Artificial neural network based metasurface inspired planar frequency reconfigurable antenna for wireless applications

Abstract

The growing interest in attaining various appealing features of wireless communication systems has motivated researchers to design innovative antennas. This paper analyzes the behavior of a low profile artificial neural network-based metasurface (MS) inspired two-layered frequency reconfigurable antenna (ANN-MSFRA) using planar technology. The designed two-layered structure is a combination of two substrates mounted one above the other, having no air gap between them. The prototype of the proposed antenna is fabricated on Rogers RO4350B material having permittivity (εr) = 3.48, and thickness (h) = 1.524 mm. The MS is represented as an artificially generated periodical array of H and I-shaped metallic inclusions. Numerically analyzed results validate the double negative properties of metamaterial. To verify the effectiveness of the projected two-layered approach, the performance of the reconfigurable antenna is analyzed through both simulations and measurements. The antenna successfully reconfigures the operating frequency from 4.86 to 5.89 GHz, covering the operating bandwidth of 1.03 GHz with a fractional tuning range of 19.1%. The presented simulated and measured outcome provides a positive correlation.The developed ANN model suggested for the estimation of specific output parameters shows minimum error during analysis and synthesis.