Edge-Enabled IoT Framework for Real-Time Heart Rate Monitoring and Anomaly Detection

Authors:

NITHISHWAR B

M.Tech Student, Department of Information Technology, Puducherry Technological University, Puducherry, India

Abstract — Cardiovascular diseases (CVDs) remain one of the leading causes of mortality worldwide, necessitating the development of intelligent, continuous, and affordable cardiac monitoring systems. This paper proposes a comprehensive real-time heart rate monitoring and anomaly detection system that integrates biomedical sensing, embedded processing, Internet of Things (IoT) communication, and advanced deep learning techniques. The system employs the ESP32 microcontroller interfaced with the AD8232 ECG sensor and the MAX30102 PPG pulse sensor to acquire multimodal physiological signals. These signals are transmitted to a cloud platform for visualization and storage, while a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) deep learning model performs intelligent cardiac anomaly detection. The CNN component extracts spatial features from waveform morphology, while the LSTM component learns temporal dependencies in cardiac rhythms. Due to practical constraints in real-time data collection, the model is trained and validated on high-quality annotated datasets from PhysioNet, comprising 14,552 ECG samples. Experimental results demonstrate an overall classification accuracy of 99.3%, a ROC-AUC score of 0.9994, a Fake Class F1-Score of 98.8% for normal and 99.5% for abnormal heart rhythms. The proposed system provides a scalable, low-cost, and portable framework for continuous cardiac health assessment across different age groups, bridging the gap between clinical-grade equipment and consumer-level monitoring devices.

Keywords: Real-time monitoring, ECG, PPG, CNN–LSTM hybrid model, ESP32, heart rate classification, age-group analysis, IoT healthcare, anomaly detection, PhysioNet.