TU Dresden presented part of the research work in the project “PHY Security Innovations for Communication Systems” (PHYSICS) at the “IEEE Conference on Communications and Network Security (IEEE CNS 2024)”. The conference paper entitled “Physical Layer Security: Learning-Aided Attack Detection based on 5G NR SRS” is a collaboration between the Chair of Transport Systems Information Technology and the Chair of Big Data Analytics in Transportation. The paper was prepared by the authors Jonas Ninnemann, Felix Rauschert, Paul Schwarzbach, Pascal Kerschke and Oliver Michler.
The paper deals with the Physical Layer Security (PLS) of 5G NR communication systems, which is particularly relevant in Vehicle ad hoc networks (VANETs), due to the necessary low latencies and high mobility. A method for channel estimation based on the Sounding Reference Signal (SRS) standardized by 3GPP was developed and implemented using commercially available 5G NR hardware. Laboratory measurements under various scenarios were used to create a dataset consisting of the Channel State Information (CSI) in the form of the Channel Impulse Response (CIR) and the Signal-to-noise ratio (SNR). For this purpose, the CSI of a trustworthy communication was recorded as well as under attack by a jammer (signal generator) with different signal strengths. The attack was detected with over 93% accuracy using various classifiers. To increase traceability and transparency, the influence of the various features on the decision was examined in more detail.
The conference took place in Taipei (Taiwan, ROC) from September 30 to October 3, 2024. Research associate Jonas Ninnemann presented the paper at the conference as part of the “Workshop on Security, Privacy, and Resilience of Next-Generation Mobile Networks”. For further details, the paper will be published in the conference proceedings on IEEE Xplore soon.
