SIA OpenIR  > 工业信息学研究室
移动传感器网络若干基础问题研究
Alternative TitleFundamental Research Issues on Mobile Sensor Networks
郑建颖1,2
Department工业信息学研究室
Thesis Advisor于海斌
ClassificationTP212
Keyword移动传感器网络 无线传感器网络 群体行为 复杂系统
Call NumberTP212/Z57/2010
Pages129页
Degree Discipline机械电子工程
Degree Name博士
2009-12-02
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract近些年来,有望应用于战场环境监控、危险环境事件检测与处理、野生动物习性追踪、海洋环境探测、太空灰尘清扫等重要领域的移动传感器网络,得到了广泛的关注和深入的研究。相比较于传统的固定传感器网络,移动传感器网络特有的移动性特征,使得系统更加有效、更加灵活、更加鲁棒,将极大地增强人类对于物理世界的感知和控制能力。移动传感器网络具有许多挑战性的研究课题,其中定位、覆盖、协作目标搜索与捕获是移动传感器网络研究的基础问题。在移动传感器网络中,由于节点的移动特征,使得节点的位置始终处于变化之中,而有效的感知信息往往都是位置相关信息(Location-Aware Information),从而使得定位成为移动传感器网络首要解决的关键问题。节点的移动特征,使得网络更加灵活,可以极大地提高网络的覆盖率,增强网络的覆盖能力,降低网络覆盖的成本,如何利用移动节点来实现这些目标是移动传感器网络需要重点关注的问题。节点的移动特征,也将为系统提供一定程度的执行能力,实现目标搜索与捕获的任务,同时移动节点与移动节点、移动节点与固定节点、固定节点与固定节点之间的协作机制和协作策略也变得更加复杂,从而使得目标搜索与捕获成为移动传感器网络研究的一类挑战问题。此外,随着移动传感器网络的规模逐渐变大、自治性逐渐增强,移动传感器网络作为一个群体,其整体性质和作用方式将发生本质变化,此时群体将以行为涌现的方式完成指定的任务或者产生期望的行为,从而使得探索大规模移动传感器网络群体行为涌现的机理成为移动传感器网络不得不关注的问题。本文将主要围绕移动传感器网络定位问题、移动传感器网络覆盖问题、移动传感器网络协作目标搜索与捕获问题,和移动传感器网络群体行为涌现问题展开工作。首先,综述了移动传感器网络定位问题、覆盖问题、协作目标搜索与捕获问题、群体行为涌现问题的国内外研究现状,系统地总结了当前的研究进展。针对于移动传感器网络定位问题,提出了一类基于相对距离约束的移动传感器网络分布式定位算法。该定位算法利用已知位置的种子节点来确定其它未知位置的传感器节点,一方面考虑了当前时刻接收到的种子节点产生的约束,另一方面充分考虑节点的运动特性,分析了节点在前后时刻接收到的种子节点之间的差异,并且把这些差异以不等式形式进行描述,从而把节点定位问题转化为不等式组求解问题。仿真结果表明:该定位算法可以减小节点定位误差,提高节点定位精度,对于种子节点的分布特征要求也很低,此外,节点定位误差随着节点运动速度的增加而减小。针对于移动传感器网络覆盖问题,提出了一种改进的基于栅格密度的移动传感器网络覆盖算法。该覆盖算法首先利用固定传感器节点,采取随机均匀部署方式,使监测区域达到一定水平的覆盖率,在此基础之上,利用移动节点来进一步提高监测区域覆盖率,移动节点的位置通过协作位置分配策略进行确定。仿真结果表明:在改进的基于栅格密度的覆盖算法中,利用少量的移动节点就可以有效地提高网络的覆盖率,从而降低网络覆盖的成本,提高网络覆盖的效率。针对于移动传感器网络协作目标搜索与捕获问题,提出了一类基于局部不确定信息决策的移动传感器网络协作搜索与捕获算法。该搜索与捕获算法考虑了目标状态的概率分布特征,通过把连续环境离散化为一系列胞元,使得目标状态只能局限于有限个胞元,然后以概率分布方式描述目标的运动模型。此外,该搜索与捕获算法还考虑了捕获者有限通信距离约束,任意时刻,捕获者只能与部分处于同一局部网络中的其它捕获者进行信息交互与协作。通过为捕获者设计合适的代价函数,可以有效地降低目标搜索与捕获的时间,提高了目标搜索与捕获的效率。仿真和实际实验结果也有力地验证了搜索与捕获算法的性能。针对大规模移动传感器网络群体行为涌现机理探索问题,提出了一种考虑通信失效的群体行为涌现模型。基于该模型,分析了群体中单个主体的通信范围和通信失效概率与群体行为一致性之间的相互关系。通过仿真实验发现:群体行为一致性主要取决于单个主体的通信范围;当主体的通信范围足够大,可以保证群体行为一致时,主体之间即使通信失效概率很高也可以使得群体产生一致的群体行为。 关键词:移动传感器网络;无线传感器网络;群体行为;复杂系统
Other AbstractThe field of Mobile Sensor Networks has recently been focused and wildly studied by many researchers. Mobile sensor networks have a lot of real applications, such as battlefield surveillance, contaminated environment detection, wild animal tracking, seabed monitoring, and space clearance. Comparing with the traditional wireless sensor networks, mobile sensor networks can offer much more characteristics, such as efficiency, flexibility, and robustness. These characteristics enhance the capability of human beings to perceive and control the real world. There are abundant research topics in mobile sensor networks. Some of these topics are considered as fundamental issues, such as localization, coverage, and target search and capture. In mobile sensor networks, the positions of sensor nodes change all the time due to the mobility. However, efficient sensing information is often aware of location. Therefore, localization becomes one of fundamental challenges in mobile sensor networks. The mobility provides sensor networks more flexibility. So it is able to enhance the coverage ratio, increase the coverage capability, and reduce the coverage cost. How to make use of these characteristics is another challenge in mobile sensor networks. In addition, the mobility can provide sensor networks not only the sensing capability but also the actuating capability. Therefore, it is very natural to make use of mobile sensor network to accomplish the tasks of target search and capture. The dissertation has surveyed the localization, coverage, and target search and capture in mobile sensor networks, and then summarized the obtained results and the drawback about these topics. The problem of locating mobile sensor node is studied, and a distributed range-free localization algorithm based on matrix inequality is proposed. In the algorithm, the seed nodes whose positions are known in advance are used to determine the positions of sensor nodes. We have two observations