Neural networks in microwave circuit design?beyond black-box models (invited article)

Abstract

Neural networks were developed into a computer‐aided approach designing microwave circuits. Researchers replaced device models with faster neural network models in microwave design, however, other ingredients of the design process remain unchanged. Our research explored two neural network applications that extended the role of neural networks beyond being black‐box models. ©1999 John Wiley & Sons, Inc. Int J RF and Microwave CAE 9: 187–197, 1999.

An effective approach to behavioral modeling of RF power amplifier based on recursive polynomials

“…Approximate l F using a multivariable polynomial, (1) can be expressed as (2) This is just the so-called recursive polynomial model used in this paper. Q1 and Q2 represent the model order of the exogenous (linear and nonlinear) and the autoregressive term, respectively.…”

“…However, for modern power amplifiers under the excitation of wideband signals with high PAPR, medium or relatively large Q1 and Q2 value which means long memory length and high-order nonlinearity are needed in (2) in order to meet the accuracy requirement of the model. For a high order multivariable polynomial, the number of terms grows rapidly with the increase of the number of variables, i.e., the Q1 and Q2 value.…”

“…The second class consists of the methods like the neural network and the neuro-fuzzy models [2]- [4]. The parameters in these models can only be identified through a nonlinear optimization or training process, and often require large amount of computations.…”