Research On Non-Invasive Measurement System Of Multi-Physiological Parameters Based On Near-Infrared

Measuring physiological parameters is of great significance for assessing health status,preventing diseases and avoiding disease progression.Traditional measurements for blood glucose and hemoglobin require blood samples,which cause trauma and pain,and cannot be monitored in real time.Therefore,it is valuable to investigate non-invasive,continuous and real-time detection technology.Near infrared(NIR)spectroscopy has been widely investigated in the field of physiological parameter measurement due to its non-invasive,rapid and accurate performance.However,it still needs to be improved in the reliability of theory,practicability of equipment and accuracy of results.In view of existing problems,the non-invasive detection technology of blood glucose and hemoglobin was investigated,and the main works and results are as follows:The feasibility of NIR spectroscopy was demonstrated by simulation research.The absorption and scattering characteristics of tissues to different wavelengths of photons were investigated,and the physical meaning and calculation methods of optical parameters of tissues were introduced.Based on anatomical features,a fingertip tissue model was established,which composed of epithelial layer,dermis layer,fat layer,muscle layer and phalanx.The transport of photons in fingertip tissue was simulated using the Monte Carlo method.Results manifest that the main interference in the measurement of blood components by NIR spectroscopy is the absorption of light by the epithelial layer;the transmitted light intensity can effectively reflect the periodic heartbeat and the change of blood components.A non-invasive measurement system for multi-physiological parameter was developed.The relationship between blood glucose,hemoglobin concentration and photoplethysmography(PPG)signals was established by mathematical model derived using the modifiedLambertBeer’s law.And a blood component analysis method based on multi-wavelength PPG signals was proposed.The non-invasive measurement system composed of PPG probe,hardware device and upper computer was developed,which realized high-precision synchronous sampling of 7-channel PPG signals,waveform visualization,human-computer interaction and measurement of multi-physiological parameters such as blood glucose level,hemoglobin concentration,respiratory rate and heart rate.Algorithms for PPG signal processing,prediction models for blood glucose level and hemoglobin concentration,and methods for respiratory rate detecting were investigated.98 hemoglobin and 76 blood glucose samples were collected for modeling using partial least squares regression,support vector regression,multilayer perceptron and random forest regression algorithms.The results show that the performance of support vector regression and multilayer perceptron model is better.The multilayer perceptron model got the best performance,on the testing set,the mean absolute error,mean relative error,root mean square error and correlation coefficient of blood glucose level and hemoglobin concentration predicted by the system were: 0.818 mmol/L and 1.146 g/dL,14.3% and 10.2%,0.984 mmol/L and1.515g/dL,0.646 and 0.718.Bland-Altman plot analysis found that 96.4% of the hemoglobin samples and 100% of the blood glucose samples fell within the consistency limits,indicating a good consistency of the system measurement results.Clarke’s error grid analysis found that 93.3%of blood glucose samples fell in zone A and 6.7% in zone B,both of which met the clinical requirements.Respiratory rate was detected using the fast Fourier transform method,and measurements on seven subjects showed that the mean absolute error was 0.46 times per minutes and the correlation coefficient was 0.976.In summary,the simulation research in this paper provides an effective basis for the noninvasive detection method of blood components based on NIR spectrum,and a non-invasive measurement system for multi-physiological parameters with high practicability was developed.The research outcomes could be applied to anemia screening,blood glucose detection and home health monitoring,etc.