A large signal elements' simulation of GaAs MESFET using neural network model

Abstract

Neural networks are important as fast and flexible tools for microwave modeling, simulation, optimization and design. A new approach of a neural-network based model is proposed to determine the large signal elements of a GaAs MESFET. To conveniently implement this in standard circuit simulators, we extracted the single-cell GaAs MESFET's bias-dependent behavior in terms of conventional small signal equivalent circuit elements. We represented seven intrinsic elements with a four-layered neural network whose inputs are the gate-to-source bias and drain-to-source bias. A well-trained neural network shows excellent accuracy and generates good extractions. The results of calculation between the new neural model and improved optimization are in good agreement.