Neural networks for microwave modeling: Model development issues and nonlinear modeling techniques

Artificial neural networks (ANN) recently gained attention as a fast and fl exible vehicle to microwave modeling and design. Fast neural models trained from measured/simulated microwave data can be used during microwave design to provide instant answers to the task they have learned. We review two im portant aspects of neural-network-based microwave modeling, namely, model d evelopment issues and nonlinear modeling. ii systematic description of key issues in neural modeling approach such as data generation, range and distr ibution of samples in model input parameter space, data scaling, etc., is p resented. Techniques that pave the way for automation of neural model devel opment could be of immense interest to microwave engineers, whose knowledge about ANN is limited. As such, recent techniques that could lead to automa tic neural model development, e.g., adaptive controller and adaptive sampli ng, are discussed. Neural modeling of nonlinear device/circuit characterist ics has emerged as an important research area. An overview of nonlinear tec hniques including small/large signal neural modeling of transistors and dyn amic recurrent neural network (RNN) modeling of circuits is presented. Prac tical microwave examples are used to illustrate the reviewed techniques. (C ) 2001 John Wiley & Sons, Inc.