Provided by: The Science and Information (SAI) Organization
Date Added: Aug 2014
In this paper, the authors present a neural network modeling approach to microwave LNA design. To acknowledge the specifications of the amplifier, mobile satellite systems are analyzed. Scattering parameters of the LNA in the frequency range 0.5 to 18GHz are calculated using a multilayer perceptron artificial neural network model and corresponding smith charts and polar charts are plotted as output to the model. From these plots, the microwave scattering parameter description of the LNA is obtained. Model is efficiently trained using agilent ATF 331M4 InGaAs/InP low noise pHEMT amplifier datasheet and the neural model's output seem to follow the various device characteristic curves with high regression.