Development of a Smart Lung Health Monitoring System Using Sensors and Data Analytics for Early Disease Detection

Abstract

This study introduces a novel multimodal wearable sensor system for real-time monitoring and analysis of respiratory and cardiac activity. The primary objective is to facilitate the early detection of cardiopulmonary abnormalities by integrating electrical (ECG) and acoustic data. A total of 30 participants, aged 25 to 50 years, were involved in controlled breathing experiments, which included deep (1000 ml, 15 breaths/min), moderate (750 ml, 20 breaths/min), and shallow (500 ml, 30 breaths/min) breathing, as well as coughing simulations. Signal processing using a 7th-order polynomial approximation yielded the lowest modeling error at 6.8%, ensuring precise waveform reconstruction. The system demonstrated a clear differentiation of respiratory patterns via Area Under the Curve (AUC) metrics, with average AUC values increasing from 1200 µV·s during shallow breathing to 3200 µV·s during deep breathing. Further analysis of the first derivative of AUC values revealed a strong correlation (r = 0.89) between respiratory volume and ECG amplitude fluctuations, highlighting robust cardiorespiratory coupling. Notably, the system achieved a 92% accuracy in detecting abnormal breathing events, such as shallow breathing and coughing fits. By combining ECG-derived heart rate variability with respiratory data, the system offers a comprehensive assessment of cardiopulmonary interaction. The key contribution of this work lies in its real-time, continuous monitoring capability using a compact wearable form factor, which distinguishes it from existing single-modality systems. This approach represents a significant advancement in non-invasive health monitoring, with strong potential for application in clinical diagnostics and home-based tracking of chronic conditions, such as asthma, COPD, and cardiac dysregulation.