摘要
本文针对物联网中时变的时钟参数,运用网络化控制理论观点,通过对时钟状态建模的本质分析,区别于"相对时钟建模",提出了全分布规模化时钟状态追踪卡尔曼滤波(Kalman filtering).考虑量测的丢失,则扩展为追踪时钟参数的修正Kalman filtering算法.我们提出了以BMU(Basic measurement unit)构建新的MMSE(Minimum mean square error)等价变换下的能观测性状态解耦量测模型,新的量测模型能够实现MMSE量测规模化扩展,且理论上分析了时钟同步的条件和计算了统计时钟同步误差的相应上界,并且在时钟同步精度与潜在的通信网络质量间作出了量化均衡.
In this paper, aiming at the time-varying clock parameters in the internet of things, the fully distributed scaled Kalman filtering for clock state tracking, which is different from the "relative clock modeling", has been proposed by using the point of view of the networked control theory through the essential analysis of the clock state modeling.Considering the loss of measurement, it can be extended to the modified Kalman filtering algorithm for tracking the clock parameters. We have proposed an observable state decoupling measurement model under a new MMSE(minimum mean square error) equivalent transformation based on BMU(basic measurement unit). The new measurement model is able to realize the scaled expansion of MMSE measurement, and theoretically analyze the conditions of clock synchronization and calculate the corresponding upper bound of statistical clock synchronization error, and quantify the balance between synchronization accuracy of clock and potential quality of the communication network.
作者
王頲
徐小权
唐晓铭
黄庆卿
李永福
WANG Ting;XU Xiao-Quan;TANG Xiao-Ming;HUANG Qing-Qing;LI Yong-Fu(Key Laboratory of Industrial Internet of Things and Net-worked Control,Ministry of Education,Chongqing 400065;School of Advanced and Manufacturing Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065;School of Automation,Chongqing University of Posts and Telecommunications,Chongqing 400065)
出处
《自动化学报》
EI
CAS
CSCD
北大核心
2021年第7期1720-1738,共19页
Acta Automatica Sinica
基金
国家自然科学基金(61972061,51605065,61403055,51705059)
重庆市教委科学技术研究项目(KJZD-K201900604)
重庆市基础研究与前沿探索(重庆市自然科学基金)项目(2017jcyjAX0453,cstc2018jcy jAX0139,cstc2018jcyjAX0691)
重庆市教委项目(KJ1600402)资助。
关键词
工业物联网
精确时钟同步
卡尔曼滤波
量测丢包
Industrial internet of things
precise clock synchronization
Kalman filtering
measurements loss
