Research On Heart Rate Estimation Algorithm Of PPG Signal Under Human Exercise Condition

Real-time heart rate values helps people understand their own health status and have a pre-emptive effect on cardiovascular disease.Therefore,monitoring the realtime heart rate values can largely avoid the risks caused during exercise.At present,Photoplethysmography(PPG)is commonly used to monitor the real-time heart rate of the human body,which uses a light source to irradiate the human skin and a light receiving device to receive the reflected light to obtain volume change information of the human subcutaneous blood vessels,and then calculate the real-time heart rate of the human body.However,in the case of motion,the PPG signal will be disturbed by motion artifact(MA)due to light leakage,thereby,seriously affecting effective information in the PPG signal.Therefore,how to remove the motion artifacts in PPG signals and extract the heart rate values accurately in PPG signals is a technical problem that nedds to be solved urgently.By studying the motion artifact elimination algorithm,a joint algorithm based on cascaded RLS-CEEMDAN is used to remove motion artifacts.The cascaded RLS adaptive filter uses the three-axis acceleration signal as the reference signal of the Recursive Least Squares(RLS)adaptive filter,improves the structure of the adaptive filter,overcomes the defects of the single-stage adaptive filter with unsatisfactory denoising,and filters out most of the motion artifacts in the PPG signal.Then,using Complete Ensemble Empirical Mode Decomposition With Adaptive Noise(CEEMDAN),the PPG signal is decomposed and denoised and reconstructed to remove the residual motion artifacts in the signal and finally get a clean PPG signal.By studying the heart rate estimation algorithm,an improved spectral peak tracking method based on Grey Wolf Optimization(GWO)optimized Support Vector Machine(SVM)is proposed based on the existing Spectral Peak Tracking(SPT)algorithm in the frequency domain.The method extracts the eigenvalues of PPG signals and triaxial acceleration signals in the time and frequency domains and inputs them into the classification model to achieve automatic classification of spectral peaks under different exercise states and locate the correct spectral peaks corresponding to the real heart rate values,which improves the accuracy of heart rate estimation.In order to verify the superiority of the algorithm proposed in this paper,a variety of existing exercise heart rate estimation algorithms are compared using publicly available data sets,and the performance analysis and comparison of the generated results with the algorithm proposed in this paper are performed.From the theoretical analysis point of view,it can be seen that the algorithm in this paper is superior.By building an experimental platform for PPG signal acquisition,the PPG signals under different exercise states are acquired,analyzed and processed,and the heart rate detection results are evaluated using evaluation metrics.The experimental results show that compared with the existing heart rate detection algorithms,the error between the heart rate value restored by the GWO-SVM-based spectral peak tracking method and the real value of heart rate is smaller,and the detection algorithm has better performance.Figure [50] Table [5] Reference [85]...