Academic Journals Database
Disseminating quality controlled scientific knowledge

A Novel Approach To Synthesize Microwave Filters Using Artificial Neural Networks

ADD TO MY LIST
 
Author(s): Thenmozhi A | Ram Prasath Deepak S | Raju S | Abhaikumar V

Journal: IETE Journal of Research
ISSN 0377-2063

Volume: 54;
Issue: 2;
Start page: 105;
Date: 2008;
Original page

ABSTRACT
Emergence of Artificial Neural Network (ANN) has provided newer avenues in the modeling and design of RF components. This paper proposes a novel synthesis procedure for the design and modeling of microwave filters using ANN. Building a feed forward neural network that corresponds to a Cauer ladder network of a planar filter is the key point of the paper. For validation purposes, the proposed approach has been used to synthesize a micro strip filter with a cutoff frequency of 1.5 GHz. This example filter design provides an insertion loss of around 2 dB in the pass band and a return loss of around 20 dB at the desired cutoff frequency. The design is verified by fabricating the filter on glass epoxy substrate (e1=4.6) with a thickness of 1.6 mm.

Tango Rapperswil
Tango Rapperswil

     Save time & money - Smart Internet Solutions