TMEC: A Trust Management Based on Evidence Combination on Attack-Resistant and Collaborative Internet of Vehicles

Ji-Ming Chen,Ting-Ting Li,John Panneerselvam

DOI: https://doi.org/10.1109/access.2018.2876153

IF: 3.9

2018-01-01

IEEE Access

Abstract:Message transmission in vehicular networks is increasing in popularity which exploits the network nodes to transmit messages using cooperative communication in a multi-hop fashion. But the increasing number of malicious nodes in the high-speed Internet of Vehicles demands additional methodologies to quickly detect the presence of such nodes to avoid serious security consequences. Early detection of malicious nodes, and accurate assessment of complex data to assess the node reliability are of absolute importance in vehicular networks. To this end, this paper proposes a security scheme that uses evidence combination method to combine local data with external evidence to evaluate the reliability of multi-dimensional data received from other peer nodes. In addition, this paper uses European Telecommunications Standards Institute standard and Decentralized Environmental Notification Message, and proposes a trust calculation method based on collaborative filtering by introducing a small-time interval to detect the changes in the node behaviors. While the former solution helps more accurate computation of the direct trust value, the latter scheme can calculate the indirect trust based on recommendations received from neighbors, ultimately to obtain the global trust value. Finally, more effective traffic data can be obtained to help traffic prediction. Experiments conducted under various network scenarios demonstrate that our proposed scheme outperforms the existing trust models, such as precision or recall and can resist bad-mouth attacks, selective-misbehavior attacks, and time-dependent attacks, especially under larger proportions of malicious nodes.