Federated Edge Intelligence for Privacy-Preserving IoT Networks in Smart Healthcare Environments: A Comprehensive Survey and Architecture Framework

Abstract

The rapid convergence of the Internet of Medical Things (IoMT) and Artificial Intelligence (AI) has revolutionized modern healthcare systems, enabling continuous patient monitoring, early clinical diagnosis, and personalized therapeutic interventions. However, transmitting sensitive patient telemetry— such as electrocardiograms (ECG), photoplethysmograms (PPG), vital signs, and electronic health records (EHR)—to centralized cloud data centers presents profound privacy risks, network latency bottlenecks, and regulatory compliance hurdles under stringent mandates like HIPAA and GDPR. Federated Edge Intelligence (FEI) emerges as a transformative paradigm that decentralizes machine learning model training by shifting compute tasks directly to resource-constrained edge gateways and IoMT nodes. By keeping sensitive health records localized and transmitting only encrypted or perturbed model parameters to a central aggregator, FEI preserves patient privacy while maintaining continuous learning capabilities. This comprehensive review critically examines the state-of-the-art in Federated Edge Intelligence tailored for privacy-preserving IoMT networks. We analyze core architecture topologies, communication-efficient federated optimization algorithms (e.g., FedAvg, FedProx, FedNova), privacy-enhancing technologies including Differential Privacy (DP), Secure Multi-Party Computation (SMC), and Homomorphic Encryption (HE), alongside edge hardware acceleration strategies. Furthermore, this survey highlights persistent technical challenges regarding non-IID (non-independently and identically distributed) medical data, resource heterogeneity, communication overhead, dynamic security threats, and model verification. Finally, we formulate a actionable taxonomy, identify crucial research gaps, and chart future research trajectories toward resilient, privacy-centric, and scalable smart healthcare environments.

KEYWORDS: Federated Edge Intelligence, Internet of Medical Things (IoMT), Privacy-Preserving Machine Learning, Differential Privacy, Edge Computing, Smart Healthcare, Federated Learning Algorithms.