Phantom Experiment Study On Approximate Floating Reference Measuring Method For Non-invasive Blood Glucose Sensing With Near-infrared Spectroscopy

In near-infrared non-invasive blood glucose measurement,the signal variation caused by glucose is very weak,vulnerable to human background,measuring instruments and surrounding environment changes,limiting the accuracy of non-invasive blood glucose sensing.Aiming at this problem,the thesis presents an approximate floating reference measuring method which is based on the theory of floating reference put forward by the research group.We use this method to correct the background variation of measurement,and study the correction effect by phantom experiments.Firstly,this thesis introduces the meaning and research progress of non-invasive blood glucose measurement,and addresses the current technical problems encountered in the study.Specially,how to eliminate the impact of background changes in the measured objects themselves becomes the focus of the current study.This thesis summarizes the transmission of light in the biological groups,according to the absorption and scattering's characteristics of light in body tissue,leads to the floating reference theory,then proves the existence of the floating reference position and the principle of background correction.Based on Monte Carlo simulation method,the thesis studies the floating reference points of palm's three-layer skin model,2% and 3% Intralipid phantom solution.Found that their floating reference points are basically similar,so try to use approximate floating reference measuring method which detect diffuse light in two different radial distances by double loop optical fiber probe.The radial distance with0.6 mm as the main position,the radial distance of 2 mm as the reference measurement location.Then derives the method's principle of light source drift's correction by formulas derivation.The performance of the experimental system are evaluated,shows that the light source will basically reach a steady state when the measurement system is turned on for about one hour later.Two photodetector circuit's dark noise are relatively small whether in a short or a long term,their drift trend are basically reaching agreement,so their impact can be ignored.The system's signal to noise ratio can achieve a relatively ideal level in a short or long term.Then analyzes the measuring system's limit of detection accuracy,makes the simulation calculation of system's limit concentrationof resolution,then validates the limit of detection accuracy whether reach the requirements by phantoms Intralipid solution experiments.With 2% and 3% concentration of Intralipid solution as the research objects,the experiment of pure Intralipid solution without glucose shows that the approximate floating reference measuring method can reduce light drift influence on the measured results,improve the repeatability and stability of the data.Through three-days' experiments,based on multiple diffuse reflection signal acquisition of 2% and 3%Intralipid solutions with different glucose content and the glucose concentration prediction models,found that the regression models' root mean square errors of prediction were reduced by 38.51%~79.98% and 29.72%~52.22%.The comparison shows that this method could effectively eliminate the common background changes at two measured positions,improve the prediction accuracy of the calibration models.