Research On Blood Oxygen Saturation Detection Algorithm Under Exercise State

With the development of society,people's life pressure has gradually increased,and the prevalence and mortality of cardiovascular diseases have also increased year by year,becoming the number one killer of deaths in the world.Therefore,people are paying more and more attention to their own physical conditions.With the development of technology,the wearable device industry has gradually emerged,which makes it possible for people to effectively monitor their physical conditions in real time.Blood oxygen saturation(Sp O2)is the physiological index that best reflects the body's heart and lung function and the health of the blood circulatory system in the body.Real-time and accurate monitoring of changes in blood oxygen saturation can effectively prevent cardiovascular diseases.However,while the wearable blood oxygen saturation monitoring device brings convenience to people,it also introduces greater motion noise,which causes the accuracy of blood oxygen saturation value calculation to be greatly reduced or even wrong,which makes it possible to monitor blood oxygen saturation in real time.Lost its proper meaning.In response to these problems,this article mainly completed the following tasks:First of all,in order to solve the problem of inaccurate blood oxygen saturation monitoring caused by the large motion noise in the pulse signal under the exercise state,an adaptive method based on variational modal decomposition(VMD)and normalized minimum mean square error(NLMS)is proposed.The algorithm combined with the filtering method filters out motion artifacts(MA)in the photoplethysmography(PPG)signal.The signal is decomposed into several intrinsic mode functions(IMF)by the variational modal decomposition algorithm,and then the threshold is determined and selected by calculating the multiscale permutation entropy(MPE)of each mode The appropriate symptom modal function is used to construct the noise reference signal,and finally,an adaptive filtering algorithm is used to filter out noises such as motion artifacts in the blood oxygen pulse signal,which improves the calculation accuracy of the blood oxygen saturation value.Secondly,in view of the difficulty in selecting the key parameters of the variational modal decomposition algorithm,and the selection based on manual experience,the decomposition results may be inaccurate.The selection of the number of modes K and the penalty factor ? two key parameters caused by the signal decomposition result is analyzed.Therefore,a parameter optimization method based on gray wolf optimization algorithm is proposed to find the best combination of parameters,and the method of parameter empirical selection is used to verify the effectiveness of the optimization method.In addition,this article also designed blood oxygen saturation monitoring equipment.The front-end equipment for collecting photoplethysmography signals was built,and the mobile app was developed.The local data management module was used to perform algorithmic filtering processing on the collected photoplethysmography signals,and the processed pulse signals and calculations the obtained blood oxygen saturation value is displayed on the mobile terminal.Finally,in order to verify the accuracy of the algorithm proposed in this paper,the photoplethysmography pulse waves of the wrists of ten test subjects in the three motion states of walking,running and cycling were collected,and the method in this paper and another acceleration-based empirical model were used respectively.Two algorithms,state decomposition and acceleration-based variational modal decomposition,denoise the photoplethysmography signal,and calculate the signal-to-noise ratio and mean square error of the denoised signal.The experimental results show that the signal-to-noise ratio of the algorithm in this paper is higher,and the mean square error is smaller.In addition,after processing the PPG signal,the blood oxygen saturation values in the three exercise states are calculated.Compared with the other two methods,the Mean Absolute Deviation(MAE)of the method proposed in this paper is reduced by 2.0%?3.0%.Mean Absolute Error Percentage(MAEP)is also reduced by 2.0%?3.0%,so the calculated blood oxygen saturation value is more accurate,which proves that the algorithm in this paper is suitable for blood oxygen in the state of wearable device exercise.Saturation monitoring.