SIA OpenIR  > 工业信息学研究室
面向目标跟踪的无线传感器网络部署与定位问题研究
Alternative TitleResearch on Deployment and Localization in Target Tracking Oriented Wireless Sensor Networks
黄艳1,2
Department工业信息学研究室
Thesis Advisor于海斌 ; 梁炜
ClassificationTP212
Keyword无线传感器网络 信息覆盖 干涉定位 目标跟踪 粒子滤波
Call NumberTP212/H79/2009
Pages113页
Degree Discipline机械电子工程
Degree Name博士
2009-01-19
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract微机电系统、先进传感器、无线通信及现代网络等技术的进步,推动了无线传感器网络的产生和发展。集数据采集、处理、无线传输等功能于一体的无线传感器网络扩展了人们的信息获取能力,将逻辑上的信息世界与真实物理世界融合在一起,将改变人类与物理世界的交互方式。 无线传感器网络具有许多挑战性的研究课题,其中部署与定位是无线传感器网络应用的基础问题。部署就是在一定区域内通过适当的策略放置传感器节点以满足某种特定需求,它决定了网络对目标的检测能力。定位在无线传感器网络中包括两层含义:节点定位和目标定位。节点定位是目标定位的前提,会影响目标跟踪的精度。论文面向应用于目标跟踪的无线传感器网络,围绕部署与定位问题展开研究。 论文分别综述了无线传感器网络部署问题、定位问题和目标跟踪问题的研究内容、评价标准、算法分类和研究现状。 从信息覆盖的角度研究满足给定指标的确定性部署问题,提出一种基于检测融合的确定性部署策略。首先采用奈曼-皮尔逊准则融合单元网格内所有传感器节点的检测信息,实现正方形和正三角形两种单元网格的高效覆盖。然后分别给出对应的监测区域网格划分方法,从而确定监测区域需要的传感器节点数量以及放置的具体位置。使用相同数量的传感器节点,基于检测融合的部署策略可以获得更大的覆盖范围,也就是在保证一定感知性能的情况下减少了成本。 针对无线电干涉定位系统的多径效应抑制问题,分别从理论和实验的角度分析多径效应对无线电干涉测距的影响,然后根据多径信号的物理特性,从理论上推导了镜反射的多径误差数学模型,在此基础之上进行无线电干涉定位系统的多径误差分析,讨论了衰减因子、传感器节点天线高度和水平距离等多径参数对多径误差的影响。通过仿真计算和分析可知,衰减因子对多径误差的大小起着决定性作用,而多径误差在定位空间上分布的复杂程度,随着天线高度的增加而增加,同衰减因子无关。 无线电干涉测距技术获取的干涉距离是两个发送节点和两个接收节点间距离的线性组合值。针对以节点间距离作为输入的传统定位算法,无法直接利用上述干涉距离进行定位的问题,提出一种基于改进粒子群优化的定位方法。借鉴遗传算法中变异的思想,在每次迭代时对更新过的粒子位置进行微扰,以维持种群的活性。利用干涉距离的实验数据,分析比较了遗传算法和改进粒子群优化在无线电干涉定位系统节点定位问题中的性能。实验结果表明,两种定位方法都能在有限的进化代数内找到节点坐标的近似值,但是基于改进粒子群优化的定位方法的平均耗费时间,远远小于基于遗传算法的定位方法,具有更高的优化效率。 针对具有簇-树型网络拓扑结构的无线传感器网络,首先给出集中式粒子滤波跟踪算法实现的具体步骤,然后提出两种分布式粒子滤波跟踪算法。综合考虑信息收益和参与协作的资源消耗,两种分布式算法分别采用不同的准则来选择参与跟踪的传感器节点。为了系统地比较跟踪算法的性能,给出跟踪精度、通信开销、能量开销和跟踪反应时间等各项指标的定义,进而构建了综合性能评价体系。通过仿真,量化比较了三种跟踪算法的性能。结果表明,两种分布式算法能够在损失很少跟踪精度的同时,大幅度减少通信开销、能量开销和跟踪反应时间。最后,仿真分析了传感器节点覆盖密度和检测阈值对跟踪算法性能的影响。 总之,论文对无线传感器网络的部署问题、定位问题和目标跟踪问题进行了研究和探讨,旨在对无线传感器网络的应用起到一定的促进作用。
Other AbstractRecent technological advances in micro-electro-mechanical system (MEMS), advanced sensor, wireless communication and modern networks have enabled the emergence and development of wireless sensor networks (WSN). The WSN, which integrates the functions of data collection, data processing and wireless communication, reinforces human’s ability in information obtaining. Inosculating the information world in logic with the physical world, the WSN will change the interaction mode between the human and the world. There are abundant research topics which are considerably challenging in WSN. As fundamental issues, deployment and localization are basis for the applications of WSN. The deployment, to place sensor nodes with proper strategy to meet some particular requests, determines WSN’s performance of target detection. Localization in WSN includes node localization and target localization. The former is the premise of the latter, and affects the precision of target tracking. The dissertation studies the deployment and localization of WSN applied in target tracking. The research contents, evaluation criteria, algorithm classification, and related works of the deployment, localization and target tracking in WSN are represented respectively in the dissertation. Deterministic deployment problem which meets the particular requests is studied from information coverage’s point of view. A deployment strategy based on detection fusion is proposed. At first, Neyman-Pearson criterion is adopted to fuse the detection information from all sensor nodes in one unit grid, to realize highly efficient coverage for square and triangle unit grid. Thereafter the grid plot methods of detection field are presented respectively for two types of unit grid. Consequently, the number