Review of Neuro-Space Mapping Method for Transistor Modeling

Abstract

This paper reviews the nonlinear microwave device modeling technology based on Neuro-Space Mapping (Neuro-SM). We mainly introduce two kinds of Neuro-SM models: the Neuro-SM model with input mapping network and the Neuro-SM model with output mapping network. Compared with the traditional equivalent circuit model, the Neuro-SM models are more accurate. Measurement data of the RF power laterally diffuse metal-oxide semiconductor (LDMOS) transistor and InP HEMT transistor are used as the application examples to verify the reviewed two Neuro-SM models can accurately reflect the characteristics of transistors with simple operation process and enhance the accuracy of the existing model.