Strong Robustness Heart Rate Estimation Using Discrete Fourier Transform and Personality Heart Rate Characteristic

摘要

Photoplethysmography (PPG) heart rate (HR) measurement technologies have the drawback of robustness. The traditional threshold method is not able to compute HR from the disorderly and unsystematic PPG signal (especially the PPG signal come from the people who has hypertension). To enhance HR estimation robustness, we present a strong robustness fusion method (SRFM) which fuses discrete fourier transform (DFT) and HR characteristics of testers to estimate HR. The commercial infrared pulse wave sensor is used to collect PPG signal. We use the db5 wavelet transform to carry out pre-processing which can remove the power frequency noise and baseline drift. The frequency signal of PPG signal is converted by DFT. According to personality, we choose the PPG frequency signals peak. HR can be compute by this peak. Experimental results show that, compared with the traditional time domain feature extraction method and extreme method, the proposed approach is improved significantly in the accuracy rate and recall rate.