Main article

João Pereira
Department of Electronics, Telecommunications and Informatics, Universidade de Aveiro, Aveiro, Portugal, 3810-193
Mariana Soares
School of Health, Polytechnic of Porto, Porto, Portugal, 4200-072
Ricardo Oliveira*
Department of Computer Science, University of Beira Interior, Covilhã, Portugal, 6201-001
roliveira@ubi.pt

Abstract

Healthcare cyber-physical systems generate sensitive longitudinal data across mutually distrustful institutions and impose latency, privacy, and audit requirements that no single technology can satisfy alone. This paper argues that blockchain, federated learning, and edge intelligence converge in healthcare to form an architecture whose properties are qualitatively different from the sum of its parts. Blockchain alone offers traceability without learning; federated learning alone offers private training without provenance; edge intelligence alone offers low-latency inference without trust anchors. Designed together as a three-tier stack, they provide privacy-preserving training, low-latency inference, and tamper-evident governance simultaneously. We synthesise the design choices that determine whether such a stack delivers on its promise, surface the latency, throughput, energy, and accuracy trade-offs that govern its operational viability, and quantify the cost-of-resilience envelope using benchmarks aggregated from recent deployments. Across a representative non-IID multi-hospital benchmark, the proposed configuration reaches a multi-class detection accuracy of 0.974 — within 0.6 percentage points of an unattainable centralised oracle — while preserving privacy and audit guarantees that no centralised baseline can match. We then propose a six-priority research roadmap for 2024-2030 and discuss the regulatory implications under emerging adaptive-AI device frameworks. The contribution is intended as a practical bridge between the distributed-systems and clinical-informatics communities.

Article details

How to Cite

Pereira, J. ., Soares, M., & Oliveira, R. . (2024). Converging Blockchain, Federated Learning, and Edge Intelligence for Resilient Healthcare Cyber-Physical Systems. Crossroads of Future Technologies, 2(2), 16-29. https://doi.org/10.63646/cft.2024.020202

Similar Articles

You may also start an advanced similarity search for this article.