GaN-AI: Intelligent PCB Design and Optimization Platform for GaN Power Electronics
This Vertically Integrated Project (VIP+) proposes to develop an AI-driven PCB design and optimization platform dedicated to Gallium Nitride (GaN)-based power electronics. The system, called GaN-AI, will leverage machine learning to analyze Gerber files and generate optimized layouts for high-frequency, high-efficiency GaN systems used in robotics, AI data centers, and power-supply applications.
The project is multidisciplinary in nature and developed continuously over several semester and academic years. This project will unite students and faculty from multiple fields under the guidance of Aleph-B Technologies, an industry partner specialized in GaN system design. The main outcomes include:
- Automatic Gerber-file interpretation: Train AI models to read and understand standard PCB Gerber formats, identifying design features that influence EMI, parasitics, and thermal behavior specific to GaN transistors.
- Intelligent design recommendations: Generate corrective design suggestions based on GaN design rules and measured performance data, closing the loop between layout and device efficiency.
- Text-to-layout generation: Enable users to input system-level specifications (e.g., voltage, current, switching frequency, board size) and automatically produce GaN-ready PCB layouts suitable for manufacturing.
- Multi-disciplinary integration and commercialization: Combine engineering and product design disciplines to create a scalable, commercial-grade tool that bridges academic research and industry deployment.
Desired disciplines
Computer Science
Data Analytics
Engineering
Marketing / Advertising
Team Leaders
Partner
Dr. Michel Khoury, Aleph-B Technologies.