Griffith Variable Step Size Sign FLANN Algorithm for Nonlinear ECG Cancellation

Abstract

The electrocardiogram (ECG) is a weak bioelectrical signal which is susceptible to noise, hence, the noise cancellation plays an important role for practical applications. Before that, various algorithms have been proposed to ECG denoising. However, the above efforts do not involve the nonlinear distortions, which may encounter in cancellation model. To address this problem, we propose a functional link artificial neural network sign algorithm (FLANN-SA) based on Griffith variable step size (VSS). Compared with existing algorithms, the proposed algorithm exhibits the improved performance than existing algorithms in terms of signal noise ratio (SNR) and the mean square error (MSE).