Development of respiratory rate estimation technique using electrocardiogram and photoplethysmogram for continuous health monitoring

Nazrul Anuar Nayan, Rosmina Jaafar, Nur Sabrina Risman

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Abnormal vital signs often predict a serious condition of acutely ill hospital patients in 24 hours. The notable fluctuations of respiratory rate (RR) are highly predictive of deteriorations among the vital signs measured. Traditional methods of detecting RR are performed by directly measuring the air flow in or out of the lungs or indirectly measuring the changes of the chest volume. These methods require the use of cumbersome devices, which may interfere with natural breathing, are uncomfortable, have frequently moving artifacts, and are extremely expensive. This study aims to estimate the RR from electrocardiogram (ECG) and photoplethysmogram (PPG) signals, which consist of passive and non-invasive acquisition modules. Algorithms have been validated by using PhysioNet’s Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC-II)’s patient datasets. RR estimation provides the value of mean absolute error (MAE) for ECG as 1.25 bpm (MIMIC-II) and 1.05 bpm for the acquired data. MAE for PPG is 1.15 bpm (MIMIC-II) and 0.90 bpm for the acquired data. By using 1-minute windows, this method reveals that the filtering method efficiently extracted respiratory information from the ECG and PPG signals. Smaller MAE for PPG signals results from fewer artifacts due to easy sensor attachment for the PPG because PPG recording requires only one-finger pulse oximeter sensor placement. However, ECG recording requires at least three electrode placements at three positions on the subject’s body surface for a single lead (lead II), thereby increasing the artifacts. A reliable technique has been proposed for RR estimation.

Original languageEnglish
Pages (from-to)487-494
Number of pages8
JournalBulletin of Electrical Engineering and Informatics
Volume7
Issue number3
DOIs
Publication statusPublished - 1 Sep 2018

Fingerprint

respiratory rate
electrocardiography
Health Monitoring
Electrocardiography
health
Health
Monitoring
artifacts
Oximeters
Sensor Placement
recording
Sensors
Deterioration
Lung
Placement
chest
sensors
Electrode
air flow
breathing

Keywords

  • Algorithms
  • ECG
  • MIMIC II
  • PPG
  • Respiratory rate

ASJC Scopus subject areas

  • Computer Science (miscellaneous)
  • Control and Systems Engineering
  • Information Systems
  • Instrumentation
  • Hardware and Architecture
  • Computer Networks and Communications
  • Control and Optimization
  • Electrical and Electronic Engineering

Cite this

@article{b3872d9d4bf84c25bf3a013f5bfafdad,
title = "Development of respiratory rate estimation technique using electrocardiogram and photoplethysmogram for continuous health monitoring",
abstract = "Abnormal vital signs often predict a serious condition of acutely ill hospital patients in 24 hours. The notable fluctuations of respiratory rate (RR) are highly predictive of deteriorations among the vital signs measured. Traditional methods of detecting RR are performed by directly measuring the air flow in or out of the lungs or indirectly measuring the changes of the chest volume. These methods require the use of cumbersome devices, which may interfere with natural breathing, are uncomfortable, have frequently moving artifacts, and are extremely expensive. This study aims to estimate the RR from electrocardiogram (ECG) and photoplethysmogram (PPG) signals, which consist of passive and non-invasive acquisition modules. Algorithms have been validated by using PhysioNet’s Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC-II)’s patient datasets. RR estimation provides the value of mean absolute error (MAE) for ECG as 1.25 bpm (MIMIC-II) and 1.05 bpm for the acquired data. MAE for PPG is 1.15 bpm (MIMIC-II) and 0.90 bpm for the acquired data. By using 1-minute windows, this method reveals that the filtering method efficiently extracted respiratory information from the ECG and PPG signals. Smaller MAE for PPG signals results from fewer artifacts due to easy sensor attachment for the PPG because PPG recording requires only one-finger pulse oximeter sensor placement. However, ECG recording requires at least three electrode placements at three positions on the subject’s body surface for a single lead (lead II), thereby increasing the artifacts. A reliable technique has been proposed for RR estimation.",
keywords = "Algorithms, ECG, MIMIC II, PPG, Respiratory rate",
author = "Nayan, {Nazrul Anuar} and Rosmina Jaafar and Risman, {Nur Sabrina}",
year = "2018",
month = "9",
day = "1",
doi = "10.11591/eei.v7i3.1244",
language = "English",
volume = "7",
pages = "487--494",
journal = "Bulletin of Electrical Engineering and Informatics",
issn = "2089-3191",
publisher = "Institute of Advanced Engineering and Science (IAES)",
number = "3",

