Design of autonomous algorithms for location privacy

Abstract

Over the past years, the number of location-based applications have been increasing due to the advancement of mobile computing technology and the sensors built in the mobile devices. These applications grant individuals to locate one another, and benefit from services and information those are delivered according to individuals’ location. However, despite the popularity of location-based applications, they are increasingly criticised for putting user privacy at risk by disclosing location data mistakenly to undesired parties. Adapting and predicting changes of user’s sharing attitude over time in order to overcome possible leakages, presents challenging problem for location-based applications and location privacy approaches. In this thesis, we extended an existing work on detecting routine and out-of-routine location events of a user based on entropy estimation, to present a control mechanism that can discover correlation between sharing behaviour and routineness of a location for individuals. It is designed as supportive mechanism that aims to prevent possible leakage induced by location privacy approaches. We additionally extended an existing policy generation algorithm to evaluate proposed control mechanism by identifying optimal privacy preferences for users. The proposed approaches were evaluated and implemented in Python environment. We ran simulations with different metrics to test the overall performance of the system.