A Method for Estimating Human Respiratory Rate and Heart Rate Using Sparse Spectrum Analysis

Monitoring human health conditions has always been a significant concern, especially in the operational safety industry, such as civil aviation. Traditional contact-based methods can accurately measure human heartbeat and other vital sign information but have certain limitations for the users have to wear devices or sensors. Therefore, research has recently been increasingly attracted to non-contact methods such as optical cameras, infrared thermography, and millimeter wave radar. This paper uses Linear Frequency Modulated Continuous Wave (FMCW) mm-wave radar for non-contact measurement of vital signs. Considering the time-varying and frequency-domain sparse characteristics of vital sign signals and the low-sampling number limitation that will be faced in realistic scenarios, the Sparse Iterative Covariance-based Estimation (SPICE) is utilized to estimate the vital sign signals. The SPICE method demonstrated outstanding performance compared to the non-sparsity methods, including Fast Fourier Transform (FFT) or Chirp Z-Transform (CZT), for estimating the human respiratory and heart rates accurately in simulations and experiments, especially under short sampling time and low-sampling number conditions.