The Research On Dynamic Measurement Of Human Physiological Parameters Based On Photoplethysmography

Cardiovascular diseases(CVDs)is usually caused by abnormalities in the heart,blood vessels and blood transport systems.According to the data released by World Health Organization(WHO)in 2015,31%of people die of this disease,which is a serious threat to human health.It is important for the diagnosis and treatment of CVDs that to mornitor heart rate,blood pressure,blood oxygen concentration and blood viscosity dynamically.At present,there are some shortcomings of traditional physiological parameters measurement methods,such as lack of creativity,inconvenience and poor measurement of comfort,which can not support the continuous measurement.Human pulse wave is an important criterion to judge the health of cardiovascular system.The research of physiological parameter measurement model is usually based on pulse wave.Compared with other pulse wave measurement methods,photoelectric measurement method measures human volume pulse wave by specific spectrum.The process of measurement is good repeatability,high comfort and support continuous measurement.In this paper,we mainly studied the relationship between photoplethysmography(PPG)and different physiological parameters.Then,we established the corresponding measurement models for heart rate,blood pressure,blood viscosity and blood oxygen saturation.The measurement results were displayed on the watch,which has good academic significance and market application value.In this paper,a series of explorations and researches have been carried out for the difficulty in the process of model establishment and complexity in calibration process of blood oxygen saturation measurement model.First,523nm green spectrum,660nm red light spectrum and 810nm infrared spectrum were selected as light source.In the design of the shape structure of the watch system,the "mesh" groove structure is introduced to eliminate the ambient light interference and improve the signal-to-noise ratio of the measured signal.In the hardware circuit design,the watch integrated low-power communications,micro-sensor technology,ambient optical sensor,low power control processor and Bluetooth Low Energy(BLE)model,which implement PPG signal amplification and filtering in hardware level.According to the Lambert-Beer law,the relationship between the parameters,such as pulse wave morphological features and waveform area,and the pulse pressure difference(systolic pressure-diastolic blood pressure)was deduced.Seoncd,based on the traditional generalized morphological baseline drift removal algorithm,this paper made some improvement to achieve a fast and low power signal processing at the expense of partial accuracy.This simplified algorithm reduced computational complexity by four times.Third,the GM(1,1)model was introduced to compensate the PPG signal distorted by motion artifacts.Then combined with acceleration model,dynamic heart rate measurement mode was established.The experiment results showed that more than 94%measurement value within the error range ?/-6;According to the Lambert-Beer law,the relationship between the PPG waveform parameters and the pulse pressure difference(systolic pressure-diastolic blood pressure)was derived.The indirect continuous measurement of diastolic pressure was achieved and the experiment results validated that the measurement error range within ?/-10,which satisfy the regulations of Association for the Advancement of Medical Instrumentation(AAMI);The key parameter K is calculated from systolic and diastolic pressure,then take K into empirical formula to implement the blood viscosity measurement.Regard the recorded data in the Multiparameter Intelligent Monitoring in Intensive Care(MIMIC)?database as experimental analysis object,the experiment results show that about 95%of the measuremt value within error range of ?/-0.5;According to the modified oxygen saturation measurement formula,the unknown coefficient is calibrated.A continuous measurement model of blood oxygen concentration with similar calibration environment was obtained.The average error of the measured results was 0.076%.At last,based on the connection between hand-held device and cloud service platform,multi-dimensional physiological information can be collected and stored.Not only to achieve remote monitoring of physiological parameters,but also to provide a basis for healthy trend analysis through deep mining information of physiological parameters.