Microstrip filter design using FDTD and neural networks
β Scribed by M. G. Banciu; E. Ambikairajah; R. Ramer
- Publisher
- John Wiley and Sons
- Year
- 2002
- Tongue
- English
- Weight
- 226 KB
- Volume
- 34
- Category
- Article
- ISSN
- 0895-2477
No coin nor oath required. For personal study only.
β¦ Synopsis
Abstract
A new design technique using the FDTD method and neural networks is applied to a microstrip filter. The total design time is reduced by two means. First, an iterative ARMA signal estimation technique is utilized to reduce the computation time for each FDTD run. Second, the number of FDTD simulations is decreased with the use of the device model provided by a neural network with the ARMA coefficients at the output. The trained network was then incorporated to an optimization procedure for a microstrip filter design. Β© 2002 Wiley Periodicals, Inc. Microwave Opt Technol Lett 34: 219β224, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.10422
π SIMILAR VOLUMES
A single neural network is developed to model the resonant frequency of rectangular patch printed on uniaxially anisotropic substrate with air gap using effective parameters in conjunction with spectral dyadic Green's function. Also, the strength of ANN models in antenna design is demonstrated by co
## Abstract Experimental designs were compared using stackedβlayer feedβforward neural networks. Several traditional threeβlevel designs and uniform designs were investigated using threeβfactor linear and nonlinear models. The prediction error was found to be inversely proportional to the number of
Figure 6 Measured axial ratio in the broadside direction against frequency for the proposed antenna in Figure 5 with left-hand CP Ε½ . radiation feed at point A ; h s 1.6 mm, s 4.4, L s 30 mm, r Ε½ . Ε½ . Ε½ . S s 6.4 mm 4.7 β , 5.0 mm 0 β , d s 11.2 mm 4.7 β , 12.2 mm Ε½ . 0 β TABLE 2 Characteristics of