Group N3: Digital Control

1
Group N3 Digital Control

• Wayne Blake
• Mary Nsunwara
• Reshun Gethers

2
Project Purpose

• Design a system that will efficiently measure
  signals produced by muscle movement and use these
  signals to operate a remote control car.

3
Current Applications

• EMG applications have two key categories
• Biofeedback
• Mainly for clinical purposes
• Utilized as muscle relaxation and treatment
• Also to prevent painful muscle conditions
• Communication
• Mostly experimental
• Muscle movements that control electronic and
  computer devices

4
Current Products

• Prices range from 175 - 2000
• Inexpensive, But
• Have no computer interface
• Generally only Useful for muscle training

5
Bodily Electrical Activity

• Contraction of all muscles is triggered by
  electrical impulses called Motor Unit Action
  Potentials (MUAPs)
• MUAPs can be created internally or externally by
  devices such as the pacemaker

6
Electromyography I

• Recording of Motor Unit Action Potentials (MUAPs)
  produced by muscle movement
• Frequency Range
• 50 500 Hz
• Voltage Range
• 0.75 2000 uV

7
Electromyography II

• Underlying Processes
• Differential Amplification
• Signal-to-Noise Ratio (SNR)
• Common Mode Rejection Ratio (CMRR)
• CMRR is the Ratio of the Differential Gain to the
  Common Mode Gain
• Both a High SNR and CMRR are Desired, in Order to
  Suppress any Common Noise Signals
• Filtering Removes High Frequency Components
• Sampling More Advanced Digital Signal Processing

8
Design Goals

• Front End
• Build and test EMG Circuit from Last Semester
• Acquire a transferable A/D converter
• Back End
• Determine best way to signal the Remote Control
  device

9
Interfacing to PC

• Digitize Signals Measured with EMG
• Analog to Digital Converter
• Comparator
• Assign Codes for System Control
• Determine Number and Degree of Muscle Movements
  to Execute Commands
• Apply Commands on Test System

10
Initial Design Issues

• EMG Circuit
• Will previous design work
• Signal Source Usage
• Which muscles to monitor
• A/D Converter Software
• Need PCI connection on A/D Card

11
EMG Circuit Diagram
12
EMG Circuit

• Design based on Older Version of the BrainMaster
  EEG monitor
• Circuit Consists of Two Amplification Stages
• An integrator circuit is utilized in this first
  stage as a low-pass filter and its main purpose
  is to provide good linearity
• Stage two has a gain of 390 for a total gain of
  19500 for the entire amplifier. Also provides a
  frequency response from 1.7 up to 34 Hz

13
Circuit Specifications

• Gain 20,000
• Bandwidth 1.7 - 34 Hz
• Input Impedance 10 M?
• CMRR 100dB

14
Signal Source Options

• Eyebrows
• Good contraction speed
• Difficult to move one eyebrow
• Low fat tissue coverage
• Jaw muscles
• Better contraction speed
• Easier to bite on one side of mouth
• Usually covered by more fat tissue thus making it
  harder to get a usable signal

15
A/D Converter

• Keithley KPCI-3107
• Maximum sampling rate 100 kb/s
• Input ranges 0 - 5V, 1V, 100mV, 20mV
• 12 gain ranges (1, 2, 4, 8, 10, 20, 40, 80, 100,
  200, 400, 800)
• 32 digital I/O lines
• Uses DriverLINX Package Software which is able to
  use up to 24 software programmable ranges
• Equipped with PCI connection

16
Amplifier Design Testing

• Duplicated Circuit designed last semester
• Tested New and Old Board
• Compared and Contrasted signals acquired from
  both boards
• Found solutions to old and new problems
• Old Board Had Loose or Non-existent Solder Joints
• Details Provided on New Board Issues

17
Carbon Copy

• Old Amplifier

• New Amplifier

18
Building Procedure

• Duplicated Amplifier From Last Semester
• Accuracy and Cost of Our Circuit Components were
  Reduced
• Okay because Board Measures Muscle Movements,
  which are Larger than Brain Waves
• 5 Resistors were used in place of 1 Resistors
• 50V Capacitors were used in place of 400 V
  Capacitors
• Had To Utilize Adapters to Connect Surface-Mount
  OP 90 Amplifier Chips to Board

19
Testing

• Electrical signals generated from various muscle
  movements such as
• Blinking
• Holding eyes closed
• Heartbeat
• Examination of Noise Artifacts
• Sensitivity of Electrode Leads
• Head Movements

20
Results - Blinking

• Old Amplifier

• New Amplifier

21
Results - Heartbeat

• Old Amplifier

• New Amplifier

22
Results - Noise

• Old Amplifier

• New Amplifier

23
Problems Encountered

• Noise Waveforms
• Head Movement, Touching Electrodes
• Reduce by Twisting Wires Together, so that Random
  Movements are Common to the Differential Inputs
• Cannot Eliminate All the Noise Output, so Handle
  Amplifier with Care
• Oscillation of Output Signals
• Caused by 60 Hz Electrical Source, E.g., Lamp
• Suggested Solving by Testing in the Dark or
  Operate circuit within an isolated black box
• Oscillation Disappeared without Conscious Effort

24
Cleaner Output
25
More Results

• Screen Captures from DriverLINX Software
• No Processing done Using Keithley Card
• However, Waveform was Successfully Transmitted to
  DriverLINX Screen
• MATLAB Simulations
• Data was Properly Digitized Using MATLAB
• MATLAB Functions could be Implemented Using DSP
  Chips for Memory, Adding, Squaring, etc

26
DriverLINX Capture

• Quiet Line

• Four Blinks

27
MATLAB Processing

• Function Created for Digital Processing of the
  Signal
• Captured Data from Oscilloscope Screen after
  Blinking Sequences
• Concatenated the input Data because Oscilloscope
  saved only one screen at a time
• Output Waveform was Purely Digital

28
Signal Digitization

• Before Digitization

• After Digitization

29
Conclusions

• Amplifier Functions Acceptably
• MATLAB Simulations Illustrate Intended
  Implementation of Amplifier
• Future Groups may Investigate Alternative Methods
  of Applying the Amplifier
• Potential Applications
• Wheelchair Control- motions of a wheelchair are
  manage without using the hands
• On-screen Mouse Control- Move mouse pointer using
  facial muscles alone