Blind Filtering at Third Parties: An Efficient Privacy-Preserving Framework for Location-Based Services

TLDR

Location‑based services are increasingly popular, yet protecting user privacy remains difficult, with TTP‑based solutions creating single‑point failures and TTP‑free approaches incurring heavy costs due to redundant POI records. This work introduces a new framework that safeguards user privacy while maintaining efficiency. The framework employs redundant POI records and a semi‑trusted proxy that uses a blind filtering protocol, called Blind Filter, to remove redundant encrypted POI records without revealing sensitive data. Compared with existing methods, the framework resists dual‑identity attacks, reduces communication and computation overhead, and is shown to be secure and highly efficient on mobile devices.

Abstract

Location-based service (LBS) has gained increasing popularity recently, but protecting users' privacy in LBS remains challenging. Depending on whether a trusted third party (TTP) is used, existing solutions can be classified into: TTP-based and TTP-free. The former relies on a TTP for user privacy protection, which creates a single-point-failure and is thus impractical in reality. The latter does not require any TTP, but usually introduces redundant point-of-interest (POI) records in query result and thus incurs significant computation and communication costs on the user side, making them unsuitable for resource-constrained mobile devices. In this paper, we propose a novel framework to protect user privacy while ensuring efficiency. Our framework also uses redundant POI records to protect privacy against LBS provider but employs a semi-trusted third party, called proxy, to filter out redundant POI records. To protect privacy against proxy, we design a novel filtering protocol, Blind filter, to allow the proxy to filter out redundant encrypted POI records in a blind way. In comparison with existing solutions, our framework is not only resilient to dual identity attack, but also incurs lower communication and computation overhead. Comprehensive analysis and experiments show that our framework is secure and highly efficient in mobile environments.