Data set for Vehicular Network

One major vertical for 5G/6G networks is the Internet of Vehicles (IoV), that promises to enhance the driving experience and make roads safer. However, IoV networks face several security research challenges that need to be addressed before deployment. Lightweight authentication mechanism is a vital security element that must be designed, not only to ensure a good security level but also to maintain high authentication rates with low overheads due to the broadcast and time criticality nature of safety messages in IoV. Previous works had adopted the Timed Efficient Stream Loss-tolerant Authentication (TESLA) scheme with Elliptic Curve Digital Signature Algorithm (ECDSA) to fulfil the security requirements of IoV networks. However, these schemes lack an understanding of the dynamic nature of IoV networks, thus they suffer from low authentication rates and high communications overheads. Therefore, in this paper we propose an intelligent lightweight authentication scheme which utilises TESLA and Bloom filters to achieve the necessary security requirements for IoV networks. Furthermore, the Multi-access Edge Computing (MEC) server that covers a certain geographical area plays the security manager role in this design that uses the supervised Machine Learning (ML) algorithms to analyse the network. As such a MEC server generates Bloom filters to be sent to all vehicles for as part of the lightweight authentication process. Additionally, a labelled data set has been generated for this work by adopting the OpenStreetMaps, SUMO and NS-3. The use of such tools ensures that the generated labelled data set is realistic and based on real world roads layouts. Finally, based on extensive simulations we prove that the proposed work achieves higher authentication rates than existing schemes (that are recommended by IEEE and ETSI) at lower costs, while maintaining high level of security.