Indian Journal of Science and Technology

Open Access Open Access  Restricted Access Subscription Access

Neural Network Based Characterizing Parameters of Coplanar Waveguides


Affiliations
1 Terna Engineering College , Plot No 2, Sector 22, Phase 2, Nerul Navi Mumbai-400706, India
 

Artificial neural networks (ANNs) has been a promising tool for microwave modeling, simulation and optimization. In this paper we present the estimation of characteristic parameters of top shielded multilayer coplanar wave-guides(MPCWs) using ANN model. For training the model is done with Levenberg-Marquardt algorithm. Our result shows that the neural network successfully calculates characteristic parameters of top shielded Microwave coplanar waveguides with the high accuracy (error is just about 0.05%). Using these models one can calculate effective relative permitivity and the characteristic impedance of the top shielded MCPWs without possessing strong background knowledge. Even if training takes a few minutes, the test process only takes a few microseconds to produce εeff and Z0 after training. It should also be emphasized that both parameters can be determined from one neural model.

Keywords

Coplanar Waveguides, Artificial Neural Networks (ANNs)
User

  • Change CN and Chen CH (1991) Full wave analysis of multilayered coplanar lines IEEE Trans. Microwave Theory Tech. 39, 747-750.

  • Fan JY (2003) A modified Levenberg-Marquardt algorithm for singular system of nonlinear equations. J.Computnl. Math. 21 (5), 625-636.

  • Gevorgian S, Linner LJP and Kollberg EL (1995) CAD models for shielded multilayered CPW. IEEE Trans. Microwave Theory Tech. 43, 326-334.

  • Jinyan (2006) Convergence properties of a self-adaptive Levenberg-Marquardt algorithm under local error bound condition. Kluwer Acad. Publ. Norwell, MA, USA. 34 (1), 47-62.

  • Levenberg K (1944) A method for the solution of certain nonlinear problems in least squares Quart. Appl. Math. 2, 164-166.

  • Marquardt DW (1963) An algorithm for least-squares estimation of nonlinear inequalities. SIAM J. Appl. Math.11, 431-441.

  • Moré JJ (1978) The Levenberg-Marquardt algorithm: implementation and theory. In: Lecture Notes in Mathematics 630: Numerical Analysis, Watson GA (Ed.), Springer-Verlag: Berlin. pp: 105-116.

  • Moré JJ (1983) Recent developments in algorithms and software for trust region. In: Methods mathematical programming: The state of art. Bachem A, Grötschel M and Korte B (Eds.), Springer: Berlin. pp: 258-287.

  • Sagiroglu S and Yhdiz C (2002) A multilayered perception neural network for a micro planar strip line. Intl. J. Electromagnetic. 22, 553-563.

  • Tai-cong Chen, Da-jian Han, Au FTK Tham LG (2003) Acceleration of Levenberg-Marquardt training of neural networks with variable decay.Proc. Neural Network. 3, 1873-1878.

  • Wen CP (1969) Coplanar waveguide: a surface strip transmission line suitable for non-reciprocal gyro magnetic device applications, IEEE Trans. Microwave Theory Tech. MTT-17, 1087-1090.

  • Yamashita N and Fukushima M (2001) On the rate of convergence of the Levenberg- Marquardt method. Computing (Suppl), 15, 239-249.


Abstract Views: 353

PDF Views: 96




  • Neural Network Based Characterizing Parameters of Coplanar Waveguides

Abstract Views: 353  |  PDF Views: 96

Authors

M. Gurumuti Laxaman
Terna Engineering College , Plot No 2, Sector 22, Phase 2, Nerul Navi Mumbai-400706, India

Abstract


Artificial neural networks (ANNs) has been a promising tool for microwave modeling, simulation and optimization. In this paper we present the estimation of characteristic parameters of top shielded multilayer coplanar wave-guides(MPCWs) using ANN model. For training the model is done with Levenberg-Marquardt algorithm. Our result shows that the neural network successfully calculates characteristic parameters of top shielded Microwave coplanar waveguides with the high accuracy (error is just about 0.05%). Using these models one can calculate effective relative permitivity and the characteristic impedance of the top shielded MCPWs without possessing strong background knowledge. Even if training takes a few minutes, the test process only takes a few microseconds to produce εeff and Z0 after training. It should also be emphasized that both parameters can be determined from one neural model.

Keywords


Coplanar Waveguides, Artificial Neural Networks (ANNs)

References





DOI: https://doi.org/10.17485/ijst%2F2010%2Fv3i3%2F29690