Non-invasive Blood Glucose Determination using Near Infrared LED

Abstract

Diabetes Mellitus in human beings is an inability to produce adequate insulin or process them which will cause a rise in blood sugar level. To maintain the blood glucose level check-up at a regular interval is required. The conventional method available currently involves pricking which irritates skin when done repeatedly. The objective of this paper is to overcome the painful process and to be noninvasive. Our proposed system is based on the NIR (near-infrared) spectroscopy method of a selected wavelength which is more sensitive to blood glucose. This helps to reduce the error factor and the system is implemented with machine learning to improve the efficiency of the output.