PRIVACY PRESERVING METHODS FOR LOCALIZATION

Abstract

Location-based Services (LBS) can provide numerous valuable applications by utilizing location information. However, the privacy threats come together with LBS, and the contextual information attached to a location trace can be derived to infer behaviors and critical decision-making. To address this challenge, this thesis proposes three privacy preservation frameworks for localization. To improve the efficiency and ensure the desired accuracy for indoor localization while preserving location privacy, we propose a hierarchical privacy-preserving indoor localization system based on differential privacy. To provide the tradeoff and privacy analysis thoroughly and systematically, we propose a generalized and theoretical location privacy preservation framework with a Quality of Experience (QoE) model and two adversary models. To consider the privacy threats caused by location correlations in trajectories, we propose a two-tier privacy preservation framework based on Pedestrian Dead-Reckoning (PDR) and local differential privacy (LDP) to preserve trajectory privacy and accelerometer data privacy.