煤矿井下无线传感器网络时间同步算法研究

发布时间：2018-03-25 19:20

本文选题：无线传感器网络　切入点：井下同步　出处：《内蒙古科技大学》2015年硕士论文

【摘要】：随着社会和科技的全面发展，煤炭工业逐渐成为我国的支撑性产业。无线通信技术的不断成熟和传感技术的日益发展，将无线传感器网络推动到国民生产和日常生活的各个领域中。针对我国煤矿分布较为广泛和开采难度较大的特殊现状，无线传感器网络的到来将会对其产生划时代的意义。煤矿井下安监系统包括多种组成部分，时间同步作为其中基础组件技术之一，弥补了GPS及其他基础时间服务信号无法为井下工作提供时间信息的约束，，为井下人员的安全和国家资源的保护提供可靠保障。传统时间同步算法虽然可以在较短时间内达到较高的同步精度，但却存在着同步误差随跳距积累、能耗大、容错性低等问题，不适用于煤矿井下的特殊应用。因此，在对无线传感器网络、井下开采和分布环境进行分析，以及对时间同步算法研究的基础上，提出了将煤矿井下无线传感器网络时间同步算法研究与实现作为本文的研究方向。首先，本文针对传统时间同步算法双向交换数据包以获得较高同步精度而导致过大能耗的现象，结合煤矿井下环境的特殊结构，在分析双向成对同步机制与单向广播同步机制的基础上，利用广播传输特性将两种算法有效结合起来，提出一种基于簇状结构的低功耗双向时间同步算法。并通过实验验证了算法的正确性，确定了该算法解决传统算法运用于煤矿井下能耗过大问题的可行性。其次，本文针对无线传感器网络对节点在精度、强壮性等方面的要求，结合传统时间同步算法在煤矿井下应用过程中对差错冗余度较低的问题，提出一种面向煤矿井下的无线传感器网络容错性时间同步算法。该算法结合井下特殊结构，在研究FTSP算法的基础上，运用动态选举根节点以及避免重复发送策略使得新加入节点得以快速收敛；结合无线传感器网络具有的拓扑动态变化性、以及通信过程中信道半双工性等特点，通过概率统计学中的残差分析理论，对线性回归算法进行改进和提升，从而利用差错判断来剔除阈值范围以外的数据点，减少其对拟合曲线的影响，抑制由异常数据点所引起的同步误差大幅度跳变，并在节点上对时钟漂移和偏移引起的误差进行补偿，对算法可靠性和容错性方面进行相应提升。最后，在由11个GainZ节点组成的时间同步测试平台上通过TinyOS操作系统进行实际的运行测试，并在同步精度、能耗、网络拓扑变化对同步误差影响的三个指标上与传统算法进行对比。实验结果表明：所提出的面向煤矿井下时间同步算法能够使用较低的能耗达到与传统算法相近的同步精度与收敛速度，且具有可适应无线传感器网络拓扑动态变化的健壮性，其精度和能耗均满足井下同步所需。
[Abstract]:With the overall development of society and science and technology, the coal industry has gradually become the supporting industry of our country. Wireless sensor network (WSN) will be promoted into the fields of national production and daily life. In view of the wide distribution of coal mines and the difficulty of mining in China, the arrival of WSN will have epoch-making significance. The underground safety monitoring system of coal mine includes many components. As one of the basic component technologies, time synchronization makes up for the constraint that GPS and other basic time service signals can not provide time information for underground work. The traditional time synchronization algorithm can achieve high synchronization precision in a short time, but the synchronization error accumulates with the jump distance, and the energy consumption is large, while the traditional time synchronization algorithm can provide reliable guarantee for the safety of underground personnel and the protection of national resources, although the traditional time synchronization algorithm can achieve a higher synchronization accuracy in a relatively short time. Therefore, based on the analysis of wireless sensor network, underground mining and distribution environment, as well as the research of time synchronization algorithm, In this paper, the research and implementation of time synchronization algorithm for underground wireless sensor networks in coal mines are proposed. Firstly, aiming at the phenomenon that the traditional time synchronization algorithm switches data packets bidirectionally in order to obtain high synchronization precision and lead to excessive energy consumption, this paper combines the special structure of underground coal mine environment. Based on the analysis of the bidirectional pairwise synchronization mechanism and the one-way broadcast synchronization mechanism, the two algorithms are effectively combined by using the broadcast transmission characteristics. A low power bidirectional time synchronization algorithm based on cluster structure is proposed, and the correctness of the algorithm is verified by experiments, and the feasibility of applying this algorithm to the problem of excessive energy consumption in coal mines is determined. Secondly, aiming at the requirements of the wireless sensor network in the aspects of accuracy and robustness, combined with the traditional time synchronization algorithm, the problem of low error redundancy in underground coal mine is discussed. This paper presents a fault tolerant time synchronization algorithm for underground wireless sensor networks in coal mines, which combines with the special structure of coal mine, and based on the research of FTSP algorithm. Dynamic selection of root nodes and avoidance of repeat transmission strategy are used to make the new join nodes converge rapidly, combined with the dynamic topology of wireless sensor networks, and the channel half-duplex in the communication process, and so on. Through the residual analysis theory in probability and statistics, the linear regression algorithm is improved and promoted, and the error judgment is used to eliminate the data points outside the threshold range and reduce its influence on the fitting curve. The synchronization error caused by abnormal data points is restrained and the errors caused by clock drift and offset are compensated on the node. The reliability and fault tolerance of the algorithm are improved accordingly. Finally, on the time synchronization test platform composed of 11 GainZ nodes, the actual running test is carried out through the TinyOS operating system, and the synchronization accuracy and energy consumption are obtained. The effect of network topology on synchronization error is compared with the traditional algorithm. The experimental results show that the proposed time synchronization algorithm for coal mines can achieve similar synchronization accuracy and convergence speed with lower energy consumption, and is robust to the dynamic changes of wireless sensor network topology. Its precision and energy consumption meet the need of underground synchronization.
【学位授予单位】：内蒙古科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TD65;TN929.5;TP212.9

【参考文献】