关键词: 道路网, 数据流, 聚集查询, 预测聚集, DynSketch索引

Abstract: The technologies of spatial-temporal data streams have been the hotspot in the research field of databases. However, there is not an efficient index applied to aggregate queries over data streams in two-dimensional non-Euclidean spatial road networks until now. In order to implement aggregate queries over moving objects in road networks about the past, present, and future, it needs to solve the problems as follows: (1) the non-Euclidean spatial problem of road networks; (2) the problem of distinct counting, non-uniform of moving objects, and predictive ag-gregate queries over moving objects in road networks. Sketch RR-tree solves the problem of distinct counting and non-Euclidean spatial. In order to solve the problem of non-uniform moving objects, using sketching-partition idea for reference, this paper proposes dynamic sketch index: DynSketch by using AMH(adaptive multi-dimen¬sional histogram) to intelligently partition static sketch, making the data in every part uniform, and to improve the approximate quality of aggregate queries in road networks. Then, based on DynSketch index, it proposes predictive aggregate queries over data streams in road networks using ES (exponential smoothing) methods.

Key words: road network, data stream, aggregate query, predictive aggregate, DynSketch index