| Heart rate monitoring as the most common examinations in medical care is one of the best way which measuring human health.Moreover,using wearable devices to monitor change of heart rate not only makes body be realized for human every time,but also provides a reference for controlling activity.It can avoid pressure for immune system because of ex-cessive exercise.However,heart rate monitoring error is large due to PPG signal affected by motion artifact easily.Therefore,how to reduced the influence of motion artifacts for PPG signal of the and improved the accuracy of heart rate monitoring would be studied in this pa-per.Two methods are proposed based on characteristics of PPG signal.The first one,with the name of adaptive filtering method based on acceleration sensor,used to extracting heart rate from PPG signal.A variable step size LMS algorithm based on the nonlinear relationship of hyperbolic sine function is proposed(as "HSFLMS").PPG signal for motion state was monitored through applying motion discrimination in framework.And then the PPG signal in motion case was filtered by adaptive method.The result showed that the PPG signal could be reconstructed effectively when HSFLMS applied in framework.Thus,heart rate could also be extracted easily.Due to problem which acceleration signal was no relationship with noise jamming signal may be appeared in the first one,a new framework of wavelet thresholding and adaptive fil-tering is proposed in the second method.The denosing PPG signal using wavelet thresholding and acceleration signal obtained by accelerometer as reference signal.A new nonlinear rela-tionship between iterative factor and error signalbased on "Lorentz" function is proposed in adaptive filtering system(as"LVS-LMS").The PPG signal could be reconstructed effectively and heart rate could also be extracted accurately compared traditional FFT method when us-ing new reference signal and "LVS-LMS" in the second framework. |
PDF Full Text Request