华南理工大学学报(自然科学版) ›› 2012, Vol. 40 ›› Issue (1): 138-145,158.

• 计算机科学与技术 • 上一篇    下一篇

公路网移动终端的KNN 查询技术

梁茹冰1,2 刘琼1†   

  1. 1. 华南理工大学 计算机科学与工程学院,广东 广州 510006; 2. 华南农业大学 理学院,广东 广州 510642
  • 收稿日期:2011-06-07 修回日期:2011-10-12 出版日期:2012-01-25 发布日期:2011-12-01
  • 通信作者: 刘琼(1959-) ,女,教授,博士生导师,主要从事计算机网络研究。E-mail: liuqiong@scut.edu.cn E-mail:liang_ru_bing@163.com
  • 作者简介:梁茹冰(1980-) ,女,在职博士生,华南农业大学讲师,主要从事语义缓存、移动计算研究.
  • 基金资助:

    国家“973”计划项目( 2007CB07100, 2007CB07106)

KNN Query Technology of Mobile Terminals in Highway Networks

Liang Ru-bing1,2  Liu Qiong1   

  1. 1.School of Computer Science and Engineering,South China University of Technology,Guangzhou 510006,Guangdong,China; 2.College of Science,South China Agricultural University,Guangzhou 510642,Guangdong,China
  • Received:2011-06-07 Revised:2011-10-12 Online:2012-01-25 Published:2011-12-01
  • Contact: 刘琼(1959-) ,女,教授,博士生导师,主要从事计算机网络研究。E-mail: liuqiong@scut.edu.cn E-mail:liang_ru_bing@163.com
  • About author:梁茹冰(1980-) ,女,在职博士生,华南农业大学讲师,主要从事语义缓存、移动计算研究.
  • Supported by:

    国家“973”计划项目( 2007CB07100, 2007CB07106)

摘要: 公路网中移动兴趣点( POIs) 的查询处理是一个难点,目前的研究多基于欧氏距离对静态POIs 进行处理,不能很好地适应移动环境下终端弱连接和频繁移动的需要. 文中在公路网移动计算场景下,设计了一种存储分区数据对象的结构来表示公路网图形模型,提出适用于移动终端的连续KNN 查询( CQ-KNN) 算法. 该算法改进了Wang 等提出的MKNN 算法,将逐层渐近探测和检索边列表结合起来进行近邻查询,避免了MKNN 算法在限定层数不够却不得不执行范围查询时所带来的开销; 同时使用缓存策略来支持移动终端提交的连续查询请求,并给出基于广播位置失效报告的缓存一致性维护策略. 仿真结果表明,CQ-KNN 算法较MKNN 算法有更快的CPU 处理速度和更短的网络响应延时,并且能支持移动终端的离线KNN 近似查询.

关键词: 公路网, 移动终端, 位置相关查询, K 近邻, 缓存, 移动计算

Abstract:

The dynamic POIs ( Points of Interest) in highway networks are difficult to query. Most current researches focus only on the static POIs with the help of the Euclidean distance metrics,which are inefficient for the weak connection and frequent movement of mobile terminals in mobile computing environments. In order to solve this problem,a structure to store cell data objects is designed to describe the highway network graph model,and a continuous KNN ( K-Nearest Neighbor) query ( CQ-KNN) algorithm for mobile terminals is presented. For the purpose of improving the existing MKNN algorithm proposed by Wang et al,CQ-KNN algorithm combines the progressive probe and the edge information list retrieval,thus saving the cost of range query execution in MKNN algorithm when fixed layers are insufficient. Moreover,CQ-KNN algorithm employs the local cache strategy to support the continuous query of mobile terminals and adopts the cache consistency maintenance strategy based on the invalid broadcast location report. Simulated results show that CQ-KNN algorithm is superior to MKNN algorithm in terms of CPU processing speed and network response delay,and that it effectively supports the off-line approximate KNN query of mobile terminals.

Key words: highway network, mobile terminal, location-dependent query, K-nearest neighbor, cache, mobile computing

中图分类号: