A Global Optimization Method for Wideband and Small Supergain Arrays Design Using Artificial Neural Network

Abstract

This paper introduces an efficient approach for compact, wideband and supergain arrays design using artificial neural network (ANN) based optimization. The proposed method optimize at the same time the distance inter-elements of the array antenna, its input impedance as well as its directivity. Such global optimization considerably improves the performances of superdirective arrays in terms of gain, bandwidth and efficiency. The proposed method is used afterwards to synthesize a three-elements array using two different unit elements. The comparison of the achieved performances and the Harrington limit reveals that the developed antennas can be qualified as supergain antennas. To our knowledge, this is the first demonstration of a wideband supergain array with more than two elements in the open literature. To validate the results, a prototype was manufactured and measured. The measurements show that the antenna has a wide impedance bandwidth of 22.6%, a peak directivity of 8.3 dBi and a total efficiency greater than 80%.

Keywords

• Superdirective array
• supergain array
• parasitic array
• directivity
• artificial neural network
• optimization algorithm