Synthesis Of MEtamaterial Using Inverse ANN and PKI-D Method

Abstract

Metamaterial has been a buzzword in the field of electromagnetics over the last few decades. Many new inventions and discoveries have been done due to the unique properties of metamaterial. These properties include –ve permeability,–ve permittivity and consequently –ve refractive index, infinite propagation at a specific resonant frequency. These properties depend on the orientation or the configuration of the structure rater than the material constitutes. These properties can be exploited to meet the demand over the field of electromagnetic and antenna for many applications. But there exist a trade off between advantages and Design, Analysis & Synthesis of metamaterial and advantages of metamaterial. Although the Methods and Formulas for calculation of metamaterial characteristics from the design parameter of metamaterial is available, the reverse is not achievable that is calculation of design parameter from the known output. One of the reason being its mathematical complexities. This thesis represents an approach towards synthesis of metamaterial unit cell using IANN (Inverse Artificial Neural Network) and PKI-D (Prior Knowledge Input with Difference ). The proposed method gives us the design parameters of a metamaterial when a desired permittivity is known to us. Here IANN with PKI-D is used to develop a CAD for the synthesis of metamaterial . Compared to traditional approach here we train the inputs of an ANN by keeping the weights fixed and the known output with a prior knowledge of output and input sets. As in the synthesis process, number of output parameters is more than the number of input parameters, more than one solution exist so for accurate results we have used PKI-D method followed by IANN.