}

TY - JOUR

T1 - Development of respiratory rate estimation technique using electrocardiogram and photoplethysmogram for continuous health monitoring

AU - Nayan, Nazrul Anuar

AU - Jaafar, Rosmina

AU - Risman, Nur Sabrina

PY - 2018/9/1

Y1 - 2018/9/1

N2 - Abnormal vital signs often predict a serious condition of acutely ill hospital patients in 24 hours. The notable fluctuations of respiratory rate (RR) are highly predictive of deteriorations among the vital signs measured. Traditional methods of detecting RR are performed by directly measuring the air flow in or out of the lungs or indirectly measuring the changes of the chest volume. These methods require the use of cumbersome devices, which may interfere with natural breathing, are uncomfortable, have frequently moving artifacts, and are extremely expensive. This study aims to estimate the RR from electrocardiogram (ECG) and photoplethysmogram (PPG) signals, which consist of passive and non-invasive acquisition modules. Algorithms have been validated by using PhysioNet’s Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC-II)’s patient datasets. RR estimation provides the value of mean absolute error (MAE) for ECG as 1.25 bpm (MIMIC-II) and 1.05 bpm for the acquired data. MAE for PPG is 1.15 bpm (MIMIC-II) and 0.90 bpm for the acquired data. By using 1-minute windows, this method reveals that the filtering method efficiently extracted respiratory information from the ECG and PPG signals. Smaller MAE for PPG signals results from fewer artifacts due to easy sensor attachment for the PPG because PPG recording requires only one-finger pulse oximeter sensor placement. However, ECG recording requires at least three electrode placements at three positions on the subject’s body surface for a single lead (lead II), thereby increasing the artifacts. A reliable technique has been proposed for RR estimation.

AB - Abnormal vital signs often predict a serious condition of acutely ill hospital patients in 24 hours. The notable fluctuations of respiratory rate (RR) are highly predictive of deteriorations among the vital signs measured. Traditional methods of detecting RR are performed by directly measuring the air flow in or out of the lungs or indirectly measuring the changes of the chest volume. These methods require the use of cumbersome devices, which may interfere with natural breathing, are uncomfortable, have frequently moving artifacts, and are extremely expensive. This study aims to estimate the RR from electrocardiogram (ECG) and photoplethysmogram (PPG) signals, which consist of passive and non-invasive acquisition modules. Algorithms have been validated by using PhysioNet’s Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC-II)’s patient datasets. RR estimation provides the value of mean absolute error (MAE) for ECG as 1.25 bpm (MIMIC-II) and 1.05 bpm for the acquired data. MAE for PPG is 1.15 bpm (MIMIC-II) and 0.90 bpm for the acquired data. By using 1-minute windows, this method reveals that the filtering method efficiently extracted respiratory information from the ECG and PPG signals. Smaller MAE for PPG signals results from fewer artifacts due to easy sensor attachment for the PPG because PPG recording requires only one-finger pulse oximeter sensor placement. However, ECG recording requires at least three electrode placements at three positions on the subject’s body surface for a single lead (lead II), thereby increasing the artifacts. A reliable technique has been proposed for RR estimation.

KW - Algorithms

KW - ECG

KW - MIMIC II

KW - PPG

KW - Respiratory rate

UR - http://www.scopus.com/inward/record.url?scp=85052689302&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052689302&partnerID=8YFLogxK

U2 - 10.11591/eei.v7i3.1244

DO - 10.11591/eei.v7i3.1244

M3 - Article

VL - 7

SP - 487

EP - 494

JO - Bulletin of Electrical Engineering and Informatics

JF - Bulletin of Electrical Engineering and Informatics

SN - 2089-3191

IS - 3

ER -