No abstract

In this work, a support vector machines (SVM) model for the small-signal and noise behaviors of a microwave transistor is presented and compared with its artificial neural network (ANN) model. Convex optimization and generalization properties of SVM are applied to the black-box modeling of a microwave transistor. It has been shown that SVM has a high potential of accurate and efficient device modeling. This is verified by giving a worked example as compared with ANN which is another commonly used modeling technique. It can be concluded that hereafter SVM modeling is a strongly competitive approach against ANN modeling.

In this article, a simple, accurate, fast, and reliable black-box modeling is proposed for the scattering (S)-parameters and noise (N)-parameters of microwave transistors using the general regression neural network (GRNN) with the substantially reduced measurements and computational cost. In this modeling method, GRNN is employed as a nonlinear extrapolator to generalize the S-data and N-data belonging to only a single bias voltage in the middle region into the entire device operation domain of the bias condition (V DS /V CE , I DS /I C , f) within the shortened human effort. The proposed method is implemented to the modeling of the two transistors BFP640 and ATF-551 M4 as study cases. Thus, comparisons are made with the multilayer perceptrons, trained by the two standard backward propagation algorithms, which are the Levenberg-Marquardt, Bayesian regularization and the 10 data mining methods recently published in the literature using the chosen training data sets in both ınterpolation and extrapolation types of generalization. All the comparisons are achieved using four criteria commonly used in the literature. It can be concluded that GRNN is found to be a fast and accurate modeling method that extrapolates the reduced amount of training data consisting of measured S-parameters and N-parameters at the typical currents of the middle bias voltage to the wide operating range.

Abstract-In this paper, we proposed an efficient knowledge-based Support Vector Regression Machine (SVRM) method and applied it to the synthesis of the transmission lines for the microwave integrated circuits, with the highest possible accuracy using the fewest accurate data. The technique has integrated advanced concepts of SVM and knowledge-based modeling into a powerful and systematic framework. Thus, synthesis model as fast as the coarse models and at the same time as accurate as the fine models is obtained for the RF/Microwave planar transmission lines. The proposed knowledge-based support vector method is demonstrated by a typical worked example of microstrip line. Success of the method and performance of the resulted synthesis model is presented and compared with ANN results.

The performance characterisation of a microwave transistor is carried out rigorously based on the linear circuit and noise theories, subject to the maximum output power and the predetermined input termination. For this purpose, the transducer gain G T is maximised analytically with respect to the input termination Z S for the output port matched, provided that Z S meets the noise figure requirement F req ≥ F min remaining within the unconditionally stable working area (USWA). Analysis is made in the z-parameter domain which facilitates a single unique crescent conditional stability configuration to replace the eight different, rather complicated stability configurations in the S-parameter domain. Finally, the compromise relations between the gain, noise figure for the output port matched are obtained with typical design configurations depending on the operation conditions of a selected high technology transistor. Incompatible noise and gain requirements can also be observed in their design configurations. Furthermore the cross-relations among the bias condition (V DS , I DS) and ingredients of the performance {F req ≥ F min , V out =1, G T ≤ G Tmax } triplets and together with their terminations {Z S , Z L = Z* out (Z S)} can be formed basis for "Performance Data Sheets" of microwave transistors to be employed for the amplifier designs of maximum output power and low noise.