Context-aware Trust Management Scheme for the IoV based on Blockchain Intelligence

The Internet of Vehicles (IoV) is a critical technology designed to mitigate traffic congestion and enhance traffic safety through Vehicle-to-Everything communication. However, the presence of malicious nodes in the IoV may propagate erroneous messages or refuse to disseminate information, thereby compromising driving decisions and endangering human lives. Existing state-of-the-art trust management schemes based on blockchain and artificial intelligence algorithms lack sufficient context-awareness, adapt poorly to dynamic environments and messages, and consume large-scale resources for consensus. This paper proposes a context-aware trust management scheme utilizing blockchain intelligence to address these challenges. The scheme employs deep learning-based trust evaluation model that incorporates environmental factors and model update in-formation as input metrics, enhancing context-awareness and adaptability to evolving contexts. Moreover, blockchain-based federated learning is utilized to enable continuous learning of trust evaluation model using newly collected data. The final trust values, along with model weights and associated information, are stored on a consortium blockchain with a efficient Proof-of-Model Training consensus mechanism. Simulation results demonstrate that the proposed scheme achieves high accuracy in detecting malicious nodes and low latency in storing trust values on the blockchain.