Citation

Chowdhury MEH, Alzoubi K, Khandakar A, Khallifa R, Abouhasera R, Koubaa S, Ahmed R, Hasan MA. Sensors (Basel) 2019; 19(12): s19122780.

Affiliation

Department of Industrial and Mechanical Engineering, College of Engineering, Qatar University, Doha 2713, Qatar. ahasan@qu.edu.qa.

Copyright

(Copyright © 2019, MDPI: Multidisciplinary Digital Publishing Institute)

DOI

10.3390/s19122780

PMID

31226858

Abstract

Heart attack is one of the leading causes of human death worldwide. Every year, about 610,000 people die of heart attack in the United States alone-that is one in every four deaths-but there are well understood early symptoms of heart attack that could be used to greatly help in saving many lives and minimizing damages by detecting and reporting at an early stage. On the other hand, every year, about 2.35 million people get injured or disabled from road accidents. Unexpectedly, many of these fatal accidents happen due to the heart attack of drivers that leads to the loss of control of the vehicle. The current work proposes the development of a wearable system for real-time detection and warning of heart attacks in drivers, which could be enormously helpful in reducing road accidents. The system consists of two subsystems that communicate wirelessly using Bluetooth technology, namely, a wearable sensor subsystem and an intelligent heart attack detection and warning subsystem. The sensor subsystem records the electrical activity of the heart from the chest area to produce electrocardiogram (ECG) trace and send that to the other portable decision-making subsystem where the symptoms of heart attack are detected. We evaluated the performance of dry electrodes and different electrode configurations and measured overall power consumption of the system. Linear classification and several machine algorithms were trained and tested for real-time application. It was observed that the linear classification algorithm was not able to detect heart attack in noisy data, whereas the support vector machine (SVM) algorithm with polynomial kernel with extended time-frequency features using extended modified B-distribution (EMBD) showed highest accuracy and was able to detect 97.4% and 96.3% of ST-elevation myocardial infarction (STEMI) and non-ST-elevation MI (NSTEMI), respectively. The proposed system can therefore help in reducing the loss of lives from the growing number of road accidents all over the world.

Language: en

Keywords

heart attack; machine learning algorithm; portable device; real time system; support vector machine