A fuzzy-neural approach for the characterisation of the active microwave devices

Abstract

Artificial neural networks are emerging as a powerful technology for RF and microwave characterization, modeling, and design. A neural modeler helps us to immediately start developing neural models for RF/microwave components and circuits and helps to provide neural models for our simulators. In this study, a novel fuzzy neural network structure is used for behavior of an active microwave device. Here, the device is modeled by a black box whose small signal and noise parameters are evaluated through a fuzzy clustering neural network based upon the fitting of both of these parameters.