Decomposition of Two-Component Multivariate Signals with Overlapped Domains of Support

Time-frequency signal analysis has been an active research area for several decades. It provides efficient tools for the analysis and processing of signals with time-varying spectral content. However, the analysis and characterization of nonstationary signals with several components having overlappe...

Full description

Saved in:
Bibliographic Details
Published in2019 11th International Symposium on Image and Signal Processing and Analysis (ISPA) pp. 324 - 328
Main Authors Brajovic, Milos, Stankovic, Isidora, Stankovic, Ljubisa, Dakovic, Milos
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.09.2019
Subjects
Concentration measures
Correlation
digital signal processing
Matrix decomposition
multivariate signals
Signal analysis
signal decomposition
Signal processing algorithms
Time-frequency analysis
Transforms
Online AccessGet full text
ISSN1849-2266
DOI10.1109/ISPA.2019.8868775

Cover

More Information
Summary:Time-frequency signal analysis has been an active research area for several decades. It provides efficient tools for the analysis and processing of signals with time-varying spectral content. However, the analysis and characterization of nonstationary signals with several components having overlapped domains of supports has been a particularly challenging topic. In this paper, we discuss the applicability of the multivariate signal paradigm to solve the problem of multicomponent signal decomposition. In particular, the paper is focused on bivariate two-component signal case, aiming to present the decomposition approach as simply as possible, demystifying the basic theoretical concepts. In the presented numerical experiments, we explore the limits of the approach subject to the additive noise.
ISSN:1849-2266
DOI:10.1109/ISPA.2019.8868775