【作者】 李敏；

【导师】 郑国强；

【作者基本信息】 河南科技大学 ， 检测技术与自动化装置， 2014， 硕士

【摘要】 随着微电子机械系统、无线通信和低功耗嵌入式技术的快速发展，具有低成本、低功耗和多功能的无线传感器网络(WSN:Wireless Sensor Network)得到了快速发展，并在国防军事、灾害预测、交通管理、医疗卫生、环境科学和日常生活等领域展现出巨大的应用前景。然而无线传感器网络的实际应用却面临着诸多挑战，如时间同步问题。在无线传感器网络中，大量的应用都要求传感器节点时间保持一致。因此，如何在能量有限且恶劣的无线信道条件和时变的拓扑环境中，为无线传感器网络提供满足应用需求的时间同步算法具有重要的意义。现有的时间同步算法一般只是针对普通无线传感器网络同步精度和同步能耗的需求，无法适用于数据量大、对网络性能要求较高的大规模无线传感器网络。因此为了解决大规模无线传感器网络同步精度低、可扩展性差的问题，必须制定一个兼顾高同步精度和良好扩展性的时间同步算法。本文摒弃了典型时间同步算法的思想，采用新型的时间同步机制，设计一种基于脉冲耦合和分布式扩散相结合的时间同步协议。首先，介绍了无线传感器网络时间同步技术的国内外研究动态和相关理论知识，全面分析和总结了现有时间同步协议的优缺点，为下文的同步算法设计做好理论基础。其次，借鉴生物学中的萤火虫同步思想，研究应用于无线传感器网络的萤火虫同步算法，重点阐述了M&S模型和RFA算法，在此基础上通过引入后相应窗口和分布式序列对RFA算法进行了优化改进，并且进行了仿真实验。最后，设计一种基于脉冲耦合和分布式扩散相结合的协作时间同步协议PCODD，将同步分为时钟节拍同步和时间同步两个阶段。时钟节拍同步阶段，通过脉冲信号耦合来改变节点相位实现节拍同步；时间同步阶段，基于双向信息交换模式，把主节点域的平均时间扩散有限的跳数，采用互扩散使节点时间近似同步到网络节点的平均时间上实现时间同步。优化设计完成后，将N-RFA算法和PCODD算法在Matlab仿真平台上进行了仿真实现。仿真结果表明，PCODD算法在不同的网络规模中均能获得接近百微秒的同步精度，并且具有较短的收敛时间、低同步误差、低数据包交换的显著优势，降低了无线传感器节点进行时间同步的算法能耗，提高了网络的可扩展性。更多还原

【Abstract】 With the help of technological advances in micro-electro-mechanical systems(MEMS), wireless communications and embedded technology, low-cost, low-power,and multi-functional wireless sensor networks (WSN) has been developed. And WSNshows broad application prospects in the fields of military defense, disaster prediction,traffic management, health care, environmental science and everyday life. However,the practical applications of WSN are faced with many challenges, such as the problemof time synchronization. Large numbers of applications of wireless sensor networksrequire each sensor node in the network to keep the consistency of the time. How todesign time synchronization algorithm for WSN in the conditions of limited energy,harsh wireless channel and topological time-varying, to meet the applicationrequirements, is the vital significance for WSN.The existing time synchronization algorithms aim at general wireless sensornetworks, are not available to WSN with large data and high network performancerequirement. Therefore, in order to ensure large WSN performing normally, trading-offbetween computation quality and good scalability must be make for synchronizationalgorithms running on wireless sensors. In this paper, a new time synchronizationprotocol was proposed based on pulse-coupled oscillators and distributed diffusioninstead of the typical algorithm thought.Firstly, we analyzed the research dynamic of time synchronization technology athome and abroad, and then compared some existing algorithms of timesynchronization, and which is theoretical basis for the synchronization algorithmdesign. At the same time, the advantages and disadvantages of existing algorithmswere studied in detail.Secondly, this makes use of the idea of firefly synchronization in biology,introduced the idea into wireless sensor network. We focus on the M&S model andRFA algorithm. Based on RFA algorithm, a late sensitivity window and a distributed sequence number mechanism are used in multi-hop network. And besides, weimplement and simulate the new protocol N-RFA.Thirdly, a novel cooperative time synchronization protocol based on pulse-coupled oscillators and distributed diffusion (PCODD) is proposed. The protocoldivides the synchronization into two phases: clock tick synchronization phase and timesynchronization phase. In the first phase, the pulse-coupled oscillator’s phases arechanged by the pulse coupled signal. The tick synchronization is achieved bydistributed diffusion. In second phase, the average time of master-node domain isspread to a limited number of hops based on two-way message exchange mechanism.It adopts the method of mutual diffusion to finish the approximately synchronizationbetween the node time and the average time of all nodes. After the optimal design iscompleted, N-RFA and PCODD algorithms are achieved on Matlab SimulationPlatform and result of that is analyzed. The experimental results show that the PCODDalgorithms can synchronize the network and obtain the synchronization precision ofnearly hundred microseconds for different scales networks. And besides, PCODDalgorithms can synchronize the network quickly with good precision convergencespeed, the energy consumption of wireless sensor nodes effectively and improve thenetwork scalability.更多还原