Show simple item record

Modifications of Prony's Method for the Reconstruction of Structured Functions

dc.contributor.advisorPlonka-Hoch, Gerlind Prof. Dr.
dc.contributor.authorKeller, Ingeborg Marlen
dc.date.accessioned2021-12-20T08:29:02Z
dc.date.available2021-12-27T00:50:08Z
dc.date.issued2021-12-20
dc.identifier.urihttp://hdl.handle.net/21.11130/00-1735-0000-0008-59C9-2
dc.identifier.urihttp://dx.doi.org/10.53846/goediss-9020
dc.language.isoengde
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/4.0/
dc.subject.ddc510de
dc.titleModifications of Prony's Method for the Reconstruction of Structured Functionsde
dc.typedoctoralThesisde
dc.contributor.refereePlonka-Hoch, Gerlind Prof. Dr.
dc.date.examination2021-04-12
dc.description.abstractengThe reconstruction and analysis of sparse signals is a common and widely studied problem in signal processing, for example in wireless telecommunication or power system theory. Hereby, most recovery methods exploit structures or special properties of the functions which are to be reconstructed. Particularly interesting are methods which aim to recover functions which possess a sparse representation in a given basis and use only a small set of sampling values. One of the most widely used methods is the so called Prony method, which is a deterministic method for the recovery of sparse exponential expansions. In recent year a generalization of Prony’s method for the reconstruction of spars expansion of eigenfunctions of certain linear operators has been introduced by Peter&Plonka in 2013. While some examples of suitable linear operators were given by Peter&Plonka, e.g., the shift operator as well as certain differential operators, the sample values needed for the reconstruction are not always accessible in practice. This leads to the following question. Can we find other suitable linear operators with meaningful structured functions as eigenfunctions and easily accessible sample values? Based on this question we investigate which structured functions can be recovered using only easily accessible sample values. Using the theory of one-parameter semigroups we derive a framework of so called generalized shift operators and their eigenfunctions, so-called generalized exponential sums, which covers all previously given examples for the generalized Prony method. Furthermore, we elaborate on the connection between generalized shift operators and linear differential operators and present a Prony based algorithm for the reconstruction of sparse generalized exponential expansion. Additionally, we present a new Prony based algorithm for the reconstruction of sparse expansions into orthogonal polynomials of length M using generating functions. Finally, we also consider the numerical analysis of the Prony method for generalized exponential sums and present a modified version of the ESPRIT algorithm and a sub-sampling based Prony method for the recovery of generalized exponential sums in the case of clustered frequencies.de
dc.contributor.coRefereePrestin, Jürgen Prof. Dr.
dc.subject.engProny methodde
dc.subject.enggeneralized Prony methodde
dc.subject.engparameter identificationde
dc.subject.engreconstruction of generalized exponential expansionsde
dc.subject.engreconstruction of sparse expansions into orthogonal polynomialsde
dc.subject.enggeneralized exponential sumsde
dc.subject.enggenerating functionsde
dc.subject.engsignal processingde
dc.identifier.urnurn:nbn:de:gbv:7-21.11130/00-1735-0000-0008-59C9-2-4
dc.affiliation.instituteFakultät für Mathematik und Informatikde
dc.subject.gokfullMathematics (PPN61756535X)de
dc.description.embargoed2021-12-27
dc.identifier.ppn1782630309


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record