Research And Implementation Of Adaptive Interference Elimination Algorithm For Dynamic ECG Signals

Wearable electrocardiogram(ECG)monitoring systems have been increasingly used in patients with cardiovascular disease(CVD).ECGs are a measure of electrical signals generated when the heart beats.It is a very common signal for medical professionals because it can indicate an individual's heart rhythm and can be further used to detect specific abnormalities in the heart,such as arrhythmia,coronary heart disease and hypertension.However,ECG signals are often corrupted by various noises,such as power line interference,motion artifacts,myoelectric interference and baseline drift.Among them,the motion artifact is caused by the relative movement between the electrode and the skin.During the daily movement of the wearer,the relative motion between the electrode and the skin will generate an electrode-skin impedance signal,thereby causing motion artifacts.It is very difficult to remove the kind of the noise,because it changes according to the wearer's movement and usually falls in the same spectrum as the ECG signal itself.This noise has no regularity and cannot be removed by traditional filtering methods.To improve the quality of ECG signals,suppress motion artifact noise,increase the signal-to-noise ratio,and reduce the complexity of operations,this paper improves the adaptive cancellation algorithm.Experiments show that compared with traditional methods,the method proposed in this paper has a significant improvement effect and is significantly better than other traditional methods.1.Aiming at the problem that traditional filtering methods cannot suppress motion artifacts,this paper proposed a wavelet adaptive elimination method based on single inertial sensing.First,a self-made wearable fabric chest strap is used,which contains two fabric electrodes for the acquisition of mixed signals of ECGs and motion artifacts.It is installed on the chest or left arm or right arm to collect human motion information.After using wavelet processing,irrelevant noise information is removed and used as the reference input signal of the adaptive filter;then an adaptive cancellation algorithm is used to process the signal.Finally,experimental results prove that the method can effectively suppress motion artifacts.Moreover,the experimental results are compared with the traditional adaptive filtering methods,and the results show that the proposed method is more effective,improving the signal-to-noise ratio and convergence rate.2.Considering the insufficient information of single inertial sensor to capture the relative motion information of the ECG electrode of the wearing chest strap,this paper further proposes a wavelet adaptive elimination method based on multi-inertial sensing fusion,which can more accurately simulate and predict motion artifacts through data fusion,thereby further improving the correlation of the reference input signal in the adaptive filter.This method uses multiple auxiliary inertial sensors with different properties from the working electrode to detect the reference signal and studies the importance and effectiveness of the reference signal in the adaptive motion artifact removal algorithm from another angle.The algorithm selects the most relevant signal from the motion information collected by inertial sensing installed in different parts of the body as the reference input signal of the adaptive cancellation algorithm,so that more information about the electrode movement can be collected through multiple inertial sensing.Relevant motion information is used to improve the correlation,and irrelevant motion information is removed by wavelet transform preprocessing to further improve the correlation of the reference input signal,thereby improving the ability of the adaptive cancellation algorithm to suppress motion artifacts in dynamic ECG.3.Test and verify the two adaptive elimination methods proposed in this paper and evaluate their performance.The test content mainly includes subjective and objective evaluations and ECG signals and exercise information acquisition tests in the real state of human motion.Performance indicators include signal-to-noise ratio,mean square error,convergence rate,and frequency spectral density.Experimental results show that the two methods proposed in this paper can effectively suppress the motion artifact noise in the dynamic ECG signal.Moreover,the proposed wavelet adaptive elimination method based on multi-inertial sensing is more effective than the proposed method based on single inertial sensing to eliminate motion artifacts.