AI, Blockchain and Vehicular Edge Computing for Smart and Secure IoV: Challenges and Directions

• Post by: Hani Sami
• 25 June 2021
• Comments off

Abstract

Internet-of-Things (IoT) is turning into an undeniably developing point of discussion in both research and industrial fields. A key part that is among the quick development in the utilization of IoT devices is the Internet-of-Vehicles (IoV). IoV is the evolvement of the Vehicular Ad-hoc Networks that allows information exchange among vehicles and infrastructures. Notably, Artificial Intelligence (AI) has been widely adopted for solving challenging vehicular problems, and managing the IoV infrastructure. Despite the advantages AI carries for IoV, its model can be affected through processing falsified data from malicious vehicles. On the other hand, Blockchain is a decentralized and distributed peer-to-peer network architecture, that empowers security and resists against undesirable data modification. Thereupon, we propose in this paper an architecture overview that operates on the Vehicular Edge Computing and employs Blockchain to overcome AI’s limitation. We then discuss the main challenges of the proposed architecture and give notice to concerned parties and stakeholders about promising directions that arises from combining both technologies for improving IoV.
% Vehicles nowadays generate data through sensors, actuators, and cameras to be processed, analyzed, and shared by the intelligent infrastructure. This makes the new generation of smart vehicles part of the Internet of Things (IoT) which we call the Internet of Vehicles (IoV). These data are utilized by the IoV to make smart decisions empowering intelligent vehicular applications. Because of the many benefits smart vehicles can bring to the communities and economies; government, companies, and researchers are spending a lot of time and resources towards enhancing the future of transportation. Despite the unprecedented success in smart vehicles, many aspects still require binding improvements. Thus, we first study in this paper the current architecture and advancement of the Internet of Vehicles (IoV) after integrating AI as a solution for its challenging decision making. This is followed by an overview of the current development of Blockchain in IoV. Regardless of the separate advantages AI and Blockchain can bring to IoV, it is still suffering from some limitations that need to be addressed. Therefore, we propose an architecture overview that combines AI and Blockchain using the help of Vehicular Edge Computing as an IoV infrastructure to overcome these limitations. This architecture benefits from the centralized intelligence of AI and decentralized, trusted, and secured data of Blockchain.

Authors’ Notes

with the increasing number of automated vehicles, the burden of security, privacy, and trust fears has spread, affecting the industry finance and customer trust. In parallel, Blockchain is a decentralized ledger that entails a high level of trustworthiness and availability. Correspondingly, the integration of Blockchain within the IoV started to gain attention lately. It was essentially integrated by scholars to maintain high data credibility among vehicles, enabling traffic safety and efficiency. Many manufacturers are considering this framework to improve the driving experience.

Further, the expected large number of smart vehicles embedding several AI models obligates the need for a rich source of computation power. Simultaneously, employing Blockchain in such an environment raises many questions regarding its feasibility due to its high demand for resources. To the best of our knowledge, no research efforts have addressed the aforementioned limitations combined.
Thereupon, a need for integrating AI’s learning process with a scalable Blockchain implementation within the same architecture arises for enabling smart, secure, and efficient IoV.

In this regard, we propose a new VEC-based architecture embedding both AI and hybrid Blockchain models for (1) increasing the accuracy of IoV intelligent applications affected by lacking computation resources and processing unreliable falsified data and (2) efficiently managing the IoV infrastructure resources to enable the successful deployment of both technologies. Regardless, this synergy arises many limitations that need to be addressed for realizing the proposed architecture. Accordingly, we elaborate on a thorough list of challenges and research directions.