Neural procedure for microwave MOSFET modelling versus bias and gate length

In this paper, a neural procedure for development of a microwave bias-dependant MOSFET model versus the gate length is presented. Artificial neural networks are applied to model the small-signal scattering parameters. The developed model is validated through comparison of the simulated S-parameters for three devices differing in the gate length with the measured data in the frequency range up to 40 GHz. The obtained results confirm the achieved good accuracy of the extracted model.