City Road Net Work Moving Object Indexing Technique Based On Costs Clustering

In moving object database with a large number of moving objects, the index is the key for storing and querying the objects. Meanwhile, for the purpose of the tracking of moving objects, they will have to send their locations from time to time to the server which results in the great consumption of the communcation bandwith.This paper investigates the moving objects location reporting and indexing technology at home and abroad. Based on the cost analysis of the location reporting and index maintenance, we design a protype system that are suitable for urban road network moving objects. This system is composed of the moving side and the server, which are interrelated. The moving side first gather the adjacent moving objects which have similar velocities, and sends the information to the server. The server adopts the improved TPR tree structure, namely S-CCTPR tree. The S-CCTPR classifies the moving objects by hash bucket S, and every speed bucket corresponds to a CCTPR tree. The updating operation is performed in batch.In the experiment, using the simulated and real moving object data, moving side first divides zone according the density of moving objects, and then generates moving objects clusters in each small area. Finally, the information of cluster block is sent to the server. The server test the index I/O performance by setting the appropriate parameters. It shows that the proposed location reporting method can largely reduce the communication cost. Compared with the indexing structure of TPR tree, our designed indexing structure has greatly improved I/O efficiency in querying.