An Effective Adaptive Filter to Reduce Motion Artifacts from ECG Signals Using Accelerometer

Abstract

Electrocardiogram (ECG) is most commonly used for the diagnosis of heart disease. Physicians utilize high quality ECG to interpret and identify physiological and pathological phenomena. However, motion artifacts are caused by the electrode-skin impedance with electrode motions that are the results of a subject's movement, so it is a classical problem to reduce motion artifacts from ECG signals during real-time heart rate measurements. To the best of our knowledge, Block Least Mean Square (BLMS) algorithm has not been considered in the context of the use of accelerometer measurement data in ECG signals as a reference signal on reducing motion artifacts. Therefore, in this paper, we propose a 3-axis accelerometer to measure the acceleration signal of the movement of the trunk as the reference input of the adaptive filter and the optimal weight of the adaptive filter is adjusted by the BLMS algorithm. Finally, we have applied this algorithm on ECG signals from the subject and compared its performance with the conventional Least Mean Square (LMS) algorithm. The results show that the performance of the BLMS algorithm is superior than the LMS algorithm. And, we have found the R wave of the filtered ECG was clearly appeared.