| The diabetes is a kind of endocrine obstacle disease, and without radical cure for it, the main way is to frequently monitor and control the blood glucose level. So the protable noninvasive blood glucose apparatus is very important.In order to reduce data redundancy, speed up computation, improve the detection accuracy and suit operation for TMS320F2812, the algorithm of measuring blood glucose was improved and optimized, based on the method of COEM (Conservation of Energy Metabolism), which estimates the blood glucose level with the metabolism rate, oxygen saturation, pulse frequency and blood flow of the finger. Firstly, by analyzing the insufficient of the method of metabolic heat integration, optimizing the theoretical hypothesis, increasing the input parameters and founding the blood glucose testing model of non-linear adaptive regression multianalysis initially on the basis of the first noninvasive blood glucose apparatus. Then, ameliorating algorithm models at all levels, including of the metabolism rate, blood flow, air saturated water vapor partial pressure, surface convection coefficient and sensor preprocessmg.Improving the algorithm of blood flow, a new one was founded on non-steady state, based on the method of temperature field distributing. According to this algorithm model, a experimental platform was designed to realize the algorithm in the case of the regulable and variational temperature, humidity and wind velocity, which size was 700mm×274mm×200mm. Including:the base, the blood flow simulator, the central experiment cavity, and the control& transmission components of sensors and receptors.To be cabinet, portable and operated easily, the apparatus adopts the systemic project based on DSP, TMS320F2812. Firstly, based on the method of COEM, a integrator of sensors was designed to measure the blood sugar level non-invasively, which size was 60mm×30mm×85.5mm, and result was optimization of the first non-invasive blood sugar measuring sensor. Including:the base, the blood flow measuring region of the finger surface, the metabolism rate measuring region of the part of finger, the oxygen saturation measuring region infrared of the part of finger, the link region of circuit and the information in-out region. Secondly, using the keyboard with prompting of system menu, the physiological parameters were received non-invasively by the apparatus, such as oxygen saturation, pulse frequency and blood glucose level. Lastly, the results were displayed on LCD.The results of experimental analysis and data processing about integrator suggest that the integrator is better than the first non-invasive blood sugar measuring sensor. Magnesium Tablet, which size was 1.5mm×5mm×55mm, was selected by the selection experiments of metal material and response time analysis of steady-state recovery. The result of optimal location analysis of sensors shows that the optimal location is 49.5mm. Output signal of integrator is slowly changing and low frequency signal, and is the most abundant in the rang of 0-0.02 Hz. At last, the results of timing analysis of integrator further show that the integrator is better than the first sensor too.The results of 30 preliminary clinical experimentations suggest that the average relative different is 0.039 and the boundary of average relative different is less than 0.5 through the statistical analysis, which show the integer part of the results is effective and accurate. The result of significance test is 0.86, greater than 0.05, which shows no significant difference. Besides, the metabolism rate is more important than other parameters through the correlative multianalysis, which correlative coefficient is up to 0.92. Then, the correlative coefficient is up to 0.8621, between the noninvasive apparatus based on COEM and Biochemical Analyzer, which shows strong correlation and good coherence, and the detection accuracy increases by 8.21%. |
PDF Full Text Request