Research On Distributed Index Of Moving Objects For Future Query

With the rapid development of database, network, wireless communication and online analytic processing techniques, moving objects database techniques have been prospected to have wide applications on medication in battlefield, real-time command and control, digital battlefield, battlefield state-prediction, and so forth. Moving objects database indexing is a key technique to moving objects database, which has gained extensive attentions of researchers both home and abroad.Existing moving objects database indices mainly consider the centralized environment, while, the management, in real applications, of huge volumn of moving objects requires extending the centralized environment to distributed environement. Peer-to-Peer network is one the most important distributed processing techniques in rencet years, which can provide abundant resources and great processing service abilitties. Due to its good extensibility and robotness, P2P network has been one effective solution for dealing with large-scale distributed moving data management. This thesis proposes a future prediction oriented moving objects indexing technique in distributed environment: PeerFOX. The major resesearch work and its contribution can be concluded as follows.1. Based on the analysis of recent research and developmental trend of moving objects indexing techniques and P2P data management, we observe that moving objects indexing technique in P2P environment is a promising trend to effectively deal with organization and management of huge volumn of moving objects data in distributed environment. Hence, we suvey the current existing work on moving objects indexing techniques, as per the categories including indexing topological structure, query-supporting tense, and moving objects representation.2. Based on reduced MOST model under assumptions, we propose a moving objects data model MOST* in distributed environment, which is employed as the representation for moving objects in distributed environment in this thesis. Comparatively analyzing the moving objects based partition and space based partition, two partition methods, and viewing the requirement of the real application in this thesis, we adopt the space based partition as the method that the severs manage distributed moving objects. On top of the space partition based method, we incorporate Hilbert curve, which is of good filling performance, to sort the sub-spaces, and thus to reduce the dimensionality.3. We propose a future prediction oriented double-layered index structure in P2P environment, PeerFOX: the upper layer is for the network topology and data structure among server nodes, while the lower layer is for the index structures inside serve nodes. Then, we design the routing rules, querying algorithm, as well as construction and maintainance algorithm for PeerFOX index structure. Further, the load balancing strategy of the system is also explained.4. Implmenting GSTD spatio-temperal data generator, we generate some types of moving objects datasets. In view of the datasets, we conceive certain simulational experitments.5. We investigate the application of moving objects database technology in P2P.Based on the abstraction and modeling of warfield environment, we discuss the distributed spatio-temperal query processing in battlefield environment.