of sensor nodes to monitor the detection field and their locations are determined. With fixed quantity of sensor nodes, the new deployment strategy offers larger coverage area with the desired detection accuracy. Multipath effect is one of the key factors which restrict the application of Radio Interferometric Positioning System (RIPS) in complicated environment. Multipath effect on radio interferometric ranging technique is analyzed from the theoretical and experimental point of view. A mathematic model of multipath error is proposed in condition of mirror’s reflecting, according to the physical characteristic of multipath signal, and then multipath error analysis of RIPS is presented. The relationships between RIPS measurement and multipath parameters including attenuation factor, antenna height and horizontal distance of sensor nodes are studied. Simulation results indicate that the magnitude of multipath error is determined by attenuation factor, but the distribution of multipath error will be more complex as antenna height increases. Radio interferometric ranging technique provides a linear combination of distances among two sending nodes and two receiving nodes, however, the existing localization methods can not be applied directly. A modified Particle Swarm Optimization (PSO) based localization method is proposed. In order to keep the diversity of the swarm, a small perturbation is added to the position of particle after updating phase, which is similar to the mutation of Genetic Algorithm (GA). Experimental data are used to compare the performance of two localization methods based on GA and the modified PSO. The results indicate that both localization methods have found the approximation of the node locations in finite numbers of iterations, but the modified PSO based localization method shows significantly quicker convergence to a reasonably good solution. For a WSN with cluster-tree network topology, the implementation details of centralized particle filter based tracking algorithm is presented firstly, thereafter two tracking algorithms based on a distributed particle filter are proposed, which adopt different methods of selecting measurements from sensor nodes to balance the information contribution against the resource consumption. Performance metrics are proposed and discussed in terms of tracking performance, communication cost, energy cost, and tracking response time. Simulations are conducted to quantitatively compare the proposed tracking algorithms using the proposed performance evaluation system. The results indicate that two distributed algorithms achieve dramatic reduction of communication cost, energy cost and tracking response time at only small cost of tracking performance. Finally, performance of three tracking algorithms has been studied in terms of detection threshold and sensor density. In summary, the dissertation focuses on the study of deployment, localization and target tracking in WSN, in order to promote the application of WSN.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/398
Collection工业信息学研究室
Affiliation1.中国科学院沈阳自动化研究所
2.中国科学院研究生院
Recommended Citation
GB/T 7714
黄艳. 面向目标跟踪的无线传感器网络部署与定位问题研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2009.
Files in This Item:
File Name/Size DocType Version Access License
10001_20051801471902(6615KB) 开放获取--Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[黄艳]'s Articles
Baidu academic
Similar articles in Baidu academic
[黄艳]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[黄艳]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.