With the rapid development of smart mobile devices and wireless location techniques, more and more users tend to attempt location-based service. Specifically, mobile users usually request continuous queries based on moving trajectories instead of traditional snap-shot queries for fixed locations. As obstacles can be found everywhere in the real-world or virtual space, more and more attentions has been paid on query processing techniques in the obstructed space. Notably, continuous reversek-nearest neighbor queries in obstructed space are widely used. This paper presents an in-depth study on the problem of moving reversek-nearest neighbor queries in obstructed spatial databases. By defining control points and split points, the processing framework for this problem is constructed. Furthermore, several pruning and verification algorithms, including data points reduction, obstacles retrieving, control points calculating and results set updating, are proposed to improve the query efficiency. Extensive experimental evaluation is conducted based on various datasets. Compared with the basic method which computes thek-nearest neighbors for each data point, the proposed methods can significantly improve CPU and I/O efficiency.%随着智能移动设备和无线定位技术的飞速发展,使用基于位置服务应用的用户越来越多。特别地,不同于传统的针对固定位置的快照查询,移动的用户往往基于移动轨迹发出连续的查询。在真实和虚拟的空间环境中,障碍物的影响都是广泛存在的,障碍空间内的查询处理技术得到了越来越多的关注,其中,障碍空间内的连续反k近邻查询处理有着重要的应用。对障碍空间中的连续反 k 近邻查询问题进行了定义和系统的研究,通过定义控制点和分割点,提出了针对该问题的处理框架。进一步地,提出了一系列的过滤和求精算法,包括剪枝数据集、获取障碍物、剪枝和计算控制点和更新结果集等处理策略。基于多种数据集对所提出的算法进行了实验评估。与针对每个数据点进行k近邻计算的基本方法相比,这些方法可以大幅度提高查询处理的CPU和I/O效率。
展开▼