Abstract
Automatic External Defibrillator is a device that automatically diagnoses the heart rhythm which requires an electric shock. In general the signal processing unit of an AED aims to immediately identify the occurrence of ventricular fibrillation and ventricular tachycardia in a patient's electrocardiograph signal. In this research, a quick and high-precision method is presented that identifies peaks and heart rate arrhythmias of the cardiac signals that do not include QRS complex. Initially, asystole and presence of extra noise are recognized in a preprocessing step. Furthermore, two support vector machine classifiers have been used in a hierarchical way to separate Ventricular Fibrillation, Ventricular Tachycardia and normal signals. As a result, a real-time algorithm is developed that matches American Heart Association standard with an accuracy of 98.6%. Finally the Raspberry pi board is used as a hardware platform to embed the proposed algorithm into an AED system. The results show that the implemented system can detect shock-required rhythms in 0.56 s; and the method may be used in an actual signal processing unit of an AED.
Original language | English |
---|---|
Pages (from-to) | 277-284 |
Number of pages | 8 |
Journal | Biomedical Signal Processing and Control |
Volume | 51 |
Early online date | 15 Mar 2019 |
DOIs | |
Publication status | Published - May 2019 |
Keywords
- Adaptive threshold
- Automatic external defibrillators
- Embedded system
- Hierarchical
- Python
- Raspberry pi
- Real time
- Support vector machine
- Ventricular fibrillation
- Ventricular tachycardia
ASJC Scopus subject areas
- Signal Processing
- Health Informatics