Preliminary Study Of The Non-invasive Measurement Of Blood Compositions Based On Dynamic Spectroscopy

Non-invasive measurement of blood compositions using near-infrared spectroscopy (NIRS) is a foreland in the biomedical engineering field that still not been totally fulfilled. The current focal points of the research are the weak signal due to low component concentration and the numerous disturbance factors in the in vivo measurement. Based on analysis and characterization of the photoplethysmography (PPG), using the inherent characterization of the artery vessel, a new method, dynamic spectroscopy, which can eliminate the interference of the individual discrepancy and measurement conditions theoretically, is come up with. And the testing system is built also.The main content of the dissertation involves:The approach of the dynamic spectrum was first proposed. The principle of eliminating the interference of the individual discrepancy and measurement conditions are presented. Two ways to acquire the dynamic spectrum are described, one is the time domain spectroscopy and the other is the spatial domain spectroscopy. High precision pure dynamic spectrum is extracted from the pulse wave in the test.The design of the hardware and software of the non-invasive measurement system of the dynamic spectrum in time domain has been conducted. The performance test of the Acousto-optic Tunable Filter AOTF has been finished. With the earlap selected as the measuring spot, a multi-layer simulation model of the earlap is built. The transmittance of the earlap model and its regularities of distribution under vary beam properties and model parameters are researched. The CDS (Correlated Double Sampling) is applied in the detection of the weak signal. The weak current signal is converted to voltage by IVC102U, a precision switched integrator tran-impedance amplifier. Using the 24 bit⊿-∑ADC, the acquisition of weak signal is realized. The rapid coding of the MCU software based on Simulink/Stateflow has been conducted. Using the modulus maxima algorithm, the singularity in the pulse wave can be successfully located and corrected, and the precision of the pulse wave signal is improved. In order to get higher accuracy of amplitude measurement, suppress the respiration interference of the pulse wave, the original pulse wave signal is processed by the characteristic parameter reconstruction based on the coherent averaging combined with the wavelet analysis. Using the traditional Statistical estimation theory and the assembly average method, the measurement accuracy can be raised by the outlier eliminating.The quantitative analysis of non-liner characteristic of the whole blood spectra corresponding to different optical path length is conducted, and the method of multi- optical path length modeling method is presented. The problem of the best optical path length selected corresponding to vary component is solved and the resolution of the blood calibration model has been raised.The dynamic spectrum method proposed a new idea for the NIR non-invasive measurement applicant in the biomedical field. The patient may abandon the invasive measurement method with high price and easy infection. It has great application value in this field. Theory and experimental results disclosed by this dissertation laid solid basis for the successful application of this method.