基于卡尔曼滤波估计的一致性时钟同步算法

发布时间：2018-05-01 04:16

本文选题：无线传感器网络 + 时钟同步　；参考：《计算机应用》2017年08期

【摘要】：针对无线传感器网络(WSN)的众多应用都需要依赖时钟同步的节点协同完成,而由于节点的晶体震荡器受自身以及外界环境的影响,使得节点时钟偏斜和时钟偏移两个参数发生变化导致时钟不同步问题,提出了基于分布式卡尔曼滤波估计的一致性补偿时钟同步算法DKFCC。该算法首先利用双向信息交换机制以及分布式卡尔曼滤波实现时钟偏斜和偏移两个参数的最优估计,然后基于时钟参数的最优估计值采用一致性补偿方法实现节点的时钟同步。实验结果表明:在100个节点随机部署的WSN中,采用虚拟全局一致性方式的DKFCC同步算法比异步一致性同步(AC)算法的同步均方根误差(SRAMSE)值降低了约95%,具有较高的同步精度;同时,所提出算法从时钟参数层面实现同步,无需频繁地进行时钟同步操作,相比AC算法更节能。
[Abstract]:Many applications of WSNs depend on the synchronization of clock nodes, but the crystal oscillators of the nodes are affected by themselves and the external environment. Two parameters of node clock skew and clock offset are changed and the clock is out of sync. A consistency compensated clock synchronization algorithm DKFCC based on distributed Kalman filter estimation is proposed. Firstly, the bidirectional information exchange mechanism and distributed Kalman filter are used to realize the optimal estimation of clock skew and offset, then the consistency compensation method is used to realize the clock synchronization of the node based on the optimal estimation of clock parameters. The experimental results show that in 100 nodes randomly deployed WSN, the synchronous root mean square error (SRAMSE) value of the DKFCC synchronous algorithm using virtual global consistency is about 95% lower than that of the asynchronous consistency synchronous algorithm, and the synchronization accuracy is higher. The proposed algorithm realizes synchronization from the clock parameter level and does not need frequent clock synchronization operation, which is more energy efficient than AC algorithm.
【作者单位】：重庆邮电大学自动化学院;
【基金】：国家自然科学基金资助项目(51605065) 重庆市教委科学技术研究项目(KJ1600402)~~
【分类号】：TN929.5;TP212.9

【参考文献】