A Spatial Cloaking Framework Based on Range Search for Nearest Neighbor Search

For nearest neighbor search, a user queries a server for nearby points of interest (POIs) with his/her location information. Our aim is to protect the user’s sensitive information against adversaries including the location-based service itself. Most research efforts have elaborated on reasonable trade-offs between privacy and utility. We propose a framework based on range search query without a trusted middleware. We design a query processing algorithm for the minimum set of candidate POIs by computing the local Voronoi diagram relevant to the cloaked region. Contrary to common belief that cloaking approaches using range search incur expensive processing and communication cost, the experimental results show that the framework incurs reasonable processing and communication overhead even for large cloaked regions.