Research On Privacy Protection In Vehicular Ad-hoc Networks

Vehicular Ad Hoc Network (VANET), contains not only mobile nodes—vehicles,but also stationary Roadside Units (RSUs), is a special instantiate of mobile ad hoc net-work. It opens a new door to facilitating road safety, traffic management, and infotain-ment dissemination for drivers and passengers. However, as a double-edge sword, it alsoattracts tremendous privacy concern, such as identity disclosure and tracing. Due to openessence of vehicular applications, high privacy sensitivity of drivers and the huge scaleof vehicle number, VANET can not adopt the existed privacy preservation method in thetraditional network scenarios. Therefore, studying on privacy preservation becomes afundamental requirement in VANET.Based on the deep understanding of critical challenges and key technologies for pri-vacy preservation in VANET, focusing on the protection of identity privacy and locationprivacy, we studied the following issues related with anonymity authentication and loca-tion privacy threaten in both of traffic safety application and vehicular Internet accessing:·To satisfy both of the performance and security requirements of vehicular appli-cations, we propose an efficient pseudonymous authentication scheme with strongprivacy preservation, named PASS. In PASS, an pseudonymous certificate issuedby a distributed certificate authority (i.e., RSU) can only be used in a short-timeperiod in the region under the control of the distributed authority, which is goodfor reducing revocation cost. Specially, PASS generates the pseudo identities of thepseudonymous certificates belonging to the same owner based on one-way hash-chain technology. It is easy to revoke the unexpired certificates held by an revokeduser by only releasing two hash seeds which is an improvement compared withtraditional pseudonym schemes. Moreover, although only the pseudonymous cer-tificates issued by a legitimate RSU are valid in vehicular communication, PASSallows a vehicle to store a large set of pseudonymous certificates issued by theTA. Based the proxy re-signature cryptography technology[1] where a semi-trustedproxy with given some information can turn a user's signature on a message intoanother user's signature on the same message, the vehicle only needs to request there-signature keys from an RSU and re-sign numbers of the certificates issued by theTA to be as same as ones issued by the RSU itself. In this way, the service over- head is almost irrelated to the number of the updated certificates. At the same time,since the RSUs don't know what certificates are held by a vehicle, PASS providesstrong privacy preservation to the vehicles against the RS Us. Extensive simulationsdemonstrate that PASS outperforms of the previously reported works.·To avoid the tracing risk derived from the high spatial and temporal correlation be-tween traffic heartbeat messages, we propose an optimal Mix-zones deployment foranti-tracing in large scale city scenario, named DT-ET deployment scheme. As thebasis of mix-zone deployment, a traffic statistics-based entropy metric is presentedto evaluate the effectiveness of a Mix-zone, and make it possible to find out all mix-zone candidates in in large scale city scenario. Furthermore, DT-ET deploymentscheme is proposed to deploy minimum-size Mix-zone to guarantee that vehiclesat any place could pass through an effective mix-zone in certain driving time (DT),and the extra overhead time(ET) of adjusting routes to across the mix-zone is small.After proving DT-ET deployment problem is NP-complete, we transform it into anequivalent instance of the set-covering problem which can be solved by a classicalapproximation algorithm. Extensive simulations demonstrate that the proposed en-tropy metric performs as well as the traditional sample based metric to evaluate theanti-tracing ability of Mix-zones, and DT-ET deployment scheme is cost-efficientto provide anti-tracing protection in large scale city scenarios.·With an eye to vehicular internet accessing,we study the location privacy protectionas the first work. We present a potential method which would be used by an serviceprovider to infer an user's location privacy by mining correlation between the RSUlocations and the accessing time intervals of the same user. Then, an RSU accessingdecision problem (named RAP) is proposed to make user obtaining the maximalperformance utility in the context of satisfying user privacy requirement. Since thedimension of resultspace equals the number of RSU(denoted asκ),we further studythe efficient solving algorithm for RAP. Focusing on the special case of RAP wherethe user holds the same privacy requirement at any RSUs, named URAP, we provethat there is a special relationship between the accessing decisions at diffident RS Us.In this way, URAP can be solved in O(κ) while RAP is in O(κ2). Finally, adoptingan real road map, plenty simulations verified the location privacy risk in vehicularInternet accessing, and RAP can guarantee an user to obtain more performance utility under a given privacy requirement.