IoT-Based System for Non-Invasive Blood Glucose Prediction, Heart Rate, and Oxygen Saturation Monitoring

Abstract

Non-communicable diseases such as diabetes mellitus and cardiovascular disorders account for over 70% of total mortality in Indonesia. Continuous physiological monitoring is essential for early disease management; however, conventional invasive blood sampling methods cause physical discomfort and lack real-time capabilities. This paper presents an Internet of Things (IoT)-based portable health monitoring system designed for simultaneous measurement of oxygen saturation () and heart rate, alongside non-invasive prediction of blood glucose concentration. The hardware incorporates a MAX30105 optical photoplethysmography sensor, an ESP32 microcontroller, and a simple linear regression algorithm for non-invasive glucose estimation. Telemetry data is transmitted over Wi-Fi to Firebase Realtime Database and displayed live on a Flutter-based Android application equipped with automated alert notifications. Experimental validation across 20 human subjects demonstrated mean absolute percentage errors of 4.86% for heart rate, 2.74% for , and 4.69% for blood glucose prediction. Real-time early push alerts were reliably triggered upon detecting physiological anomalies. The integrated prototype provides a practical, accurate, and sustainable solution for continuous personal health telemetry and non-invasive screening.