Research On Networked And Noninvasive Blood Oxygen Monitoring System

With the increase of aging population,along with the spread of various diseases,more and more attention has been paid to people’s health.It is very important to establish a sound networked medical monitoring system.Blood oxygen saturation is one of the most important parameters to reflect the physiological state of human body,so this paper focuses on the research of networked blood oxygen monitoring systemFirstly,the improved algorithm of peak valley extraction,sliding window method with variable window length,has been proposed.With apllying jumping detection method,calculation quantity is reduced.Due to real-time update of jumping step,the algorithm is not only efficient but also has good adaptability to the change of heart rate,which lays the foundation for motion artifact removal algorithm and estimation of blood oxygen.Secondly,in order to remove motion artifact in PPG signal,two kinds of algorithms are discussed.The first one is a new method which combines the independent component analysis algorithm and least mean squares adaptive filter for reducing MA in corrupted PPG signals.The novelty of the method lies in the fact that a synthetic reference signal for an adaptive filtering process is generated internally from the MA-corrupted PPG signal itself instead of using any additional hardware.The combination of periodic moving average algorithm and ICA can separate the PPG component and the MA component better than the ICA only method.And the underlying PPG signal could be extracted from the MA contaminated PPG signals automatically by using independent component selection algorithm after applying ICA.Then the amplitude information of the PPG signals could be recovered by using adaptive filters.The second one is MA detetion and removal algorithm.The statistics of the segmented signals is detected by setting a threshold and high-quality signal segments will be extracted with MA segments eliminated.By using high-quality signal segments to calculate the saturation of blood oxygen instead of the whole signal,accuracy of blood oxygen detection can be improved.Thirdly,feature extraction algorithm based on linear regression has been introduced.Correlation between red and infrared PPG signals has been proved.When the number of samples is near the sampling frequency,the value of R will have smaller deviation and linear regression algorithm is more stable than peek-valley algorithm.Calibration test concludes that the value of R is highly correlated with blood oxygen.Finally,the realization of the whole networked blood oxygen detection system is described.The background and Realization of the networking are firstly illustrated.Then the hardware system is introduced,including the selection of oxygen sensor,the description of WiFi module and the function of the acquiring and processing module.At last the signal analysis software system based on Android system has been introduced,and its main functions include data communication,real-time waveform display and calculation of physiological parameters.