基于稀疏化压缩感知的无线传感网数据融合研究

发布时间：2018-01-26 18:20

本文关键词： 压缩感知 LDPC 稀疏随机矩阵能量均衡无线传感器网络　出处：《西南大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着大数据和人工智能时代的到来,物联网再次成为了研究者关注的焦点。物联网不仅给人们平时的生活带来了很大的便利,同时在医疗健康、环境监测、军事侦查以及工业领域都有很大的应用。无线传感器网络,作为物联网的核心技术之一,是由大量随机分布的传感器节点负责感知、收集、处理和分析数据信息,从而得到有效的信息并传递给用户。但是由于节点的电量有限、存储空间受限,因此,如何能够融合这些大量的数据,并且在传输的过程中减少能量的消耗,延长网络寿命变得尤为重要。此外,在一些实时性要求较高的场景下,如何将收集到的信息以最短的延迟尽快地传递到汇聚节点也是主要研究内容之一。在无线传感器网络中,传统的数据融合方式主要以减少数据冗余为目的,而不能够大幅度减少数据包的传输量与传感器的通信消耗,因此,本文利用所采集数据的时空相关性和传感器网络的自身特点,通过稀疏设计测量矩阵,提出基于稀疏性压缩感知的数据融合方法,从而减少网络的传输数据包的数量量和能量消耗。本文的主要工作内容如下:首先,针对传统的压缩感知的收集方式,本文提出了一种基于确定性二值的测量矩阵的测量算法,该算法构造过程简单快速。测量矩阵中的每一行代表一次测量过程,每次测量相互独立。由于测量矩阵的稀疏性特点,对应矩阵中的非0元素的节点参与每次测量,参与同一个测量的节点的数据被融合成一个数据包传递到汇聚节点。当汇聚节点收集到所有的测量值的时候,可以准确地恢复原始数据。其次,针对传感网中时延长和能量不均衡的问题,本文提出了一种基于稀疏随机测量矩阵的融合算法。该算法在保证恢复原始数据的前提下,将测量过程分解为多个融合树,单个融合树是由部分节点参与。在传递数据的过程中,本文提出了一种减少时延的传输策略。另外,由于所设计矩阵的随机性和稀疏性的特点,节点的能量消耗能够达到均衡,从而可以延长网络寿命。最后,本文系统分析了所提出的算法,并对算法进行实验验证。实验结果表明:对于能在频域上稀疏表示的信号,采用确定性二值矩阵能够有效地减少网络能量消耗,而基于稀疏矩阵的低延迟且能量均衡的数据融合算法可以减少通信时延,均衡通信消耗。
[Abstract]:With the arrival of the era of big data and artificial intelligence, the Internet of things has become the focus of attention again. Internet of things not only brings great convenience to people's normal life, but also in the medical health, environmental monitoring. Wireless sensor network, as one of the core technologies of the Internet of things, is a large number of randomly distributed sensor nodes responsible for sensing and collection. Processing and analyzing the data information to get effective information and transfer to the user. However, because of the limited power of nodes, storage space is limited, so how to integrate these large amounts of data. And in the process of transmission to reduce energy consumption, extended network life has become particularly important. In addition, in some real-time requirements of the scene. How to transfer the collected information to the convergence node with the shortest delay is also one of the main research contents. In wireless sensor networks, the traditional data fusion mainly aims to reduce data redundancy. However, the communication consumption between the data packet and the sensor can not be greatly reduced. Therefore, based on the spatio-temporal correlation of the collected data and the characteristics of the sensor network, the sparse measurement matrix is designed. A data fusion method based on sparse compression sensing is proposed to reduce the amount of data packets and energy consumption. The main work of this paper is as follows: first. In view of the traditional collection method of compression perception, this paper presents a measurement algorithm based on deterministic binary measurement matrix. The algorithm is simple and fast. Each line in the measurement matrix represents a measurement process. Each measurement is independent of each other. Because of the sparsity of the measurement matrix, the nodes corresponding to the non-zero elements in the matrix participate in each measurement. The data of the node involved in the same measurement is fused into a packet and delivered to the sink node. When the sink node collects all the measured values, the original data can be recovered accurately. Secondly. In order to solve the problem of time prolongation and energy imbalance in sensor networks, a fusion algorithm based on sparse random measurement matrix is proposed, which can restore the original data. The measurement process is decomposed into multiple fusion trees, and a single fusion tree is joined by some nodes. In the process of data transfer, this paper proposes a transmission strategy to reduce delay. Because of the randomness and sparsity of the designed matrix, the energy consumption of the nodes can reach equilibrium, which can prolong the network life. Finally, the proposed algorithm is systematically analyzed in this paper. The experimental results show that the deterministic binary matrix can effectively reduce the network energy consumption for the signals which can be represented sparsely in frequency domain. The low delay and energy equalization data fusion algorithm based on sparse matrix can reduce the communication delay and equalize the communication consumption.
【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;TP202

【参考文献】