RF-Based Noncontact Respiratory Rate Monitoring With Parametric Spectral Estimation

Abstract

Respiratory rate (RR) monitoring of an adult or an infant during sleep or in steady position can be lifesaving, especially in home care systems. In this paper, we examine the potentials of wireless radio frequency (RF) signals to monitor the RR. A new noncontact RR monitoring system is proposed. The system includes a simple motion detection algorithm based on adaptive thresholding to eliminate the effects of the large-scale body movements on the RR estimation. In the proposed system, the high resolution subspace-based parametric spectral estimation approaches, estimation of signal parameters by rotational invariance technique (ESPRIT) and multiple signal classification (MUSIC), are presented as the RR estimation algorithms. According to our knowledge, the ESPRIT algorithm, which estimates the spectrum without searching, is used for the first time in this paper for RR estimation. It is shown that ESPRIT is computationally efficient and works approximately 49 times faster than the MUSIC algorithm. It is also shown with various experiments conducted with ten volunteers that the proposed noncontact RR monitoring system attains 0.13 breath per minute (bpm) error rate with the limited number of data and outperforms the periodogram method commonly used as the benchmark.