Microstrip circuit design using neural networks

Abstract

A novel technique that uses neural networks to efficiently design microstrip circuits is presented. A full-wave analysis is employed to rigorously characterize a microstrip circuit, which results in a finite set of pairs of input and output parameter vectors. The neurons, arranged as a three-layer network, are used to learn the mappings from input to output and then give accurate approximations for the output vectors at any arbitrary input. It is emphasized that a three-layer neural network is capable of performing any mapping if the right connections among the neurons can be made. A real example involving microstrip corporate feed design is given to illustrate the potential power of this technique.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>