Department of Computer Science, HKUST
We are witnessing in recent years growing interest for location-dependent information services among mobile users. This presentation examines the issue of processing location-dependent queries in a mobile broadcast environment. Different from a traditional environment, mobile users are concerned with not only access latencies but also power conservation. For Nearest-Neighbor queries, a grid-based index structure is proposed which partition the original space into disjoint grids. For each grid, we index all the objects that are potential nearest neighbors of a query point inside the grid. For general case location-dependent queries, a new index data structure, called D-tree, is devised. The basic idea is to index data regions based on the divisions between them. We describe how to construct the new indexes, how to process location-dependent queries based on those index structures, and how to broadcast the index information on the wireless channel.