ECG Dynamical System Identification Based on Multi-scale Wavelet Neural Networks

Abstract

In order to identify the Electrocardiograph (ECG) dynamical system more accurately, a system identification method based on multi-scale wavelet neural networks is proposed in this paper. Firstly, the stationary wavelet transform is used to remove the baseline drift and high frequency noise of ECG signal; Secondly, wavelet theory, radial basis function neural networks and grid points are used to design the multi-scale wavelet neural networks architecture; Finally, in order to facilitate the iteration of discrete data, the discrete difference equation is used to replace the continuous differential equation in the system identification algorithm, and the value range of gain parameters is proved by Z-transform. In this paper, the effectiveness of this method is verified by using three-dimensional ECG signals from PTB database, which also opens up a new research method for the identification of ECG dynamical system.