Machine Learning for Trust in Internet of Vehicles and Privacy in Distributed Edge Networks

As the Internet of Vehicles (IoV) continues to evolve, the imperative for advanced algorithms capable of managing increased network demands, ensuring data security, and boosting overall system efficiency becomes crucial. This article introduces a novel suite of algorithms designed to enhance IoV system performance across multiple metrics. Our comprehensive simulations contrast the proposed system with three contemporary approaches the two-layer computing resource management (TCRM) model, the federated edge learning (FEL) approach, and the blockchain-based trust-value management (BTVM) approach. We demonstrate significant improvements: a latency reduction to as low as 90 ms, compared to 118 ms in TCRM, 125 ms in FEL, and 120 ms in BTVM; reliability in packet delivery with an enhancement from an initial 98% to 99.9%, compared to 98.5% in TCRM, 97.8% in FEL, and 99.5% in BTVM; resource utilization efficiency that surpasses baseline models by maintaining rates up to 85%, compared to their 60–65% in TCRM and FEL, and 75% in BTVM; and swift network response times peaking at just 50 ms, against 60 ms in TCRM, 65 ms in FEL, and 50 ms in BTVM. Additionally, our algorithms maintain robust data security levels, consistently achieving 100% effectiveness, compared to 99.2% in TCRM, 98.9% in FEL, and 99.5% in BTVM. These results underscore the proposed system’s potential to significantly outperform existing solutions, paving the way for more resilient and efficient IoV architectures. The integration of these algorithms into real-world IoV applications can substantially contribute to the advancement of intelligent transportation systems.