In the non-invasive blood glucose sensing based on near-infrared spectroscopy, the skin scattering variation during the long-term blood glucose monitoring would be a big challenge to get an accurate measurement result. We present a scattering and absorption separating method for the near-infrared diffuse reflectance spectra of scattering media. And we use the extracted absorption part, which is the medium’s effective attenuation coefficient (EAC) spectrum, to improve the accuracy for long-term glucose monitoring. We optimized the light source-detector separation (SDS) setup to realize the maximum sensitivity for the EAC spectra-based measurement. The measurement uses two SDSs to perform a differential on their diffuse reflectance spectra, as the differential could help to reduce the light drift during the long-term in vivo tissue monitoring. The SDS setup optimization for the two positions was tested by the Monte-Carlo (MC) simulation of tissue. The human oral glucose tolerance test (OGTT) with the optimized SDSs also shows a satisfactory blood glucose prediction result. In conclusion, this method shows a good application potential in the non-invasive blood glucose sensing.
|