Performance improvement of pulse oximetry-based respiration detection by selective mode bandpass filtering

Abstract

In this paper, an improved method to detect respirations by pulse oximetry during exercise is proposed. As a method for robust respiration detection, fixed bandpass filtering to block the heart beat signals is commonly utilized. But the fixed bandpass filtering cannot guarantee reasonable performances when the HR(Heart Rate) is varied highly. Therefore, the respiration detection performance is degraded. In the proposed algorithm, the HR information is used to estimate the RR(Respiration Rate). Using the RR, the corresponding bandpass filter(BPF) is selected to detect respiration points. The selection of the passband makes the proposed algorithm possible to guarantee the performance during exercise. Our test results show that the overall estimation error of the proposed algorithm was 20.32% during exercise. © Springer-Verlag Berlin Heidelberg 2007.