Bias-dependent scalable modeling of microwave FETS based on artificial neural networks

A simple and efficient procedure for modeling of scattering and noise parameters for a class of microwave transistors manufactured in the same technology is presented in this article. It is based on multilayer perceptron artificial neural networks (ANN), whose inputs are device gate width, biases, and frequency that produce scattering and noise parameters at their outputs. After the ANN training, the scattering and noise parameters' prediction under different operating conditions for any device from the class requires only calculation of the ANN response, without changes in the ANN structure. Numerical examples for S- and noise parameters modeling for one specific series of pHEMT devices are presented to show the validity and effectiveness of this approach. (C) 2006 Wiley Periodicals, Inc.