WEMA15: Revolutionizing RF AI: Integrating EM Solvers with NVIDIA Omniverse for Real-Time Synthetic Data

Abstract
This session presents an innovative methodology to accelerate AI/ML solutions for RF applications using physics-based synthetic data generated by a GPU-accelerated Shooting and Bouncing Ray solver, Ansys Perceive EM. By integrating accurate electromagnetic (EM) solutions into the NVIDIA Omniverse platform, this approach supports industrial-scale digitalization and enables the creation of physically accurate virtual worlds for AI training in radar sensing and communication systems. Leveraging technologies such as OpenUSD, RTX rendering, and AI, the methodology facilitates advanced applications including human activity detection, drone detection, and communication channel aging prediction, showcasing the role of synthetic data in revolutionizing next-gen EM system design