Design, Development, and Validation of the Smart Mattress Cover System: A Compact and Scalable Textile-Based Platform for Unobtrusive At-Home Multi-Parametric Sleep Monitoring

The transition toward Healthcare 4.0 marks a paradigm shift from hospital-centered ters in daily-life conditions has become a cornerstone of preventive and patient-centered care. Among the enabling technologies driving this transformation, wearable and textile-based systems stand out for their ability to merge sensing capabilities with comfort, ergonomics, and usability–making them ideal for long-term health monitoring outside clinical environments. Sleep is one of the most critical indicators of health, occupying one-third of human life and influencing physical recovery, cognitive function, and emotional balance. Poor sleep quality and disorders—such as insomnia and obstructive sleep apnea (OSA)—are highly prevalent and strongly associated with chronic cardiovascular and metabolic diseases. Consequently, accurate, long-term monitoring of sleep-related arameters is essential for both clinical management and large-scale preventive healthcare. However, current monitoring solutions present significant limitations. Clinical gold standards such as polysomnography (PSG) ensure precision but are invasive, time-consuming, and restricted to short clinical sessions. Questionnaires and actigraphy extend monitoring duration but provide only subjective or motion-based data, lacking physiological insight. Commercial devices offer accessibility but rely on proprietary, non-transparent algorithms and unvalidated outputs. Research prototypes, while technically advanced, often employ dense, rigid sensor arrays that compromise comfort and scalability. It is therefore clear that there is a need for compact, ergonomic, and scientifically validated flexible systems capable of reliable sleep monitoring under real-world conditions. The main objective of this doctoral research was to enable unobtrusive in-bed monitoring of sleep-related parameters through a compact, textile-based and scalable sensing system validated under realistic conditions. This dissertation presents the design, development, and validation of the Smart Mattress Cover System (SMCS) and its custom algorithms. The SMCS integrates textile-based sensing, embedded electronics, and intelligent signal processing into a single, unobtrusive platform for continuous sleep monitoring. The SMCS comprises two main components: the Smart Mattress Cover (SMC), a flexible cover positioned on the bed surface, and the Bedroom Box Hub (BBH), an external off-bed unit. The SMC integrates a 4×10 low-density piezoresistive pressure matrix and two embedded accelerometers for ballistocardiography (BCG) analysis, within a compact sensing area of only 40×50 cm. From the SMC data, key patient-related parameters (PP)–including heart rate (HR), breathing rate (BR), body movement, and bed occupancy status–are extracted. The BBH serves as the intelligent hub, acquiring environmental parameters such as temperature, humidity, air quality, light, and noise, while managing data storage and wireless transmission. This integrated rchitecture enables the simultaneous acquisition of physiological and environmental data, supporting integration with cloud-based infrastructure and providing a comprehensive view of sleep dynamics. The system development followed a structured engineering cycle–from requirements analysis to hardware realization and algorithmic validation–addressing technical, clinical, and data-driven constraints. From a manufacturing perspective, the SMCS validates the feasibility of large-scale, textile-integrated sensor fabrication through standardized and reproducible processes. The pressure matrix layers and the SMCS multi-layer assembly were designed for mass production using industrial sewing, screen printing and laser-cutting techniques. The system’s architecture leverages commercially available electronic components and 3Dprinted enclosures, reducing costs and facilitating rapid prototyping and deployment. Over 100 units have already been produced and deployed within the EU-funded TOLIFE project on Chronic Obstructive Pulmonary Disease (COPD) patients, confirming both scalability and readiness for real-world applications. Although deployed within TOLIFE, the SMCS operates as a stand-alone platform, easily adaptable to other clinical or research contexts. Custom algorithms were developed to extract PPs under realistic, sleep-like conditions and were validated across multiple subjects, mattresses, breathing patterns, postures (supine, prone and lateral both left and right sides), and movement frequencies. The HR estimation algorithm, based on an adaptive fusion of accelerometer and pressure matrix data, achieved a mean root mean square error (RMSE) of 2.02 bpm, mean absolute error (MAE) of 1.60±1.28 bpm, Pearson correlation R = 0.999 (p<0.01), a bias of -0.19 bpm, and limits of agreement of ±4.6 bpm in the Bland–Altman analysis, maintaining stable performance across all performed postures. Similarly, the BR estimation algorithm, relying exclusively on pressure matrix data, achieved RMSE of 3.75 acts/min, MAE of 2.05±3.15 acts/min, Pearson correlation R = 0.943 (p<0.01),ma bias of 0.33 acts/min, and limits of agreement of ±7.3 acts/min in the Bland–Altman analysis, confirming robustness across different breathing rates and postures. Finally, the body movement and bed occupancy classifier, based on a custom Artificial Neural Network (ANN) exploiting both pressure and accelerometer data, achieved a global accuracy of 95.09±2.84% across random sub-sampling runs and 98.39% in the bestperforming model, accurately distinguishing among in-bed stillness, in-bed movement, and off-bed conditions. This thesis provides both scientific and technological contributions to the field of unobtrusive sleep monitoring. Scientifically, it demonstrates that accurate estimation of HR, BR, and body movement can be achieved using a textile-based system. To the best of our knowledge, the SMCS represents the most compact low-density system capable of simultaneously detecting multiple parameters. Technologically, it introduces a modular and scalable architecture that combines sleep-related and environmental sensing while preserving comfort, flexibility, and usability. Compared to state-ofthe-art systems, the SMCS achieves comparable or better accuracy with significantly fewer sensors, confirming the feasibility of our approach. In line with the principles of Healthcare 4.0, the SMCS establishes the foundation for advanced sleep analysis, including posture detection, sleep staging, and sleep quality assessment, bringing the system closer to medical-grade performance standards. Overall, the SMCS advances the field of unobtrusive sleep monitoring by providing a scientifically alidated, ergonomic, and scalable solution. It contributes to the broader vision of Healthcare 4.0, enabling continuous assessment of sleep and related parameters, with significant implications for decentralized, continuous healthcare and real-world deployment.