Cooperative approach for data-centric and node-centric misbehavior detection in VANET

Abstract

Vehicular Ad Hoc Network (VANET) has risen as a paramount technology for efficiently providing traffic management, safety and infotainment services to road users. Vehicles are allowed to use pseudo identities during vehicular network access to preserve their privacy. This property makes VANET vulnerable to Sybil attack, performed by exploiting the set of pseudo identities to send messages. Detecting a Sybil attack solely by verifying the accuracy of messages received is challenging, as the messages sent through Sybil identities can appear plausible. Current data-centric and certain machine learning-based approaches only identify Sybil attacks within a local context. It is necessary to find the connection between the Sybil nodes both locally and at the Road Side Unit (RSU) level to effectively mitigate this attack. Hence, we introduce a novel cooperative and hybrid misbehavior detection framework for Sybil attack detection in VANET. It does not only detect Sybil identities but also establishes connections between them by analyzing their speed time series with the Dynamic Time Warping (DTW) technique. Furthermore, it confirms the association between Sybil nodes through node-centric detection using Dempster Shafer Theory (DST) at RSU. This advanced detection can help the Linkage Authority (LA) to find and revoke the actual node responsible for carrying out Sybil attack globally. This is the first framework in its category which can provide accurate detection at both local and RSU level in different scenarios. We acquired a higher detection rate by assessing performance with an existing dataset and a generated real-time Sybil attack dataset.