计算机应用 ›› 2015, Vol. 35 ›› Issue (10): 2852-2857.DOI: 10.11772/j.issn.1001-9081.2015.10.2852

• 网络与通信 • 上一篇    下一篇

基于控制角度的无线传感器网络时钟同步优化算法

曾培, 陈伟   

  1. 重庆邮电大学 自动化学院, 重庆 400065
  • 收稿日期:2015-04-13 修回日期:2015-05-20 出版日期:2015-10-10 发布日期:2015-10-14
  • 通讯作者: 曾培(1990-),女,河南驻马店人,硕士研究生,主要研究方向:网络控制、无线传感器网络时钟同步,498990356@qq.com
  • 作者简介:陈伟(1990-),男,河南新乡人,硕士研究生,主要研究方向:网络控制、工业物联网网络安全。

Improved algorithm of time synchronization based on control perspective of wireless sensor network

ZENG Pei, CHEN Wei   

  1. School of Automation, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
  • Received:2015-04-13 Revised:2015-05-20 Online:2015-10-10 Published:2015-10-14

摘要: 针对无线传感器网络(WSN)时间同步过程中易受干扰,易发生通信延迟所造成的同步精度不高、收敛速度不快的问题,从控制的角度提出一种时钟同步优化算法。该算法首先建立时钟同步状态模型,然后通过现代控制理论的思想,引入中心控制策略,建立基于控制的时间同步状态模型。该控制策略是通过全局的时钟状态信息进行设计,在卡尔曼滤波最优估计前提下,使控制满足使性能指标函数最小的条件下,得到最优控制。仿真结果表明,所提出的时钟同步优化算法和无线传感器时钟同步协议(TPSN)相比,从第6步时钟同步开始,前者的同步误差逐渐比后者的同步误差小;在实现同一较高精度的同步需求时,前者需要的同步次数是后者所用的同步次数的20%左右;由时钟同步误差收敛均值的方差对比值显示,前者比后者的同步误差均方差小了两个数量级,因此所提出的时钟同步优化算法比时钟同步协议同步精度高、收敛速度快、网络通信负荷低。

关键词: 无线传感器网络, 时钟同步, 状态模型, 控制理论, 中心控制

Abstract: Focusing on the issue of low synchronization accuracy and slow convergence speed caused by the process of time synchronization which is susceptible to disturbance and communication delay in Wireless Sensor Network (WSN), an improved time synchronization algorithm was proposed from control perspective. Firstly, the clock synchronization state model was established, then, through the thought of modern control theory, a centralized control strategy was introduced, and the time synchronization state model based on control strategy was established. The centralized control scheme was designed based on global clock status information, and optimal control was obtained under the condition of minimizing the performance index function and the optimal estimation of Kalman filtering. Comparison simulation were carried out on the proposed clock synchronization optimization algorithm and Timing-sync Protocol for Sensor Networks (TPSN). The results showed that, from the 6th step of the clock synchronization, the synchronization error of the former was gradually smaller than that of the latter;when achieving the same relative high synchronization precision performances, the former required steps is about twenty percent of the latter;the variance of synchronization error convergence of the former was two orders lower than that of the latter. The results prove that the proposed algorithm of time synchronization has higher synchronization accuracy, faster convergence speed and lower communication load than TPSN.

Key words: Wireless Sensor Network (WSN), time synchronization, state model, control theory, centralized control

中图分类号: