University of Peloponesse
Μελέτη με ημιαναλυτικές τεχνικές ή μεθόδους τεχνητής νοημοσύνης της ακτινοβολίας κυκλικής βροχοκεραίας παρουσία ή μη ένθετων διηλεκτρικών κυκλικών κυλινδρικών ράβδων: ευθύ και αντίστροφο πρόβλημα ακτινοβολίας: σκέδαση κυλινδρικών διατάξεων
Abstract
dc:descriptionObject of this dissertation is to study the circular loop antenna radiation in the presence or absence of inset circular cylindrical dielectric structures. The problem is addressed by the use of semi-analytical or artificial intelligence techniques.Chapter 1 presents a brief introduction of the techniques used to solve the various electromagnetic problems and we attempt to define the area of interest of this dissertation. Additionally, applications of the aforementioned techniques in similar problems and the key points of this dissertation are presented.Chapter 2 introduces elements of artificial intelligence techniques, by focusing on artificial neural networks and neuro-fuzzy systems. We present the topic of artificial neural networks, their architectures, and the learning process algorithms. In addition the adaptive neuro-fuzzy inference system and the hybrid training algorithm are presented.Chapter 3 addresses the forward circular loop antenna radiation problem with two different soft computing techniques, artificial neural networks and adaptive neuro-fuzzy inference systems (ANFISs). The model of the problem is presented, the generation of the training data are discussed and the learning process metrics are computed. The necessary error indicators extracted during the test process and the suitability of each model is discussed. At the end of the chapter there is the necessary comparison between the two soft-computing techniques. On the one hand, with the analytical solution and on the other hand, with the 2 different artificial intelligence techniques.Inverse problems are by nature complex and often it is not possible to find a solution in a closed form. Thus, Chapter 4 addresses the inverse radiation problem of circular loop antenna with three different artificial neural network architectures while exploring the efficiency of different learning algorithms. The problem modeled and the learning process completed by investigating the efficiency of different learning algorithms. Subsequently, during the test process, the necessary error metrics were calculated and the suitability of each model was discussed. By closing the chapter, the numerical results are presented, the different architectures are compared and the necessary conclusions are drawn.Chapter 5 addresses the circular loop antenna radiation problem when in the far field of the antenna are one or more (inserted or not) circular cylindrical dielectric structures. The problem is solved by two different techniques. Firstly, the equivalent scattering problem is solved in order to calculate the scattered electric and magnetic field and the radar cross-section of the structure. Then the problem is solved by using neural networks that are trained with data from the semi-analytical solution.
Degree
thesis:*- Grantor dc:publisher
- University of Peloponesse
- Year dc:date
- 2019
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kapetanakis, Theodoros
- Καπετανάκης, Θεόδωρος
Subjects
dc:subject × 12- Ημιαναλυτικές τεχνικές
- Τεχνητή νοημοσύνη
- Ακτινοβολία κυκλικής βροχοκεραίας
- Διηλεκτρικές ράβδοι
- Semi-analytical
- Artificial intelligence
- Circular loop antenna radiation
- Scattering dielectric rods
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
dc:rights- Statement dc:rights
-
- BY_NC_ND
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/46251