Non-invasive blood glucose measurement performance analysis through UWB imaging

Diabetes is a silent killer disease worldwide. Regular blood glucose level monitoring and healthy life management is the only remedy as this cannot be cured. In conventional way, need to draw a drop of blood by pricking fingertip and measure the glucose level through a device. This creates mental trauma among most of the diabetes patient specially the children. To overcome this problem, a non-invasive and patient friendly way of measurement is essential, which is the focus of this paper. Some investigation have been conducted in non-invasive field but none could stand for a durable period in terms of accuracy, comfort and safety. This paper presents an Ultra-wide band (UWB) microwave imaging and Artificial Intelligence (AI) based a solution with performance analysis to detect blood glucose concentration level non-invasively (i.e., without taking any blood sample). The system consists of a pair of small UWB biomedical planar antennas, UWB transceiver as hardware and an artificial neural network (ANN) with signal acquisition and processing interface as software module. The UWB signal with center frequency of 4.7 GHz was transmitted through one side of left hand and forward scattering signals were received from another side. Characteristic features from received signal were extracted for pattern recognition and detection through ANN. The system exhibits 81% accuracy to detect glucose concentration in blood plasma. Besides, it is user friendly, affordable, safe and can be used with comfort in near future.