基于主成分分析的无线传感网数据融合

发布时间：2018-03-13 14:10

本文选题：无线传感器网络　切入点：数据融合　出处：《西南大学》2017年硕士论文　论文类型：学位论文

【摘要】：在无线传感器网络(WSNs,Wireless Sensor Networks)中,为了对感知区域进行实时有效地监测以及让采集的信息更加精准与全面,网络往往会布置大量的不同种类的传感器节点。因此,无线传感器网络中感知的数据具有大量、多样和高速的特点。此外,无线传感器节点是随机分布在感知区域内,所以网络中大部分节点存在感知区域重叠,这造成了感知数据冗余。其次,由于无线传感器节点的能源大多来源于微型电池,电池的能量有限,而冗余数据的传输会造成额外能量的消耗,所以消除传输中的数据冗余,降低能量消耗显得非常重要。数据融合是一种有效消除数据冗余的方法,因此,为解决无线传感器网络传输中数据冗余会消耗大量能量的问题,本文主要采用主成分分析方法来消除簇内融合度高的数据,减少能量消耗,提高能量效率。本文的主要内容如下:针对无线传感器网络中数据传输的冗余问题,本文提出了一种基于数据相似性的数据融合算法;其原理是数据集越相似,数据的压缩比越小。首先,为把数据相似度高的节点划分到一个簇内,提高数据的融合度,本文提出了一种新的分簇方法,同时为避免簇内节点的传输冲突,提出了限定簇内节点数量的能量模型,实现了簇内能量均衡的消耗。随后,为了最大限度地消除传输过程中的数据冗余,本文在每个簇的簇头采用了基于主成分分析的数据融合算法,最终减少了冗余数据的传输以及能量的消耗。最后,实际仿真结果表明:与LEACH-PCA与Kmeans-PCA算法相比,本文的算法能分别有效减少21.1%和13.4%的数据传输量。其次,相比LEACH-PCA算法,本文的算法能有效延长7.8%的网络生存时间。当节点数目大于300时,算法的能量消耗明显低于其它两种算法。因此,所提出的数据融合算法和分簇算法能明显提高网络的性能。
[Abstract]:In wireless sensor networks (WSN), in order to effectively monitor the perceptual region and make the collected information more accurate and comprehensive, the network often has a large number of sensor nodes of different types. In addition, wireless sensor nodes are randomly distributed in the perceptual region, so most of the nodes in the network have overlaps of perceptual regions. This results in sensing data redundancy. Secondly, because most of the energy of wireless sensor nodes comes from microbatteries, the energy of the batteries is limited, and the transmission of redundant data results in additional energy consumption, so the data redundancy in transmission is eliminated. It is very important to reduce energy consumption. Data fusion is an effective method to eliminate data redundancy. Therefore, in order to solve the problem that data redundancy will consume a lot of energy in wireless sensor network transmission, In this paper, the principal component analysis (PCA) method is used to eliminate the data with high degree of convergence, reduce energy consumption and improve energy efficiency. The main contents of this paper are as follows: aiming at the redundancy of data transmission in wireless sensor networks, In this paper, a data fusion algorithm based on data similarity is proposed. The principle is that the more similar the data set, the smaller the compression ratio of data. Firstly, in order to divide the nodes with high data similarity into a cluster, the fusion degree of data can be improved. In this paper, a new clustering method is proposed. In order to avoid the transmission conflict of nodes in the cluster, an energy model is proposed to limit the number of nodes in the cluster, which realizes the consumption of energy balance within the cluster. In order to eliminate the data redundancy in the transmission process to the maximum extent, this paper adopts a data fusion algorithm based on principal component analysis in the cluster heads of each cluster, which ultimately reduces the transmission of redundant data and the consumption of energy. Finally, The simulation results show that compared with the LEACH-PCA and Kmeans-PCA algorithms, the proposed algorithm can effectively reduce the data transmission by 21.1% and 13.4% respectively. Secondly, compared with the LEACH-PCA algorithm, the proposed algorithm is more efficient than the LEACH-PCA algorithm. When the number of nodes is greater than 7.8%, the energy consumption of the algorithm is obviously lower than that of the other two algorithms. Therefore, the proposed data fusion algorithm and clustering algorithm can obviously improve the performance of the 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】：TN929.5;TP212.9

【参考文献】