Improvement of empirical mode decomposition based on correlation analysis

Abstract

Empirical mode decomposition (EMD) is an effective method for analyzing nonlinear and non-stationary signals. However, it can be found that there are two problems in this method. One problem is that it may lead to unnecessary redundant decompositions and bring interferences to the results of the analysis. Moreover, unnecessary decompositions may increase the computational time of the EMD algorithm. Another problem is that a signal belongs to the same one of intrinsic mode functions (IMFs) may be decomposed into other IMF components in EMD method. In order to determine the reasonable number of IMFs and reconstruct the signal belongs to the same IMF, an improved method of empirical mode decomposition based on correlation analysis is proposed in this paper. In this method, a simple but effective stopping criterion was proposed to reduce the number of decompositions and remove undesirable redundant decompositions. In order to avoid over decomposition, we reconstruct the signal belongs to the same IMF by using correlation analysis. Finally, the method we proposed is contrasted with the EMD method by analyzing simulation and real signals. The experimental results indicate that our method is very effective in removing redundant IMFs and the running time of our method is reduced by about 70% compared with EMD method. Meanwhile, a similar or better decomposition results can be obtained compared with the EMD method.

RDsLINet: A Novel Lightweight Inception Network For Respiratory Disease Classification Using Lung Sounds

“…From Table VII,it is clear that the suggested RDsLINet, has outperformed the existing research works. For task 2, though [2] provides decent amount of accuracy, the overall algorithm including the preprocessing steps, is heavily computationally expensive since it includes EMD, which is computationally heavy due to iterative decomposition of multicomponent signal [35], thereby increases the preprocessing time and the same has been discussed by Subho et.al [2]. Likewise, for task 3, we have outperformed the only existing work by Altan et al [13] by a margin of 7%.…”

A Vibration Feature Extraction Method Based on Time-Domain Dimensional Parameters and Mahalanobis Distance

“…To lessen the damage of machinery and keep the equipment performing at its best, people usually use the vibration signals to detect the faults of machine parts. Furthermore, different fault diagnosis methods have been developed [5][6][7]. Of course, many algorithms have also been proposed and applied to different fields, and a wealth of research results have been achieved [8][9][10][11].…”