对流层时间序列的 Kalman 滤波组合方法

摘要

as the final troposphere products provided by IGS were delayed for at least one week,even though the ultro-rap-id products also delayed for three hours,they all affected the real-time meteorological application,and also it’s rare to see the research papers on sequential or real-time combination for troposphere time series,most were focused on the experimental pro-jects in Europe.Thus,on this paper the method of high precision and sequential combination for short-term troposphere time series provided by several analysis centers was discussed in detail,with the technique of robust Kalman filter and least-squares variances component estimation,a real-time or lagging sequential combination of GPS or VLBI short-term time series can be manipulated.First the bias and weight factor for each analysis center solution were estimated,and then the combined solution and its standard deviation can be obtained by Kalman filtering,the average accuracy can approach 0.85mm.%由于目前 IGS 提供的对流层最终产品滞后约一星期到四星期，即使是超快速产品也滞后约3小时，这严重影响了目前实时气象的应用与需求，而且对流层产品的连续和实时组合的相关研究较少，大多集中在欧洲的试点项目。基于此，本文主要研究了一种对流层时间序列的短期高精度连续组合方法，利用各个分析中心给出的对流层时间序列，利用抗差 Kalman 滤波和最小二乘方差分量估计的原理，进行 GPS 或 VLBI 的短期时间序列的实时的或事后的连续组合。先估计各个分析中心产品偏差，同时计算出其权因子，然后利用 Kalman 滤波技术进行产品组合可以获得滞后的或实时的组合解及其标准差，组合解的平均精度达到0．85 mm。