Removing subject dependencies on Non-Invasive Blood Glucose Measurement using Hybrid Techniques

Diabetes mellitus is a major health problem which needs regular glucose monitoring for management of the disease. Invasive blood glucose measuring systems with acceptable accuracy are currently used for measurements. Several non- invasive blood glucose measurement techniques are reported in research literature, but most of them are not commercially available due to low accuracy or dependency on individual physiological parameters. This paper presents a hybrid technique to measure blood glucose level non-invasively using a combination of multi-wavelength Near Infrared (NIR) spectroscopy and bio-impedance spectroscopy. Physiological parameters of individuals and other environmental factors that affect non-invasive glucose measurements have been identified and compensated to make the device work on any person without calibrations. The measured parameters, along with the glucose level of the subject obtained from a commercial blood glucose meter is used to train a Random Forest regression algorithm. A training set of 315 data samples were used for the development of the system. The accuracy of predictions was tested using a testing set of 80 data samples. The trained system can predict the blood glucose levels non-invasively with 90.7% accuracy.