Enhancement of misbehavior detection scheme for vehicular ad-hoc networks

Abstract

Vehicular ad hoc networks (VANETs) will facilitate various safety and non-safety applications to be deployed in the future. A vehicle in a VANET can misbehave by sending false or inaccurate information to other vehicles. Detection of such misbehavior is an important research problem. In this thesis, we study and improve an existing scheme for misbehavior detection. In that scheme, if a vehicle X generates an incorrect alert, then the nearby vehicles report the misbehavior of X to Road side unit (RSU). Upon receiving such a report, RSU imposes a fine on vehicle X. It is possible for a malicious vehicle to send a false report implicating X, even if X has generated a correct alert. As a result, the RSU may inadvertently fine an honest vehicle, potentially discouraging it from sending true alerts in the future. In this thesis, we propose a modified RSU detection algorithm to avoid honest vehicles from being fined due to malicious reports. We perform a simulation of the modified scheme and show that it identifies misbehaving vehicles with high accuracy