Transfer Learning-based DDoS Detection and Mitigation in a MATLAB-Simulated 5G Network Environment
Shilpa Shashikant Chaudhari, Jyoti Tolanur, Keerthana Balaji, Kritika Sah, Lakshay Mahindro · 2025
The rise of 5G networks has significantly increased the threat of cyberattacks, especially Distributed Denial-of-Service (DDoS) attacks that target real-time service availability. This paper proposes a real-time DDoS detection and mitigation framework for a MATLAB-simulated 5G environment, using transfer learning with a fine-tuned DistilBERT model. A custom sender-attacker-receiver architecture is used to simulate traffic and classify it using a lightweight classifier. Detected malicious traffic is blocked through automated IP-based mitigation. The system achieves a detection accuracy of 96%, low false positive rates, and real-time latency of less than 0.5 seconds. Experimental results confirm the effectiveness of the approach under varying network load conditions, validating its suitability for deployment in latency-sensitive and resource-constrained 5G infrastructures.