| With the rapid development of wireless communications and GPS, the moving objects databases techniques towards location-based services cannot meet the needs of users. The existing disk-based indexing techniques for moving objects mainly include the indexing for the historical information of moving objects, the present and future information of moving objects. The indexing methods for current and future positions of moving objects have been focused on by many researchers. On the basis of the comprehensive analysis for present research situation at home and abroad, this paper has further deep research to the indexing technology for the current and future positions of moving objects.Firstly, the paper analyses the Bottom-up updating method, which is proposed by M.L.Lee, and the HTPR-tree indexing structure, which is proposed by LiaoWei. Aiming at the problem of update localized of indexing methods, combining with existing index technology, ULTPR-tree is proposed for moving objects with frequent updates, which is based on update-log. As the continued updating ranges increasing, ULTPR-tree is able to guarantee the query performance, and also effectively improve the dynamic updating performance of indexing structure.Secondly, in the situation of frequent update, a large number of updating problems will be caused by the change of moving objects'position. BTPR-tree is put forward for moving objects with frequent updates, which is based on the TPR-tree, supplemented by insert-buffers in disk and delete-buffer in memory. BTPR-tree is not only effectively support query operation of the TPR-tree, but also effectively solve the problem of the low updating throughput of indexing in the case of having lots of moving objects.Finally, the indexing method of moving objects proposed above is compared with the existing indexing methods by experiment testing. As a result, the the effectiveness and practicality of ULTPR-tree and BTPR-tree are verified in a frequent updating environment. |
PDF Full Text Request