基于时空关联性的无线传感器网络网关异常检测研究
发布时间：2018-07-24 21:48
【摘要】：随着通信技术、检测技术、嵌入式技术的快速发展，无线传感器网络在各领域中得到了广泛的应用。网关作为无线传感器网络的管理者，承上启下地控制着整个网络运行，其管理权限和可靠性要求都是最高的。由于无线传感器网络通常工作在复杂、恶劣甚至敌对环境中，而网关作为无线传感器网络最主要的现场设备，其工作中必然出现一些不可避免的重要或紧急问题，因此，无线传感器网络网关的故障检测已引起了学术界和工程界的高度重视。无线传感器网络中不同的传感器节点所感知的信息不仅在时间上高度关联，而且在空间上也具备了关联特性，这就形成了无线传感器网络的一个新的“时空关联性”特性。通过研究无线传感器网络时空变化规律，能够在网络故障检测、网络调度、网络节能等方面发挥巨大的作用。因此基于时空关联性的网关异常检测研究具有重要的理论意义和实际应用价值。 本文在总结国内外研究成果的基础上，，针对目前网关故障研究所存在的问题，提出了一种“现场级”网关故障检测架构。着重研究了无线传感器网络时空关联特性理论，基于时空关联性的网关异常检测实现方法，时空关联性的MAC协议，面向适用于工业现场并构建了时空关联无线传感器网络，进行了工业现场的性能测试和验证。主要研究工作总结如下： 1.根据无线传感器网络在监测区域内信息融合的过程，从信号的相关性理论出发，研究构建时空关联性的基本架构，提出无线传感器网络时间关联性和空间关联性的基本概念。构建了无线传感器网络的时间关联性函数，空间关联性函数，时间关联性系数，空间关联性系数，网关状态模型，以及用于网关状态模型评价的节点贡献评价函数及参数贡献评价函数。为网关异常检测及时空关联MAC协议的研究提供理论基础。 2.分析了无线传感器网络网关异常检测存在的问题，针对网关当前状态和历史状态信息难以比对的问题，提出了改进型的滑动窗口法，实现了网关历史特征状态信息的特征提取；并在此基础上，综合了误检率、漏检率、计算复杂度等多个指标，分析比较了多种异常检测方法，提出了一种改进型的基于距离的离群点检测技术，实现了网关状态的离群点检测。 3.根据无线传感器网络时空关联性的理论架构，研究了一种时空关联性的MAC协议。该协议融合了时空关联的方法，针对自适应入网、功率调节、时隙分配、网关异常检测及异常处理等问题，提出了相应的入网自适应机制、功率自适应机制、时隙动态分配机制、网关异常检测及网络异常控制机制的具体方法，为时空关联性的网关异常检测技术在无线网络中的应用提供了具体的实现途径。 4.创新构建一套具有网关异常检测功能的时空关联性的网络系统，提出了基于时空关联性规则的传感器节点及诊断节点的具体实现方法；研究设计了整个网络系统中传感器节点，诊断节点和网关的具体软硬件实现方式，以及相应的工作状态机；提出了一种面向工业应用的应用层实现方法，为网络系统推向实际工业应用奠定基础。 5.将具有时空关联性的无线传感器网络系统实际应用于612冷轧连退机组设备检测与诊断系统中，并进行了传感器节点，诊断节点的传输距离测试，功率自适应测试，时隙长度分配测试，网关异常检测测试及网络异常处理测试。全面评估网络性能，验证网络系统在工业现场运行的可行性，并在实际冷轧连退机组中构建具有一套示范系统。全文研究成果为网关的异常检测的实现方法，技术手段提供重要的理论支持和实验基础，为提高无线传感器网络的可靠性研究提供一种新思路。
[Abstract]:With the rapid development of communication technology, detection technology and embedded technology, wireless sensor networks have been widely used in various fields. As the manager of wireless sensor networks, gateways are the most important for the management of the whole network, and the power and reliability of the wireless sensor network are the highest. As the most important field equipment in wireless sensor networks, the gateway is the most important field device in the wireless sensor network, and there must be some unavoidable and urgent problems in its work. Therefore, the fault detection of the wireless sensor network gateway has caused great attention in the academic and industrial circles. The information perceived by sensor nodes is not only highly correlated in time, but also in space, which forms a new "spatio-temporal correlation" characteristic of wireless sensor networks. By studying the temporal and spatial variation of wireless sensor networks, it can be used in network fault detection, network scheduling, and network energy conservation. Therefore, the research of gateway anomaly detection based on spatial-temporal correlation is of great theoretical significance and practical application value.
