1. Real-Time Control of a Multi-Fingered Robot Hand Using EMG Signals
SDSU
Real-Time Control of a Multi-Fingered Robot Hand Using EMG Signals
Master’s Thesis By
Luenin Barrios
Supervisor: Marko Vuskovic
Department of Computer Science
San Diego State University
June 29, 2010

2. Outline Introduction to Research
SDSU
Outline
Introduction to Research
Multi-Fingered Robot Hands and Prostheses
Measurement of EMG Signals
Feature Extraction and Classification
Synergy and Robot Control System
Hardware Description
Implementation
Observations and Results
Summary

3. Introduction Goal of Research:
SDSU
Introduction
Goal of Research:
To implement a program that uses the EMG Classifier output to control the grasp motions of the SDSU robot hand in real-time
Grasp modes:
Chris Miller Master’s Thesis 2008.

4. Prosthetic Hands Overview
SDSU
Prosthetic Hands Overview
Early Models
Restrictions and Limitations
Degrees of Freedom
EMG Signal Control
Otto Bock Grasp Pincher
TAP Version 3 Prototype
SDSU Robot Hand

5. Overall Schematic SDSU Saksit Siriprayoonsak 2005 This Project
Chris Miller 2008

6. EMG Signals Electromyography
SDSU
EMG Signals
Electromyography
EMG potentials: 50 μV and up to 20 to 30 mV
Source:

7. Forearm Muscle Anatomy
SDSU
Forearm Muscle Anatomy
Chris Miller Master’s Thesis 2008

8. EMG Amplifier Device Saksit Siriprayoonsak 2005 4 Bipolar Channels
1 Reference Channel
Surface Electrodes

9. EMG Amplifier Device Con’td
SDSU
EMG Amplifier Device Con’td

10. EMG Classifier Program Signal Detection
SDSU
EMG Classifier Program Signal Detection
Bonato Method
Onset of Movement

11. Classifier Signal Processing
SDSU
Classifier Signal Processing
Feature Extraction
Methods:
Waveform Length (Farry et al., 1996)
Spectral Moments (Vuskovic et al., 2005)

12. EMG Signal Processing Feature Extraction Method 1
SDSU
EMG Signal Processing Feature Extraction Method 1
Waveform Length

13. EMG Signal Processing Feature Extraction Method 2
SDSU
EMG Signal Processing Feature Extraction Method 2
Spectral Moments
I-coefficients

14. Feature Classification
SDSU
Feature Classification
Mahalanobis Distance(Mahalanobis, 1936)
Sample Feature Vector Space

15. Feature Log Transformation
SDSU
Feature Log Transformation
Box Cox Transformation (1964)

16. Robot Joint Control System
SDSU
Robot Joint Control System
PID Controller and Actuator

17. Joint Control System Acutuator Model SDSU Variable Description Unit Ra
Terminal or Armature Resistance
3.38 Ohm Ka
Torque Constant
8.11 mNm/A Jm
Rotor Inertia
1.27 gcm2 Kg
Gear Transmission Ratio - Thumb
Gear Transmission Ratio - Finger
1:26
1:19 Ga
Driver Gain 1 Kb
Speed or Proportionality Constant
1180 rpm/V V0
Nominal Voltage
12 Volt ω0
No Load Speed
13900 rpm

18. Joint Control System PID Controller SDSU
e = qmd - qm; // Get controller error
qmdot = (qm-qmold)/_Ts; // Get derivative of error
ei = eiold + e * _Ts; // Get integral of error
u = _Kp*e - _Kv*qmdot + _Ki*ei; // Control law
qmold = qm;
eiold = ei;

19. Synergetic Motion Synergetic Mapping
SDSU
Synergetic Motion
Synergetic Mapping
θj = fj (m, D) where j = 0, 1…5 and m = 1…4  
Approximation Function (Vuskovic and Marjanski)
am,j = γm,j
cm,j = αm,j

20. Synergetic Training Joint Angle SDSU
θ1 (cm,1 + D1) = am,1 bm,1 - am,1 D1

21. Object Shapes and Sizes
Spherical
Point
Cylindrical
Lateral

22. Calibration and Training
SDSU
Calibration and Training
Sample Training for Point Objects:
Sample Positions for Lateral, Cylindrical and Spherical

23. SDSU
Robot Hardware
Servo To Go Board
Signal Transition Box

24. Servo To Go Interface Board
SDSU
Servo To Go Interface Board
Encoder Input A/B signal
Analog Input/Output

25. Signal Transition Box Central hub for signals/cables
SDSU
Signal Transition Box
Central hub for signals/cables
Relays information
Example: Joint 0
P3 DAC
EnIn A 14
EnIn B 17
DB50
EnOut A 35
EnOut B 34
DB25
AnalogIn

26. EMG Robot Hand SDSU User Interface/Motion Command Interpreter
Client/Server
TCP/IP
Real-time EMG/User Commands
Grasp modes: Cylindrical, Spherical, Point, Lateral

27. SDSU
EMG Robot Hand Cont’d
Examples:
Command: g 0 45
Command: o 3 9

28. Overall Runtime Flow Chart
SDSU
Overall Runtime Flow Chart

29. Overall Runtime Flow Chart Cont’d
SDSU
Overall Runtime Flow Chart Cont’d

30. SDSU
System Execution
Step 2
Step 3
Step 1
Step 4

31. Summary Multi-fingered Robot Hands and EMG Signals
SDSU
Summary
Multi-fingered Robot Hands and EMG Signals
Collection of EMG Signals
Feature Extraction and Classification
PID Controller
Synergetic Motion
Overall System Diagram and Transition Box
Real-time control of Robot Hand using EMG
Signals

32. Conclusions/Future Work
Feasibility of EMG Signal for Hand Control
Synergetic Grasp Motions
Classifier for real-time control
Combine projects so they reside on same machine
Improve arm/amplifier device contact
Wireless electrodes/sensory network
Improve time delays in Classifier

33. SDSU
Questions/Comments?