Levenberg–Marquardt-Based Non-Invasive Blood Glucose Measurement System

The present work focuses on the development of non-invasive blood glucose measurement device to revolutionize diabetes management and reduce severe complications associated with it. A low cost, painless and non-invasive blood glucose measurement system is designed using near-infrared (NIR) LED and four photodiodes for the purpose. NIR light emitted by LED passes through the skin and is detected by photodiodes after attenuation. The detector converts the attenuated light into a voltage signal. The interference due to background noise generated by human skin is removed by taking floating or internal reference. The voltage signal obtained from the photodiodes is calibrated using Levenberg–Marquardt-based Artificial Neural Network to obtain the glucose concentration. The accuracy of proposed prototype was examined by comparing non-invasively predicted data with invasively measured reference data. It is observed that all measurements lie in A and B zones of Clarke error grid and thus clinically accurate.