Non-Contact Measurement of Blood Glucose based on Artificial Neural Network

Abstract

Diabetes mellitus is a disease caused by metabolic disorder that result a lack of insulin in human body. Deficiency of the insulin hormone causes abnormal blood sugar level fluctuation. These conditions should be treated appropriately to prevent the occurrence acute metabolic disorder and chronic complication, so an appropriate treatment can be done by medical personal. This paper develops a non-contact device to measure blood sugar level based on Artificial Neural Network (ANN). ANN in this device convert temperature difference between tragus and antihelix to an index of blood glucose. Tragus and antihelix are the name of two sections of human ear. Using infra red sensor that called thermopile, temperature of these section is measure. So there is no contact between human body and the device. Weight and bias of ANN determine by Back Propagation trained using data from conventional measurement. Experiment result using Matlab Simulink and Peripheral Component Interconnect (PCI) interfacing showed that the device can measured blood glucose and can be used for measurement easier, faster and less intimidating as occurs in conventional measurements.