Privacy-Preserving Mutual Authentication Protocol for Federated Learning in Intelligent Transportation Systems

Federated Learning (FL), which enables collaborative model training across distributed nodes, can play a significant role in Intelligent Transportation Systems (ITS). Intelligent vehicles can work collaboratively in FL networks to improve not just traffic flow but also road safety, environmental sustainability, and urban mobility. However, ensuring secure authentication among entities participating in the FL process remains a critical challenge. This paper proposes a privacy-preserving authentication protocol for FL in ITS, ensuring secure, efficient, and privacy-preserving participant verification. The proposed protocol leverages the concepts of privacy-preserving Decentralized Identifiers (DIDs) and Verifiable credentials (VCS) together with lightweight cryptographic operations. Experimental results demonstrate that the proposed protocol enhances the security and privacy protections of FL in ITS while maintaining efficient authentication compared to other existing protocols, making it a viable solution for real-world deployment in next-generation intelligent transportation systems. Additionally, its scalability makes it suitable for integration into large-scale vehicular environments.