1. Smart In-Home Rehabilitation System via 3D Printing Augmentation
Matthew Stafford, Feng Lin, Wenyao Xu
Department of Computer Science and Engineering, University at Buffalo, SUNY, Buffalo, NY
We have used a multiple 3-D prints to provide a variety of workout types.
• Cup
• Bowl
• Key
3D Printed Containers
• TheraPutty
• TheraBand
• Tailwind
• Rejoyce
Problems
• Expensive
• Low adaptability
• Limited or no feedback
• Not personalized
• Data not saved
Current Methods
Smart Watch
• Detect Hold Accuracy
• Provide Visual Feedback
Smart Phone
• Audio and Visual Feedback
3D Printed Container
• Standardize Workout
• Provides interface for
multiple workouts and difficulties
System Overview
• The Smart Watch is
used to ensure the
patient maintains good
holding posture.
• Red indicates that the
patient’s hand in some
axis has rotated in such
away that is out of
bounds.
Smart Watch Interface
Smart Phone Interface
• Patients select their workout.
• Workouts track whether users choose to use their left
or right hand.
• Each Workout gives users a dynamic display.
• Repetitions are tracked and made audible.
Smart Phone Application Overview
Workout Data
Start
Workout History
Workout
Selection
Left or Right Hand
Hand Accuracy
Detection (Y/N)
List View
Graph View
List View
Workout Activity
• After institutional therapy, sufferers of stroke must
continue their exercise at home with materials
provided by their outpatient care.
• Over time compliance to these workout programs
weakens and sufferers tend to stop entirely.
• This is often due to the overly-simplified nature of the
tools given.
• Our solution is a smart phone based rehab system that
gives patients meaningful feedback.
Introduction
3D Printer
Smart Phone
Software
Sensor Data
Analysis
Feedback
Interface
Smart Watch
3D Printer
Smart Phone
• Data from workout tracked
• Can view workout by itself or
compared to past workouts on a
graph to better understand
progress
Post Workout Data
Fig1. TheraPutty
Fig2. Tailwind
Fig3. System Overview
Fig4. Unlock Workout Fig5. Workout SelectionFig5. Workout Selection Fig6. Pitcher of Beer Pour
Workout
Fig7. Smartphone App
Overview
Fig8. Smart Watch Interface
Fig9. 3-D printed Bowl Fig10. 3-D printed lock Fig11. 3-D printed Cup
• “Easy to use and understand, something
I would do.”
Testing With Stroke Patients
• “This is
something I
would use and
something I
would continue
to use as long as
I am getting
positive
feedback.”
Fig12. Stroke survivor testing system
Fig13. Workout Info Screen Fig14. Historical Workout Info
Screen
• During workouts users movements
are analyzed using a Naturalized
jerk score algorithm (see fig.12)
• An average is taken at the end of
the workout and that becomes the
users ‘score’.
• Preliminary results during a ten cup pick
up activity show effective feedback can
improve quality of movement.
• (a) activity data and jerk score before
providing feedback;
• (b) activity data and jerk score after
providing feedback.
• The bottom graph (b) is clearly much
smoother. This suggests a less ‘jerky’
movement, what we would consider an
improvement.
Fig16. Acceleration (top) and jerk score (bottom) data
Fig15. equation for normalized jerk score
Accuracy of Movements