An Approach for Ensuring Robust Support for Location Privacy and Identity Inference Protection

The challenge of preserving a user’s location privacy is more important now than ever before with the proliferation of handheld devices and the pervasive use of location based services. To protect location privacy, we must ensure k-anonymity so that the user remains indistinguishable among k-1 other users. There is no better way but to use a location anonymizer (LA) to achieve k-anonymity. However, its knowledge of each user’s current location makes it susceptible to be a single-point-of-failure. In this thesis, we propose a formal location privacy framework, termed SafeGrid that can work with or without an LA. In SafeGrid, LA is designed in such a way that it is no longer a single point of failure. In addition, it is resistant to known attacks and most significantly, the cloaking algorithm it employs meets reciprocity condition. Simulation results exhibit its better performance in query processing and cloaking region calculation compared with existing solutions. In this thesis, we also show that satisfying k-anonymity is not enough in preserving privacy. Especially in an environment where a group of colluded service providers collaborate with each other, a user’s privacy can be compromised through identity inference attacks. We present a detailed analysis of such attacks on privacy and propose a novel and powerful privacy definition called s-proximity. In addition to building a formal definition for s-proximity, we show that it is practical and it can be incorporated efficiently into existing systems to make them secure.