A Comparative Study on Decomposition of Test Signals Using Variational Mode Decomposition and Wavelets

Abstract

The decomposition of signals into their primitive or fundamental constituents play a vital role in removing noise or unwanted signals, thereby improving the quality and utility of the signals. There are various decomposition techniques, among which the linear wavelet technique and the Variational Mode Decomposition (VMD) are the most recent and widely used ones. This paper presents a comparative study of the decomposition of spatially inhomogeneous test functions namely Doppler and Bumps used by statisticians. An effort is made in this article to compare the efficiency of the noise removal in the resulting decompositions at various approximation levels using wavelets and by varying the number of reconstruction modes in VMD. Surprisingly it is found that the VMD technique yields better results with more accuracy for a specific set of parameters irrespective of the spatial character of the function.