Privacy-aware query processing in vehicular ad-hoc networks

Recently there is a research trend to integrate the cloud and the vehicular ad-hoc networks to enhance each other on the cooperative urban sensing applications. However, it remains a challenging problem on how to forward queries to search the results and then to route back the query results, as the vehicular nodes dynamically move on the road. Also, the dark side of vehicular cloud is often ignored. Getting the owner’s identity or locations of a given vehicle, or getting the queries submitted by vehicles would put the data privacy at risk. In this paper we propose a privacy preserving query processing scheme for vehicular sensor networks. The scheme consists of two phases: 1) at the index mapping phase, sensed data are stored locally at road side units (RSU), and a global index is maintained at the cloud; 2) at the query processing phase, queries are transformed based on a mapping function, and then are routed to the cloud. The cloud computes the set of RSUs that have the query result based on the mapping index, and then directs the query to the RSUs. Then the result forwarding is transformed into a query result forwarding problem at the RSU, where the data of query results are routed back at their best paths either through the 4G channel or through the DSRC (Dedicated Short Range Communication) channel. Also, the query results are encrypted and diffused to a set of RSUs that the query requester would travel along, so the requester could fetch the results before the query is outdated. Extensive experiments demonstrate the effectiveness of the proposed algorithm in vehicular sensor networks. The ratio of successful query delivery is higher than existing query schemes, while at the same time preserving the privacy of query requesters and the data owner in vehicular sensor networks.