A resource-aware approach for authenticating privacy preserving GNN queries

Abstract

Nowadays many location service providers (LSPs) employ spatial databases outsourced from a third-party data owner (DO) to answer various users’ queries, e.g., group nearest neighbor (GNN) queries that enable a group of users to find a meeting place minimizing their aggregate travel distance. Along with the benefits from LSPs and DO, protection of location privacy and authentication of query results become two major concerns for users while assessing GNN queries. This paper proposes a resource-aware approach that supports effective location privacy preservation and efficient query result authentication with a less storage, communication and computation overhead. Specifically, two centroid-based techniques are investigated to generate a centroid point, which initiates GNN query on behalf of the group members. Then, an authentication algorithm based on Voronoi diagram is proposed for spatial queries. Finally, we demonstrate how our approach is resistant to various attacks, and evaluate its performance by comparing with three competitive approaches. The results show the proposed approach is better and more economical in terms of resource overhead, while considering both privacy preservation and query authentication.