Multidimensional signal-noise neural network model

Abstract

Signal and noise behaviours of microwave transistors are modelled through the neural network approach for the whole operating ranges including frequency, bias and configuration types. Here, the device is modelled by a black box whose small-signal and noise parameters are evaluated through a neural network based upon the fitting of both of these parameters for multiple bias and configuration. The concurrent modelling procedure does not require the solving of device physics equations repeatedly during optimisation, and by this type of modelling the signal (S) and noise (N) parameters can be predicted not only at a single operation frequency around the chosen bias condition for a configuration, but at the same time for the whole operation frequency band for the same operating conditions, with good agreement compared to the measurements.