A Low-Cost Sensorized Mask for Respiratory Monitoring: Calibration and Proof-of-Concept Evaluation

In recent years, remote monitoring devices have achieved a central role worldwide, driven by their ability to collect and transmit data in a continuous and non-invasive way. The potential to provide remote clinical support and facilitate early intervention in emergency scenarios has led to increased interest in the healthcare system. However, respiratory monitoring is still less integrated into portable solutions than other physiological measurements. In fact, currently available solutions have some significant limitations including accessibility, high costs and ease of use. These shortcomings limit a large scale adoption of this technology. To overcome these limitations, this work presents a low-cost system based on a sensorized face mask. The device integrates a miniaturized environmental sensor (BME280) which detects cyclical variations induced by breathing, enabling the reconstruction of volume-time curves and the estimation of key respiratory parameters. Respiratory volume is obtained through a calibration procedure based on a linear regression model, using a commercial spirometer (Spirobank-G) as a reference. Preliminary validation on a single subject yielded promising results, with a Root Mean Square Error (RMSE) equal to 0.11 L and a coefficient of determination R2 of 0.91 for respiratory volume estimation. In addition, the estimated respiratory rate (RR) shows a deviation of less than 1%, with the reference values, while inspiratory (tINSP) and expiratory (tEXP) times exhibit a relative error of approximately 5%. This suggest a good segmentation of respiratory cycles. Conversely, the estimated tidal volume (TV) and minute ventilation (VE) values show a slight underestimation compared to the reference. Future works will focus on validating a larger cohort and optimizing the volume-estimation algorithms