MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202621048322 A) filed by Symbiosis International Deemed University on April 15, 2026, for A System And Method For Efficient Parameter Extraction In Gan Hemt Compact Models Using Machine Learning Optimization.
Inventors include Dr. Mohankumar N; Aarya Gourkar; Arya Dashputra; and Ankush Nagwekar.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: ABSTRACT A SYSTEM AND METHOD FOR EFFICIENT PARAMETER EXTRACTION IN GAN HEMT COMPACT MODELS USING MACHINE LEARNING OPTIMIZATION The present invention discloses a system (100) and method for efficient parameter extraction in Gallium Nitride High Electron Mobility Transistor (GaN HEMT) compact models using machine learning optimization techniques. The system comprises a virtual data generation module (110) for generating synthetic GaN HEMT datasets, a GaN HEMT simulation module (120) for modeling device behavior, an optimization engine (130) implementing Particle Swarm Optimization (131) and Bayesian Optimization (132) algorithms, a benchmarking module (140) for systematic algorithm comparison, and a user interface module (150) for interactive operation. The optimization engine includes an objective function calculator (133) computing Root Mean Square Error (RMSE) between simulated and target data. The method involves generating virtual datasets, initializing simulation parameters, executing parameter search through selected optimization algorithms, and benchmarking multiple algorithm-dataset combinations to identify the optimal methodology for GaN HEMT parameter extraction, thereby accelerating compact model development for millimeter-wave MMIC applications. [
Disclaimer: Curated by HT Syndication.
