Misbehaviour detection in vehicular networks using logistic trust

Abstract

The quality of data and the performance of nodes in a dynamic network, such as vehicular networks (VANETs), is crucial for the applications running on the nodes of these networks. The data in such networks is vulnerable to various types of attacks, amongst which false information dissemination and on-off attacks offer the biggest threats to the applications. As the data depends on the events, it is necessary for any detection mechanism to determine the correct events before identifying the possible attacks. Therefore in this work, first the correct event is learned using information from the most trusted sources including the observations of the receiver itself. Later this information is used to identify the behaviour of the nodes by using the receiver's own observation which is complemented by the opinions of other nodes, in a logistic trust algorithm. It is observed that logistic trust results in a high accuracy of over 99% and a very low error of less than 2% even when the majority of the nodes are malicious.