Real Time Control of Hand Prosthesis Using Surface EMG

Transcription

Real Time Control of Hand Prosthesis Using Surface EMG
Real Time Control of Hand Prosthesis
Using Surface EMG
Or
2
Dicker ,
Aviv
1
Peleg ,
Tal
2
Shnitzer ,
Oscar
1
Lichtenstein
1 Faculty of Biomedical Engineering, Technion - IIT, Haifa, Israel
2 Faculty of Electrical Engineering, Technion - IIT, Haifa, Israel
Introduction
Objective
-There are 1.8 million, below the elbow amputees.
Development of a control mechanism for 3D printed prostheses:
• Real Time: Less then 250ms from command to execution.
• Reliability: Classification success rate of more then 96%.
• Portability: Fully portable.
• Variety of gestures: A range of 6 different gestures.
• Short setup time: Less then 1 minute of calibration.
-Current hand prostheses capabilities range from single
action to expensive (>30,000$) multi-action prostheses.
-3D blueprints for affordable single action prosthesis are
currently available as “open source”.
-With today’s technology, affordable multi-action prostheses
can be developed for all in need.
Methods
Algorithm workflow
1. Calibration:
1.A Recording and labeling the EMG signals ,fig.2, using the MYO ,fig.3
Fig.2:Raw data
Fig.3:MYO armband
Fig.1:Algorithm workflow
1.B Feature extraction in the time domain,
using Mean Absolute Value, fig.4.
1.C Pattern recognition using
K Nearest Neighbors algorithm (K=3)
2. Real time hand control:
2.A Recording.
2.B Real time classification
2.C Convert classification to servo
movements fig.5
Command
Mechanical Output
duration
Servo 1
Fig.4:Data after Feature extraction
Servo 2
position
Servo 3
pulse width
modulation
gesture
Thumb
Index finger
Digits
2
2000
1500
500
Index finger
Closed
Open
Closed
3
2000
500
500
fist
Closed
Closed
Closed
4
500
500
500
Thumb up
Open
Closed
Closed
Fig.7: Prototype
3. Movement
Fig.9:Microcontroller Intel Edison
Fig.5:convertion table of classification to servo command to
mechanical output
2.D Pre action threshold is used to
rectify the classification Fig.6.
Fig.8: The six
different hand
gestures: fist,
pinch, point, rest,
Frisbee catch,
like.
Fig.6:Effect of pre action threshold shown in red
Results
Real Time
Reliable
Portability
Set goal
<250 msec
>96%
Fully portable
Variety
6 gestures
set up time <1 minute
Achievement
Fully achieved
Fully achieved: 99%
No PC needed. Electricity
portability was not validated
Fully achieved
Fully achieved: around 30 sec.
Acknowledgments:
 Mr. Nimrod Peleg
 Prof. Yoav Medan
 Mr. Shlomi Dach
 Mr. Yair Herbst
 Mr. Oren Forkosh
 Mr. Vasily Vitchevsky
Conclusions
-In regard to affordability,
current price stands at
345$. Future price is
estimated at 180$.
-Nearly all the objectives were met,
proving the feasibility of our solution.
-Further work in regard to price reduction
and redesigning the electro-mechanics
is needed to reach a complete product.
References
 Mr. Sharon Ishar
 Mr. Koby Kohai
 Mr. Gal Pressman  Mr. Vasily Vitchevsky
 Mr. Johanan Erez
[1] https://www.myo.com [2] M. B. I. Reaz, M. S. Hussain and F. Mohd-Yasin, Techniques of EMG signal analysis: detection, processing, classification and applications,
Biological Procedures Online , 2006 [3] [2] G. Vannozzi , A. M. Sabatini , P. Dario, Improving detection of muscle activation intervalsIEEE Engineering in Medicine ,
2002