Privacy-preserving [formula omitted] nearest neighbor query with authentication on road networks

Abstract

k nearest neighbor (kNN) queries are frequently used in location-based services (LBSs), by which we wish to get k closest points of interest (POIs) given a certain point. Since the cloud computing is developing fast, LBS providers are tended to outsource spatial databases to the cloud. However, cloud servers are often untrusty, so that ensuring the spatial query integrity as well as the spatial query privacy is critical. We present a verifiable privacy-preserving kNN query scheme, which can be used on road networks. Our work makes use of the network Voronoi diagram and several cryptographic primitives including pseudo-random functions, Paillier cryptosystem, condensed RSA digital signature, and so on. It can simultaneously preserve the privacy of spatial data and kNN queries, and verify the reliability of query results. The effectiveness and practicability of our scheme are validated by our experimental results. We further analyzed the security of our scheme under the adaptive chosen-query attack via rigorous proof.