Monte Carlo simulation-based thinning and calculating method for noninvasive blood glucose sensing by near-infrared spectroscopy

Abstract

Previous results show that the floating reference theory (FRT) is an effective tool to reduce the influence of interference factors on noninvasive blood glucose sensing by near-infrared spectroscopy (NIRS). It is the key to measure the floating reference point (FRP) precisely for the application of FRT. Monte Carlo (MC) simulation has been introduced to quantitatively investigate the effects of positioning errors and light source drifts on measuring FRP. In this article, thinning and calculating method (TCM) is proposed to quantify the positioning error. Meanwhile, the normalization process (NP) is developed to significantly reduce the error induced by light source drift. The results according to TCM show that 7μm deviations in positioning can generate about 10.63% relative error in FRP. It is more noticeable that 1% fluctuation in light source intensity may lead to 12.21% relative errors. Gratifyingly, the proposed NP model can effectively reduce the error caused by light source drift. Therefore, the measurement system for FRPs must meet that the positioning error is less than 7μm, and the light source drift is kept within 1%. Furthermore, an improvement for measurement system is proposed in order to take advantage of the NP model.