Respiratory Rate Estimation from the Photoplethysmogram Combining Multiple Respiratory-induced Variations Based on SQI

Currently measuring respiratory rate (RR) accurately and conveniently is still a difficult problem. The photoplethysmogram (PPG) obtained from pulse oximetry contains a variety of physiological information related to human health, such as heart rate, blood pressure, oxygen saturation and RR. This paper presents an algorithm to estimate respiratory rate based on the characteristic parameters of PPG. Firstly, signal quality index (SQI) is used to identify the low-quality PPG signal. Secondly, respiratory waveform is modeled by four features extracted from PPG signal. Thirdly, Autoregressive modelling (AR) is used to obtain frequency of the four respiratory-induced variations. Fourthly, RR is estimated by a data fusion method combine the four respiratory-induced variations. Finally, the proposed method has been tested on physiological recordings from some adults. The accuracy of the RR is assessed using root mean square (RMS) error. The median RMS error obtained for RR is 0.75 breaths/min showing that the proposed algorithm has achieved high accuracy and robustness.