Transcript Slide 1

Portable Heart Rate Monitor &
Arrhythmia Detector
Enyinnaya Egejuru
December 3, 2008
Senior Design Project
Instructor: Prof. Swenson
TA: Bob Schoonover
Section 1
INTRODUCTION
Objective
Design a portable monitoring system able to acquire
a biological signal (ECG) and detect abnormality from
the regular sinus rhythm
•
•
•
•
Provide Immediate notification of abnormalities
Low cost and low power
Accessibility
User friendliness
Project Significance/Market
This portable system will be useful whenever there is
need for a low cost and low power monitoring
system.
Possible markets:
• Third world countries
» Poor Medical Infrastructure, facilities and equipment
» Unreliable power supply
• Quick, portable monitoring
» Warfront
The ECG Signal
The Natural Pacemaker
• The Sino-Atrial node (SA
node) is a group of cells
positioned on the wall of the
right atrium
• Cells in the SA node
depolarize simultaneously
resulting in contraction
• Atrial contraction leads to
the P-wave and the
ventricular contraction leads
to the QRS complex
Section 2
SYSTEM DESIGN (FRONT END)
Block Diagram
MSP430FG4618
AD6
22
LM358
ALARM
Electrodes (Signal Acquisition)
• Interface between body and
measuring/monitoring device.
• Serve as transducer to change
ionic current in body to
electrical current
• Electrolyte gel (AgCl) used to
maintain good contact to skin
• Pair used to measure potential
difference between 2 parts of
the body
ECG Signal Amplifier
AD622 Instrumentation amplifier chosen for initial
signal amplification. Heart beat signal ranges from
1mV – 4mV, need amplification to about 1V
ECG Amplifier Requirements
AD622 Instrumentation Amp
Low DC Offset

Low Drift

Low Noise

High Open Loop Gain
 (Range 2 – 1000)
High CMRR

High input Impedance

Low Power/Cost
 (2.6V – 15V)
ECG Signal Amplifier Cont’d
• Gain of 10 from signal
source, RG should be
5.62KΩ:
G  1
50.5K 
(2.7 K   2.7 K )
Biosignal from
electrodes
Signal from
Instrumentation
Amplifier
Analog Filters
Filters are used to attenuate certain frequencies
• 0.05Hz High Pass Filter: Filters voltage drift to reduce
baseline wander and amplifier saturation
• 130Hz Low Pass Filter: Remove high frequency noise
and muscle artifact
Cutoff Frequency determined by:
1
Fc 
2 RC
Analog Filters Cont’d
• For High Pass: RC ≈ 3.
Hence, Rhp = 100KΩ, Chp =
30µF
• For Low Pass: RC ≈ .001.
Hence, Rlp = 10KΩ Clp =
100pF
• Further amplification
(factor of 100) is done
using a LM358 OP AMP
1M 
G  1
 100
10 K 
ECG Signal After Amplification
This signal shows a real BioSignal coming from a test subject
MSP430FG4618
•
•
•
•
•
•
•
•
Low supply-voltage range, 1.8V-3.6V
Ultralow power consumption
Five power saving modes
Wake up from standby mode in less than 6µs
16 bit RISC architecture
12 Bit A/D and 12 Bit D/A converters
16 bit Timer_A
116KB+256B Flash or ROM memory, 8KB RAM
Section 3
SYSTEM DESIGN (CORE/SOFTWARE)
MSP430FG4618 MCU Software
The figure above is a block diagram showing the function of the MCU
Digital Filter Design
The MCU uses a high pass and a low pass filter
• Fastest Deflection is the QRS complex. It lasts for
about 20ms and has a frequency of 6Hz-30Hz
• FIR filters were used because they are more stable
• 17 tap FIR filter was implemented. Although higher
taps would be better, we are working with limited
resources
• Filter design done using Matlab
Digital Low Pass Filter
This filter was built using the filterbuilder functionality of Matlab.
Coefficients are determined and convolution is performed
Digital High Pass Filter
This filter was built using the filterbuilder functionality of Matlab.
Coefficients are determined and convolution is performed
Heart Rate Calculation
• Sample rate of 512 Hz. Allowing up to 10 samples or
QRS complex.
• Tracking of timescale using a variable (pulseperiod)
• Output of QRS discriminator compared against
threshold
• Due to real-time need, pulseperiod is accumulated
over 3 beats
Heartrate per minute = 1/[pulseperiod/(3 x 512 x 60)] = 92160/pulseperiod
Section 4
SYSTEM DESIGN (BACKEND)
NOTIFICATION
•
•
•
•
•
•
The Winbond® ISD17120 ChipCorder® is a highquality,
fully integrated, single-chip multimessage voice record
and playback device Ideally suited to a variety of
electronic systems
120sec messages at 8Khz sampling frequency
PWM class D speaker driver
LED blinking notification during playback
Multiple message addressing
Playback enabled by signal from MCU
Great sound quality during voice playback using 54c6
ISD17120
Section 5
TESTING AND SIMULATION
ECG Simulation
• Was able to get ECG simulated WAV files
• Testing of circuit for tachycardia and bradycardia could be
easier as database has records of such arrhythmias
• WAV files are transferred to circuit via a modified speaker
phone. The right and left ear pieces correspond to the
right arm and left arm
LCD Heart Rate Display
• New feature added to design
• LCD displays a readable heartrate to user. This helped
me in testing and figuring out what was going on in
the MCU
Other Testing
• Oscilloscope reading at each amplification stage
• Setting test levels for tachycardia and bradycardia to
check response of notification pin
• Using an LED to tell when heartbeat has gone above
or below normal
• Artifact from movement and electrodes
Section 6
SUCCESSES, CHALLENGES, AND
RECOMMENDATIONS
Successes
•
•
•
•
•
•
•
•
Adequately acquire Biosignal
Adequately amplify Biosignal
Proper ADC conversion
Semi Functional Heart Rate counter
Proper notification functionality
Adding new feature (LCD for rate display)
Using Low Power components for battery operation
Used a new device in ECE445 (MSP430) and also performed
core functionality in MCU
• Low power. System powered with a 3V battery (2x1.5)
Challenges
• Sophisticated QRS/Thresholding algorithm with the
amount of resources we have in a microcontroller
• Poor Skin/Electrode interface
• Artifact
• Amplifier issues
» Offset voltages
» Open Lead wires
» Electric field interference
Recommendations
• Digital 60Hz Notch Filter to reduce power line noise
• A more sophisticated algorithm for QRS detection
and heart rate calculation exists, but requires more
resources (which will affect battery life) and is
challenging to implement in a microcontroller
Thanks to…
•
•
•
•
•
•
•
Dr. Gary Swenson
Bob Schoonover
Kenneth Gentry
ECE Parts Shop personnel
ECE Stores personnel
Spencer Brady
My Fellow ECE445-its
Instructor
TA
ECE415 Instructor
ECE420 TA
QUESTIONS…