<![CDATA[The decline in lung function associated with Chronic Obstructive Pulmonary Disease (COPD) requires an accurate, portable, and cost-effective monitoring system to support the early detection of respiratory disorders. However, conventional spirometry devices are often limited in terms of portability, affordability, and accessibility for community-based screening. This study aims to develop a portable lung volume measurement and classification system based on an ESP32 microcontroller and an MPXV7002DP differential pressure sensor. Respiratory pressure signals were converted into airflow to estimate Forced Vital Capacity (FVC) and Forced Expiratory Volume in One Second (FEV1), while the FEV1/FVC ratio was calculated using Body Temperature and Pressure Saturated (BTPS) correction to improve measurement accuracy. A Naïve Bayes classifier was employed because of its low computational complexity, fast inference, and low memory requirements, making it suitable for resource-constrained embedded systems such as the ESP32. The classification model was trained using the NHANES dataset and evaluated using 40 prototype measurement samples obtained from 40 participants. The proposed system achieved 95% accuracy, 90.5% precision, 100% recall, and a 95% F1-score. These findings indicate promising classification performance in the prototype evaluation; however, the results should be interpreted with consideration of the sensor-level measurement limitations identified in this study. Further validation involving a larger and more diverse participant population and reference spirometry measurements is required before the system can be considered for broader respiratory screening applications.]]>
Copyrights © 2026
