Abstract
With the growing demand for health monitoring, non-contact vital signs monitoring technology has garnered widespread attention. While traditional health monitoring methods are accurate, they have limitations in terms of non-contact and non-invasive capabilities. This paper proposes a non-contact vital signs monitoring method based on frequency modulated continuous wave (FMCW) millimeter-wave radar, named MRVS, to enhance both convenience and accuracy. The method consists of three steps: signal processing, decomposition, and reconstruction. Firstly, the millimeter-wave radar is used to detect chest movements, extracting both respiration and heartbeat signals. Then, combining signal superposition with phase difference techniques effectively eliminates static clutter and respiratory harmonic interference, enhancing the signal. Next, discrete wavelet transform (DWT) is utilized to suppress clutter and noise further, performing signal decomposition. The reconstruction module employs an adaptive Kalman filter (AKF) combined with square root normalization for accurate heart rate estimation. Experimental results demonstrate that this method achieves an estimation error rate of less than 7% under different distances, angles, and postures, showcasing high accuracy and robustness and providing a new solution for non-contact vital signs monitoring.