Heart Monitor Application for Android Craig Santangelo Advisor: Professor Hedrick.

Heart Monitor Application for Android

Craig Santangelo

Advisor: Professor Hedrick

Background Information

According to the American Heart Association, each year:

785,000 Americans will have a new coronary attack470,000 will have a recurrent attack195,000 silent myocardial infarctions

Every 25 seconds someone will have a coronary event

Every 60 seconds someone will die of one

Background Information

Quick response time is crucial to surviving a heart attack because:

Cardiac arrest normally occurs within the first 2 hours of the heart attack

If victim is not treated within the first 2 hours, permanent heart damage can occur

Also, many infarctions can occur without the victim knowing

General System Block Diagram

Specifications

Comfortable enough for around-the-house use

Quality signal and heart attack detectionMinimize errorLow noiseSample rate high enoughEfficient algorithm

Signal processing time frame

Heart sensors

Kendall Q-Trace 5400 Resting ECG Electrode

Silver/Silver Chloride (Ag/AgCl) sensing system

Pre-equipped with conductive adhesive hydrogel

Low impedance for a clearer signal

Project Break Down: 2 parts

1. Analog Circuit

Developing, constructing, and testing the following: Instrumentation

Amplifier Input Driver Optocoupler Output Driver High Pass Filter Low Pass Filter Output gain stage

2. Algorithm

Research how a heart attack affects an ECG waveform

Develop an algorithm to detect the occurrence of a heart attack

Translate the algorithm into MATLAB code for testing

Analog Circuit:Block Diagram

9 V Battery Power Supply (9 V)

Analog Circuit: Main Components

Instrumentation Amp AD620 chip Gain = 11 V/V

Output Gain Stage LM248 quad op-amp chip Non-inverting amp Gain = 100 V/V

High Pass Filter LM248 quad op-amp

chip Cutoff frequency: 0.05

Hz

Low Pass Filter LM248 quad op-amp

chip Cutoff frequency: 160 Hz

A/D Converter

Vernier Sensor DAQ

USB data acquisition

Can sample up to

48 kHz

13 bits single-ended

Common Heart Signal

Typical time spans:

PR interval: 120-200 ms

PR segment: 50-120 ms

QRS complex: 80-120 ms

ST segment: 80-120 ms

QT interval: 300-430 ms

Sample rate: 1000 Hz

Heart Attack Algorithm:ST Elevation

Most common and most damaging form of a heart attack

Requires immediate treatment

Heart Attack Algorithm Necessities

1. Establish a baseline (average of all samples in the PQRST waveform)

2. Detect QRS peak

3. Make sure the signal returns to the baseline directly after the peak within approximately 100 ms (width of ST segment)

4. If signal does not return to the baseline within 100 ms of QRS complex, alert the victim and dial 911

Testing

MCI Patient Simulator

Physionet.org

MATLAB

Possible Problems

Muscle activity

Extra common mode noise

Solution: Better amplifiers with higher CMRR

Use of Right Leg Driver

Power-line noise

Solution: notch filter at 60 Hz

Special Thanks

Professor Takashi Buma

Professor Hedrick

EE Department

Sources Kuangwei Hwang, Diane Limsui, Lianhan Zhao, “Wireless Heart Detector

with GPS”, Fall 2004, University of Illinois at Urbana-Champaign.

T. Buma, class notes for ECE 386, Department of Electrical Engineering, Union College, Winter 2012

ABC News, “What does an EKG show during a heart attack?” http://abcnews.go.com/Health/HeartDiseaseScreening/story?id=4223255#.T0MFeCNWq3w

MedicineNet.com, “Electrocardiogram (ECG or EKG)” http://www.medicinenet.com/electrocardiogram_ecg_or_ekg/article.htm

R. Spengler, “How does my doctor interpret my EKG,” 2002, http://yalenewhavenhealth.org/library/healthguide/en-us/support/topic.asp?hwid=tu6258abc.html

Vernier.com, “SensorDAQ”, http://www.vernier.com/products/interfaces/sdaq/

Helen Kim, Choon Yik Lee, “Walking Stick with Heart Attack Detection”, Spring 2005, University of Illinois at Urbana-Champaign.

Questions?