Privacycentric modeling and management of context information

Abstract—Context-aware computing has been an intensively researched topic for years already. Consequently, there exists a plethora of usage scenarios for context-aware applications as well as several approaches for the modeling and management of a user’s context information, many of which focus on the efficient and scalable distribution of the latter. With the ongoing rise of smartphones as everyday mobile devices and their steadily increasing amount of sensing and communication capabilities, we finally find ourselves at the edge towards a widespread usage of these techniques. However, apart from technical issues such as how to reliably determine a user’s current context, privacy still remains a crucial factor for these systems ’ acceptance rate. Therefore, inspired by earlier works on privacy in context-aware computing and the authors ’ beliefs in the necessity to put users in control, this paper presents a novel approach for modeling and managing a mobile user’s context information in a user-centric and privacy-preserving way. To this end, this work’s contribution is twofold: First, based on widely recognized requirements for privacy in context-aware applications, we propose a privacy-centric context model which allows for an intuitive and context-dependent definition of a user’s privacy preferences, directly integrating privacy aspects into the context model itself. Second, we present a generic and flexible architecture for the management and distribution of context information in a privacy-preserving way fit for a multitude of different usage scenarios. Keywords—context-awareness; context modeling; privacy-by-design; context-dependent privacy policies; context obfuscation. I