ANN for noise estimation of microwave FETs from S-parameters

Measurements of microwave transistors’ noise require complex and expensive equipment and can be time consuming. In this paper, we are proposing an application of artificial neural networks for fast estimation of noise parameters from the device scattering parameters. This method requires measured values of the device noise parameters for the network training only. Once the network is trained noise parameters can be estimated from the scattering parameters without additional noise measurements. Devices working under different temperature conditions are considered.