Indexing the Past, Present and Future Positions of Moving Objects in Urban Traffic Networks

Advance in wireless sensor networks and positioning technologies enable new data management applications that monitor continuous streaming data. In these applications, efficient management of such data is a challenging goal due to the highly dynamic nature of the data and the need for fast, on-line computations. An efficient indexing structure for moving objects is necessary for supporting the query processing of these dynamic data. Existing work can not index the past, current and future positions of moving objects at the same time. In this paper, a novel index technique is proposed to support querying the past, present and future positions of moving objects in urban traffic networks. First, a simulation based location prediction model for the vehicle future trajectory is presented, which is more accurate than the traditional linear prediction model in the TPR-tree. Moreover, exploiting the feature of traffic networks, it presents a dynamic structure termed AU (adaptive unit) and develops it to an R-tree based index named current-AU. Finally, by naturally extending the AU, the past-AU is proposed, which is capable of indexing historical trajectory and at the same time avoiding the dead space that is inevitable in the TB-tree. Experimental studies indicate that the AU-index outperforms the traditional TPR-tree and TB-tree.