Detection of malicious vehicles with demerit and reward level system

Abstract

In Vehicular Ad hoc Networks (VANETs), to ensure the reliability of safety applications, each vehicle periodically receives crucial information from neighboring vehicles, in its transmission range, about their respective current state including position, direction, etc. Some vehicles will want to inject false information on the network, whether accidentally or intentionally. These vehicles, that we will call malicious vehicles, may affect the timeliness and the pertinence of safety related data. Our proposed scheme called TrustLevel, ranks a vehicle sharing inaccurate information repeatedly as malicious. To ensure system reliability, the TrustLevel scheme is able to promptly detect the malicious vehicles and ensure the accuracy of the vehicle perception of its environment. To do this, TrustLevel collects data for each vehicle on its environment. Thereafter, it creates a perception map for each vehicle representing its surroundings. TrustLevel can then cross these maps in order to increase each vehicle's perception. When doing so, the information sent by malicious vehicles will be incoherent, so TrustLevel will be able to detect these inconsistencies and isolate the faulty vehicles. The isolated vehicles will then be ignored by the other vehicles when exchanging their perceptions, increasing the system reliability. Simulation results demonstrate the efficiency of the TrustLevel scheme under high density of malicious vehicles.