Study On The Measurement Error Of Floating Reference Point

With the ever increasing number of diabetic,the demand for near infrared noninvasive blood glucose detection technology becomes more and more urgent.However,the complicated and real-time changes from such as the stability of measurement conditions,measurement instrument and physiological background of the measured subject make it extremely difficult to extract the signal that specific to glucose concentrations,which severely restrict the measuring accuracy as well as the wide application in clinical of this technology.To solve these problems,the floating reference measuring method is proposed by our group to remove or eliminate the influence on spectra caused by the aforementioned background changes.Based on Monte Carlo(MC)simulation,this study mainly focus on the measuring errors of a custom-built multipath receiving fibers system which is used for the measurement of the floating reference positions.The conclusions of this study are expected to further amend the measuring accuracy and universal application of floating reference method,and make this method to apply to in vivo non-invasive glucose sensing as soon as possible.Firstly,the optimum measuring conditions were studied which mainly including: the best measuring time,the most suitable solution concentration for experiment,etc.According to the experimental results,we come to the conclusion that in order to minimum the gross error caused by the instability of light source,the system need to be preheated 4 hours before measurement.Considering the precision of the instrument and the consistency of the MC simulation and in vitro experiments,the solutions with glucose concentrations ranging from 1000 mg/d L to 2500 mg/d L are used in this study.Secondly,the Thinning and Calculating Method(TCM)was put forward to realize the quantitative analysis of the measuring errors caused by the positioning inaccuracy on the basis of MC simulation.TCM has two prominent superiorities.On the one hand,it can effectively guarantee the uniqueness of the error source,that is,the specified positioning error.On the other hand,the work efficiency can be improved significantly due to the flexibility in multiple sets of errors analysis by using only one certain MC simulation data.According to the investigation results based on TCM,a deviation of 7 ?m in positioning could approximately lead to a 10.63% relative error in the position of floating reference point.Hence,it comes to the conclusion that the positioning inaccuracy should be kept within 7 ?m for achieving more precise measurement.Afterwards,the effects of light source drift were discussed quantitatively based on MC simulation.The random drift of the light source was obtained by changing the number of the incident photons stochastically.The results showed that the light source drift had a dramatic effect on measurement accuracy of floating reference positions.A fluctuation of 1% in light source intensity might result in a 12.21% relative errors in floating reference positions.In order to reduce the error caused by the light drift,a Normalization Process(NP)was the proposed.By calculation,this model could effectively reduce this error by improving the relative error from 12.21% to 0.14%.Finally,the comprehensive errors caused by positioning inaccuracy and light source drift were studied through both MC simulation and in vitro experiment.The results suggested that the synthesis influence from these two factors was more complicated than the influence from their own respective,which is not just the simple superposition of them.Gratifyingly,the Normalization Process showed good effect on reducing the error caused by light source drift during the measurement process of the floating reference positions.