Project Janus: The anomaly detection demo (MWC 2024)

Explore the transformative world of vRAN operations in this insightful video, where frequent updates and complex troubleshooting meet the power of AI and machine learning. Discover how defining a normal operational state in vRAN is challenging due to real-time constraints and how generative AI models are revolutionizing problem detection by learning and identifying deviations proactively. We demonstrate a distributed unsupervised machine learning model that collects extensive telemetry and deploys real-time anomaly detection, using over 600 metrics and edge computing for precise issue resolution. Witness how this AI-driven approach simplifies the “single point of management” dilemma, automates troubleshooting, and promises a future of optimized wireless networks with unparalleled efficiency. This is a joint demo with the University of Edinburgh, running on top of Capgemini 5G RAN with Intel FlexRAN.

Youtube link: https://youtu.be/eVD-nTy4AO4 (opens in new tab)