ETD

• 5G/6G Network
• AI-Driven Security
• Man in the Middle
• Zero-touch

The evolution of mobile networks towards 5G and 6G, characterized by disaggre-
gated and Service-Based Architectures (SBA), introduces never seen before attack
surfaces that render traditional security techniques inefficient. The most critical
threat is the Man-in-the-Middle (MitM). Those attack pose a significant risk to
the confidentiality and integrity of user data, particularly on unattended network
interfaces such as the backhaul. This thesis proposes the design, implementation
and validation of an AI-based security solution, containerized and orchestrated via
microservices, for the automated detection of MitM attacks in 5G networks.
The research activity begins with a State of the Art analysis, focusing on the
taxonomy of 5G MitM attacks, a review of AI/ML models for Intrusion Detection,
and a critical evaluation of available public datasets. Based on the acquired in-
formation, the work proceeds with the design and practical implementation of the
solution, structured according to a "closed-loop" architectural paradigm. In the
end, the machine learning model trained on a 5G dataset enable the system to de-
tect threats in near real time and trigger automated responses using the standard
OpenC2.