TY  - JOUR
T1  - Fusion of Induced Variations Using Quality Metrics to Estimate
Respiratory Rate from Photoplethysmography Signal
AU - Anuar Nayan, Nazrul AU - Mohamad Rosli, Nur Azhani 
JO  - Journal of Engineering and Applied Sciences
VL  - 13
IS  - 21
SP  - 9101
EP  - 9105
PY  - 2018
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2018.9101.9105
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2018.9101.9105
KW  - Respiratory rate
KW  -photoplethysmogram
KW  -algorithm
KW  -signal quality indices
KW  -estimation
KW  -critical
illnesses
AB  - Among the vital signs of acutely ill hospital patients, Respiratory Rate (RR) is a highly accurate
predictor of health deterioration. The most common method for measuring RR in hospitals is transthoracic
Impedance Pneumography (IP). The drawback of IP which measures impedance at the electrocardiogram
electrodes is the injection of high-frequency alternating current into the tissue through drive electrodes. Thus,
IP becomes an active electronic device. The usage of IP may also cause natural breathing disturbance in
patients and eventually contributes to discomfort. This study aims to evaluate the RR from passive and
noninvasive acquisition module, Photoplethysmogram (PPG) signals. Algorithms comprise signal quality
indices. The RR estimation method for extracting three respiratory signal-induced variations of PPG was
described. The three respiration rates were combined through a weighted average using quality metrics for each
signal. The weights were determined using good quality MIMIC II benchmark datasets. PPG signal and
reference breathing signal using nasal air flow sensor of 20 healthy subjects have also been recorded and the
RR has been combined. The Mean Square Error (MSE) was 0.86 breath/min compared with the reference RR.
The proposed methodology could replace the manual counting method of RR, uncomfortable nasal airflow
sensor, chest band and IP which are often used in hospitals. Given its simple setup, the future system can
increase the efficiency of the RR monitoring frequency for patients with critical illnesses.
ER  -