【作者】 张金荣；

【导师】 曹长修；

【作者基本信息】 重庆大学 ， 控制理论与控制工程， 2008， 博士

【摘要】 无线传感器网络系统(WSNs)已经成为当前国内、外的重要的研究领域之一,它综合了传感器技术、嵌入式计算技术、现代网络及无线通信技术、分布式信息处理技术等,使人们可以在任何时间、任何地点和任何环境条件下,通过传感器节点的数据采集和处理,获取大量详实而可靠的信息.无线传感器网络在军事监控,地震与气候预测、地下、深水以及外层空间探索等许多方面都具有广泛的应用前景。在民用领域如生态环境监测、基础设施安全、先进制造、物流管理、医疗健康、工业传感、智能交通控制等方面也已经有应用。可以说无线传感器网络是信息感知和采集的一场革命,是21世纪最重要的技术之一。一般认为短距离的无线低功率通信技术WPAN和LR-WPAN最适合传感器网络使用.目前,ZigBee是部署无线传感器网络的新技术.它的技术特性决定它将是短距离、低速率无线网络技术的较佳选择.本文结合横向科研项目"基于ZigBee的民用无线燃气抄表系统",对无线传感器网络的能效和安全进行了分析和研究。能量是无线传感器网络研究的一个核心内容,其它各种内容的研究都必须围绕能量节省和均衡来进行,对无线传感器网络的安全研究也不例外。本文在分析无线传感器网络体系结构的基础上,主要在能效和安全这两个方面做了如下研究:①针对无线传感器网络中的能量有限性问题,提出了一种组合加权能量均衡的分簇与路由算法CW-EBCR,其组合权值综合考虑了节点的度、节点能量水平、节点到其邻居节点的平均距离、以及节点当选簇首的累计时间、距离sink基站的距离、最优簇数等因素.给出了簇通信半径的调整规则,实现了簇首的冗余递补和簇的自维护,顾及了簇间的能量均衡.仿真分析表明,本算法可以很好地实现簇内和簇间能量均衡,降低了网络的复杂度,延长了网络的生命周期.②针对无线传感器网络系统应用加密算法的密钥生成与配置的难题,提出一种基于矩阵QR分解的分布式传感器网络密钥生成配置方案(QR-KM).在该方案中,节点广播的是标识符和正交Q矩阵的行向量而不是密钥本身,攻击者无法获取明文密钥信息;节点在收到其它节点发送的密钥建立信息后再进行正交性验证。分析表明该方案增强了网络的安全性.③针对无线传感器网络的数据特点,将灰预测理论引入到无线感器网络数据流预测中,在改进G(1,1)模型的基础上,使用组合优化策略进行系统参数辨识.从而建立了一种无线传感器网络数据流预测算法CO-GA,并给出了其通用框架.分析和仿真表明:该算法可以实时地预测传感器的数据流、伸缩性强、能耗低、能够在小样本数据空间上完成,符合无线传感器网络资源匮乏又需快速反应的特点.该方法既可以在每个传感器节点使用,也可以在sink节点使用,因而具有广泛的适用性.④针对无线传感器网络面临的安全威胁以及现有入侵检测方法不能直接用于无线传感器网络的特点,结合模糊K均值聚类方法,利用粒子群算法(PSO)全局寻优、快速收敛的特点对其优化,提出了一种应用于传感器网络的分布式入侵检测算法(DP-EFKM)。仿真和分析表明, DP-EFKM算法建模快、运算量小,满足无线传感器网络实时检测、快速反应需求;本地检测与全局检测相结合,能有效减少能量开销;对未知类型的攻击有较高的检出率和较低的误报率。⑤结合IEEE802.15.4协议和ZigBee的特点,利用PIC18F4620处理器和CC2420无线芯片,设计开发了基于ZigBee的无线传感器网络无线燃气抄表系统原型,实现了数据的采集和传输.设计中综合考虑了效率和灵活性的需要.更多还原

【Abstract】 Wireless sensor networks (WSNs) have become one of the important research fields in the world, which integrate the technologies of sensors, embedded computation, modern networks, distributed processing, etc. WSNs are composed of low cost sensor nodes that can communicate with each other in a wireless manner, have limited computing capability and memory and operate with limited battery power.The main goal of such networks is to perform distributed sensing tasks, particularly for applications like military affairs, climate precision, smart spaces. In civil fields, such as medical systems, environmental monitoring, smart traffic control, industry control and family residence intelligence, WSNs have become one of the most active encouraging measures.Generally, WPAN and LR-WPAN is best for WSNs. At present, ZigBee is the new technology to deploy WSNs. Its technical characteristic determines that it is the better choice of short-distance and low-rate WSNs.Integrated with the crosswise project "Embedded Wireless Auto Meter Reading System Based on ZigBee", this paper mainly researches on the energy efficiency and security of WSNs. Energy efficiency is the basis of WSNs and security is a kernel issues. On the basis of analyzing network architecture of WSNs, this paper mainly makes the following researched around these two aspects:①To the energy limitation of WSNs, a combination weight energy balance clustering and routing algorithm (CW-EBCR) is proposed. This algorithm takes into consideration the following elements: the node degree and energy level, the average distance between node and its neighbors, the accumulative time of node being cluster head, the distance to sink and the optimal cluster number. Then a regulating rule of cluster communication radius and a method of selecting redundancy substitute of cluster head are developed. At the same time, the self-maintenance tactics of cluster is put forward. Simulation and analysis show that this algorithm can make energy balanced both in inner-cluster and in inter-clusters, can postpone the network life time.②A key pre-deployment scheme based on matrix QR decomposition (QR-KM) is proposed. In this scheme, what the node broadcasts are its ID and the row vector of Q orthogonal matrix, not the key itself. After receiving the row vector, node can confirm the identity by orthogonality validation. Therefore, the attacker can not access the key itself and can not decrypt the cryptography text. Analysis result shows that this scheme strengthens the WSNs security.③Aiming to the data characteristic of WSNs, grey prediction method is introduced to WSNs data stream predictions. On the basis of modifying G(1,1) model, a CO-GA algorithm for WSNs is proposed. And a combination optimization tactic is used to system parameters identification. The results of analysis and simulation show that this algorithm can predict data stream real-timely only by small-sample, meeting with the WSNs requirements for low resource occupation and for quick response. This algorithm can be used in both sensing node and sink node.④According to the security threats of WSNs, a distributed intrusion detection algorithm (DP-EFKM) is put forward. Based on modifying fuzzy K-means, this algorithm uses particle swarm optimization (PSO) method to find the optimal parameters values. The results of analysis and simulation show that this method is feasible, and it is characterized by its simple structure, small calculating amount and fast speed. Moreover, its distributing character makes integration of local detection and global detection, reducing the energy spending, improving the detection rate and decreasing the false positive rate.⑤A proto type system of WSN base on ZigBee and IEEE802.15.4 is designed which aims to realize“wireless auto gas-meter reading system based on ZigBee”. Using PIC18F4620 and CC2420 to make up the wireless gas-meter node, this proto system realized correct communication between two nodes. This project has been checked and accepted.更多还原