Non-Invasive Blood Glucose Monitoring using Deep Learning

Monitoring of blood glucose levels is an essential part of diabetes management. However, conventional monitoring techniques rely on invasive methods such as finger prick testing, which can cause discomfort and reduce patient compliance. In this study, a non-invasive blood glucose monitoring system is proposed using photoplethysmography (PPG) signals acquired from a MAX30102 sensor. The collected signals are processed to remove noise and extract important physiological features such as mean, standard deviation, peak count, and signal ratio. These extracted features are used as a feature vector for blood glucose estimation. A machine learning model, specifically a Random Forest Regressor, is employed to predict glucose levels based on the extracted features. The hardware implementation of the system is carried out using an ESP32 microcontroller, and the predicted glucose values are displayed on an OLED screen for real-time monitoring