Published May 7, 2026 | Version v1

A Standards-Aligned IoMT Federated Learning and Blockchain Framework for MITM Mitigation and Risk Quantification using NS-3* [Appendix]

  • 1. Consortium for Creativity in the Community for Health and Wellbeing
  • 2. ROR icon University of South Wales
  • 3. Smart Solutions
  • 4. ROR icon Romanian-American University
  • 5. ROR icon Institute of Electrical and Electronics Engineers
  • 6. The Worshipful Livery Company of Wales (WLCoW)

Description

The increasing reliance on Internet of Medical Things (IoMT) systems introduces critical vulnerabilities to advanced cyber threats such as Man-in-the-Middle (MITM) attacks, which simultaneously compromise data integrity, availability, and clinical safety. This paper presents a standards-aligned security framework integrating a Federated Learning-based Intrusion Detection System (FL-IDS) with blockchain to enable real-time detection, mitigation, and risk quantification of MITM attacks in IoMT networks. The framework is implemented within an NS-3 simulation environment, where MITM attacks are injected and mitigated using distributed, privacy-preserving anomaly detection across heterogeneous devices. Experimental results reveal that MITM attacks produce selective, high-severity disruptions rather than uniform degradation, with critical flow collapse and hybrid anomaly patterns that evade traditional intrusion detection approaches. To quantify attack impact, derived performance metrics—Throughput Reduction Ratio (TRR), Packet Loss Ratio (PLR), Network Delay Increase (NDI), and Jitter Variation Increase (JVI)—are aggregated into a device-weighted Attack Severity Index (wASI). This enables fine-grained, flow-level assessment of clinically relevant degradation. The framework aligns with ISO/IEEE 11073, ISO 14971, ISO/IEC 27001, ISO/IEC 27005, and IEC 80001-1 standards, achieving 94% concordance across clinical, cybersecurity, and governance risk classifications. The proposed approach bridges the gap between fragmented compliance standards by providing a unified, actionable framework for AI-driven intrusion detection, mitigation, and clinically interpretable risk assessment in IoMT environments.

Files

Cyber-AI_2026_paper_0667_CR_3.pdf

Files (276.8 kB)

Name Size
md5:9a2cddbca58cabd6939acf0a4856cef1
276.8 kB Preview Download

Additional details

Dates

Available
2026-05-07
Published for the IEEE Cyber-AI 2026 Conference