On the basis of summarizing the research results at home and abroad, this paper puts forward a "field level" gateway fault detection architecture in view of the existing problems of gateway fault research, and focuses on the study of the spatio-temporal correlation characteristic theory of wireless sensor networks, the realization method of gateway anomaly detection based on spatio-temporal correlation, and the MAC protocol of spatio-temporal correlation, In view of the application of the industrial site and building a spatio-temporal wireless sensor network, the performance test and verification of the industrial site are carried out. The main research work is summarized as follows:
1. according to the process of information fusion in the monitoring area of wireless sensor networks, based on the correlation theory of the signal, the basic architecture of constructing spatiotemporal correlation is studied, and the basic concepts of time correlation and spatial correlation of wireless sensor networks are proposed. The time correlation function and spatial correlation function of the wireless sensor network are constructed. Time correlation coefficient, spatial correlation coefficient, gateway state model, and node contribution evaluation function and parameter contribution evaluation function used to evaluate gateway state model. It provides a theoretical basis for the research of gateway anomaly detection and spatio-temporal Association MAC protocol.
2. the existing problems of abnormal detection of wireless sensor network gateway are analyzed. In view of the problem that the current state and historical state information of the gateway are difficult to compare, an improved sliding window method is proposed, which realizes the feature extraction of the historical feature state information of the gateway, and on this basis, the error detection rate, the leak detection rate, the computational complexity and so on are integrated. A variety of anomaly detection methods are compared and compared, and an improved distance based outlier detection technique is proposed, which realizes the outlier detection of the gateway state.
3. according to the theoretical framework of spatio-temporal correlation of wireless sensor networks, a spatio-temporal MAC protocol is studied. The protocol combines the method of spatio-temporal correlation, and proposes a corresponding adaptive mechanism, power adaptive mechanism, and adaptive mechanism, power adaptive mechanism, for adaptive network entry, power adjustment, time slot allocation, gateway anomaly detection and exception handling. The gap dynamic allocation mechanism, the gateway anomaly detection and the specific method of network exception control mechanism provide a specific implementation approach for the application of space-time associated gateway anomaly detection technology in wireless network.
4. a set of spatio-temporal network system with gateway anomaly detection function is set up. The specific implementation method of sensor nodes and diagnostic nodes based on spatiotemporal association rules is proposed, and the specific software and hardware implementation methods of sensor nodes, diagnostic nodes and gateways in the whole network system are designed and designed, as well as the corresponding work. As a state machine, a method of application layer for industrial application is put forward, which lays the foundation for pushing the network system to practical industrial application.
5. the wireless sensor network system with spatiotemporal correlation is applied to the detection and diagnosis system of the 612 cold rolling mill unit, and the sensor node, the transmission distance test of the diagnosis node, the power adaptive test, the time slot length distribution test, the gateway anomaly detection test and the network anomaly processing test. The performance of the network proves the feasibility of the network system running in the industrial field, and has a set of demonstration systems in the actual cold rolling mills. The full text research results provide an important theoretical support and experimental basis for the implementation of the abnormal detection of the gateway, and the technical means. It provides a kind of reliability research for the high wireless sensor network. New ideas.
【学位授予单位】：上海大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5
[Abstract]:With the rapid development of communication technology, detection technology and embedded technology, wireless sensor networks have been widely used in various fields. As the manager of wireless sensor networks, gateways are the most important for the management of the whole network, and the power and reliability of the wireless sensor network are the highest. As the most important field equipment in wireless sensor networks, the gateway is the most important field device in the wireless sensor network, and there must be some unavoidable and urgent problems in its work. Therefore, the fault detection of the wireless sensor network gateway has caused great attention in the academic and industrial circles. The information perceived by sensor nodes is not only highly correlated in time, but also in space, which forms a new "spatio-temporal correlation" characteristic of wireless sensor networks. By studying the temporal and spatial variation of wireless sensor networks, it can be used in network fault detection, network scheduling, and network energy conservation. Therefore, the research of gateway anomaly detection based on spatial-temporal correlation is of great theoretical significance and practical application value.
