基于数据融合的无线传感器网络目标检测研究
发布时间：2018-10-09 19:50
【摘要】：无线传感器网络是一种由大量空间分布式自治设备组成的无线网络,通常利用位于不同位置的传感器节点相互协作监测某一物理或者环境条件,过去的20年已经从理论概念发展成一种新兴的现代技术。目标检测作为无线传感器网络的一种具体应用,即判断目标对象是否出现在监测区域,可应用于军事指挥、医疗诊断和天气预测等。近年来,将数据融合技术应用于目标检测,极大地提高了目标检测的精确度和可靠性。本文主要研究了基于数据融合的无线传感器网络目标检测问题,主要工作如下:1.介绍了无线传感器网络和数据融合技术的相关内容,并对数据融合的功能模型和结构模型进行着重介绍,本文主要针对检测融合结构进行研究。多传感器检测系统通常采用集中式和分布式两种检测结构,本文对基于两种检测结构的目标检测问题分别进行了分析研究。2.介绍了多传感器目标检测的基础理论,包括二元假设检验问题、贝叶斯准则和Neyman-Pearson准则,这是论文的理论基础。然后分析了两种经典的分布式目标检测融合规则,Chair-Varshney融合规则和CountingRule,并通过仿真对比了两种融合规则的检测性能。3.研究了一种基于顺序传输的传感器选择策略,传统的集中式目标检测中,监测区域内的所有传感器节点需要将接收观测数据发送至融合中心,通信开销较大。文章从减少参与判决传感器节点数量进而减少通信开销的角度分析了一种基于顺序传输的集中式目标检测策略。通过该策略,融合中心仅利用监测区域内的部分传感器节点就能进行判决,从而减少通信开销和能量消耗。4.基于可靠性理论中的Bathtub-Shaped失效率,进一步研究了无线传感器网络存在失效传感器节点的分布式目标检测问题。虽然考虑Bathtub-Shaped失效率的并行拓扑结构优于经典的Chair-Varshney融合规则,但当存在大量失效传感器节点时,并行拓扑结构的检测性能下降较快。针对这一问题,利用目标辐射能量衰减模型,文章给出了一种基于目标辐射能量的串行目标检测融合结构,并推导了相应的决策融合规则,最后通过仿真对比所提融合规则与考虑传感器失效率的并行拓扑结构的检测性能。
[Abstract]:Wireless sensor network (WSN) is a wireless network composed of a large number of spatial distributed autonomous devices. It usually uses sensor nodes located in different locations to cooperate with each other to monitor a physical or environmental condition. The past 20 years have evolved from a theoretical concept to a new modern technology. As a specific application of wireless sensor networks, target detection can be used in military command, medical diagnosis and weather prediction. In recent years, data fusion technology has been applied to target detection, which greatly improves the accuracy and reliability of target detection. In this paper, the problem of target detection in wireless sensor networks based on data fusion is studied. The main work is as follows: 1. The related contents of wireless sensor network and data fusion technology are introduced, and the function model and structure model of data fusion are emphatically introduced. Multi-sensor detection systems usually use centralized and distributed detection structures. In this paper, the target detection problem based on two detection structures is analyzed and studied respectively. This paper introduces the basic theory of multi-sensor target detection, including binary hypothesis test, Bayesian criterion and Neyman-Pearson criterion, which is the theoretical basis of this paper. Then, two classical fusion rules of distributed target detection, Chairperson Varshney fusion rule and CountingRule, are analyzed, and the detection performance of the two fusion rules is compared by simulation. 3. A sensor selection strategy based on sequential transmission is studied. In the traditional centralized target detection, all sensor nodes in the monitoring area need to send the received observation data to the fusion center, and the communication overhead is high. This paper analyzes a centralized target detection strategy based on sequential transmission from the point of view of reducing the number of nodes involved in the decision sensor and thus reducing the communication overhead. Through this strategy, the fusion center can use only some sensor nodes in the monitoring area to make decision, thus reducing communication overhead and energy consumption. Based on the failure rate of Bathtub-Shaped in reliability theory, the distributed target detection problem with invalid sensor nodes in wireless sensor networks is further studied. Although the parallel topology considering the failure rate of Bathtub-Shaped is superior to the classical Chair-Varshney fusion rule, the detection performance of the parallel topology decreases rapidly when there are a large number of failed sensor nodes. In order to solve this problem, a serial target detection fusion structure based on target radiation energy is presented by using the radiation energy attenuation model of the target, and the corresponding decision fusion rules are derived. Finally, the performance of the proposed fusion rule and the parallel topology considering the sensor failure rate are compared by simulation.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5
本文编号：2260567
[Abstract]:Wireless sensor network (WSN) is a wireless network composed of a large number of spatial distributed autonomous devices. It usually uses sensor nodes located in different locations to cooperate with each other to monitor a physical or environmental condition. The past 20 years have evolved from a theoretical concept to a new modern technology. As a specific application of wireless sensor networks, target detection can be used in military command, medical diagnosis and weather prediction. In recent years, data fusion technology has been applied to target detection, which greatly improves the accuracy and reliability of target detection. In this paper, the problem of target detection in wireless sensor networks based on data fusion is studied. The main work is as follows: 1. The related contents of wireless sensor network and data fusion technology are introduced, and the function model and structure model of data fusion are emphatically introduced. Multi-sensor detection systems usually use centralized and distributed detection structures. In this paper, the target detection problem based on two detection structures is analyzed and studied respectively. This paper introduces the basic theory of multi-sensor target detection, including binary hypothesis test, Bayesian criterion and Neyman-Pearson criterion, which is the theoretical basis of this paper. Then, two classical fusion rules of distributed target detection, Chairperson Varshney fusion rule and CountingRule, are analyzed, and the detection performance of the two fusion rules is compared by simulation. 3. A sensor selection strategy based on sequential transmission is studied. In the traditional centralized target detection, all sensor nodes in the monitoring area need to send the received observation data to the fusion center, and the communication overhead is high. This paper analyzes a centralized target detection strategy based on sequential transmission from the point of view of reducing the number of nodes involved in the decision sensor and thus reducing the communication overhead. Through this strategy, the fusion center can use only some sensor nodes in the monitoring area to make decision, thus reducing communication overhead and energy consumption. Based on the failure rate of Bathtub-Shaped in reliability theory, the distributed target detection problem with invalid sensor nodes in wireless sensor networks is further studied. Although the parallel topology considering the failure rate of Bathtub-Shaped is superior to the classical Chair-Varshney fusion rule, the detection performance of the parallel topology decreases rapidly when there are a large number of failed sensor nodes. In order to solve this problem, a serial target detection fusion structure based on target radiation energy is presented by using the radiation energy attenuation model of the target, and the corresponding decision fusion rules are derived. Finally, the performance of the proposed fusion rule and the parallel topology considering the sensor failure rate are compared by simulation.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5
【参考文献】
相关期刊论文 前3条
1 蒋铁珍;;分布式检测多目标融合算法研究[J];中国电子科学研究院学报;2010年06期
2 王声才;李艾华;肖秋生;王涛;;无线传感器网络中基于区域决策的距离均值融合算法[J];传感技术学报;2009年09期
3 王骐;王殊;孟中楼;;分布式入侵检测系统的融合算法[J];华中科技大学学报(自然科学版);2009年09期
相关硕士学位论文 前2条
1 倪静;无线传感网络下基于数据融合的目标检测[D];电子科技大学;2015年
2 曹啸;无线传感器网络分布式目标检测研究[D];南京邮电大学;2012年
，本文编号：2260567
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2260567.html
