Modeling of HEMT Devices Through Neural Networks: Headway for Future Remedies

Abstract

Small-signal and large-signal modeling of high elec-tron mobility transistors (HEMTs) are developing day-by-day where accurate model extractions rely on characterizing the behaviour of transistors appropriately. Determining the suitable and optimal model structure with component values is not straightforward and requires significant effort especially at high frequencies. This modeling task is becoming more sensitive to numerical errors and convergence issues and needs careful consideration. Recently, neural networks (NNs) prove their ben-eficial applications in the radio frequencies design leading to accurate modeling. In this framework, this paper devotes to provide the comprehensive literature review around the various methods employed to modeling HEMT transistors through NNs. By referring to this review, radio designers can get a general view of HEMT modeling in one glance and can select the most suitable scheme for their applications. © 2023 IEEE.