Edge-Optimized Hybrid Intrusion Detection: Combining CNN-GRU Deep Learning with Signature-Based Rules for IoT Security

@article{1722543,
      author = {Faisal Rahman, Ingrid Solberg, Vivek Malhotra},
      title = {Edge-Optimized Hybrid Intrusion Detection: Combining CNN-GRU Deep Learning with Signature-Based Rules for IoT Security},
      journal = {Iconic Research And Engineering Journals},
      year = {2024},
      volume = {8},
      number = {4},
      pages = {982-987},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1722543.pdf},
      abstract = {We present an edge-optimized hybrid intrusion detection system that pairs a CNN-GRU deep engine with a signature-based Suricata rule set for IoT security. The neural engine captures temporal attack patterns while the rule layer catches known signatures at negligible cost, and a priority-fusion policy reconciles their verdicts. Across N-BaIoT, CICIDS2017, and UNSW-NB15 the system reaches up to 99.1% F1 with a false-positive rate under 1%, sustaining low latency even under heavy traffic.},
      month = {October},
  }