Privacy-Preserving Intelligence for Electric Vehicle Charging Networks: Federated Learning, Edge Analytics, and Grid Integration

Le résumé fourni par la source

The rapid growth of electric vehicle (EV) adoption is transforming charging infrastructures into large-scale, data-intensive cyber-physical systems that must simultaneously optimize operational efficiency, maintain grid stability, and safeguard user privacy. While machine-learning-based intelligence has enabled advances in load forecasting, charging coordination, and vehicle-to-grid integration, prevailing centralized architectures raise persistent concerns regarding privacy leakage, cybersecurity risk, scalability, and regulatory compliance. This review adopts a hybrid narrative-systematic methodology to synthesize evidence from approximately 70 peer-reviewed studies published over the past decade on privacy-preserving intelligence in EV charging networks, with particular emphasis on federated learning and edge analytics. The retained literature is examined through a system-level analytical framework encompassing learning paradigms, deployment layers, privacy mechanisms, and application domains, alongside critical trade-offs among privacy protection, learning performance, communication overhead, and scalability. The synthesis shows that federated and edge-based approaches can achieve learning performance comparable to centralized models while substantially reducing raw data exposure, improving resilience, and enabling grid-aware decentralized decision-making. However, the review also identifies unresolved challenges related to communication and energy overhead, adversarial robustness, interoperability across heterogeneous infrastructures, governance and trust, and the scarcity of large-scale real-world deployments, particularly in resource-constrained settings. By consolidating fragmented research across energy systems, artificial intelligence, and privacy engineering, this review provides a unified taxonomy and outlines priority research and policy directions for the design and deployment of secure, scalable, and privacy-preserving intelligent EV charging networks that support sustainable and integrated electrified mobility.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.