A method to suppress the end-effects of emd based on lstm network

Abstract

The invention belongs to the technical field of signal processing, and in particular relates to a method for suppressing EMD end point effect based on LSTM network. The method effectively combines the prediction of the signal by the LSTM network and the application of EMD decomposition in the single-channel blind source separation technology, and finds the time step of the LSTM network according to the characteristics of the signal, which effectively improves the prediction accuracy, thereby effectively It solves the end effect problem of EMD method in single-channel blind source separation, and achieves good results in solving the sound separation problem in complex factory environment. This method can effectively solve the problem of end-effects in nonlinear non-stationary time series analysis, and can be applied to the separation of single-channel blind source mixed signals.

Description

technical field [0001] The invention belongs to the technical field of signal processing, and in particular relates to a method for suppressing EMD end point effects based on an LSTM network. Background technique [0002] Blind source separation technology refers to the process of separating each source signal by only using the observed signal according to the statistical characteristics of the source signal when both the transmission channel and the source signal are unknown. In fact, most of the collected signals are single-channel signals, and single-channel blind source separation is an extreme case of the underdetermined blind source separation problem. It only uses the characteristic information of a single-channel observation signal to separate multi-channel source signals. It is very difficult to solve. In 1998, Norden E. Huang of NASA first proposed a signal processing method to first perform empirical mode decomposition on a column of time series data, and then pe...

AI Technical Summary

Problems solved by technology

The endpoint effect in cubic spline interpolation and Hilbert transform has a great influence on the time-frequency analysis method based on EMD. If this problem is not handled properly, the effect of this time-frequency analysis method will be affected or invalidated.

However, in its application, the selection of LSTM network time step parameters has a great influence on the prediction results, so determining the LSTM time step to improve the prediction performance of data is the key issue to suppress the EMD endpoint effect

Images