Abstract: Vehicular networks are experiencing a significant transformation driven by integrating connected vehicles and Intelligent Transportation Systems (ITS). The Internet of Vehicles (IoV) is a rapidly evolving domain within ITS that connects vehicles, infrastructure, and smart devices to facilitate seamless data exchange. This data encompasses vital information, and by leveraging it, vehicles can make informed decisions and adapt to real-time situations. For instance, traditional vehicle maintenance practices often rely on reactive approaches, addressing issues after failures occur, which can lead to safety risks, costly repairs, and disruptions. Thus, there is a pressing need for proactive solutions to identify vehicle failures before they escalate. Accordingly, we propose the IoV-TwinChain framework integrating Digital Twin (DT), Machine Learning (ML), and blockchain for performing predictive maintenance of vehicles in the IoV by monitoring vehicle operating conditions to prevent road breakdowns and failures. DT provides real-time monitoring of vehicle operating conditions, while ML facilitates data-driven predictions for the predictive maintenance of vehicles. The IoV-TwinChain framework utilizes blockchain for data integrity and traceability within physical and twin environments. We implement a Proof of Concept (PoC) of the IoV-TwinChain framework using Microsoft Azure DT, ML models such as Random Forest and XGBoost, and Ethereum blockchain. Additionally, we formally verify the IoV-TwinChain framework correctness using High-Level Petri Nets and Bounded Model-Checking methods. With our PoC implementation and formal verification, we demonstrate that the IoV-TwinChain framework effectively enhances the predictive maintenance capabilities of vehicles in the IoV and ensures the reliability and accuracy of the system.
Authors: Mubashar Iqbal, Sabah Suhail, Raimundas Matulevičius, Faiz Ali Shah, Saif Ur Rehman Malik, Kieran McLaughlin