基于压缩网络编码的高能效WSN数据汇集方法研究

发布时间：2018-07-15 10:04

【摘要】：作为21世纪十大新兴技术之一，无线传感器网络（Wireless Sensor Network，WSN）通过对传感技术、微电子工艺，以及无线网络的融合，实现了物理世界与信息网络世界的连接，极大的扩展了人类认识和改造世界的能力。但在当前的WSN应用中，所部署的传感器节点大多无法支持能量捕获，仅采用能量有限的电池供电且很难得到补充，同时网络中各节点往往处于能耗不均衡状态，这些现状严重制约了WSN的运行寿命以及网络的扩容。在低成本电池技术、新型能源技术等取得突破之前，能否有效提高WSN的能量效率是保证其广泛、可靠部署与应用的关键，其根本问题是如何减小WSN节点通信模块所产生的通信能耗。本文针对上述问题，主要从如何设计高效的数据汇集方法的角度分析并解决问题。主要研究思路为：通过数据压缩减少待汇集的数据总量，通过优化传输机制提高数据汇集的效率。与此同时，压缩感知和网络编码作为新兴的信息技术，分别从数据压缩和优化传输的角度为高能效数据汇集方法的设计提供了新思路。但是，这两种技术单独应用于WSN数据汇集仍存在较大的缺陷与不足。本文通过深入分析网络编码与压缩感知的内在联系，利用WSN各节点感知数据的相关性以及无线传输的广播特性，创新性的将网络编码与压缩感知结合在一起，从联合信源网络编码的角度提出了新型的高能效WSN数据汇集方案，称之为压缩网络编码方案（Compressed Network Coding，CNC）。本文首先对WSN中应用压缩感知与网络编码的可行性进行了论证，建立了CNC数据汇集方案的系统框架；基于该系统框架，以分布式数据传输效率提升和网络整体能耗降低为研究目标，给出了CNC的一种具体实现方法——OMP-RL-CNC（Orthogonal Matching Pursuit Random Linear Compressed Network Coding）；进而，，本文通过分析各个节点能耗情况，以增加一定的集中式调度操作为代价，给出了两种在降低整体能耗的同时能实现能耗均衡的CNC实现方法——E-OMP-RL-CNC（Enhanced OMP-RL-CNC）、E2-OMP-RL-CNC（two-times Enhanced OMP-RL-CNC），这两种方法综合解决了能耗不均所致的关键节点过早失效的问题，有效地延长了网络运行寿命；最终，本文搭建了实际硬件平台，对所提出的CNC方案进行了实现与验证。本文的主要工作可归纳为如下几个方面：（1）建立基于CNC的WSN数据汇集方案的系统框架。通过分析WSN数据汇集的研究现状，分析了现有数据汇集方法的优缺点；通过分析广义蝶形网络的编码增益、传感器节点感知数据的相关性，论证了基于CNC进行WSN数据汇集的可行性。在此基础上，提出了CNC数据汇集方案的系统框架，其中包括网络结构、节点功能、节点工作模式等。（2）给出了一种以分布式高效数据传输和网络整体能耗降低为目标的CNC数据汇集方案的具体实现方法——OMP-RL-CNC。基于CNC数据汇集方案的系统框架，设计了随机线性实现的网络编码向量，使得网络编码的全局编码矩阵满足RIP（Restricted Isometry Property）特性要求。并且，将面向压缩感知重构的OMP解码算法结合进来，有效的解决了基于有限域的随机线性网络编码（Finite Field based Random Linear Network Coding，FFRL-NC）所存在的“全有全无”问题以及对有限域依赖所导致的重构效率较低的问题。该方法在待汇集数据量相同的情况下，相对已有方法在全网范围内降低了数据包传输次数，进而降低了整网能耗。（3）通过改进OMP-RL-CNC，进一步给出了两种能实现能耗均衡的CNC实现方法——E-OMP-RL-CNC、E2-OMP-RL-CNC。这两种方法以最大化网络寿命为目标，以增加一定的集中式调度操作为代价。通过与已有数据汇集方法中具有代表性的汇聚树路由协议（Collection Tree Protocol，CTP）、传统网络编码（Network Coding，NC）数据传输方案进行仿真对比分析，验证了E-OMP-RL-CNC、E2-OMP-RL-CNC可综合解决WSN数据汇集的能效问题：不仅能有效节约整体网络能耗、还能实现能耗均衡。（4）通过硬件测试平台对所提方法进行了实现与验证。搭建了室外无线传感器网络环境，并在该环境下实现了本文提出的三种CNC数据汇集方法。基于实测数据分析了网络中各节点的能耗情况，实验验证了所提方法解决WSN能效问题的有效性。
[Abstract]:As one of the top ten emerging technologies in the 21st century , Wireless Sensor Network ( WSN ) has realized the connection between physical world and information network world through the fusion of sensing technology , microelectronic process and wireless network .

In order to improve the energy efficiency of WSN , it is the key to ensure its extensive , reliable deployment and application before the breakthrough of low cost battery technology and new energy technology . The fundamental problem is how to reduce the communication energy consumption of WSN node communication module .

In this paper , we deeply analyze the relationship between network coding and compression perception , use the relativity of the sensor data of WSN and the broadcast characteristics of wireless transmission , combine the network coding with the compression perception , propose a new high energy efficiency WSN data collection scheme from the angle of joint source network coding , which is called the compressed network coding scheme ( CNC ) . Firstly , the feasibility of applying compression sensing and network coding in WSN is demonstrated , and the system framework of CNC data collection scheme is established .
Based on the framework of the system , the distributed data transmission efficiency and the overall energy consumption of the network are reduced to the research target , and a concrete realization method of CNC _ OMP - RL - CNC ( Orthogonal Matching Random Linear Compressed Network Coding ) is given .
Furthermore , by analyzing the energy consumption of each node , this paper presents two CNC realizing methods _ E - OMP - RL - CNC ( Enhanced OMP - RL - CNC ) , E2 - OMP - RL - CNC ( two - times Enhanced OMP - RL - CNC ) , which can reduce the whole energy consumption while reducing the whole energy consumption .
Finally , the practical hardware platform is set up , and the CNC solution is implemented and verified .

The main work of this paper can be summarized as follows :

( 1 ) establishing a system framework of a CNC - based data collection scheme of the WSN , analyzing the research status of the WSN data collection , and analyzing the advantages and disadvantages of the existing data collection method ;
Based on the analysis of the correlation between coding gain and sensor node perception data of generalized butterfly network , the feasibility of data collection based on CNC is demonstrated . On this basis , a system framework of CNC data collection scheme is proposed , including network structure , node function , node operation mode , etc .

( 2 ) A concrete implementation method _ OMP - RL - CNC based on distributed high - efficiency data transmission and reduction of total energy consumption of network is presented . A random linear implemented network coding vector is designed based on CNC data aggregation scheme .

( 3 ) By improving OMP - RL - CNC , two kinds of CNC realizing methods _ E - OMP - RL - CNC , E2 - OMP - RL - CNC , which can realize energy consumption equalization , are given .

( 4 ) The method is realized and verified by the hardware test platform . The outdoor wireless sensor network environment is built , and the three kinds of CNC data collection methods presented in this paper are realized under the environment . Based on the measured data , the energy consumption of each node in the network is analyzed , and the proposed method is verified to solve the problem of energy efficiency of WSN .
【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9

【参考文献】