Privacy for Location-based Services

Description

Sharing of location data enables numerous exciting applications, such as location-based queries, location-based social recommendations, monitoring of traffic and air pollution levels, etc. Disclosing exact user locations raises serious privacy concerns, as locations may give away sensitive information about individuals' health status, alternative lifestyles, political and religious affiliations, etc. Preserving location privacy is an essential requirement towards the successful deployment of location-based applications. These lecture notes provide an overview of the state-of-the-art in location privacy protection. A diverse body of solutions is reviewed, including methods that use location generalization, cryptographic techniques or differential privacy. The most prominent results are discussed, and promising directions for future work are identified. Table of Contents: Introduction / Privacy-Preserving Spatial Transformations / Cryptographic Approaches / Hybrid Approaches / Private Matching of Spatial Datasets / Trajectory Anonymization / Differentially Private Publication of Spatial Datasets / Conclusions Preface Acknowledgments Introduction Privacy-Preserving Spatial Transformations Two-Tier Spatial Transformations Three-Tier Spatial Transformations Discussion Cryptographic Approaches A Primer on Computational PIR Spatial Queries with PIR Protocols for Approximate and Exact NN with PIR Comparison with Geometric Transformations Hybrid Approaches Privacy Model System Overview Private Evaluation of Point-Rectangle Enclosure Private Evaluation of Point-Convex-Polygon Enclosure Private Matching of Spatial Datasets Problem Formulation Dataset Mapping Join Processing Enhancing Privacy by Use of Chebyshev Distance Trajectory Anonymization Publishing Independent Location Samples Publishing Individual Trajectories Differentially Private Publication of Spatial Datasets A Primer on Differential Privacy Publication of Static Datasets of Locations Data-Independent Decompositions: Quad-trees Data-Dependent Decompositions: kd-trees Publication of Trajectory Datasets Conclusions Bibliography Author's Biography