Author, Institution: Dovilė Karalienė, Kaunas University of Technology
Science Area, Field of Science: Technological Sciences, Informatics Engineering – 07T
Scientific Advisors:
since 2016; 2009-2014 –Scientific Supervisor: Prof. Dr. Zenonas NAVICKAS (Kaunas University of Technology, Technological Sciences, Informatics Engineering, 07T).
2009–2014: Prof. Dr. Stanislovas SAJAUSKAS (Kaunas University of Technology, Technological Sciences, Electrical and Electronic Engineering, 01T).
Dissertation Defence Board of Informatics Engineering Science Field:
Prof. Dr. Rimvydas SIMUTIS (Kaunas University of Technology, Technological Sciences, Informatics Engineering, 07T) – chairman;
Prof. Dr. Adas GELŽINIS (Kaunas University of Technology, Technological Sciences, Informatics Engineering, 07T);
Prof. Dr. Eugenijus KANIUŠAS (Vienna University of Technology, Physical Sciences, Informatics, 09P);
Prof. Dr. Algimantas KRIŠČIUKAITIS (Lithuanian University of Health Sciences, Biomedical Sciences, Biophysics, 02B);
Prof. Dr. Dalius NAVAKAUSKAS (Vilnius Gediminas Technical University, Technological Sciences, Informatics Engineering, 07T).
The Doctoral Dissertation is available at the libraries of Kaunas University of Technology (K. Donelaičio St. 20, Kaunas) and Vilnius Gediminas Technical University (Saulėtekio al. 14, Vilnius).
Annotation:
The identification algorithms, i.e. the methods of mathematical modeling for signals approximated by linear recurring sequence is examined in the dissertation.
Some of the most popular methods found in scholarly literature for signal identification are Fourier and Prony’s methods. Prony’s method describes signals by a fixed number of exponential function linear derivatives with constant coefficients. However, a major drawback of these methods is the unknown number of model components (a model order) which identify the signal; moreover, it is limited by the exponential function coefficients of stability, which, in some cases, results in increasing the signal approximation error. Nevertheless, Prony’s method is applicable in many fields, such as biomedicine, non-destructive material control, genetics, financial, etc. fields. Not surprisingly, a wide variety of modifications of Prony’s method are developed and applied for the above outlined areas of analysis.
The aim of the research is to modify Prony’s method for the identification of signals that can be approximated by the sum of exponential functions with polynomial coefficients by using the optimal number of model components based on approximation errors and convergence speed.
The modify Prony’s method and an algebraic signal interpolation algorithm (determines the optimal number of components of the exponential function that describe the signal) are presented in this thesis. The developed algorithm is applied for the analysis of algebraic estimates of electrocardiographic and ultrasonic signals.
