红外多波长无创人体血糖检测阵列模型的研究

A Model Study on Noninvasive Blood Glucose Measurement with Multiwavelength Infrared Array

无创血糖监测不仅可以减少患者的痛苦 ,还能进行连续测量 ,从而降低并发症的发生 ,它是一种不需收集血样进行体内血糖浓度测量的新技术。本文根据朗伯—比尔定律 (The L ambert- Beer L aw)原理 ,为克服无创血糖检测研究中的难点和缺陷 ,详细分析了血糖红外光谱的吸收特性 ,利用阵列技术设计了红外多波长无创血糖检测传感器阵列模型 ,对关键技术进行了讨论。用 ME算法建立了信号检测和标定数学模型 ,结合了人体各种影响因素 ,使红外无创血糖检测的精度和稳定性得到了改善。列举了详细的实验实施步骤 ,同时叙述了系统调试的详细过程 。

The concentration of glucose in the blood may soon be measured noninvasively by near infrared multiwavelength sensor array without the painful puncture for obtaining a drop of blood. For overcoming the limitation of low measuring accurate degree and unstable working state, according to the Lambert Beer Law, the authors analyzed the feature of blood adsorption spectroscopy and designed an infrared multiwavelength blood glucose measuring sensor array to surmount the difficulties in noninvasive blood glucose measurement. The key technique, most suitable for detecting site and the influencing factors from human body were discussed, and the Mixture of Expert(ME) algorithm was adopted in building calibration model with multiple parameters of human body. It can overcome the existing problems and get more exact blood glucose information from the weak changes in spectral signals. Also presented and addressed in this paper are the detailed implementing steps of ME and the system,as well as the problems need to be solved.