An authentication and plausibility model for big data analytic under LOS and NLOS conditions in 5G-VANET

The exchange of correct and reliable data among legitimate nodes is one of the most important challenges in vehicular ad hoc networks(VANETs). Malicious nodes and obstacles, by generating inaccurate information, have a negative impact on the security of 5G-VANET. The big data generated in the vehicular network is also an issue in the security of VANET. To this end, a security model based on authentication and plausibility is proposed to improve the safety of network named ‘AFPM’. In the first layer, an authentication mechanism using edge nodes along with 5G is proposed to deal with the illegitimate nodes who enter the network and broadcast wrong information. In the authentication mechanism, because of the growth of the connected vehicles to the edge nodes that lead to generating big data and hence the inappropriateness of the traditional data structures, cuckoo filter, as a space-efficient probabilistic data structure, is used. In the second layer, a plausibility model by performing fuzzy logic is presented to cope with inaccurate information. The plausibility model is based on detection of inconsistent data involved in the event message. The plausibility model not only tackles with inaccurate, incomplete, and inaccuracy data but also deals with misbehaviour nodes under both line-of-sight(LOS) and non-line-of-sight(NLOS) conditions. All obtained results are validated through well-known evaluation measures such as F-measure and communication overhead. The results presented in this paper demonstrate that the proposed security model possesses a better performance in comparison with the existing studies.