基于典型相关性分析的无线传感网数据融合

发布时间：2018-04-24 10:06

本文选题：数据融合 + 分簇　；参考：《西南大学》2017年硕士论文

【摘要】：无线传感器网络作为物联网的重要组成部分,是由随机分布在感知区域内的成百上千的廉价的微型传感器节点构成的,通过无线通信的形式构成的一个多跳的、自组织的网络系统。它以感知、收集和处理无线传感器网络覆盖区域的数据信息,并将最后的结果发送给用户为主要目的。部署在感知区域的传感器节点的能量十分有限,且不能增补。因此,如何降低节点能量的消耗以尽可能延长网络的生命周期是目前无线传感器网络研究的重点。为了全面地覆盖感知区域,传感器节点通常部署十分地密集,这会使得相邻传感器节点感知的数据存在时间和空间的相关性,从而导致明显的冗余。为了消除冗余,数据融合成为了一种非常有效的能消除冗余、最小化传输中的数据量、节约能量的方法。在先前的工作中,很多学者提出了很多有效的消除数据冗余的方法。但是,这些方法有的能量效率较低,有的复杂度较高。在基于簇的网络结构中,能量主要被消耗在从簇头节点到汇聚节点的传输过程中。因此,本文主要考虑通过在簇头节点进行数据融合来降低传输的数据量,从而达到节约能量的目的。基于这些考虑,本文提出了一种在网络总延时约束下的无线传感器网络中的基于典型相关性分析的数据融合方法。首先,为了平衡簇与簇之间的能量消耗,我们在总延时约束下将整个网络中的总能量消耗作为最小化目标,并通过拉格朗日对偶的方法获得网络中的最优簇的数量。其次,为了避免传输过程中的拥塞,我们采用“分时隙”的分簇算法来构造网络中的最优融合树,并保证在传输过程中一次传输只占用一个时隙。再者,在簇头节点处,我们提出了一种基于典型相关性分析的数据融合方法,它可以低复杂度的处理不同类型的多维数据,消除簇头节点处的数据冗余,从而最大限度的减少传输的数据量,节约网络中的能量。最后的仿真结果表明,与现有的方法相比较,我们的方法不仅可以减少传输的数据量,节约网络中消耗的能量,还可以减少网络的延时,延长无线传感器网络的生命周期。
[Abstract]:As an important part of the Internet of things, wireless sensor networks are made up of hundreds of cheap sensor nodes randomly distributed in the perceptual region, and a multi-hop in the form of wireless communication. Self-organizing network system. Its main purpose is to perceive, collect and process the data information of the wireless sensor network coverage area, and send the final results to the user. The energy of sensor nodes deployed in the sensing area is very limited and cannot be supplemented. Therefore, how to reduce the energy consumption of nodes in order to prolong the lifetime of wireless sensor networks as much as possible is the focus of research on wireless sensor networks. In order to completely cover the perceptual region sensor nodes are usually deployed in a very dense manner which results in temporal and spatial correlation between sensing data of adjacent sensor nodes and resulting in obvious redundancy. In order to eliminate redundancy, data fusion has become a very effective method to eliminate redundancy, minimize the amount of data in transmission, and save energy. In previous work, many scholars have proposed many effective methods to eliminate data redundancy. However, some of these methods have lower energy efficiency and higher complexity. In cluster-based network architecture, energy is mainly consumed in the transmission process from cluster head node to sink node. Therefore, in this paper, the data fusion in cluster head node is mainly considered to reduce the amount of data transmitted, so as to save energy. Based on these considerations, this paper proposes a data fusion method based on canonical correlation analysis in wireless sensor networks with total delay constraints. First, in order to balance the energy consumption between clusters, we take the total energy consumption in the whole network as the minimization objective under the total delay constraint, and obtain the optimal number of clusters in the network by Lagrange duality method. Secondly, in order to avoid the congestion in the transmission process, we use the "time slot" clustering algorithm to construct the optimal fusion tree in the network, and ensure that only one time slot is used at a time during the transmission process. Furthermore, at the cluster head node, we propose a data fusion method based on the canonical correlation analysis, which can deal with different types of multidimensional data with low complexity and eliminate the data redundancy at the cluster head node. In order to minimize the amount of data transmission, save energy in the network. The simulation results show that compared with the existing methods, our method can not only reduce the amount of data transferred, save the energy consumed in the network, but also reduce the delay of the network and prolong the life cycle of the wireless sensor network.
【学位授予单位】：西南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5

【参考文献】