Identification and classification of cylindrical targets above infinite perfectly conducting flat surfaces by artificial neural networks

Recognition of radar targets from their scattered signals has been an interesting area in electromagnetics. The electromagnetic field scattered by an object contains the characteristic information about the object. Recognition of an object from its scattered field is a complicated task. Because of their ability to learn and generalize, artificial neural networks are attractive for numerous engineering applications and in this research, there are many applications of neural networks for target recognition. In the first chapter of the thesis, there is an introduction in to the topic. In the second chapter, the Method of Images and the Moment Method are used to calculate the total scattered E-field from a cylindrical target located above a perfectly conducting flat surface and the surface is chosen to be infinitely long. Frequency data points belonging to the sum of the electric fields are obtained by calculating electric fields in particular frequency steps.Thus, a set of features utilized by the artificial neural network is obtained by the calculation of the scattered electric field. The database contains a finite number of scattered field values. A portion of the database is used to train the network and the rest is used to test the performance of the neural network for target identification and classification. In the third chapter, Multilayer Perceptron (MLP), Radial Basis Functions (RBF), Probabilistic Neural Networks (PNN) and Generalized Regression Neural Network (GRNN) are used for target identification and classification. This part aims to find the heights measured from the surface and radiuses of the targets for identification of targets and determine the right target for classification of targets from the scattered field values. In chapter four , the Method of Images and the Moment Method are used to calculate the total scattered E-field from two cylindrical target located above a perfectly conducting flat surface for measuring the performance of the neural networks about the identification of multiple cylindrical targets. In chapter five, GRNN is designed and applied to the group of test targets which are at the same height from the infinite flat surface. This part aims to find the radiuses of the targets and the distance between the targets located above perfectly conducting flat surface from the scattered field values in both the frequency domain and the time domain. In the last chapter, there is a result part pointing out the uses of this research for target recognition. Keywords: Electromagnetic target identification, electromagnetic target classification, Method of Images, Moment Method, scattered field, artificial neural networks.