Record:   Prev Next
作者 Goli, Asghar
書名 Multi-component non-stationary signal processing
國際標準書號 9780549732693
book jacket
說明 189 p
附註 Source: Dissertation Abstracts International, Volume: 69-08, Section: B, page: 4929
Thesis (Ph.D.)--Clarkson University, 2008
This thesis presents a study on instantaneous frequency (IF) estimation in the context of multi-component non-stationary signals. It aims at creating the proper framework for the use of an adaptive signal processing technique in estimating the instantaneous frequency of non-stationary signals potentially comprising a multiplicity of individual components
For mono-component non-stationary signals, i.e. those mono-component signals the spectral content of which potentially varies with time, the IF is a parameter of practical importance in a wide variety of applications. However, the notion of the IF as understood by different researchers in different contexts is not unique. The ambiguity of the meaning of the IF becomes even more aggravated in the case of multi-component non-stationary signals, i.e. those signals that comprise a multiplicity of non-stationary mono-components. However, the notion of the instantaneous frequency, though ambiguous in its meaning, is widely used in various contexts as a useful estimation parameter
A sinusoid tracking algorithm (STA), developed at the University of Toronto and Clarkson University over the past several years, offers much promise in the analysis of non-stationary signals. It is a nonlinear adaptive algorithm capable of estimating parameters of each mono-component signal, i.e. its amplitude, frequency and phase, while effectively tracking parameter variations over time. Its success in a variety of applications provides concrete substantiation of its superior performance predicted by the extensive theoretical analyses reported in the literature. While the use of the STA in problems involving IF estimation presents significant advantages, no rigorous study has been reported on the meaning of the IF as afforded by the STA
This thesis presents a framework for the application of the STA in instantaneous frequency estimation. Various techniques of IF estimation in both cases of mono- as well as multi-component signals have been studied and most commonly used definitions of the instantaneous frequency have been re-visited. The meaning of the IF as provided by the STA is studied. It is observed that while the meaning of the IF as provided by the STA in the case of mono-component signals has a predecessor in the literature, proper framework is needed to be developed for the IF estimation using the STA in the case of multi-component signals. To this end, a scheme for the decomposition of a multi-component signal into its individual mono-components is more rigorously studied and a framework for the use of the STA in IF estimation of non-stationary signals is developed. Within this framework, the STA-based decomposition and IF estimation techniques find numerous applications in diverse areas of engineering and science. A few examples of such applications are developed and presented to more concretely provide a substantiation of the ideas presented in this thesis
School code: 0049
Host Item Dissertation Abstracts International 69-08B
主題 Engineering, Electronics and Electrical
0544
Alt Author Clarkson University
Record:   Prev Next