CkNN monitoring based on parallel pre-computing

The problem of k nearest neighbor (kNN) queries plays an important role in spatial information retrieval. The continuous k nearest neighbor query is a variation of kNN query which is aimed to find the kNN in a given path for a query point continuously. Recently, The problem of CkNN queries over moving objects in road networks has caught more and more researchers' attention due to its various applications. In this paper, we report on a pre-processing based approach to answer CkNN queries with light online computation cost. We evaluate our approach on a real data set. The evaluation results validate the effectiveness and efficiency of our approach. Besides, we design a prototype system for monitoring and navigating the urban taxis based on CkNN queries. In our system, we utilize the parallel pre-computing and approximation techniques to support a large number of moving objets. Through a web-based graphical interface, both taxi drivers and pedestrians can access our system and query for their CkNN.