Abstract
In the non-invasive blood glucose concentration (BGC) sensing, the measurement based on near infrared spectroscopy has been a promising technology since it had acquired dozens of satisfactory results in short-term glucose monitoring tests. However, it’s still necessary to improve the measurement precision because it has challenges of the reduced precision in a long-term test when a lot of variables in the test would exist. Considering the requirement of multivariable analysis, the signals of diffuse reflectance spectra should include enough absorption information from glucose. However, the sensitivity of diffuse light intensity to the absorption variation at different source detector separations (SDSs) could be different. We present an analysis method using Monte-Carlo (MC) simulation and the diffuse equation for reasonably selecting proper SDS to get a satisfactory glucose measurement precision when there are multivariable disturbances. In the case of measuring glucose in a tissue phantom using the waveband of 1000-1340 nm, we show the SDS optimization result by using this analysis method. The experiment was designed to measure the diffuse reflectance spectra at 0.1-3.0 mm with the step of 0.1 mm, and the phantom solutions with different glucose concentrations and hemoglobin concentrations are tested. The glucose prediction precision was evaluated using the root mean squared error of prediction (RMSEP) for the all SDSs of 0.1-3.0 mm, and the SDSs with the lower RMSEP were selected for use. Moreover, the selected SDSs in the experiment shows a similar conclusion from the MC simulation. This work could be referenced to the in vivo BGC measurement.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call