Results - Details

Search command : Author="Ταρουδάκης"  And Author="Μιχαήλ"

Current Record: 9 of 23

[Add to Basket]
Identifier	000412922
Title	Bayesian inversion technique based on the statistical characterization of acoustic signals with applications in Ocean Acoustic Tomography
Alternative Title	Μπαεσιανή τεχνική αντιστροφής βασισμένη σε στατιστικά χαρακτηριστικά ακουστικού σήματος με εφαρμογές στην ακουστική ωκεανογραφία
Author	Σίσκογλου, Ουρανία
Thesis advisor	Ταρουδάκης, Μιχαήλ
Reviewer	Μακράκης, Γεώργιος Σκαρτσούλης, Εμμανουήλ
Abstract	The purpose of this thesis is to use a statistical characterization of acoustic signals and a bayesian inversion method for the estimation of the sound speed profile in a marine environment. The problem of estimating the sound speed profile in the water column using data derived from the recordings of an acoustic signal that has been propagated in the marine environment is a typical problem of Ocean Acoustic Tomography. The forward acoustic propagation problem in a typical marine environment is presented and analyzed in 1.2. For the inverse problem we apply bayesian inver-sion based on the statistical characterization of the signal (1.4). According to our knowledge, this is the first time that the bayesian inverse methodology has been applied in connection with the statistical signal characterization scheme. The observables are the statistical characteristics of the wavelet sub-band coef-ficients. The parameters that are retrieved by this method are the coefficients of the Empirical Orthogonal Functions (EOF) which uniquely define the sound speed profile. As prior assumption, the coefficients are assumed to follow Gaus-sian distributions within their upper and lower limits (search space) and to be independent of each other. The likelihood probability is also assumed to follow gaussian distribution. We are considering a simulation of a real-world acoustic tomography problem to test our method. An algorithm has been created to perform all the necessary procedures for the inversion scheme. In Chapter 4, we present some results de-rived through the application of our algorithm, at first for noise-free signals and then for noisy signals. When a noisy signal is considered, a denoising procedure is applied. The algorithm is written in MATLAB. For the forward model we use 'model' [1] and for the denoising the algorithm presented in [2].
Language	English
Issue date	2018-03-23
Collection	School/Department--School of Sciences and Engineering--Department of Mathematics and Applied Mathematics--Post-graduate theses
	Type of Work--Post-graduate theses
Permanent Link	https://elocus.lib.uoc.gr//dlib/1/c/e/metadata-dlib-1512212105-882185-20487.tkl
Views	695

Digital Documents
	Download document View document Views : 53