A Non-Invasive Blood Glucose Detection System Based on Photoplethysmogram With Multiple Near-Infrared Sensors

Area of Science:

• Biomedical Engineering
• Medical Devices
• Sensor Technology

Background:

• Non-invasive blood glucose detection is crucial for diabetes management.
• Photoplethysmogram (PPG) methods excel at baseline glucose prediction but miss daily fluctuations.
• Near-infrared (NIR) spectroscopy effectively tracks short-term glucose changes but is sensitive to external factors.

Purpose of the Study:

• To develop and evaluate a novel non-invasive blood glucose detection system.
• To combine photoplethysmogram and multiple near-infrared (m-NIR) sensing techniques.
• To overcome the limitations of individual sensing methods for improved glucose monitoring.

Main Methods:

• A hybrid system integrating PPG and m-NIR sensors was designed.
• A lightweight deep learning model was employed for data analysis.
• The system was tested on 10 participants during oral glucose tolerance tests (OGTTs).

Main Results:

• The combined system achieved a root mean squared error (RMSE) of 0.242 mmol/L.
• Achieved 100% accuracy within Zone A of the Parkes error grid.
• Demonstrated robust performance across approximately 7000 data segments.

Conclusions:

• The novel hybrid system effectively combines PPG and m-NIR sensing for non-invasive glucose detection.
• The developed system shows significant potential for accurate and reliable continuous glucose monitoring.
• This approach addresses key challenges in current non-invasive glucose sensing technologies.