Non-invasive Blood Glucose Estimation Using Multi-sensor Based Portable and Wearable System

Currently, invasive glucometers are widely used for monitoring blood glucose level. It is required to prick a finger of the subject to collect the blood sample. In this paper, a cost-effective wearable system is proposed to estimate the glucose level in a non-invasive way. Data are collected using different sensors such as photoplethysmography(PPG), galvanic skin response(GSR) and temperature sensor, and fed to a convolutional neural network (CNN) model. The model estimates the blood glucose level, which is comparable with the glucose level obtained with conventional invasive technique. This system is believed to help in monitoring blood glucose, which leads to control of diabetes and other related diseases.