Intelligent route planning on large road networks with efficiency and privacy

Abstract

When using Location-Based Services (LBS), intelligent route planning becomes crucial to improve service quality and user experience. The state-of-the-art G-tree structure enables efficient route planning on large road networks, but lacks usability and intelligence. In this paper, we first propose a comprehensive service framework, called BCloud-IFog, which consists of blind cloud servers and intelligent fog servers. Then, we propose an Outsourced Real-time Route Planning (OR2P) scheme, where the search index is built as a G*-tree structure and each G*-tree leaf node is split into a set of non-confidential outsourced graphs. Compared with the G-tree structure, our work has the following advantages: (1) Higher usability. Unlike the G-tree structure requiring the user to perform all calculations in the search process, it delegates the most of computations to cloud servers. (2) Privacy Protection. Unlike the straightforward solution that directly outsource the G-tree structure, it outsources only the non-confidential graphs such that cloud servers cannot infer the original road network or user trajectory. (3) Better intelligence. Unlike the G-tree structure handles only static route planning, it allows fog servers to make route plans based on the dynamic real-time traffic status. Extensive experiments on real data sets demonstrate the effectiveness of our work.