Improved EMD algorithm based on polynomial

Abstract

The invention provides an improved EMD algorithm based on polynomial, which comprises the following steps: 1, selecting a window with the start point of the signal as the center point, and carrying out polynomial fitting on the signal value in the window; 2, taking the center point of the window as an average point to construct a discrete Gaussian signal; 3, multiplying that fitting sequence valuein the window and the correspond Gaussian signal value and summing to obtain the denoising value of the center signal point of the window; 4, sliding the window in steps of 1; repeating Step 1-3, completing denoising of all signal points of that signal; 5, Empirical mode decomposition being performed on the de-noised signal to obtain a reconstructed signal. The invention improves the traditionalEMD algorithm, decomposes the EEG signal with the algorithm of the invention, obtains the intrinsic mode function signal with relatively concentrated frequency, distinguishes the intrinsic mode function of each frequency band, greatly suppresses the problem of the end effect when decomposing, and improves the problem of the mode aliasing at the same time.

Description

technical field [0001] The invention relates to an improved EMD algorithm based on polynomials, and belongs to the technical field of intelligent information processing. Background technique [0002] Empirical mode decomposition (EMD for short) is a new adaptive signal time-frequency processing method proposed by E. Norden, Huang et al. in 1998, which is suitable for analyzing and processing nonlinear and non-stationary time series. The difference from the traditional signal analysis method is that it does not need to select the basis function in advance. According to the characteristics of the signal itself, the algorithm can adaptively generate suitable intrinsic mode functions (IMFs) through repeated screening, and these intrinsic modes The state function can well reflect the empirical mode decomposition of the local frequency characteristics of the signal at any time. [0003] When performing empirical mode decomposition, there are two issues that are most concerned: on...

AI Technical Summary

Problems solved by technology

Take the right end point as an example, if the point is the maximum value (minimum value), then the upper envelope (lower envelope) can take it as the right end point, and no large swing will occur; but for the lower envelope (upper envelope) Envelope) Since the right end point is not a minimum value (maximum value), and the right end point of the lower envelope (upper envelope) cannot be determined, a large swing will occur, and this swing will gradually move towards the Inner pollutes the entire data sequence

Images