TrustChain-FL: A Reputation-Aware Blockchain-Enabled Hierarchical Federated Learning Framework for Secure Autonomous Vehicle Networks

Abstract

This paper proposes TrustChain-FL, a conceptual framework integrating hierarchical federated learning, lightweight blockchain, reputation-aware client selection, and secure aggregation to improve trust, privacy, and resilience in autonomous vehicle networks. The framework is designed for evaluation using public datasets and simulators. Autonomous Vehicles (AVs) generate massive volumes of real-time data that can improve driving intelligence through collaborative machine learning. However, centralized model training raises significant challenges related to data privacy, communication overhead, and vulnerability to cyberattacks. Although Federated Learning (FL) enables decentralized model training without sharing raw data, it remains susceptible to model poisoning and malicious participant attacks. This paper proposes TrustChain-FL, a blockchain-enabled hierarchical federated learning framework that integrates smart contracts, reputation-based trust management, and secure aggregation to validate model updates and enhance system security. The proposed framework aims to improve trustworthiness, privacy preservation, scalability, and communication efficiency in autonomous vehicle networks. It provides a secure and reliable foundation for next-generation Intelligent Transportation Systems (ITS) and Vehicle-toEverything (V2X) communication environments.