线性加权组合Kalman滤波在钟差预报中的应用

Clock Bias Prediction Based on Linear Weighted Combination Kalman Filter

建立原子钟运行模型,实时预报钟差,在时频工作以及卫星导航定位中有着非常重要的意义.目前,Kalman滤波是一类重要的钟差预测模型.为了充分利用各种Kalman滤波模型的特点,提出了线性加权组合Kalman滤波模型.因此详细讨论了如何对每种Kalman滤波模型赋权,并且给出了两种简单而又实用的赋权方法.最后,以IGS(International GNSS Service)精密铷钟数据为例,运用该模型计算了预报误差.结果表明,线性加权组合Kalman滤波模型有利于提高钟差预报的准确性和可靠性.

Establishing the operation model of atomic clock to predict the clock bias and errors plays an important role in the time and frequency community.Currently,Kalman filter（KF）is one kind of the most important prediction models.In order to make full use of serval KF models＇ characteristics and combine these models＇ results,the linear weighted combination KF（LWCKF）model is put forward.The essence of LWCKF model is the weighted average of all local KF models.Therefore,how to determine the weight of every local KF is discussed in detail.As a simple and effective method,every local KF model is directly given with the same weight regardless of the differences among them.On the contrary, variational weight is built based on relation between the residual error and precision matrix.Finally,to demonstrate the efficiency of LWCKF,the rubidium atomic clock data downloaded from IGS（International GNSS Service）website are taken as an example.The results show that the accuracy and reliability of the atomic clock prediction are improved with the LWCKF model.