Synthesis of intelligent antenna array using radial basis function networks

Abstract

This paper considers antenna array synthesis for linear amplitude and phase distribution of the element excitations. Following the concept of human brain, large number of neurons is assumed and radial basis neural network with exact solution is used. Detailed analyze of performances are presented for different values of the number of training samples and different number of antenna elements. First regular antenna array is presented and then the investigation is broadened to linear amplitude distribution. A small mean error of the amplitude and phase at the output of the network are concluded, showing the ability of the network to perform the synthesis. This analyze is strong basis for guidance in a future synthesis of a more irregular antenna array.