The smart mattress cover system for unobtrusive home sleep monitoring and automated sleep quality assessment

Introduction Long-term, non-invasive sleep monitoring is increasingly relevant for the management of chronic conditions; still, existing solutions face a trade-off between clinical accuracy and real-world usability. Most unobtrusive in-bed systems rely on proprietary algorithms, limiting transparency and reproducibility. This study introduces and validates a complete framework for unobtrusive home-based sleep monitoring based on the Smart Mattress Cover System (SMCS), integrating a textile-based pressure matrix and two triaxial accelerometers. Methods Sensors' informations are combined in a reproducible signal-processing pipeline extending the Cole-Kripke algorithm with adaptive thresholding for automated estimation of Total Sleep Time (TST), Wake After Sleep Onset (WASO), number of nocturnal awakenings (nWASO), and Sleep Efficiency (SE). An ablation analysis investigated the individual and combined contributions of accelerometric and pressure-derived signals and adaptive vs. fixed thresholding. The algorithm was developed using data from 13 healthy subjects monitored for 3-13 consecutive nights at home and evaluated against actigraphy. Results The pipeline achieved strong performance for TST ( r = 0.71, ICC = 0.74), moderate performance for nWASO ( r = 0.47, ICC = 0.55), and lower agreement for WASO ( r = 0.37, ICC = 0.22) and SE ( r = 0.29, ICC = 0.29). The ablation analysis showed complementary contributions of accelerometric and pressure-derived signals, with the clearest benefit of their combination for nWASO, while adaptive thresholding effectively accounted for inter-subject and inter-night variability compared to fixed configurations. Discussion These findings provide a promising initial assessment of SMCS feasibility in sleep parameters estimation, highlighting complementary information from the two modalities, particularly for sleep fragmentation. Further development incorporating physiological information and validation against polysomnography in larger, clinically representative cohorts are warranted.