A Systematic Review of Machine Learning-based Small-Signal Modeling Approaches for Gallium Nitride High Electron Mobility Transistors: Performance Analysis and Algorithmic Insights

Abstract

Machine learning (ML)- based modeling is an evolving and thriving research field that must be kept up to date with technological advancements. This paper presents an in-depth analysis of Gallium Nitride High electron mobility transistor (GaN HEMT) behavioral modeling using ML techniques. Through a comparative analysis of ML-based approaches, we explore the development of HEMT models, encompassing scattering parameters, C-V and I-V characteristics, thermal profiles, and more. It also explores conventional techniques and their limitations, emphasizing the advantages of ML applications. This study systematically identifies, analyzes, summarizes, and reports the current state of utilization of ML in the modeling of GaN HEMT. The study critically assesses various ML techniques, including regression, optimization, Artificial Neural Network (ANN), Support Vector Regression (SVR), Decision Tree (DT), Particle Swarm Optimization (PSO), and genetic algorithms (GA), etc. considering precision, complexity, and computational efficiency. Intended for engineers and researchers in electronic and semiconductor devices, this paper serves as a crucial resource, fostering cross-disciplinary collaboration and aiding in the selection of appropriate modeling algorithms in this rapidly progressing field, thereby contributing significantly to the existing literature.