Our 6G-XCEL team did a powerful demo at EuCnC &6G Summit 2025. The demo highlighted a real-time use case where the system collects real-time data from a live 5G O-RAN deployment and uses an LLM to make slice-level resource allocation decisions. While the demonstration is focused on a 5G/6G use case, the underlying framework, powered by the containerized architecture, is flexible and can be deployed across diverse environments.
As described by Hojjat Navidan and Adnan Shahid from UGent-imec:
“The growing complexity of modern networks and physical systems has made traditional management and control methods insufficient. Integrating AI and machine learning into these environments can provide dynamic, real-time decision-making capabilities. This work’s novelty lies in utilizing large language models (LLMs) to analyze real-time data from physical systems and enable closed-loop control. We create a closed-loop digital twin with AI-enabled capabilities to analyze real-time data from the physical environment and support decision-making functionalities.
This demonstration aims to showcase a system that uses LLM to analyze live data from a live 5G network, enabling real-time, autonomous decision-making and closed-loop control. This framework enables autonomous control of physical and network systems through AI-powered data analysis, showcasing the potential for wide applications in industries ranging from telecommunications to smart cities.”