Context Optimized and Spatial Aware Dummy Locations Generation Framework for Location Privacy

摘要

Location-based services (LBS) involve customizing service offerings based on users' current location. Location privacy is a major concern with location-based services since LBS services involve the user sharing their true location with the LBS servers and are exposed to risk associated with the possible misuse of the user's true location. Dummy locations technique is one of the well-known approaches for addressing location privacy concerns in LBS services. However, most of the existing methods for generating dummy locations have problems addressing scenarios where the true location is part of a large parcel area or if the true location is in a remote area with no building structures nearby. In this paper, we propose a novel approach to dummy locations generation for location privacy that is context optimized and spatial aware. We evaluated the proposed solution using real city parcel data and outlined and geo-visualized our results at each step. Our results indicate that considering the spatial context of the true location in the dummy generation process can generate dummy locations that are more difficult to distinguish from the true location and thus achieve better location privacy. Our proposed solution also reduces the total number of dummy locations while still achieving location privacy, thus saving computational and communication resources involved in executing queries for dummy location on LBS servers. Built and evaluated on real-time geospatial data, our proposed solution paves the way to further leverage real-time geospatial analytics for preserving location privacy.