Models & Optimisation and Mathematical Analysis Journal
Volume 12, Numéro 1, Pages 28-33
2025-12-31

Federated Learning–enabled Privacy Protection And Cybersecurity In Distributed Iomt Environments

Authors : Zioual Tahar . Bouazza Abdelhammid . Hoceini Mohammed . Kacher Abdelhafidh . Silarbi Ibrahim . Horri Mourad .

Abstract

The proliferation of the Internet of Medical Things (IoMT) has fundamentally transformed contemporary healthcare delivery by facilitating continuous patient monitoring, intelligent clinical decision support, and seamless integration of intercon- nected medical devices. Nevertheless, the inherently sensitive nature of medical data, coupled with the architectural hetero- geneity of IoMT environments, introduces substantial challenges for data privacy, cybersecurity, and system reliability. Conven- tional centralised machine learning methodologies necessitate the aggregation of patient data within a unified repository, thereby elevating vulnerability to data breaches, contravening stringent medical privacy regulations, and constraining resilience against evolving cyber threats. Federated Learning (FL) has emerged as a compelling paradigm to address these challenges by enabling multiple healthcare institutions and IoMT devices to collaboratively train machine learning models without exchanging raw patient data. This paper investigates an FL-based cybersecurity framework for distributed IoMT environments, with particular emphasis on privacy-preserving intrusion detection. Each participating client conducts local training on its proprietary network traffic data and transmits exclusively model updates to a coordinating server, thereby ensuring that sensitive medical information remains within institutional boundaries. The proposed framework is evaluated utilising the WUSTL- EHMS-2020 IoMT dataset and a deep Multilayer Perceptron (MLP) classifier trained via Federated Averaging (FedAvg). Ex- perimental results obtained under controlled conditions demon- strate that federated intrusion detection achieves robust detection performance whilst preserving data privacy.

Keywords

Federated Learning ; Internet of Medical Things ; Data Privacy ; Cybersecurity ; Distributed Learning ; Intrusion Detection ; Privacy-Preserving Artificial Intelligence