomous fashion.  However, if an interactive robot is controlled by a human, it is possible to have motivational benefits of the technology for children with ASD, while still targeting the disorder’s characteristic social deficits.

Objectives: We will showcase a free, publicly available multimedia database that supports our therapeutically-relevant control system (DOMER, Villano et al., 2011) for a NAO robot (Aldebaran Robotics).  For the past two years, the DOMER control system has been tested with 20 children with ASD.  NAO-Base is a compendium of therapeutically-relevant robotic movements we have created for the NAO so that ASD researchers, therapists, schools, and centers can more rapidly implement our socially-assistive robot therapy system.  All of the behaviors in NAO-Base can be utilized with DOMER.  

Methods:  We created DOMER, a graphical user interface that allows anyone to wirelesslycontrol a NAO robot  during an Applied Behavior Analysis (ABA) session. With DOMER, the interactive responses of the robot allow for the simulation of real-time conversation between the child and the robot.  We also created NAO-Base, which is a collection of all of the therapeutically-relevant robot behaviors developed during the course of our research.  NAO-Base, which will be made available via a website, includes:  (a) detailed description of each movement, (b) a file that contains the NAO movement commands needed to program the behavior, and (c) a video clip of the behavior so that users can see the movement being performed. 

Results:   NAO-Base is organized by categories of behavior.  Currently there are five categories including Combination (61 behaviors), Non-Verbal (57 behaviors), Verbal (19 behaviors), Specialty Movements (24 behaviors), and Movements in Development (5 behaviors). Using DOMER to control the behaviors in NAO-Base allows the robot to respond quickly to social attempts made by the participant, with an average response time well under a second.  Using this system and the NAO-Base behaviors, we have gathered clinical data, some of which has been presented at IMFAR previously (e.g., Tang et al., 2011; Klinepeter et al., 2012), indicating that many participants make notable gains in the social behaviors targeted for them during therapy.  Moreover, we have been able to transfer this technology to two non-profit ASD programs that now use the technology.

Conclusions:  Our free, publicly available NAO-Base greatly accelerates the process of utilizing DOMER to implement socially-assistive robot therapy because it provides access to a number of pre-programmed behaviors that have been tested in ABA therapy for children with ASD.

| More