Modelisation des transistors FET appliqueé aux dispositifs radiofrequences en utilisant l’approche neuronale

Abstract

The main theme of this work is the modeling of power MOSFETs in the neural approach. The aim is to model the drain current, which represents the largest non-linearity of the transistor. The neural model is developed based on the multilayer perceptron, trained by the spread of retro gradient with momentum and gave good results in comparison with the theoretical model.