EMiL investigates and implements real-time and AI-based methods for detecting and eliminating faults and attacks in 5G/6G networks.
The customer
The challenge
Project duration: 01.01.2023 - 31.12.2024
Wireless transmission of information and data is more susceptible to physical interference and reflections than wired communication. This can lead to significant packet loss and allow attackers to interrupt or intercept data transmission unnoticed.
One example of such manipulation is jamming, in which a jammer makes successful transmission difficult or even impossible. The spectrum of jamming attacks ranges from relatively simple, constant interference signals to complex strategies that cannot be easily distinguished from natural interference.
The research project is intended to enable the detection of attacks and failures within 5G/6G networks and the initiation of suitable countermeasures in real time. Especially in industry-relevant scenarios such as factory automation, the failure safety of wireless communication is to be significantly improved through automation.
Solution
Results & effects
Since wireless communication networks change dynamically, monitoring and predicting network behavior is a challenging task. While such scenarios often cannot be modeled explicitly, machine learning methods can help identify relevant patterns in the transmission data and thus implement reliable real-time detection algorithms.
Merantix Momentum contributes here with experience in anomaly detection to the success of EMiL. The analysis and prediction of temporal data as well as techniques from anomaly detection play an important role in this context. Our goal is to develop a general-purpose detection method that can be applied to new use cases with little training data. Deep learning methods as well as self-supervised learning are used for this purpose.
.png)
.png)
.png)
