NDBC2009 传感器网络中多近似连续范围查询处理技术.docVIP

传感器网络中多近似连续范围查询的处理技术 何文麟1,2 曹芳菲1,2 陈红1,2 1（中国人民大学信息学院 北京 100872） 2（中国人民大学数据工程与知识工程教育部重点实验室 北京 100872） (hewl@) Multi-query Processing Technology of Approximate Continuous Queries in Wireless Sensor Networks * He Wenlin1,2, Cao Fangfei1,2, Chen Hong1,2 1(Information School, Renmin University of China , Beijing 100872) 2(Key Laboratory of Data Engineering and Knowledge Engineering, Ministry of Education, Beijing 100872) Abstract Wireless sensor networks opens up a fresh research area of database, where efficient using of sensors’ limited energy is the primary goal. If multiple Approximate Continuous Range Queries (ACQ) are injected into the networks and run independently without optimization, it will cause a serious problem that sensors are likely to send reduplicate data for different queries and therefore shorten the networks’ lifetime. This paper studied the technology of energy-efficiently processing multiple ACQ, and designed a structure called rq-kd-tree to index multi-dimensional range queries, which is used to get the common required data of different queries (query intersecting areas). Based on a concept called Query Similarity Degree, original queries are rewrote and issued to the networks. Finally, detailed experiments are conducted to prove our proposed methods achieving energy-efficient in processing multiple queries. Key words sensor networks; approximate continuous range query; multi-query optimization 摘 要 无线传感器网络为数据库的研究开辟了新的研究领域，高效利用节点的有限能量是当前研究的主要目标。如果发布到网络中多个近似连续范围查询不经过优化处理而独立执行，会造成节点为不同查询重复发送相同感知数据，从而增加节点的能耗，降低网络寿命。本文针对近似连续范围查询研究了多查询优化技术，设计了一种索引多维范围查询的多叉索引树结构rq-kd-tree，通过获取多查询的公共查询区域（查询相交区域）以及基于查询相似度的概念合并相交区域上的多个查询、重写查询。最后，实验证明了本文所提的算法可以实现能量有效的多查询处理过程。 关键词 传感器网络；近似连续范围查询；多查询优化 中图法分类号 TP393 1 引 言 无线传感器网络是由部署在监测区域内大量的廉价微型传感器组成，通过无线通信方式形成的一个多跳自组织的网络系统，其目的是协作地感知、采集和处理网络覆盖区域中感知对象的信息[1, 2]。由于它在军事国防、工业农业、城市管理、生物医疗、环境监测、抢险救灾等许多领域中有重要作用，因此近年来得到工业界和学术界的广泛关注。 无线传感器网络中的查询具有以下三个方面的特征：①查询的近似特征。由于传感器节点本身的硬件约束，使得节点的感知和通信能力有限，造成节点感知读数和网络中传输的数据几乎不