自适应卡尔曼滤波的精密时钟同步方法  

Precision Clock Synchronization Method Based on Adaptive Kalman Filter

作　　者：张巍然 侯艳红[1] Zhang Weiran;Hou Yanhong(School of Computing,Shaanxi Institute of Technology,Xi’an 710300,China)

机构地区：[1]陕西国防工业职业技术学院计算机学院,西安710300

出　　处：《单片机与嵌入式系统应用》2023年第5期59-64,共6页Microcontrollers & Embedded Systems

基　　金：2021年陕西省重点研发计划项目—基于IEEE1588精密时钟同步技术研究(2021GY-110);2019年陕西国防工业职业技术学院科研计划项目—基于北斗的小型化高精度网络时钟同步终端设计(Gfy19-02)。

摘　　要：为减少因网络延时抖动引起的时钟波动,减小因频率不一致和漂移造成时钟偏差,建立了频率偏差模型和基于卡尔曼滤波的同步系统;为避免卡尔曼滤波器靠经验估计测量噪声的问题,并减少Sage-Husa算法因测量周期不一致造成的时钟波动,修正了Sage-Husa算法权值系数和测量噪声计算公式;为减小野值对强跟踪算法的影响和提高强跟踪算法的准确率,设计了野值检测和处理算法并修正了强跟踪算法公式。最后,实验验证了该方案的鲁棒性、跟踪性、时钟精度和稳定性。In order to reduce clock fluctuation caused by network delay jitter and clock deviation caused by frequency inconsistency and drift,a frequency deviation model and a synchronization system based on Kalman filter are established.In order to avoid the problem of empirical estimation of measurement noise by Kalman filter and reduce the clock fluctuation caused by inconsistent measurement period of Sage-Husa algorithm,the weight coefficient and measurement noise calculation formula of Sage-Husa algorithm are modified.In order to reduce the influence of outlier on strong tracking algorithm and improve the accuracy of strong tracking algorithm,an outlier detection and processing algorithm is designed and the formula of strong tracking algorithm is modified.The experiment results show that the scheme is effective in the aspects of robustness,tracking,clock accuracy and stability.

关 键 词：IEEE1588 卡尔曼滤波 延时抖动 频率偏差 

分 类 号：TN7[电子电信—电路与系统]