Analysis of non-stationary multicomponent signals with a focus on the Compressive Sensing approach

TL;DR

This paper explores the analysis and classification of non-stationary multicomponent signals, especially in communication and musical contexts, using compressive sensing techniques to improve efficiency and hardware implementation.

Contribution

It introduces a novel approach for decomposing and classifying multicomponent signals with compressive sensing, including hardware implementation of sparse signal reconstruction algorithms.

Findings

Effective separation of signal components in the time-frequency domain.

Application of compressive sensing improves signal acquisition and transmission.

Hardware implementation of reconstruction algorithms enhances practical usability.

Abstract

The characterization of multicomponent signals with a particular emphasis on musical and communication signals is one of the problems studied in the dissertation. In order to provide an efficient analysis of the multicomponent signals, the possibility to separate signal components is observed. The procedure for decomposition and classification of the signal components whose energy and physical characteristics differ in the time-frequency domain is proposed in this work. A special focus in the dissertation is on the application of the compressive sensing approach in multicomponent signals. The compressive sensing method becomes popular in the field of signal processing until recently, and its application in various fields can increase the acquisition and transmission speed, reduce the complexity of devices, and reduce energy consumption. The procedure that applies the compressive sensing in the classification of the wireless communication signals is proposed. The algorithms for reconstruction of the compressive sensed signals are intensively developing, and therefore special emphasis in the dissertation is devoted to the hardware implementation of one of the algorithms for sparse signal reconstruction.