Visible to the public k-Nearest Neighbours classification based Sybil attack detection in Vehicular networks

Titlek-Nearest Neighbours classification based Sybil attack detection in Vehicular networks
Publication TypeConference Paper
Year of Publication2017
AuthorsGu, P., Khatoun, R., Begriche, Y., Serhrouchni, A.
Conference Name2017 Third International Conference on Mobile and Secure Services (MobiSecServ)
KeywordsAd hoc networks, composability, computational complexity, data privacy, driving pattern similarity, Eigenvalues and eigenfunctions, ETSI, IEEE, intelligent transportation systems, Intrusion detection, k-nearest neighbours classification, kNN classification algorithm, machine learning, Metrics, pattern classification, Peer-to-peer computing, privacy, pseudonymous based privacy protection mechanism, pubcrawl, Resiliency, road vehicles, runtime complexity optimization, Sybil attack, Sybil attack detection, Sybil attack vulnerability, sybil attacks, Symmetric matrices, traffic engineering computing, Transmitters, vehicle classification, Vehicle Driving Pattern, vehicle location privacy, vehicular network, vehicular networking, Wireless sensor networks
Abstract

In Vehicular networks, privacy, especially the vehicles' location privacy is highly concerned. Several pseudonymous based privacy protection mechanisms have been established and standardized in the past few years by IEEE and ETSI. However, vehicular networks are still vulnerable to Sybil attack. In this paper, a Sybil attack detection method based on k-Nearest Neighbours (kNN) classification algorithm is proposed. In this method, vehicles are classified based on the similarity in their driving patterns. Furthermore, the kNN methods' high runtime complexity issue is also optimized. The simulation results show that our detection method can reach a high detection rate while keeping error rate low.

URLhttps://ieeexplore.ieee.org/document/7886565/
DOI10.1109/MOBISECSERV.2017.7886565
Citation Keygu_k-nearest_2017