1. Rehabilitation Research at the NSF
Wendy Nilsen, PhD
Program Director, Smart and Connected Health

2. Background NSF
Rehabilitation research can be found in many areas in NSF and within the mission of several cross-directorate initiatives
It is a case of use-inspired basic research. The scientific advances in basic science can be in computing, information science, engineering or social or behavioral science. The benefit to rehabilitation research is important, but second to the advances in basic science.
Three major homes for this research:
Smart and Connected Health
Cyber-physical Systems
National Robotics Initiative

3. Smart & Connected Health (SCH) Inter-Agency Program National Science Foundation National Institutes of Health NSF Solicitation NSF Wendy Nilsen, PhD Program Director, Smart and Connected Health Computer and Information Sciences and Engineering, NSF

4. Application Inspired: Consideration of Use
Pasteur’s Quadrant
Neils Bohr
Luis Pasteur
Thomas Edison
Steve Jobs
Quest for Basic Understanding
Application Inspired: Consideration of Use
Donald E. Stokes, Pasteur's Quadrant – Basic Science and Technological Innovation, Brookings Institution Press, 1997

5. Smart and Connected Health Research Areas
Integration of EHR, clinical and patient data
Access to information, data harmonization
Semantic representation, fusion, visualization
Digital Health Information Infrastructure
Informatics and Infrastructure
Datamining and machine learning
Inference, cognitive decision support system
Bring raw image data to clinical practice
Data to Knowledge to Decision
Reasoning under uncertainty
Systems for empowering patient
Models of readiness to change
State assessment from images video
Empowered Individuals
Energized, enabled, educated
Assistive technologies embodying computational intelligence
Medical devices, co-robots, cognitive orthotics, rehab coaches
Sensors, Devices, and Robotics
Sensor-based actuation
The fundamental research issues appear to naturally cluster into four broad research areas. I would like to emphasize that these are not mutually exclusive and do not represent a unique classification. Rather this classification should be used as a guide in selecting fundamental questions to be addressed by the SHB research.
Digital health information infrastructure is associated with continuous accrual and integration of Electronic Health Records (EHR), pharma and clinical research data in a distributed but federated system. The ultimate goal of this research area is to bring data such as EHR to where it is needed, when it is needed.
The second area Data-to-Knowledge-to-Decision comprises research concerned with making the best possible use of the data in support of evidence-based healthcare.
The third area is focused on how technology could empowered individuals to participate in their own healthcare that could lead to better and more affordable care.
Sensors, devices, and robotics represent technology for sensing and intervention that enables closing the loop using intelligent technologies
I will describe briefly each of these four areas in more detail

6. NSF Directorates Participating in SCH
Office of the Director
Engineering (ENG)
Geosciences (GEO)
Mathematical and Physical Sciences (MPS)
Budget, Finance Award Management
Computer & Information Science and Engineering
(CISE)
Biological Sciences (BIO)
Diversity and Inclusion
Social, Behavioral and Economic Sciences (EBS)
Education and Human Resources (EHR)
General Counsel
Information & Resource
Management
Legislative & Public Affairs
National Science Board
Office of Inspector General
Cyber-infrastructure
Integrative Activities
International Science and Engineering
Polar programs

7. NIH Institutes Participating in SCH
OBSSR
NCI
NIBIB
NIA
NHGRI
NICHD
National Human Genome Research Institute

8. Computing Robot Motions for Home Healthcare Assistance
Motivation:
Over 10 million Americans currently need assistance with activities of daily living (ADLs), and this number is growing.
Robots could empower older adults and individuals needing ADL assistance to remain in their own homes rather than be transferred to costly institutions or nursing homes.
New software and algorithms are needed to control home healthcare robots for autonomous, safe assistance with ADLs.
Robots autonomously performing tasks in home-like environments
Technical Approach:
Learn robust metrics for ADL task motions from kinesthetic demonstrations provided by healthy humans. The computed metric serves as a guideline for fast motion planning for interactions with new care recipients.
Develop fast algorithms for real-time motion computation in uncertain, dynamic, and cluttered environments. Achieve fast performance using novel algorithms and harnessing the compute power of multi-core CPUs and many-core GPUs.
PI Ron Alterovitz, University of North Carolina at Chapel Hill, NSF Grant #

9. James Schmiedeler, Aaron Striegel, & Charles Crowell
Use of Gaming Peripherals in Acute Rehabilitation of Balance Following Stroke
Motivation:
Restoration of balance after stroke is critical determinant of patient’s long-term assistive needs.
Optimizing use of limited therapy time, particularly in acute phase shortly after injury, facilitates functional recovery.
High cost of most balance feedback systems limits clinical access & potential for in-home use after discharge.
Technical Approach:
Compare types of visual feedback provided based on center of pressure data from Nintendo Wii Balance Board.
Model human control of lateral weight shifting to identify changes associated with balance deficits.
Manipulate visual feedback during balance therapy tasks to facilitate rehabilitation of specific deficits.
James Schmiedeler, Aaron Striegel, & Charles Crowell
University of Notre Dame IIS

10. Brenna Argall, Northwestern University
SCH EXP: Collaborative Research: A Formalism for Customizing the Control of Assistive Machines
Motivation:
For those with severe upper limb motor impairments, caregivers are still relied on for manipulation tasks like meal preparation or personal hygiene.
Robotic arms hold much promise, however traditional devices for teleoperation like joysticks become tedious or untenable to control these higher degrees of freedom systems.
Customiziation of control sharing
functions to the user (U) and task (T)
Technical Approach:
A formalism that customizes how users share control with intelligent autonomous assistive devices, based on user ability and preference.
Customization to the user and task, and based on the confidence that the user's goal has been predicted correctly.
Customization by the autonomy and by the user.
Brenna Argall, Northwestern University
Siddhartha Srinivasa, Carnegie Mellon University
NSF Grant # 1R01EB

11. University of Southern California, NSF Grant #1117279
Socially Assistive Human-Machine Interaction for Improved Compliance and Health Outcomes
Motivation:
Our approach is focused on socially assistive robotics (SAR) and is motivated the following domains:
Post­stroke rehabilitation
Physical and cognitive exercise for older adults
General exercise encouragement
Technical Approach:
Affective feedback, praise, encouragement, and relationship building in SAR exercise coach and buddy systems
Personalization of motivational character backstory
Use of deviation (cheating) detection for user engagement
PI: Maja J Matarić,
University of Southern California,
NSF Grant #

12. Useful Website: www.nsf.gov

13. “You miss 100% of the shots you never take.”
Wayne Gretzky

14. Thank you!
Wendy Nilsen