【作者】 曾志文；

【导师】 陈志刚；

【作者基本信息】 中南大学 ， 计算机应用技术， 2010， 博士

【摘要】 无线传感器网络是当前在国际上备受关注的、涉及多学科高度交叉、知识高度集成的前沿热点研究领域,具有十分广阔的应用前景和实用价值,被认为是对21世纪产生巨大影响力的技术之一。无线传感器节点只配备有限的电源,往往工作在恶劣的环境,在目前的条件下,通过提高电池容量和人工更换电池的方法来更新能源是不大可行的。因此,能量是无线传感器网络中最稀有的资源,有效地提高能量效率以延长网络寿命是无线传感器网络研究中的重要问题。而无线传感器网络数据收集的特性,使得传感器网络往往会由于局部能量消耗的不均衡而形成“能量空洞”现象,从而导致网络失效。本文主要针对平面数据收集无线传感器网络的“能量空洞”现象,分别对节点发射功率可调的无线传感器网络、带缓冲区的节点非均匀部署的网络、基于移动Sink的网络的能量空洞避免策略进行了深入研究,并提出一种非均匀部署节点的方法来避免能量空洞的产生。研究内容和成果如下：(1)基于数据发送速率、能量消耗与数据收集延迟的相关性,提出了平面数据收集无线传感器网络基于可调节点数据发送速率的能量空洞避免策略。通过理论分析得到了网络中不同区域数据转发量与发射半径r之间的关系,从而推导出网络不同区域间的能量消耗情况。通过让能量消耗低的节点以较高的数据发送率,能量消耗高的节点保持较低的数据发送率发送数据,达到能量消耗均匀且不增加延迟的目的。在此基础上,将原发往能量空洞区域节点的一部分数据转而发往能量消耗较低的区域,均衡网络中节点间的数据承担量与能量消耗,进一步提高网络寿命。数据仿真显示,该方法提高了网络寿命17.48%,降低网络延迟34.76%,达到了较好的效果。(2)针对平面数据收集网络,建立了能量消耗模型,提出一种理论上可使节点能量消耗均匀的节点非均匀部署的“能量空洞”避免策略。采用微元分析方法,得到网络不同区域能量消耗的理论结果,从而给出节点非均匀部署的密度曲线,按此密度曲线部署节点,理论上能均衡网络的能量消耗,达到避免能量空洞的目的；仿真实验表明,该方法能使网络剩余能量降至20%以下。(3)针对Sink沿固定缓冲区移动的平面无线传感器数据收集网络。给出了缓冲区位置的优化取值,以及节点不均匀部署的密度函数,可有效提高网络寿命。采用微元分析方法,建立带缓冲区节点均匀分布网络的能量消耗模型,利用该模型计算出每一缓冲区、每一发射半径所对应的能量消耗,从中取出网络总能量消耗最小、网络寿命最长的缓冲区位置与对应的能量发射级别作为网络缓冲区位置与节点发射半径,然后,按选定的缓冲区位置、发射级别计算各处的能量消耗,并转换成部署节点的密度曲线,依此密度曲线部署的网络能均衡节点能量消耗。(4)针对不规则凸形网络中Sink移动问题,提出启发式算法来确定Sink的移动位置,算法能够较好地避免能量空洞,提高网络寿命。用微元分析法分析Sink位置固定时网络中任意节点的能量消耗,然后网格化网络(将网络划分成许多足够小的网格),计算出每个Sink位置对应的每个网格的能量消耗。若Sink在某个位置收集k轮数据,则Sink的下一个位置如此选取：每个网格当前已消耗的能量加上下一个位置k轮数据消耗能量,得到该位置下所有网格的能量消耗,选取所有网格中最大的能量消耗值作为Sink在该位置网络的能量消耗,使得网络能量消耗最小的位置,即为Sink的下一个位置。实验结果表明,该方法比静止Sink的网络寿命提高5倍以上,比Sink沿网络边界移动的网络寿命提高6%,网络死亡时其剩余能量降低20%以上。更多还原

