Artificial neural network (ANN) approach for synthesis and optimization of (3D) three-dimensional periodic phased array antenna

Abstract

This paper provides a precise analytical formula that permits to draw the radiation pattern characteristics corresponding to three-dimensional (3D) periodic array antennas. The given systems are implemented in Matlab, Based on the known linear array antenna in 1D and 2D configurations. The often analytical solutions are not available for complex real systems, so that the computational cost of a single analysis can be prohibitive; hence the design strategy has to be very effective and flexible. Optimization algorithms have given an important role as reliable techniques for electromagnetic designs. Then, this work focuses on using an efficient artificial neural network (ANN) approach for the modeling and synthesizing of the uniformly spaced linear phased array antenna in any given configuration (1D, 2D and 3D), and describes the basics of artificial neural network. So, assuming the network's training database contains a finite number of samples of targets at certain angles are available. A part of the database is used to train the network and the rest is used to test its performance for target, identification and classification. Used neural networks are multi-layered perceptron (MLP) with a back-propagation training algorithm. The given synthesis approach assured considerable improvements in terms of performances, computational speed (convergence's time) and software implementation. Simulation results using different numbers for three-dimensional array antenna are given. The antennas arrays synthesis and optimization by neural networks are again presented and discussed. However ANN is able to generate very fast the results of synthesis by using generalization with early stopping method. To validate this work, several examples are shown.