AI-Driven Fuzzing
Project Profile

AI-Driven Fuzzing

A highly scalable framework for vulnerability analysis in 5G traffic steering algorithms. It employs NSGA-II multi-objective optimization to isolate hidden protocol weaknesses before they compromise network stability.

Security Optimization

Vulnerability Analysis in Traffic Steering

AI-Driven Fuzzing is a highly scalable security framework designed to isolate and mitigate adversarial weaknesses within 5G traffic steering controllers. By introducing automated state-based testing, the framework ensures next-generation routing policies are resilient against protocol-level attacks and edge-case exploits.

Core Methodology

  • Multi-Objective Optimization: Uses the NSGA-II genetic algorithm to explore the massive input-state space of traffic steering controllers, finding scenarios that simultaneously maximize routing conflict and packet loss.
  • Protocol Fuzzing: Synthesizes irregular RAN control messaging packets, simulating adversarial edge environments to test fallback reliability.
← Back to all projects