Deep Learning-Based Radio Frequency Identification of False Base Stations

• Published in: 2023 Workshop on Microwave Theory and Technology in Wireless Communications (MTTW) pp. 45 - 49
• Main Authors: Bolcek, Jan, Kufa, Jan, Harvanek, Michal, Polak, Ladislav, Kral, Jan, Marsalek, Roman
• Format: Conference Proceeding
• Language: English
• Published: IEEE 04.10.2023
• Subjects: 4G/5G SRS RAN; 5G mobile communication; 5G Open RAN; Base stations; Computer architecture; Deep Learning; I/Q-data; Microwave measurement; Object recognition; Radio frequency; RF measurement; Wireless communication
• DOI: 10.1109/MTTW59774.2023.10320078

Advances in mobile and wireless communications allow to handle the continuously increasing demands on the data volume and connectivity of users. The 5G Open Radio Access Network (RAN) concept offers a flexible and inter-operable solution enabling network operators to select equipment from different vendors. However, such a step can potentially increase security risks due to emergence of the false base stations (FBS) operated with a purpose to steal private information about mobile equipment users. In this paper, we introduce a simple deep-learning (DL) based classification method, working directly with In-phase and Quadrature (I/Q) data of a radio frequency (RF) signal, to identify a device working as FBS. To operate the legitimate as well as the FBS, the srsRAN open-source software suite from Software Radio Systems (SRS), connected to three distinct software defined radio (SDR) devices, is used.