用卡尔曼滤波方法进行氢钟钟差预报的方法与结果

Hydrogen Clock Bias Prediction Method and Its Result Based on Kalman Filter Model

钟差预报是实现时间同步的重要基础，本文以原子钟时间模型为理论基础，研究主动型氢原子钟与协调世界时UTC（k）的钟差预报方案。线性最小二乘模型与卡尔曼滤波模型均可进行氢钟参数估计和钟差预报。线性最小二乘模型只能估计钟的确定性参数，不能及时适应氢钟信号的变化；卡尔曼滤波模型不仅可以滤除测量噪声等部分噪声，减小频差和频漂波动，提高信号的短期稳定度，及时适应氢钟信号的变化，准确预报时差值、频差值和频率漂移值，而且有助于提高时间溯源精度和稳定度。

Clock bias prediction is the foundation for time synchronization. Based on the atomic clock time model, the clock bias prediction schemes of Hydrogen Clock SOHM - 4 and UTC （ k ） are studied in this paper. Both linear least square model and Kalman filter model can estimate Hydrogen clockg parameters and predict clock bias. However, the former model can only esti- mate the definite parameters of clock, but not adjust to the change of clock signal in time. The latter one can filter some noises like measuring noise, reduce fluctuations of frequency difference and drift, and improve the short - term stability of clock signals. In addition, it can adjust to the change of clock signal in time, and accurately predict values of clock bias, frequency difference and frequency drift. It comes to the conclusion that Kalman filter model is helpful in improving the accuracy and stability of time tracing.