Research On Blood Glucose And Cholesterol Detection Technology Based On Bioelectrical Impedance

Diabetes and hyperlipidemia are both common chronic metabolic diseases,which are prone to cardiovascular diseases and other complications,seriously endangering people’s health.The incidence rate is increasing year by year in recent years.Diabetes and hyperlipidemia do not have clinical symptoms in the early onset,easy to miss the best treatment opportunity.Therefore,early detection and continuous monitoring of blood glucose and blood lipid is the key to the treatment of diabetes and hyperlipidemia.At present,the detection methods of blood glucose and blood lipids are mostly invasive or minimally invasive,which cost a lot and bring pain to patients.Cholesterol is one of the important indicators to detect blood lipid,therefore,many researchers begin to research on noninvasive detection of blood sugar and cholesterol,and has achieved initial results,but there are also many insufficiencies.For example,due to the complex environment in the human body,as well as physiological information collected less,some of the methods proved to be feasible in vitro simulation experiments have great errors in actual anthropometry.Non-invasive detection equipment is expensive,bulky and not suitable for real-time monitoring.In view of the above shortcomings,this paper carried out non-invasive detection of blood glucose and cholesterol based on bioelectrical impedance technology.In order to meet the requirements of non-invasive detection,bioelectrical impedance spectroscopy method is adopted to design the detection platform.The platform mainly includes signal excitation module,signal conditioning module and detection module.For blood glucose measurement and cholesterol measurement,the bioelectrical impedance spectrum should be measured in the frequency range of 10MHz～60MHz and100KHz～300KHz respectively.Therefore,frequency synthesis technology is studied and used to design a signal excitation module that meets the requirements.In view of the problem of signal noise interference,as well as the signal detection circuit requirements,filter design,amplifier circuit design and constant current source design and other technologies are studied to design the signal conditioning circuit,used to complete the signal filtering,amplification,constant current conversion and other processing;Aiming at the characteristics of amplitude and phase change of frequency signal after passing through human body,a detection module was designed to obtain the complex impedance information of human tissue indirectly.Aiming at the problem that temperature has influence on the non-invasive blood glucose detection results by bioelectrical impedance spectroscopy,a temperature measurement circuit was designed to realize the simultaneous measurement of body surface temperature and ambient temperature.In order to obtain the relationship between the data collected by the hardware platform and blood glucose,and the relationship between the data and cholesterol,data preprocessing method based on Grubbs criterion,feature selection method based on principal component analysis,and modeling method based on multiple linear regression are studied for the establishment of blood glucose prediction model and cholesterol prediction model.The blood glucose prediction model was solved and verified by collecting bioelectrical impedance spectrum,temperature and standard blood glucose data in human experiments.The results showed that the blood glucose prediction model could reflect the trend of blood glucose concentration well.The feasibility and good repeatability of bioelectrical impedance spectroscopy in cholesterol noninvasive detection were verified.