Design of wideband Wilkinson dividers using neural network

Design of wideband Wilkinson dividers using neural network

Paper in proceedings (conference paper)

This document is focused on design of symmetrical wideband splitters which are determined for distribution signals from one common source to two sinks. The ratio of minimal and maximal frequency of bandwidth is approximately about 0.43. That is exactly why this complex design of splitter represents wideband solution. This paper contains fundamental theory for designing wideband splitters and solution by using neural networks. The problem is defined by two input variables – values of frequency to define the frequency band. The neural network is set up to compute parameters of planar structure. Designed splitters were manufactured and measured. Obtained S parameters, such as transmission, isolation or reflections are discussed in the closing part of the paper.

This document is focused on design of symmetrical wideband splitters which are determined for distribution signals from one common source to two sinks. The ratio of minimal and maximal frequency of bandwidth is approximately about 0.43. That is exactly why this complex design of splitter represents wideband solution. This paper contains fundamental theory for designing wideband splitters and solution by using neural networks. The problem is defined by two input variables – values of frequency to define the frequency band. The neural network is set up to compute parameters of planar structure. Designed splitters were manufactured and measured. Obtained S parameters, such as transmission, isolation or reflections are discussed in the closing part of the paper.

Ultra-wideband Power Splitter; Wilkinson Power Divider; S-parameters; Neural Network

Ultra-wideband Power Splitter; Wilkinson Power Divider; S-parameters; Neural Network

TEJMLOVÁ, L.; ŠEBESTA, J.

2014

18.04.2013

978-1-4673-5517-9

Proceedings of 23rd International Conference Radioelektronika 2013

204

208

5