Gold Medal · IIIEX 2025 2nd Place (IoT) · KMIPN VII 2025 IYSA Special Award
GLUVER: Non-Invasive Blood Glucose Monitor
Needle-free blood glucose monitoring system using a MAX30102 optical sensor, ESP32 microcontroller, machine learning model, and Flutter mobile app.
Role Mobile App & ML API Developer (Team)
Release date 2025-10
Category Mobile, IoT, AI
Technology
Flutter Firebase ESP32 Python scikit-learn Flask
Project Overview
GLUVER (Glucose Observer) is a non-invasive self-monitoring device for blood glucose estimation. The system integrates photoplethysmography (PPG) optical sensing, microcontroller processing, machine learning regression inference, and a real-time mobile app interface.
Developed collaboratively with fellow students at Politeknik Negeri Semarang, the project earned international and national awards in Internet of Things and medical engineering.
Problems
- Pain and Needle Trauma: Diabetic patients often avoid routine daily checks due to the necessity of finger-prick needles and lancets.
- Cost of Disposable Test Strips: Traditional glucometers require recurring consumable strips that create long-term financial burden.
- Manual Record Keeping: Standard meter readings are frequently lost or unorganized when patients visit their physician.
Target Users
- Diabetes Patients (Type 1 and Type 2): Individuals who require regular glucose level monitoring throughout the day.
- Prediabetic Individuals and Seniors: Users needing a comfortable and painless screening tool at home.
- Families and Caregivers: Reviewing daily fluctuation trends to guide dietary and physical activity adjustments.
Technical Decisions
- Optical Signal Acquisition with MAX30102: Uses infrared and red light transmission to capture fingertip arterial blood volume variations.
- ESP32 Data Streaming to Firebase: Microcontroller filters early signal noise and sends raw PPG data to the cloud wirelessly.
- Isolated ML Model via Flask API: Regression models built with scikit-learn in Python wrapped in a microservice endpoint predicting glucose concentration (mg/dL).
- Real-Time Flutter App: Displays live readings, trend charts, risk indicators, and lifestyle recommendations.
Challenges
- Signal Noise Sensitivity: Subject finger motion and skin thickness alter light absorption curves, requiring baseline calibration and filtering.
- IoT Connection Reliability: Ensuring robust data streaming from the hardware module to the mobile app under high network density during live jury evaluations.
Measurable Results
- Gold Medal: Electronics & IoT category at the Indonesia International Invention Expo (IIIEX) 2025 organized by IYSA and Polines (August 2025).
- IYSA Special Award: Special honor of innovation presented by the Indonesian Young Scientist Association.
- 2nd Place (IoT): National Polytechnic Informatics Competition (KMIPN VII 2025) hosted by Politeknik Negeri Padang (October 2025).