BDS/GLONASS非组合精密单点定位模型与算法

The models and algorithms of uncombined precise point positioning with BDS/GLONASS data

在GPS非组合精密单点定位(PPP)模型的基础上,针对BDS和GLONASS系统各自的特点,分别提出了适用于BDS和GLONASS系统的非组合PPP模型,并在此基础上构建了BDS/GLONASS联合处理的函数模型.新模型中考虑了BDS系统GEO卫星伪距长周期多路径效应引起的系统性偏差(BDS GEO Multipath Bias,BGMB),将其作为参数进行估计;另外,新模型中还考虑了GLONASS系统伪距频间偏差(inter-frequency-biase,IFB),将其参数化为卫星频率号的线性函数,并通过参数重组得到了满秩的函数模型和可估参数的形式.选取了2015年年积日200~230共一个月的11个MGEX跟踪站数据来验证新模型与算法的正确性和有效性,结果表明:将BDS系统伪距BGMB当作参数估计能够显著提高BDS单系统非组合PPP的收敛速度,并能减小伪距残差;通过线性函数来模型化GLONASS伪距IFB能够显著提高GLONASS单系统PPP的收敛速度,并能在一定程度上减小伪距残差;1个月BDS/GLONASS非组合PPP定位误差RMS在北、东、高三个方向分别为6.9 mm、9.1 mm和19.3 mm,表明提出的BDS/GLONASS非组合PPP模型与算法具有良好的定位性能.

On the basis of the GPS uncombined precise point positioning （PPP） model, new suitable models for BDS and GLONASS uncombined PPP are proposed respectively, considering the different characteristics of BDS and GLONASS. And based on this, the function model for joint BDS/GLONASS processing is constructed. In this new model, systematical code biases caused by long-period multipath effects of BDS GEO satellites are considered and estimated as parameters. Additionally, GLONASS code inter-frequency-biases （IFB） are taken into consideration and modeled as a linear function of frequency numbers. A full-rank function model and estimable parameters are obtained through a re-parameterization process. To validate the correctness and effectiveness of the new model, one month of data from DOY （day of year） 200~300, 2015 of 11 MGEX stations are processed. Results show that parameterizing code BGMB （BDS GEO Multipath Bias） can improve convergence speed significantly and reduce code residuals in the BDS uncombined PPP. Besides, the linear function model for GLONASS code IFB can also improve convergence speed dramatically and reduce code residuals to some extent in the GLONASS PPP. RMS of the joint BDS/GLONASS positioning errors using one month of data are 6.9 mm, 9.1 mm and 19.3 mm, for north, east and up components respectively, which indicates that the proposed joint BDS/GLONASS uncombined PPP models and algorithms have good positioning performance.