FlowRAG: Retrieval-Augmented Intrusion Detection for UAV Networks

Intrusion detection for UAV networks where the topology, traffic, and threats change in real time. FlowRAG uses retrieval-augmented generation so the detector generalizes to unseen attacks without constant retraining.

My role: end-to-end — problem framing, model design, training, and adversarial evaluation.

Stack: Python, PyTorch, retrieval-augmented generation, federated learning.

Links

  • Paper (IEEE EDGE 2026): arXiv preprint
  • Code: TODO — add GitHub repo link