Current connected and autonomous vehicles will contribute to various and green vehicular services. However, sharing personal data with untrustworthy Navigation Service Providers (NSPs) raises serious location concerns. To address this issue, many Location Privacy-Preserving Mechanisms (LPPMs) have been proposed. In addition, several quantification methods have been designed to help understand location privacy and illustrate how location privacy is leaked. However, their assessment is insufficient due to the incomplete assumptions about the adversary's model. In particular, users tend to request the same navigation routes from home to workplace and acquire traffic information along the route. An adversary can collect the coordinates of adjacent locations and infer the two true locations. In this paper, we provide a formal framework for the analysis of LPPMs in navigation services. Our framework captures extra information that is available to an adversary performing localization attacks. By formalizing the adversary's performance, we also propose and justify two new metrics to quantify location privacy in navigation services, namely accuracy and visibility. We assess the efficacy of two popular LPPMs for location privacy, i.e., differential privacy and k-anonymity. Experimental results demonstrate that the adversary can recover users' locations with a high probability.
|Number of pages||9|
|Journal||IEEE Transactions on Green Communications and Networking|
|Early online date||Jan 2022|
|Publication status||Published - 1 Sept 2022|