【Abstract】 Recently people focused their attentions on a hot international research frontier area the-Wireless sensor network (WSN), which is a multi-disciplinary and highly overlapping research area. It integrates highly with several subjects of knowledge and has the extremely broad application prospect and the practical value. Thus it is considered as one of the huge influential technique to the 21st century.As the wireless sensor node only provides limited power source and works in a bad environment, under the present condition, it is impossible to renew the energy greatly through the method of enhancing the battery capacity or shift the battery manually. Therefore, the energy is the rarest resources in the WSN. The most important issue in the WSN research is how to enhance the energy efficiently to lengthen the network lifetime. But the characteristics of the WSN data collection determine the formation of Energy hole (EH) phenomenon in the network because of imbalanced energy consumption, which causes the network failure. This dissertation aims at energy-hole avoidance in flat data-gathering WSN and deeply researched on the nodes in WSN with adjustable transmitting power, non-uniform node density WSN with buffer area and the EH avoidance strategy based on a mobile sink respectively. The research content and the achievement are as follows:1) Based on the correlation of data transmission speed, the energy consumption and the delay, it proposed a strategy of EH avoidance in the flat data-gathering WSN.The paper obtained the relationship between the quantity of the data and the transmitting radius r, then inferred the energy consumption in different network regions. It proposed a new method for balancing the energy consumption without increasing delay. The method is that the nodes with lower energy consumption adopt a higher data transmission rate, while the nodes with higher energy consumption maintain a lower data transmission rate. In this foundation, the system transmits part of the data primary toward the energy-hole region to the lower energy consumption region to balance the quantity of the data and the energy consumption, and it further lengthens the lifetime of the network. The data simulation demonstrated that the lifetime of the network has been lengthened by 17.48%, and the delay of the network reduced by 34.76%.2) In view of the flat data collection network, it has established the energy consumption model and proposed an EH avoidance strategy by deploying the node unevenly.Using the infinitesimal analysis method, the theoretical result of the network energy consumption in different regions is obtained. Therefore the uneven density curve of the node can be given. If the WSN’s nodes were deployed on this density curve, the WSN energy consumption would be balanced theoretically. The simulation experiment indicated that this method can lower the remain energy to 20% below.(3) For the flat data-gathering WSN that its sink moves along the fixed buffer zone, optimized position of the buffer and the non-uniform node deployment density function are given, which can improve the network lifetime effectively.Using the infinitesimal analysis method, this paper established the energy consumption model for the WSN on which its sink moves along the fixed buffer zone. Based on the model, if the launch radius of the node and the position of the buffer zone were given, the energy consumption of the node on each point in the WSN can be calculated. We computed the energy consumption by each radius and each position. The optimization position of the buffer and the launch radius of the node lead the minimum energy consumption of the net, so we got them. Later we changed the minimum energy consumption to the node density function. Networks deployed on this density curve can balance node’s energy consumption.(4) In view of the issue of the sink moving in the irregular convexity network, the heuristic algorithm has been proposed to determine the motion position of the sink,. This algorithm can avoid the energy hole and lengthen the lifetime of the network effectively.The paper analyzed the energy consumption of the node with the infinitesimal analytic method when the sink position is known. Then divided network into many small enough grids, calculated the energy consumption of each grids in the correspondence position of the sink. If the sink have collected data k rounds in some position, the next position of the sink can be selected as follows:1) obtain the total amount of the energy consumption that each grid has consumed, named it E1; 2) compute the amount of the energy consumption that each grid will consume when the sink is in one of the grids for collecting data k rounds, named it E2; 3) select the maximum from (E1+E2), named it E3.; 4) do 2) and 3) for the sink in every other grid. select the minimum form E3.named it E4. The sink position corresponding to E4 is the next position. The experimental result indicated that by this method the network life is 5 times longer than the ones with static sink about 6% longer than the ones whose sinks move on the borderline of the WSN.更多还原