Preserving Users’ Location Privacy in Mobile Platforms

Mobile and interconnected devices both have witnessed rapid advancements in computing and networking capabilities due to the emergence of Internet-of-Things, Connected Societies, Smart Cities and other similar paradigms. Compared to traditional personal computers, these devices represent moving gateways that offer possibilities to influence new businesses and, at the same time, have the potential to exchange users’ sensitive data. As a result, this raises substantial threats to the security and privacy of users that must be considered. With the focus on location data, this thesis proposes an efficient and socially-acceptable solution to preserve users’ location privacy, maintaining the quality of service, and respecting the usability by not relying on changes to the mobile app ecosystem. This thesis first analyses the current mobile app ecosystem as to apply a privacy-bydesign approach to location privacy from the data computation to its visualisation. From our analysis, a 3-Layer Classification model is proposed that depicts the state-ofthe- art in three layers providing a new perspective towards privacy-preserving locationbased applications. Secondly, we propose a theoretically sound privacy-enhancing model, called LP-Cache, that forces the mobile app ecosystem to make location data usage patterns explicit and maintains the balance between location privacy and service utility. LP-Cache defines two location privacy preserving algorithms: on-device location calculation and personalised permissions. The former incorporates caching technique to determine the location of client devices by means of wireless access points and achieve data minimisation in the current process. With the later, users can manage each app and private place distinctly to mitigate fundamental location privacy threats, such as tracking, profiling, and identification. Finally, PL-Protector, implements LP-Cache as a middleware on Android platform. We evaluate PL-Protector in terms of performance, privacy, and security. Experimental results demonstrate acceptable delay and storage overheads, which are within practical limits. Hence, we claim that our approach is a practical, secure and efficient solution to preserve location privacy in the current mobile app ecosystem