الفهرس 
Microwave TransistorsOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
This thesis is concerned with a vital topic in microwave electronics, which is the MESFET small-signal modeling. Several models have been proposed in the literature for this purpose. All these models differ in shape and configuration, but the major issue is how each model can approximate the MESFET performance with small signals. Working with small signals is very important for the device performance analysis, because at small signals, linearity assumptions and Kirchhoff’s laws are preserved. In this thesis, microwave transistors, their structures and principles of operation are <first reviewed. After that, different published MESFET small-signal models are discussed concentrating on the physical meaning of the circuit elements in the models. Some methods for the evaluation of circuit element values of the models are also discussed. Theconcepts of neural network modeling of microwave devices are also discussed in the thesis. The process of S-parameters measurement for the MESFET is discussed concentrating on the errors that may evolve during the measurement process. The sources of these errors and the methods of calibration to compensate for these errors are alsodiscussed. The residual errors after calibration and error compensation may be veryeffective in the MESFET small-signal modeling process. A new technique is proposed in this thesis for MESFET small-signal modelingusing neural networks. This technique is basically based on the combination of the Melfrequency Cepstral Coefficients (MFCCs) with the different discrete transforms such as the
Discrete Cosine Transform (DCT), the Discrete Sine Transform (DST), and the Discrete Wavelet Transform (DWT) of the inputs to the neural networks. The input data sets to br>traditional neural systems for FET small-signal modeling are the scattering parameters and the corresponding frequencies in a certain band, and the outputs are the circuit elements. In the proposed approach, these data sets are considered as forming random signals. The MFCCs of the random signals are used to generate a small number of features characterizing the signals. In addition, other vectors are calculated from the DCT, the DST, or the DWT of the random signals and appended to the MFCCs vectors calculated from the
signals. The new feature vectors are used to train the neural networks. The objective of
using these new vectors is to characterize the random input sequences with much more
features to be robust against measurement errors. There are two benefits for these approaches; a reduction in the number of neural networks inputs, and hence a faster
convergence of the neural training algorithm and robustness against measurement errors in the testing phase. Experimental results show that the techniques based on the discrete transforms are less sensitive to measurement errors than using the traditional and MFCCs methods.