Design and Validation of a Multimodal Non-Contact Sleep Monitoring System Based on Pressure, Acoustic, and Weight Fusion

Sleep monitoring is crucial for health assessment, but traditional polysomnography (PSG) remains too intrusive and complex for home use. This study introduces a non-contact monitoring system that combines pressure, acoustic, and weight sensors to track heart rate, respiration, snoring, and bed presence. Using bandpass filtering, envelope analysis, and multimodal signal fusion, the system reliably extracts physiological signals from background noise. Compared with PSG reference, the system achieved a mean absolute error (MAE) of 1.10 bpm and a correlation coefficient of 0.92 for heart rate, a MAE of 0.76 bpm and correlation coefficient of 0.90 for respiration rate with 33 participants. The proposed system achieved 98.9% recall in snore detection, and 1.2second off-bed response. The proposed system offers an accurate, low-cost, and fully contactless solution suitable for home-based sleep monitoring and elderly care.