The Smart Mattress Cover System: a compact textile-based system for daily monitoring of sleep-related parameters

This dataset was developed within the EU-funded Tolife project, aimed at validating an artificial intelligence solution for analyzing patient data collected through unobtrusive sensors during daily-life. The primary objectives are to support personalized treatment, evaluate health outcomes, and improve the quality of life for individuals affected by chronic obstructive pulmonary disease. In this study, we employed one of Tolife’s daily-life monitoring devices: the Smart Mattress Cover System (SMCS), a compact, non-invasive, and scalable solution for real-world sleep monitoring. The SMCS comprises two components: the Smart Mattress Cover (SMC), placed over the mattress to acquire patient-related parameters, and the Bedroom Box Hub (BBH), a smart off-bed unit responsible for data acquisition and integrating environmental sensors. The SMC consists of a compact, low-density, textile-based pressure matrix (PM) and two embedded accelerometers (ACCs). Its raw data enables the extraction of heart rate (HR), breathing rate (BR), body movements and bed occupancy status. Here, we present the dataset only related to the SMC’s validation. Specifically, BR can be estimated from PM data by tracking mattress pressure variations induced by chest volume changes during respiration; HR can be derived from ACC data via ballistocardiographic analysis of heart-induced vibrations; body movements and bed occupancy can be detected using both PM and ACC signals. Eleven participants completed three dedicated protocols targeting (i) HR, (ii) BR, and (iii) movement and bed occupancy detection (MOV). Each protocol required subjects to lie in three randomly ordered positions: supine, prone, and lateral (left or right, selected at random for each trial). Reference HR and BR measurements were obtained using the Shimmer3 ECG unit (Shimmer Research, Ireland), worn on the chest with a 4-lead thoracic band configuration. Reference for MOV were obtained from the protocol itself. The dataset is organized into folders containing SMC data and reference signals, with individual .csv files corresponding to each sensor and protocol. The dataset is structured in two folders: the first stores data acquired from the SMC, while the second one contains the reference data. Within the folders, the organization follows the structure outlined in the "readme" file. The posture's sequences assumed by subjects during the protocols are stored in the three separate file for each protocol (position_HR.csv for HR, position_BR.csv for BR, position_MOV.csv for MOV).