Application of neural networks for linear/nonlinear microwave modeling

Abstract

This paper presents an overview of emerging artificial neural network (ANN) techniques for linear and nonlinear microwave modeling. ANN based models can automatically learn the microwave component or circuit behaviors with satisfactory accuracy, and the trained ANN models are able to implement into commercial circuit simulators for efficient design and optimization. ANN modeling techniques are successfully applied on EM, nonlinear device, and circuit behavior modeling, with speed and accuracy advantages over conventional techniques.