LAUSR

High-quality sleep is a fundamental guarantee for human physical health. Sleep monitoring is crucial for understanding sleep quality and structure, but traditional sleep monitoring methods often require specialized lab environments and uncomfortable sensors. This article proposes a noncontact sleep monitoring method based on millimeter-wave radar, which accurately measures breathing and heartbeat rates during sleep by preprocessing reflected signals and using variational mode extraction (VME) algorithm and simultaneously detects snoring events. In order to improve the accuracy of snoring detection, a sliding window algorithm based on data density and amplitude is introduced, effectively filtering out low-amplitude noise and accurately identifying snoring segments. Experimental results show that this method can detect snoring events with an average error of 0.10 s, and the accuracy of breathing rate and heart rate (HR) detection is approximately 0.72 and 2.56 bpm, respectively. This method demonstrates high real-time performance and reliability and is expected to find broad applications in the diagnosis of clinical sleep disorders and home health monitoring.