Fail-safe communication technology research funded with 1.4 million euros

Three research groups from FAU involved in large EU project

Reliable communication networks are indispensable for critical infrastructure such as airports, banks, or hospitals. Scientists at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) are researching how these technologies can be made more fail-safe and sustainable. The project is funded with 1.4 million euros as part of the EU initiative CELTIC NEXT.

Improving communication networks with machine learning and digital twins

At FAU, three groups are involved in the project. Prof. Dr. Vasileios Belagiannis and his team at the Chair of Multimedia Communication and Signal Processing are developing methods that allow companies to determine when important communication devices, or subsystems or modules, will fail. These devices are often not located in easily accessible places but rather in isolated areas. Optimized maintenance planning can save money in this regard. As part of the project, error detection should be further improved to predict potential errors before they occur.

“We want to predict with the help of machine learning which devices at which location may not be functioning in a month’s time, for example. This means we want to detect anomalies or deviations from the normal functioning of these devices,” says Belagiannis. To foresee such failures, Professor Belagiannis and his team have datasets from several thousand measurements at their disposal. “The data is all anonymized and comes from various communication devices, some of which have failed at some point. We do not know whether the data is from a hospital or an airport. The data from this network is provided to us by Nokia Solutions and Networks GmbH & Co. KG and is from high-performance communication networks, not mobile networks.”

Two further Chairs at FAU are collaborating in the project

At the Chair of Computer Science 1 – IT Security Infrastructures, the team led by Prof. Dr. Felix Freiling is researching how the protection of confidential data can be ensured in digital computing processes. “Confidential computing” allows data to be protected even during processing in communication networks so that even the network provider has no access to it. Through collaboration with project partners such as Nokia Solutions and Ruhr-Universität Bochum, the project is bridging research and industrial practice.

The Chair of Computer Science 7 – Computer Networks and Communication Systems, led by Prof. Dr. Reinhard German, will conduct research on digital twins, i.e., digital models of a real object, as part of SUSTAINET. Researchers can use digital twins to simulate changes such as disruptions within the networks, and examine the system’s response without affecting the operation of the real networks.

The coordination of all three groups and communication with other SUSTAINET project groups is handled by Prof. Dr. Vasileios Belagiannis.

Further information:

Prof. Dr. Vasileios Belagiannis
Professorship for Machine Learning in Signal Processing
vasileios.belagiannis@fau.de
https://lms.tf.fau.de