基于小世界模型的无线传感器网络研究

发布时间：2019-06-04 18:38

【摘要】：由于传感器节点体积较小、能量有限且不能补充,所以无线传感器网络研究时面临的关键问题便是资源有限,通信、计算、储存及带宽传输等能力受到限制。小世界网络具有集聚系数较高而平均最短路径长度较短的特点,存在于众多现实网络中。在无线传感器网络中适当添加长程链接即捷径,构造具有小世界特性的无线传感器网络,能够有效地提升网络的性能。因为无线传感器网络所处环境复杂多变,节点失效是不可避免的,所以研究节点失效后网络正常运行的情况并基于此改进网络拓扑结构是至关重要的。网络抗毁性描述的就是网络在部分节点失效后维持正常工作的能力,因此需要确定一种衡量无线传感器网络抗毁性能的测度。本文提出了一种基于分区的均衡能耗模型(PUECM模型),在无线传感器网络中通过添加捷径构建小世界模型。该模型主要针对无线传感器网络的两个特点展开:其一是网络中所有节点都要将感知处理过的信息发送给Sink节点;其二是距离Sink节点较近的节点能耗比其他节点快,其失效极易导致整个网络崩溃。基于此,PUECM模型创建的捷径都是指向Sink节点,且捷径的端点会随着网络的运行周期地轮换于不同高级节点间。其中,捷径端点的选择取决于网络中高级节点所处的区域、剩余能量及指向Sink节点的角度。首先由高级节点地理位置确定备选节点,然后选择其中剩余能量最高的节点最捷径端点。若存在至少两个备选节点同为能量最高者,则由指向Sink节点的角度确定捷径端点。仿真结果表明PUECM模型不仅具有良好的小世界特性,还可以均衡网络中节点的能耗,降低数据通信延迟,提高网络的整体效率。针对网络节点极易失效问题,本文还提出了一种基于节点重要度熵的抗毁性评估测度,即确定一种衡量基于小世界模型的无线传感器网络抗毁性的标准。首先评估网络中节点的重要性,主要考虑的是小世界理论中衡量节点重要性的介数概念及邻居节点和邻边对节点重要性的影响。与其他节点重要度衡量标准相比,本文提出的节点重要性评价标准是最全面准确的。然后借助于熵的概念,确定网络中节点重要度的均匀程度,即网络的重要度熵,以此表示网络的抗毁性。仿真对比了采用不同抗毁性测度衡量不同类型节点失效后网络抗毁性的结果。结果显示,与基于节点度熵和最短路径数的抗毁性测度比较,基于节点重要度熵的抗毁性评估测度更能准确而全面的评估网络的抗毁性能。
[Abstract]:Because the volume of sensor nodes is small, the energy is limited and can not be replenished, the key problems in the research of wireless sensor networks are limited resources, communication, computing, storage and bandwidth transmission. Small-world networks have the characteristics of high agglomeration coefficient and short average shortest path length, and exist in many real networks. Adding long-range links to wireless sensor networks is a shortcut to construct wireless sensor networks with small-world characteristics, which can effectively improve the performance of the network. Because the environment of wireless sensor networks is complex and changeable, node failure is inevitable, so it is very important to study the normal operation of the network after node failure and improve the network topology based on it. Network invulnerability describes the ability of the network to maintain normal work after the failure of some nodes, so it is necessary to determine a measure to measure the anti-destruction performance of wireless sensor networks. In this paper, a partition-based balanced energy consumption model (PUECM model) is proposed to construct a small-world model by adding shortcuts in wireless sensor networks. The model mainly aims at two characteristics of wireless sensor network: one is that all the nodes in the network should send the perceptual processed information to the Sink node; The other is that the energy consumption of the node close to the Sink node is faster than that of the other nodes, and its failure can easily lead to the collapse of the whole network. Based on this, the shortcuts created by the PUECM model point to Sink nodes, and the end points of shortcuts rotate among different advanced nodes according to the running cycle of the network. Among them, the choice of shortcut endpoint depends on the area of the advanced node in the network, the residual energy and the angle pointing to the Sink node. Firstly, the alternative node is determined by the geographical location of the advanced node, and then the shortest endpoint of the node with the highest residual energy is selected. If there are at least two alternative nodes with the highest energy, the shortcut endpoint is determined by the angle pointing to the Sink node. The simulation results show that the PUECM model not only has good small-world characteristics, but also can balance the energy consumption of nodes in the network, reduce the delay of data communication and improve the overall efficiency of the network. In order to solve the problem that network nodes are easy to fail, this paper also proposes an evaluation measure of invulnerability based on node importance entropy, that is, to determine a standard to measure the invulnerability of wireless sensor networks based on small-world model. Firstly, the importance of nodes in the network is evaluated, and the concept of medium to measure the importance of nodes in small world theory and the influence of neighbor nodes and adjacent edges on the importance of nodes are mainly considered. Compared with other node importance measurement criteria, the node importance evaluation criteria proposed in this paper are the most comprehensive and accurate. Then, with the help of the concept of entropy, the uniformity of node importance in the network, that is, the importance entropy of the network, is determined to represent the invulnerability of the network. The results of using different invulnerability measures to measure the network invulnerability of different types of nodes are simulated and compared. The results show that compared with the invulnerability measure based on node degree entropy and the shortest path number, the invulnerability evaluation measure based on node importance entropy is more accurate and comprehensive than the one based on node degree entropy and the shortest path number.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP212.9;TN929.5

【相似文献】