Applications such as traffic management and resource scheduling for location-based services commonly need to identify regions with high concentrations of moving objects. Such queries are called dense region queries in spatiotemporal databases, and desire regions in which the density of moving objects exceeds a given threshold. Current methods for addressing this important class of queries suffer from several drawbacks. For example, they may fail to find all dense regions, provide ambiguous answers, impose restrictions on size, or lack a notion of local density. We address these issues in this paper, starting with a new definition of dense regions. We show that we are able to answer dense region queries completely and uniquely using this definition. Dense regions in our approach may have arbitrary shape and size, as well as local density guarantees. We present two methods, the first, an exact method, and the second, an approximate method. We demonstrate through extensive experiments th...
Jinfeng Ni, Chinya V. Ravishankar