Eye Controlled Typing

1
Eye Controlled Typing

• Group 37
• Xiaoyu Ma, Hung-Yuan Wu
• ECE 445 Senior Design
• April 27, 2006

2
Introduction

• Provides a way to communicate for the paralyzed
• Takes advantage of eye movements the only
  method of control

3
Objectives

• EOG electrodes to sense eye movement
• EMG signals detects for blinking
• Non-standard keyboard layout for easier typing
• Text display window
• Text-to-Speech feature
• Low cost

4
System Overview

• Hardware
• High gain amplifier, filter, summing amplifier,
  calibration indicator
• Buffer, power supply
• Software
• PIC programming, RS232 communication
• User interface

5
Hardware Schematic
6
Hardware circuit on protoboard
LED indicator
Instrumentation Amplifier
Summing Amplifier
Buffer
7
Hardware Description

• Amplifier
• Provides gain of 1000
• Filter
• Blocks noise greater than 32Hz
• Summing amplifier
• Shifts signal to valid input range of PIC
• Calibration indicator
• LED to indicate calibration

8
Hardware Additions

• Buffer Circuit
• Provides lower output impedance
• /- 5V power supply
• Input from 9V battery
• Portability

9
Amplifier Testing

• Check for gain of 1000
• Input sine wave of 1mVPP Outputs sine wave of
  1VPP

10
Amplifier Testing

• Check for flat frequency response

11
Summing Amplifier

• Controlled by potentiometer
• Able to add 5V to -5V to signal

12
Electrode output
Response after looking up, right, down, and left.
13
Power Supply

• Takes input from 9V battery
• Outputs 5V, GND, and -5V with maximum current of
  1A
• Supplies all amplifiers, PIC, MAX232
• Circuit taken
  from ECE 445
  Course website

14
Power Supply

• Ripple of 21.3mV at 5V output
• Average output of 5.0645V
• Output level fairly constant and stable

15
Software
16
Software Description

• PIC Microcontroller
• Performs 10 bit ADC on vertical and horizontal
  input voltage
• Interprets eye movement based on voltage range
• Output through RS232 communication
• User Interface
• Read input through serial port
• Update text window and keypad
• Text to Speech

17
PIC Microcontroller
j
MAX232
PIC16F877A
Fox F1100 20Mhz Oscillator
18
PIC Microcontroller
19
User Interface
20
User Interface

• Programmed in Visual Basics 6.0
• Non-conventional keyboard layout, similar to
  those on a cell phone
• Communicates with PIC through RS232
• Low system requirements

21
Text to Speech
22
Problems and Challenges

• Hardware
• Low signal transfer into PIC
• Signal drift due to sensitivity of electrodes
• Software
• Interpreting diagonal movements
• Detecting blink
• Transition values
• Text to Speech

23
Results

• Works but not as reliable as planned
• Can't be used for long periods of time
• User must remain fairly still during use
• Project parts cost of 42.14 with each battery
  
  lasting about
  20 hours

24
Sample Screenshots
25
Improvements

• Signal processing
• Uses differences in voltage average to detect eye
  movement
• Solves drifting in signal
• Goggle
• Ties wires to electrodes together
• More portable
• Re-useable
• Word Prediction
• Faster typing
• User interface on hand held devices
• More portable

26
Acknowledgements

• Dwayne Hagerman for guidance and suggestions
• Professor Ray Fish for supplying plenty of
  electrodes
• ECE Parts Shop for most of circuit devices
• Analog Devices and Microchip for providing free
  samples