Scholarlink

The expanding use of the Internet of Medical Things (IoMT) devices leads to dealing with vast amounts of sensitive data, raising important cyber-attack threats for patient safety. While Edge Computing technology improves IoMT services, maintaining maximum security and privacy persists as a challenge. AI-driven Intrusion Detection Systems (IDS) enhance risk prevention, along with efforts focusing on distributed IDS using Federated learning mechanisms to protect patient data. This paper outlines a new hybrid approach that combines 1D CNN, Bi-LSTM, and Dense Layers models to improve network attack detection. This approach adopts the Wustl-EHMS-2020 dataset for approved test scenarios, achieving accuracy measures of 92,42% (5 users), 93,30% (10 users), and 92,21% (15 users), which outperforms other IDS-based DL methods.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid Deep Fed-IDS-Based Edge Computing Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Hamza Rafik,
  • Abdelaziz Ettaoufik,
  • Abderrahim Maizate

摘要

The expanding use of the Internet of Medical Things (IoMT) devices leads to dealing with vast amounts of sensitive data, raising important cyber-attack threats for patient safety. While Edge Computing technology improves IoMT services, maintaining maximum security and privacy persists as a challenge. AI-driven Intrusion Detection Systems (IDS) enhance risk prevention, along with efforts focusing on distributed IDS using Federated learning mechanisms to protect patient data. This paper outlines a new hybrid approach that combines 1D CNN, Bi-LSTM, and Dense Layers models to improve network attack detection. This approach adopts the Wustl-EHMS-2020 dataset for approved test scenarios, achieving accuracy measures of 92,42% (5 users), 93,30% (10 users), and 92,21% (15 users), which outperforms other IDS-based DL methods.