Preserving Location Privacy with Semantic-Aware Indistinguishability

Fengmei Jin, Boyu Ruan, Wen Hua, Lei Li, Xiaofang Zhou

Abstract: The rapid proliferation of location-based services (LBSs) has facilitated the collection of extensive location data by potentially untrustworthy servers, raising privacy concerns. Conventional solutions provide location privacy but often fail to fulfill the substantial data utility requirements inherent in LBSs. Thus, effective privacy protection for location data –models that provide theoretical guarantees while delivering high-quality services– has become an urgent demand. Particularly, semantic information, often expressed by the categories of points of interest (POI), is vital for the functionality of various LBSs. In response to this gap, we introduce two types of semantic-aware indistinguishability that protect location privacy by mathematically selecting indistinguishable alternatives from geospatial and/or semantic perspectives. Our well-designed mechanisms rigorously adhere to the new privacy standards, thus safeguarding precise locations while preserving semantically useful information. Experimental results validate our method’s superiority in affording robust privacy protection without compromising semantics.