Measuring In Vivo Tissue Spectra For Long-term Non-invasive Glucose Monitoring And Preliminarily Analyzing The Spectral Characteristics Of The Multiple Measurement Interference

In the non-invasive blood glucose measurement(NBGM)based on Near-infrared spectroscopy,achieving stable in vivo tissue spectra is one of the primary tasks to realize the long-term blood glucose monitoring.This thesis is trying to find an effective index for the biological spectrum using Near-infrared spectroscopy to figure out the effects from blood glucose and other multiple variables during the measurement.And then these results would be beneficial to reduce the influence from the variables on the precision in long-term measurement.In this thesis we proposed a new method to measure the effective attenuation coefficient(EAC)of the biological tissue,and the spectrum of EAC is applied to demonstrate the skin's current state at that measurement moment.This method uses the differential diffuse reflection spectroscopy based on two light source-detector distances,and it can reduce the instrument drift during the longterm monitoring or repeated measurements.Using this method,the variable in vivo skin states can be directly monitored and the multiple effects from the different individuals or different skin conditions can be analyzed,and then we can use the calibration method to reduce the influence from them.The main work and results of this paper include:Firstly,the theory about measuring the effective attenuation coefficient(EAC)spectrum of biological tissue using the differential diffuse spectroscopy was introduced.The measurement equations are derived by Diffuse Equation(DE).Furthermore,the experiment results have validated the feasibility of this method.Then,the detection limit of the blood glucose concentration's measurement is discussed based on the acquired glucose EAC spectra.Secondly,the influence from some variations of the tested skin during the longterm spectra measuring was investigated.These variations may be induced by lots of sources,which can be different individuals or different skin area even different skinprobe contacts.Specifically,the spectral characteristics induced by skin scattering coefficient variation,blood hemoglobin concentration difference,skin temperature variation and contact pressure between skin and probe were tested.Monte-Carlo simulation and phantom experiment are used to get the spectral data under these variations.Moreover,the method of multivariable correction is applied to calibrate the glucose prediction model to reduce the effect of these interferents from in vivo test on the EAC spectra.The results show that the method can help to get a robust prediction model for the in vivo glucose measurement.In summary,this thesis presents a method of measuring biological tissue EAC spectrum based on the differential diffuse reflection spectroscopy.And the EAC spectra characteristics for the in vivo skin with variable conditions are investigated.We believe that the proposed EAC measurement method and the in vivo test results would be helpful to monitor and analyze the skin condition during long-term glucose measurement.Moreover,this method also would be referred to other non-invasive blood components measurement.