Authentication at Physical Layer in 5G/6G Systems

The 3rd Generation Partnership Project (3GPP) continually endeavours to enhance spectral efficiency, reduce latency, refine protocols, and increase system security. These advancements have led to greater energy efficiency and resilience against attacks. However, the emergence of Open-RAN (O-RAN) and low-cost hardware capable of implementing the entire system have made spoofing attacks relatively easy to execute, often serving as a precursor to more significant security breaches. During cell selection, users are vulnerable until mutual trust is established through cryptographic means. Even after trust is established, UE-controlled mobility relies on cell re-selection based solely on signal strength, posing a potential weakness. To address this, a method is proposed to authenticate base stations at the physical layer by leveraging triple-lock characterization involving hardware, environment, and geographical location. This method, akin to Integrated Sensing and Communications (ISAC), facilitates binary classification (authentication) without necessitating characterisation of the range profile, modifications to the gNB or alterations to the physical layer signals. The proposed method achieves high classification accuracy (balanced accuracy), ranging from 94.4 (96.6)% to 97.2 (98.3)%, across various channel characteristics and base station hardware utilized in spoofing attacks. It distinguishes between transmitters experiencing similar environmental impacts and those with distinct environmental footprints. Importantly, the method achieves classification without computationally expensive training, exploiting cross-layer information and relying on modest ground truth/evaluated datasets collected over relatively brief periods. Furthermore, its suitability for deployment on portable devices enhances its practical applicability.