移动对象位置预测的索引方法  

An Access Method for Prediction of Moving Objects' Location

作　　者：赵卿松[1] 卢炎生[1] 

机构地区：[1]华中科技大学计算机科学与技术学院

出　　处：《计算机科学》2006年第8期170-172,205,共4页Computer Science

摘　　要：当前对移动对象位置预测查询的研究中,索引结构的查询性能成为关注的热点,而忽视其更新代价。针对现有方法中存在的更新缺陷,本文以TPR-tree为基础提出两种索引方法(ETPR-tree和Bi R-tree)。实验结果表明,采用辅助索引结构的Bi R-tree具有最优的查询和更新性能。Currently, proposed access methods for prediction of moving objects＇ location pay more attention to the performance of window query but neglect cost of update operations. In this paper we perform an analysis to explain a defect which is not considered by many existing approaches and thus affects the performance of numerous update operations. Motivated by this, we propose two new index structures, called the ETPR-tree and BiR-tree, which both extend the TPR-tree to support object identifier query by different way. Using experimental comparison, we illustrate that the BiR-tree which uses B^＋ -tree for object ids as auxiliary index structure has the best window query and update performance under all conditions.

关 键 词：时空索引 移动对象 位置预测 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]