Non-Invasive Glucose Monitoring Using PPG, AI, and IoT-Driven Mobile Integration for Real-Time Diabetes Management

Diabetes mellitus patients must regularly monitor their blood glucose levels to manage glycaemia, typically requiring capillary tests at least three times daily and laboratory tests one to two times per month. These conventional methods involve finger pricking, causing significant discomfort and stress. This study introduces an innovative non-invasive glucose monitoring approach by integrating photoplethysmography (PPG) technology with an artificial intelligence (AI) algorithm, complemented by a mobile application using Flutter, IoT systems, and Firebase cloud for real-time data access. Among the AI models tested, polynomial regression demonstrated superior accuracy in glucose prediction, achieving a Mean Squared Error (MSE) of 14.76 and an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.729. Despite challenges such as motion artifacts and ambient light interference affecting PPG signals, the system provides an advanced, user-friendly solution for patients and healthcare teams to monitor glucose levels effectively and conveniently.