A Low Power Wearable Device for Real-Time Electrocardiogram Monitoring and Cardiovascular Arrhythmia Detection for Resource Constrained Regions

Abstract

Electrocardiography is a non-invasive technique for obtaining the electrical activity of heart by placing electrodes on various thoracic points of human body. The obtained electrical signals are then used to detect various cardiovascular abnormalities i.e., arrhythmias, Myocardial Infarction (MI) conditions, hypercalcemia and hypocalcemia etc. by comparing critical features of normal ECG signals with the abnormal one. In this work, we propose a low power wearable device with two transmission channels that runs an intuitive algorithm and processes the data from a single lead ECG front end (extended upto 3 Leads) and reports various arrhythmic conditions based on R�R interval. Pan Tompkins Algorithm is used as the basic R peak detection scheme and extending detections of P wave, T wave, Q, S and J points of the signal. To verify the proposed algorithm, the MIT/BIH arrhythmia database is used as a source of the ECG signals and the reference for R peak annotations. The R peak detection algorithm provides metric of False Detection Rate to be 1.289%, Sensitivity to be 99.492% and Positive Predictivity to be 99.293%. The energy requirement for ISM band enabled wearable device is 1.4165 J and bluetooth enabled wearable device is 2.7548 J respectively and the device can operate for 12.3 days on nRF and 4.2 days on bluetooth on a coin cell and can prove to be effective for remote monitoring applications. Results for arrhythmia, I degree atrioventricular block (AV block) and atrial fibrillation (AF) have been obtained. The probable cases of bundle branch block (BBB) are also categorized by the algorithm. It is concluded that the proposed low power wearable device design can be a useful tool in the resource constraint regions in Asia and Africa where health care is a major concern and remote monitoring can prove to be a useful alternative.