Research And Application Of Heart Rate Estimation Based On PPG Against Motion Artifacts

Real time monitoring of heart rate during exercise can help people to understand the physical condition and reduce the risk of sports in time.PhotoPlethysmoGraphy(PPG)has the advantages of easy to wear and low power consumption.It is the first choice for real time heart rate estimation of wearable smart bracelet.The PPG signal acquisition on the intelligent hand ring needs a light source and a light receiving device.By irradiating the human artery,the intensity of the light intensity along with the pulse change is received,thus the pulse wave signal is collected and the heart rate is calculated.In view of this,the real-time estimation of heart rate in the movement of smart Bracelet based on PPG signal has become the focus of current research.However,due to the non contact wearing characteristics of the hand ring,the light path of the PPG sensor irradiated the skin in the movement will cause the irregular change to cause the leakage,and the reflected light intensity contains random interference,which makes the collected PPG signal including the Motion Artifact(MA),especially in the violent motion,the motion artifact will be even more.Strong.Therefore,how to accurately estimate the real-time heart rate in the PPG signal from the strong motion artifact jamming is a practical problem to be solved.In order to solve the above problems,this thesis proposed an anti-motion artifact PPG signal heart rate estimation algorithm.The algorithm mainly included three parts:signal denoising,time-frequency conversion and spectral peak tracing.(1)The signal denoising stage removes most of the motion artifacts from the motion PPG signal.In view of the non-stationary characteristics of PPG signals in motion,this thesis used Variational Mode Decomposition(VMD)to divide the PPG signal into several modalities,and removed large signals by removing the features in the frequency domain of acceleration signals;(2)In the time-frequency conversion stage,even a relatively clean PPG signal is still difficult to perform heart rate estimation in the time domain,so the PPG signal is converted from the time domain to the frequency domain.In this stage,the frequency spectrum of the PPG signal is calculated using the Fast Fourier Transform.(3)The spectrum peak tracing phase is to find the frequency of the heartbeat from the spectrum of the relatively clean PPG signal to estimate the heart rate value.In the algorithm preprocessing stage,there may be a small amount of interference for the stationary PPG signal during the initialization phase of the algorithm.The stationary PPG heart rate estimation based on the Fourier transform may sometimes estimate the failure problem.This thesis used the Hilbert-Huang transform.Transform,HHT)obtains a high-resolution marginal spectrum and optimizes the initial heart rate estimate.In order to verify the performance of the proposed algorithm,the simulation results are compared with a variety of existing motion heart rate estimation algorithms on open datasets.Secondly,in order to obtain more data in the form of movement,this thesis setted up a motion PPG signal acquisition platform and collected experimental data under four common motion conditions in life.This set of data includes:synchronous PPG signals,acceleration signals,and ECG signal(as a reference for true heart rate values).Based on this set of data,the algorithm proposed in this thesis and the classical exercise heart rate estimation algorithm were compared.The experimental results showed that the algorithm proposed in this thesis can obtain better heart rate estimates for both the 12 or 10 sets of public data sets and the data sets collected by the experimental platform under the four motion states.