On the basis of summarizing the research results at home and abroad, this paper puts forward a "field level" gateway fault detection architecture in view of the existing problems of gateway fault research, and focuses on the study of the spatio-temporal correlation characteristic theory of wireless sensor networks, the realization method of gateway anomaly detection based on spatio-temporal correlation, and the MAC protocol of spatio-temporal correlation, In view of the application of the industrial site and building a spatio-temporal wireless sensor network, the performance test and verification of the industrial site are carried out. The main research work is summarized as follows:
1. according to the process of information fusion in the monitoring area of wireless sensor networks, based on the correlation theory of the signal, the basic architecture of constructing spatiotemporal correlation is studied, and the basic concepts of time correlation and spatial correlation of wireless sensor networks are proposed. The time correlation function and spatial correlation function of the wireless sensor network are constructed. Time correlation coefficient, spatial correlation coefficient, gateway state model, and node contribution evaluation function and parameter contribution evaluation function used to evaluate gateway state model. It provides a theoretical basis for the research of gateway anomaly detection and spatio-temporal Association MAC protocol.
2. the existing problems of abnormal detection of wireless sensor network gateway are analyzed. In view of the problem that the current state and historical state information of the gateway are difficult to compare, an improved sliding window method is proposed, which realizes the feature extraction of the historical feature state information of the gateway, and on this basis, the error detection rate, the leak detection rate, the computational complexity and so on are integrated. A variety of anomaly detection methods are compared and compared, and an improved distance based outlier detection technique is proposed, which realizes the outlier detection of the gateway state.
3. according to the theoretical framework of spatio-temporal correlation of wireless sensor networks, a spatio-temporal MAC protocol is studied. The protocol combines the method of spatio-temporal correlation, and proposes a corresponding adaptive mechanism, power adaptive mechanism, and adaptive mechanism, power adaptive mechanism, for adaptive network entry, power adjustment, time slot allocation, gateway anomaly detection and exception handling. The gap dynamic allocation mechanism, the gateway anomaly detection and the specific method of network exception control mechanism provide a specific implementation approach for the application of space-time associated gateway anomaly detection technology in wireless network.
4. a set of spatio-temporal network system with gateway anomaly detection function is set up. The specific implementation method of sensor nodes and diagnostic nodes based on spatiotemporal association rules is proposed, and the specific software and hardware implementation methods of sensor nodes, diagnostic nodes and gateways in the whole network system are designed and designed, as well as the corresponding work. As a state machine, a method of application layer for industrial application is put forward, which lays the foundation for pushing the network system to practical industrial application.
5. the wireless sensor network system with spatiotemporal correlation is applied to the detection and diagnosis system of the 612 cold rolling mill unit, and the sensor node, the transmission distance test of the diagnosis node, the power adaptive test, the time slot length distribution test, the gateway anomaly detection test and the network anomaly processing test. The performance of the network proves the feasibility of the network system running in the industrial field, and has a set of demonstration systems in the actual cold rolling mills. The full text research results provide an important theoretical support and experimental basis for the implementation of the abnormal detection of the gateway, and the technical means. It provides a kind of reliability research for the high wireless sensor network. New ideas.
【学位授予单位】：上海大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5
【参考文献】
相关期刊论文 前10条
1 鲍必赛;伍健荣;楼晓俊;刘海涛;;无线传感器网络信息时空融合模型与算法研究[J];传感器与微系统;2012年04期
2 张书奎;崔志明;樊建席;张文哲;;基于伸展树的无线传感器网络事件区域检测[J];电子学报;2010年S1期
3 吴丽华;;物联网业务网关组网方案研究[J];电信快报;2012年11期
4 闫少华;张巍;滕少华;;基于密度的离群点挖掘在入侵检测中的应用[J];计算机工程;2011年18期
5 陈默;于戈;谷峪;贾子熙;王艳秋;;An Efficient Method for Cleaning Dirty-Events over Uncertain Data in WSNs[J];Journal of Computer Science & Technology;2011年06期
6 郑国强;李建东;周志立;;无线传感器网络MAC协议研究进展[J];自动化学报;2008年03期
7 周四望;林亚平;张建明;欧阳竞成;卢新国;;传感器网络中基于环模型的小波数据压缩算法[J];软件学报;2007年03期
8 张琼;肖金超;;第二十九章 工业无线WIA网络冗余网关设计与实现[J];仪器仪表标准化与计量;2011年05期
9 曾鹏;;工业无线技术的标准化与应用[J];中国仪器仪表;2008年03期
10 王平;王泉;王恒;魏e
本文编号：2142755
本文链接：https://www.wllwen.com/kejilunwen/wltx/2142755.html
