Creating neural network based microwave circuit models for analysis and synthesis

Abstract

We have previously developed a process to synthesize microwave circuits with a neural network learning process. In this paper we describe a systematic approach to convert conventional circuit models into neural network models for our reverse modeling process. The development of an HBT amplifier model and its applications are demonstrated.