Blockchain-based federated learning for the IoV: case of electric vehicle energy management

Abstract

The growing use of electric vehicles (EVs), unmanned aerial vehicles (UAVs), and other advanced vehicle solutions necessitates a scalable, secure, and privacy-preserving Internet of Vehicles (IoV) ecosystem. Traditional centralized approaches struggle with issues such as single point of failure, scalability, and trust. Decentralized systems for the Internet of Energy (IoE), vehicle communications, data sharing, and traffic control are essential. Blockchain (BC) combined with federated learning (FL) offers a solution by enhancing security and transparency in energy transactions while enabling collaborative model training without raw data sharing. These technologies provide an effective framework for managing decentralized energy networks. This paper presents a comprehensive review of BCdriven FL frameworks, with a focus on EV energy distribution and UAV networks. We compare state-of-the-art solutions based on system architecture, consensus mechanisms, BC models, and FL types, identifying three distinct system models: FL systems, BC architectures, and application domains. We analyze the proposed solutions, highlighting novel advancements in the Internet of Things (IoT), especially in the field of energy management. Our analysis reveals key design patterns and trade-offs across these dimensions, offering critical insights into scalability, energy efficiency, latency, and robustness. Additionally, we provide key findings and future research directions to guide the development of decentralized energy management systems.