【作者】 王伟；

【导师】 汪秉文；

【作者基本信息】 华中科技大学 ， 控制理论与控制工程， 2011， 博士

【摘要】 无线传感器网络的根本任务是快速准确获取物理世界中有价值的信息。但无线传感器网络面临节点计算能力、存储能力、通信能力有限,特别是电源能量十分有限的问题。本文就无线传感器网络的三大关键技术进行研究,主要包括路由协议、数据融合和网络安全技术的问题。研究的主要目的是为了达到优化网络结构、节省网络能量、提高网络中节点能量负载均衡性、延长网络生命期以及提高网络的安全性和保证网络的可靠性。文中针对分簇网络结构的特点以及存在的一些问题,结合树型网络结构的优点,提出了一种基于分簇和树型混合结构的路由算法,该算法通过最小生成树的策略将网络中的簇头节点以树型结构组织起来,形成新的路由方式,最优化网络中直接与基站进行通信的节点的数目,减少簇头节点的能量消耗,提高网络能量利用率,延长节点的使用寿命。论文针对节点分布密度较大的网络的特征和存在大量冗余节点的问题,分析了网络中冗余节点的判定方法,提出了一种基于分簇的监测节点子集划分方法和簇内节点调度算法,其主要思想是：网络分簇完成后,根据簇内节点的分布情况,分析节点的覆盖率,根据冗余节点判定方法,将簇内节点分成若干个监测节点子集,各个监测节点子集按照调度规则进行轮换工作,减少每轮中发送数据的节点的数目,节省了网络中节点的能量消耗,延长网络的使用寿命。针对无线传感器网络中的数据融合可靠度问题,文中分析了数据融合的时间可靠度和空间可靠度,提出了一种自适应的基于时间和空间可靠度的数据融合算法,算法的主要思想是：通过控制簇内发送数据的源节点的数量来调整簇头节点执行数据融合操作的空间可靠度,通过控制源节点发送数据的频率来调整数据融合操作的时间可靠度,该算法利用网络中监测事件特征的数据融合观测可靠度来逐步调整数据融合可靠度来达到期望可靠度,达到在满足监测需求的前提下使网络能量消耗最优化。文中针对无线传感器网络的故障节点诊断问题,考虑执行网络故障节点诊断的高效和准确性,提出了一种基于分簇网络结构的实时节点故障诊断算法,该算法的主要思想是：在分簇完成后,簇头节点执行节点故障诊断算法,算法利用一种基于测试结果比较的方法来判定节点是否是故障节点,各簇簇头在网络中并行检测各个簇内的节点。该算法简化了故障诊断算法的复杂度,大大节省了诊断所需时间,提高了故障诊断的效率,同时改善了网络的能量消耗和改善网络监测的可靠性。论文最后对全文进行总结,指出论文的创新点以及提出了今后可继续研究的工作。更多还原

【Abstract】 The fundamental task of the wireless sensor network is to quickly and accurately access valuable information in the physical world. However, sensor nodes in wireless sensor networks are severely constrained by the limited computing power, limited storage capacity and limited communication capabilities. Therefore, energy efficient is the most important issue of concern in wireless sensor networks. In this paper, three key technologies of wireless sensor networks are studied, including routing protocols, data aggregation and network security technology. The main purpose is to optimize network structure, save the network energy, improve energy load balancing of nodes in the network, prolong the network lifetime, improve network security and ensure the reliability of the network.The characteristics of clustering network structure and the existence of some problems are analyzed, combined with the advantages of tree network structure, we proposed a tree-based and cluster-based hybrid structure of the routing algorithm. Using the strategy of minimum spanning tree to organize the cluster head nodes of the network into a tree structure to form a new routing. Algorithm can optimize the number of nodes which communicating directly with the base station, to reduce the energy consumption of cluster head nodes, and improve energy efficiency and prolong the lifetime of nodes.The characteristics of distribution density of nodes in a larger network are analyzed, in this case, there will be a large number of redundant nodes in the network. And then, the methods to determine redundant nodes are introduced, a cluster-based division method for subset of monitoring nodes and a scheduling algorithm for cluster head nodes are proposed for the network. The main ideas are as follows:when clustering is completed, according to the distribution of cluster head nodes to compute the coverage of nodes, and according to the determining method of redundant nodes to divide the cluster members into some subsets. Each subset of monitoring nodes take turns to work in accordance with the scheduling rules. So that can reduce the number of nodes which sending data in each round, save the energy consumption of nodes and prolong the lifetime of the network.For the problem of reliability of data aggregation in wireless sensor networks, the temporal reliability and spatial reliability of data aggregation are described. Then, an adaptive temporal and spatial reliability data aggregation algorithm is proposed, its main ideas are:to control the spatial reliability by adjusting the number of source nodes in this cluster, and to control the temporal reliability by adjusting the sending data frequency of the source nodes. Algorithm using the observed reliability of features of the events in the network to be gradually adjusted to achieve the desired reliability of data aggregation. To optimal energy consumption of the network under meeting the monitoring requirements.The fault diagnosis problem in wireless sensor networks is introduced, and a cluster-based real-time fault node diagnosis algorithm is proposed, which can ensure the efficiency and accuracy of fault diagnosis in the network. The main idea of this algorithm are:when the clustering is completed, each cluster head implements fault diagnosis, algorithm using a comparison-based method to test results to determine whether the node is a fault one. Each cluster head can diagnosis their members in parallel in the network. Algorithm simplified the complexity of fault diagnosis algorithm, reduced the diagnosis time, improve the efficiency of fault diagnosis, and improve the energy consumption and reliability of monitoring.Finally, a summary is written for the research work, and the innovations as well as the future work are introduced.更多还原