原子钟数据消噪方法比较研究  

A Comparative Study of Noise Reduction Method for Atomic Clock Data

作　　者：李滚[1] 袁海波[1] 

机构地区：[1]中国科学院国家授时中心

出　　处：《测试技术学报》2006年第6期512-516,共5页Journal of Test and Measurement Technology

基　　金：国家自然科学基金资助项目(60272067);中国科学院知识创新工程重要方向资助项目(KJCX2-SW-T07)

摘　　要：主要研究了小波方法、V ondrak、K a lm an滤波等7种方法在时间尺度计算中原子钟数据消噪的应用,和未消噪的序列相比较,其RM S分别从1.26 ns提高到0.66 ns,0.77 ns,0.80 ns;利用这几种不同的方法消噪后重新计算了时间尺度,其修正的A llan方差从7.29D-15提高到5.18D-15,5.63D-15,5.67D-15(=τ1d),证明小波消噪方法获得的时间尺度更加稳定,其他几种滤波方法也能达到比较好的消噪效果.This paper mainly investigates the application of seven methods （the wavelet, Vondrak filter, Kalman filter etc. ） for noise reduction of atomic clock data in time scale computation. Comparing with the raw data, their RMS is improved from 1.26ns to 0. 66ns, 0. 77ns and 0. 80ns respectively. Secondly, one more stable time scale is presented by using the different method of noise reduction. For, the modified Allan deviation is improved from 7. 29D-15 to 5. 18D-15, 5.63D-15 and 5.67D-15. The results indicate that the wavelet is one useful method for atomic clock data noise reduction in time scale computation and the other methods can also obtain the effective results in noise reduction.

关 键 词：时间频率 时间尺度 原子钟数据 小波 消噪 

分 类 号：TM935[电气工程—电力电子与电力传动]