附加相位延迟到弯曲角反演过程中的误差传递模型

为了研究全球卫星导航定位系统无线电掩星探测中的误差传递过程,文中讨论了掩星探测中附加相位延迟到弯曲角的反演过程中,附加相位延迟随机误差对反演弯曲角误差的影响。在对全球卫星导航定位系统和低轨道卫星进行精密定轨的情况下,忽略轨道误差对反演结果的影响,将轨道参数视为常数,考虑附加相位延迟△L存在随机误差δ△L时对反演弯曲角α带来的弯曲角误差δα。利用EGOPS仿真软件,采用数值模拟的方法模拟了无电离层影响的大气附加相位延迟数据,利用CHAMP真实轨道数据,MSIS-90大气模型和全三维射线追踪器进行前向建模,从中选取了一个掩星事件对误差传递模型进行了验证。结果表明:附加相位延迟的变化会给弯曲角的反演带来相应的误差:当随机误差为5 cm左右时,反演的弯曲角误差在±0.2 mrad范围内;当随机误差为1 cm左右时,反演的弯曲角误差在±4μrad范围内;当随机误差为1 mm左右时,反演的弯曲角误差在±40μrad范围内。最后从模拟的掩星事件中随机抽取了20个掩星事件进行统计分析,给出了20个掩星事件存在1 cm附加相位延迟测量误差引起弯曲角误差的均值误差,其误差特性虽略大于个例研究结果,但与模型完全吻合,反演的弯曲角误差也在±40μrad范围内,进一步验证了模型的准确性。

An error propagation model during the retrieval process of excess phase delay to bending angle

In order to study the error propagation process of the GNSS radio occultation , the impact of random errors of excess phase delay on retrieved bending angle were deduced theoretically during the retrieval process of excess phase delay to bending angle in this work. For GPS and LEO satellites, under the condition of precise orbit determination, neglecting the impact of orbit error to the retrieval results, treat orbit parameters as constant, consider the impact of random errors of excess phase delay on retrieved bending angle ?s errors . In this research, EGOPS simulation software were employed to simulate the data of atmospheric excess phase delay. CHAMP real orbit data, MSIS-90 atmospheric model and full 3D ray tracing program were using in the Forward Model, and one occultation event was selected to validate the error propagation model. The research results show that the changes of excess phase delay can bring corresponding bending angle errors. Random errors of 5cm, 1cm and 1mm are added on the excess phase delay and then the bending angles are retrieved, it is found that the corresponding bending angle errors are less than , and respectively.For the purpose of statistical analysis, 20 occultation events were selected randomly, and the bending angles mean bias were given that induced by excess phase which 1cm random error were added. The results are slightly larger than the error characteristics of case studies, but entirely consistent with the model,the retrieval bending angle errors are within .Through the statistical analysis processing, the model we have derived are validated.