Optical Sensor Based Continuous Blood Glucose Estimation Using Lightweight Distributed Architecture

Diabetes is a non-communicable disease and people face many health issues due to diabetes. Worldwide, more than 500 million people have diabetes today. Currently, most devices measure blood glucose levels from a person’s blood, collected by pricking a fingertip. It is necessary for people with Type-1 diabetes to regularly check their blood glucose levels to prevent major health issues. Several times pricking a fingertip in a day feels very painful. This situation increases the demand for non-invasive blood glucose measurement devices, and a certain amount of research has been reported in recent years. This paper proposes a less complex, near infrared-based non-invasive device that measures glucose level continuously with higher accuracy. This paper also proposes a lightweight distributed architecture that utilizes the readings of the proposed device to estimate blood glucose level. The proposed architecture is validated using root mean square error (RMSE), mean absolute difference (MAD), and mean absolute relative difference (mARD) which are calculated as 6.08 mg/dL, 1.79 mg/dL, and 1.51%, respectively.