The accuracy of respiratory rate estimation using electrocardiography and photoplethysmography

Respiratory rate can often be extracted from electrocardiogram (ECG) or (photoplethysmogram) PPG by analyzing its' beat morphological feature or the variability of time interval. In this paper, the respiratory rates were derived from ECG and PPG signals by using two algorithms based on beat morphology and time interval, respectively. The four groups of estimation results were evaluated by comparing with the reference respiratory rate collected using a piezoelectric sensor. In the in-situ experiments, ECG, PPG and reference respiratory signals were collected from ten subjects. To evaluate the performances of each algorithm at different respiratory rates, the subjects were asked to breathe at the natural rhythm, at the controlled respiratory rates and hold the breath during the experiment. The results demonstrated the method based on ECG peak feature was more accurate and stable among these methods.