What use cases are being implemented using the 5G testbed here at the institute?

The testbed serves as a flexible platform for various research topics. We’re developing and evaluating methods for AI-assisted network monitoring and management as well as investigating AI-native network architectures and dynamic resource and handover management. By doing so, we’re laying the foundation for self-optimizing networks that can respond in real time to changing requirements, such as peak network loads or faults. Our work also focuses on industrial applications. Specifically, our future demo production line – which is currently under construction – will be wirelessly networked. This will allow us to test new approaches to distributed control, process monitoring, and mobile robotics. We also plan to integrate use cases from flexible manufacturing and smart intralogistics.

Who can use the testbed?

Initially, we’re using the testbed for internal research projects and to develop our own expertise. The next step will be to make it available to external partners as well. As part of industrial projects, we want to work hand in hand with companies to conduct feasibility studies, develop individual networking solutions, and evaluate these solutions in a secure, isolated environment. This is of particular interest to companies planning a 5G campus network, who will benefit from practical, flexible testing scenarios and our experience in planning, optimizing, and operating application-specific networks.

What are the technical features of the testbed?

One feature of our testbed is its completely self-managed infrastructure. We deploy open RAN architecture and will operate five indoor cells and one outdoor cell in the future, making use of open-source software for core and RAN functionality. This gives us unrestricted access to all system components, allowing maximum flexibility for research and development. The architecture follows modern cloud-native principles, enabling a highly flexible, automated, and scalable environment thanks to containerization, Kubernetes-based orchestration, infrastructure as code, GitOps workflows, and the use of cloud-native monitoring tools. This is complemented by predominantly commercial off-the-shelf (COTS) components, facilitating rapid adaptation and the integration of new technologies. With these characteristics, the testbed provides the ideal basis for research and development work on AI-native network management and innovative 5G applications. In the future, the aim is for AI functionality not to act simply as an add-on, but to be integrated deep within the architecture. A system of this kind could also be used for commercial application scenarios, resulting in lower costs relative to the complete solutions currently available on the market.

What do you think will be the most exciting applications of 5G over the coming years, and which aspects of its potential do you feel are yet to be fully leveraged?

We believe that 5G has its greatest potential in industrial applications – from wireless, latency-critical networking of machines to flexible production systems, autonomous robots, and networked logistics chains. In many cases, the development of these applications is only just beginning. Above all, features such as adaptive resource management or dynamic network slicing – the flexible allocation of network resources to different applications in real time – have so far seen only scant implementation in commercial networks but offer huge potential for innovation. In combination with edge computing, AI-assisted orchestration techniques, and future developments toward 6G, we expect to see this technology advance in leaps and bounds over the next few years – and our testbed and application-oriented research is set to play a significant role in these advances.