which can form two types of constraints on location in mobile sensor networks. One is created by the seed nodes received at the current time. The other is created by the difference of seed nodes between the current time and the previous time. Both of these constraints are described by the inequalities so that the localization problem is transformed into the problem of solving the matrix inequality. Finally, Simulations have demonstrated that the proposed localization algorithm can reduce the localization errors and enhance the positioning accuracy. In addition, the localization errors reduce with respect to the increase of sensor nodes’ speed and the localization errors are also very low even when the initial distribution of seed nodes is not uniform. The problem of how to enhance the coverage performance with mobile sensor nodes is studied, and a coverage algorithm based on Grid Density is proposed. In the algorithm, static sensor nodes with random deployment are first used to monitor the given area. Then mobile sensor nodes are used to further improve the coverage performance. In order to determine the locations of mobile sensor nodes, the monitoring area is discretized into a number of cells. Those cells with low densities of sensor nodes will be selected as the possible locations for mobile sensor nodes. Simulations have demonstrated that the coverage performance is increased greatly with a small number of mobile sensor nodes and the cost of monitoring the given area is reduced significantly. The problem of searching for and capturing a single target with mobile sensor networks is studied, and a cooperative search and capture algorithm based on probabilistic strategies is proposed. The target information is perceived and obtained by static sensor networks deployed in advance, and then transmitted to the mobile sensor nodes. The mobile sensor nodes attempt to find and capture the target according to the obtained information from static sensor networks as well as coordinate the relationship with other mobile sensor nodes. Through designing the cost functions, the two above aspects are combined perfectly. Simulations have demonstrated that the proposed algorithm can reduce the time of finding and capturing the target greatly and enhance the efficiency of search and capture of single target. The mobile sensor network system will be transformed into a typical multi-agent complex system when all the sensor nodes become autonomous and mobile and the number of sensor nodes becomes very large. In the case, the emerged group behavior is much more important than the behavior of single agent. Therefore, we have built a group behavior emergence model based on explicit communication. The model assumes that each agent in the group can obtain other agents’ information by way of communication. Then the communication range and the probability of successful communication are considered as two major impact factors on the consistency of the group behavior. With simulations, the following conclusions are obtained: the consistency of multi-agent systems is largely determined by the communication range; when the communication range is large enough, the group can also obtain the consistency even when the probability of successful communication is very small. In summary, the dissertation focuses on the study of localization, coverage, target search and capture in mobile sensor networks as well as the emergence of group behavior in multi-agent complex systems. Key words: Mobile Sensor Networks; Wireless Sensor Networks; Group Behavior; Complex Systems
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/6868
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
郑建颖. 移动传感器网络若干基础问题研究[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
LWZJY01.pdf(2279KB) 开放获取LicenseApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[郑建颖]'s Articles
Baidu academic
Similar articles in Baidu academic
[郑建颖]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[郑建颖]'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.