A spatiotemporal model of obfuscation strategies and counter strategies for location privacy

Abstract

Safeguarding location privacy is becoming a critical issue in location based services and location-aware computing generally. Two drawbacks of many previous models of location privacy are: 1) the models only consider a person’slocation privacy protection, but not the invasion of location privacy by external agents; and 2) the models are static and do not consider the spatiotemporal aspectsof movement. We argue that, to be complete, any model of location privacy needs to enable the analysis and identification of techniques both to protect and toinvade an individual’s location privacy over time. One way to protect an individual’s location privacy is to minimize the information revealed about a person’s location, termed obfuscation. This paper presents an explicitly spatiotemporalmodel of location privacy that models a third party’s limited knowledge of a mobileindividual’s location. We identify two core strategies that a third party canuse to refine its knowledge, so potentially invading that mobile individual’s locationprivacy. A global refinement strategy uses the entire history of knowledgeabout an agent’s location in a single step. A local refinement strategy iterativelyconstructs refined knowledge over time.We present a formal model of global andlocal refinement operators, and show how this formal model can be translatedinto a computational model in a simulation environment.