摘要
为提高分布式控制系统的时钟同步精度,分析了影响时钟同步精度的主要因素,对时钟同步过程建立了含控制量的数学模型,提出了基于改进卡尔曼滤波的实时以太网时钟同步优化算法。针对传统时钟同步算法中驱动层加盖的时间戳精度低、链路延时抖动严重等不足,通过补偿软件时间戳精度,使驱动层获取时间戳更接近物理层获取的时间戳;采用对卡尔曼增益分段的方法改进卡尔曼滤波算法,并使用改进的卡尔曼滤波算法修正时钟偏差值,使时钟偏差测量过程更趋平稳。实验发现,当主从时钟经路由器相连,调整合适卡尔曼滤波参数并补偿软件时间戳精度后,时钟同步精度可达14μs。
In order to improve the clock synchronization accuracy of distributed control system, the main factors affecting the clock synchronization accuracy were analyzed, and the mathematical model containing control quantity was established for the clock synchronization process. Besides, a real-time ethernet clock synchronization optimization algorithm based on Kalman filtering was proposed. The timestamp stamped by the driver layer in the traditional clock synchronization algorithm has low precision, and link delay jitter is serious. For the problems, through compensating the precision of the software timestamp, the timestamp obtained by the driver layer was closer to that obtained by the physical layer. And the Kalman filter algorithm was improved based on piecewise Kalman gain, and the improved Kalman filtering algorithm was used to correct the clock deviation so that the clock deviation measurement process became more stable. It was found that the clock synchronization accuracycould reach 14 μs when the master and slave clocks were connected by the router, the appropriate Kalman filtering parameters were adjusted and the software timestamp precision was compensated.
作者
史仲渊
张华
王旭浩
SHI Zhongyuan;ZHANG Hua;WANG Xuhao(Faculty of Mechanical Engineering & Automation,Zhejiang Sci-Tech University,Hangzhou 310018,China;College of Energy Engineering,Zhejiang University,Hangzhou 310007,China;Zhejiang Dun’an Artificial Environment Co.,Ltd.,Hangzhou 310053,China)
出处
《浙江理工大学学报（自然科学版）》
2019年第5期650-656,共7页
Journal of Zhejiang Sci-Tech University(Natural Sciences)
基金
国家自然科学基金项目(U1609205,51675488,51307151)
浙江省自然科学基金项目(LY18E070006,LY18E050016)
