Study Of The Non-invasive Arterial Blood Signal Detection Base On Near-infrared Dynamic Spectrum

Near-infrared spectroscopy for the determination of human arterial blood information,has non-invasive, continuous, real-time detection of the characteristics, receives global attention in Areas of the body signal detection. Frequency Domain Extraction of Dynamic Spectrum eliminated from the principle of individual differences and measurement conditions on the spectral detection. The method is based on Lambert-Beer Law, extract the corresponding pulse of arterial blood absorbance. The Spectrum formed by these absorbance, after Fourier transform to extract the spectrum of the signal, could identify the arterial pulse wave. In the experiment, a photoplethysmography sensor has been designed which used MSP430 as core of the system. By this sensor, detects the signal of Finger arterial blood with extracting and processing. The arterial pulse wave identified has been identified and Verifying the theory of dynamic spectral detection in the arterial blood on the feasibility of non-invasive detection by wearable sensor in a real environment.The near-infrared blood sensing technology in the experiment, although the photoelectric pulse sensor takes advantage in rich payload of effective information, which is able to provide effective sampling data to dynamic spectral detection theory, its practical application is limited by individual diversity. In order to overcome the sensing problem of different thickness of the finger-sides, the relationship between the pressure and the DC component and AC component in test result is analyzed. Through experiments of testing finger-sides with different thickness, the existence of an optimal detection region which allows the pulse wave detected by the sensor in most ideal state is pointed out.