Share Email Print

Proceedings Paper

Modeling of frequency agile devices: development of PKI neuromodeling library based on hierarchical network structure
Author(s): P. Sanchez; J. Hinojosa; R. Ruiz
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Recently, neuromodeling methods of microwave devices have been developed. These methods are suitable for the model generation of novel devices. They allow fast and accurate simulations and optimizations. However, the development of libraries makes these methods to be a formidable task, since they require massive input-output data provided by an electromagnetic simulator or measurements and repeated artificial neural network (ANN) training. This paper presents a strategy reducing the cost of library development with the advantages of the neuromodeling methods: high accuracy, large range of geometrical and material parameters and reduced CPU time. The library models are developed from a set of base prior knowledge input (PKI) models, which take into account the characteristics common to all the models in the library, and high-level ANNs which give the library model outputs from base PKI models. This technique is illustrated for a microwave multiconductor tunable phase shifter using anisotropic substrates. Closed-form relationships have been developed and are presented in this paper. The results show good agreement with the expected ones.

Paper Details

Date Published: 30 June 2005
PDF: 11 pages
Proc. SPIE 5837, VLSI Circuits and Systems II, (30 June 2005); doi: 10.1117/12.607975
Show Author Affiliations
P. Sanchez, Univ. Politecnica de Cartagena (Spain)
J. Hinojosa, Univ. Politecnica de Cartagena (Spain)
R. Ruiz, Univ. Politecnica de Cartagena (Spain)

Published in SPIE Proceedings Vol. 5837:
VLSI Circuits and Systems II
Jose Fco. Lopez; Francisco V. Fernandez; Jose Maria Lopez-Villegas; Jose M. de la Rosa, Editor(s)

© SPIE. Terms of Use
Back to Top