Date of Award


Publication Type

Doctoral Thesis

Degree Name



Computer Science

First Advisor

Arunita Jaekel

Second Advisor

Sherif Saad


Basic Safety Messages, Intelligent Transportation Systems, Privacy, Privacy Assessment Framework, Pseudonym Changing Scheme, VANET




Vehicular Ad-hoc Network (VANET) is envisioned as an integral part of the Intelligent Transportation Systems as it promises various services and benefits such as road safety, traffic efficiency, navigation and infotainment services. However, the security and privacy risks associated with the wireless communication are often overlooked. Messages exchanged in VANET wireless communication carry inferable Personally Identifiable Information(PII). This introduces several privacy threats that could limit the adoption of VANET. The quantification of these privacy threats is an active research area in VANET security and privacy domains. The Pseudonymisation technique is currently the most preferred solution for critical privacy threats in VANET to provide conditional anonymous authentication. In the existing literature, several Pseudonym Changing Schemes(PCS) have been proposed as effective de-identification approaches to prevent the inference of PII. However, for various reasons, none of the proposed schemes received public acceptance. Moreover, one of the open research challenges is to compare different PCSs under varying circumstances with a set of standardized experimenting parameters and consistent metrics. In this research, we propose a framework to assess the effectiveness of PCSs in VANET with a systematic approach. This comprehensive equitable framework consists of a variety of building blocks which are segmented into correlated sub-domains named Mobility Models, Adversary Models, and Privacy Metrics. Our research introduces a standard methodology to evaluate and compare VANET PCSs using a generic simulation setup to obtain optimal, realistic and most importantly, consistent results. This road map for the simulation setup aims to help the research \& development community to develop, assess and compare the PCS with standard set of parameters for proper analysis and reporting of new PCSs. The assessment of PCS should not only be equitable but also realistic and feasible. Therefore, the sub-domains of the framework need coherent as well as practically applicable characteristics. The Mobility Model is the layout of the traffic on the road which has varying features such as traffic density and traffic scenarios based on the geographical maps. A diverse range of Adversary Models is important for pragmatic evaluation of the PCSs which not only considers the presence of global passive adversary but also observes the effect of intelligent and strategic 'local attacker' placements. The biggest challenge in privacy measurement is the fact that it is a context-based evaluation. In the literature, the PCSs are evaluated using either user-oriented or adversary-oriented metrics. Under all circumstances, the PCSs should be assessed from both user and adversary perspectives. Using this framework, we determined that a local passive adversary can be strong based on the attacking capabilities. Therefore, we propose two intelligent adversary placements which help in privacy assessment with realistic adversary modelling. When the existing PCSs are assessed with our systematic approach, consistent models and metrics, we identified the privacy vulnerabilities and the limitations of existing PCSs. There was a need for comprehensive PCS which consider the context of the vehicles and the changing traffic patterns in the neighbourhood. Consequently, we developed a Context-Aware \& Traffic Based PCS that focuses on increasing the overall rate of confusion for the adversary and to reduce deterministic information regarding the pseudonym change. It is achieved by increasing the number of dynamic attributes in the proposed PCS for inference of the changing pattern of the pseudonyms. The PCS increases the anonymity of the vehicle by having the synchronized pseudonym changes. The details given under the sub-domains of the framework solidifies our findings to strengthen the privacy assessment of our proposed PCS.

Available for download on Thursday, July 29, 2021