Fed-IoMT-Block: A Privacy-Preserving Framework for Secure Federated Learning in Consumer-Centric Internet of Medical Things

The rapid proliferation of the Internet of Medical Things (IoMT) has enabled sophisticated, timely, and remote healthcare monitoring by leveraging advanced networking technologies. However, despite ongoing research, the IoMT paradigm still faces significant security and privacy challenges. To address these, this study introduces Fed-IoMT-Block, an end-to-end framework for securing IoMT devices. The framework first authenticates IoMT entities using the Quantum Authentication Protocol (QAP), which identifies communicating entities and secures their exchange using unique quantum particles, providing stronger protection than conventional protocols. Once authenticated, entities share data through federated learning, leveraging the Attention Capsule Network (Att-CapsNet) and a homomorphic algorithm. Network monitoring is then handled by a machine learning agent, the Vulnerability Analysis Agent (VAA), which inspects external traffic and classifies it using the Light Gradient Boosting Machine (LGBM). The associated risk is mathematically quantified based on multiple parameters. Finally, role-based access control is enforced via a Flexible Neuro-Fuzzy Inference System (FNFIS) to prevent unauthorized access. The proposed framework demonstrates notable improvements in packet delivery, security, and scalability compared to existing methods. These improvements include a 26% reduction in malicious traffic, 84.88% accuracy in disease severity detection, 12% lower latency, and a high packet throughput ratio. This study highlights the potential of federated learning and blockchain to create secure, scalable, and privacy-preserving IoMT systems for consumer healthcare.