(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Application of neural networks in antenna performance detection/Yeo Swee Wei

Application of Neural Networks in Antenna Performance Detection_Yeo Swee Wei_E3_2007_NI
Since the early 1990s, many electromagnetic problems have been tackled using artificial neural networks. This project is intended to study the potential of artificial intelligence in the field of electromagnetic as well as communication. A neural network is proposed to determine the performance of antenna based on four antenna parameters, i.e. gain, radiation intensity, directivity and efficiency. Several neural networks will be developed and their performances are compared to identify the most suitable neural network that will be used to model the antenna performance determination system based on antenna parameters. Two types of neural networks that will be developed are Multilayer Perceptron network (MLP) and Radial Basis Function (RBF) network. From the performance analysis, the MLP network trained with Levenberg-Marquardt (LM) algorithm gives the best result, with 97.06 % of accuracy. The proposed system yield to a few advantages compare to conventional method. It provides faster result due to the cutback of some calculation and it has high accuracy. Sejak awal tahun 1990-an, banyak masalah elektromagnetik telah diselesaikan dengan menggunakan rankaian neural buatan. Projek ini bertujuan untuk mengkaji potensi rangkaian neural buatan dalam bidang elektromagnetik serta komunikasi. Satu rangkaian neural akan dicadangkan untuk menentukan prestasi antena berasaskan empat parameter antena, iaitu perolehan, kekuatna radiasi, kearahan dan kecekapan. Dua rangkaian neural akan dibina dan prestasi mereka akan dibandingkan untuk mangecamkan rangkaian yang paling sesuai untuk digunakan sebagai model bagi sistem penentuan prestasi antena berasaskan parameter antena. Dua rangkaian neural adalah rangkaian Perceptron Berbilang Lapisan (MLP) dan rangkaian Fungsi Asas Jejarian (RBF). Dari analisa prestasi, rangkaian MLP yang dilatih dengan Levenberg-Marquardt (LM) memberikan keputusan terbaik, dengan kejituan mencapai 97.06 %. Sistem yang dicadangkan menghasilkan beberapa kebaikan berbanding dengan cara tradisional. Sistem tersebut memberikan keputusan dalam masa yang lebih singkat kerana pengurangan perhitungan dan ia mempunyai kejituan yang tinggi.
Contributor(s):
Yeo Swee Wei - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 
Language:
English
Subject Keywords:
Radial Basis Function (RBF); Multilayer Perceptron network (MLP); antenna performance determination system
First presented to the public:
4/1/2007
Original Publication Date:
1/4/2018
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 95
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2018-01-04 14:23:45.348
Date Last Updated
2019-01-07 11:24:32.9118
Submitter:
Nor Hayati Ismail

All Versions

Thumbnail Name Version Created Date
Application of neural networks in antenna performance detection/Yeo Swee Wei1 2018-01-04 14:23:45.348