顾及动力学噪声特性与模型补偿的北斗卫星实时滤波精密定轨方法

Because of the low accuracy of dynamic model w.r.t GPS and frequent abnormal orbit states such as attitude switching, eclipse and orbit maneuver etc., the traditional real-time precise orbit determination (POD) method, i.e. prediction of the resent past least squares solutions, cannot determine continuous and reliable precise orbits for satellites of BeiDou Navigation Satellite System (BDS). The filtering-based real-time POD method processes real-time observations at each epoch and is able to response quickly to dynamics changes and tune the weights of observations and dynamic models, which will contribute to BDS real-time POD. However, the two crucial issues that the accuracy degradation (or filter divergency) caused by data saturation and filter abortion caused by irregular satellite dynamics need to be solved presently. To solve these issues, this proposal will study the theory and algorithms for BDS real-time POD. The research will focus on: 1) establishing the stochastic models for different types of BDS satellites based on the long-term observations; 2) model compensation methods in the circumstances that the dynamic model accuracy degrades and anomaly happens; 3) development of the filtering-based real-time POD software for BDS and data processing system validation based on the real-time BDS observation data from the network of MGEX, iGMAS and the National BDS Augmentation Service System (NBASS). The researches of the proposal is hoped to complement the real-time POD theory and bring significant improvement to the real-time POD of BDS-2 and BDS-3.

北斗导航卫星存在动力学模型精度不如GPS系统、运动状态异常变化（姿态切换、地影、轨道机动）较为频繁等问题，常用的轨道预报方法难以获得连续可靠的实时高精度轨道。实时滤波定轨方法可以更为及时快速的调节观测信息与动力学信息的权重，对提升北斗实时精密轨道处理性能具有重要意义。然而，“数据饱和”引起的长时间滤波轨道精度退化（滤波发散）、运动状态异常引起的处理异常（滤波崩溃）是北斗实时滤波定轨亟需解决的难题。本课题旨在建立适用于北斗系统的实时可靠的滤波精密定轨方法，针对以上难题重点开展：基于长期观测数据的不同类型北斗卫星动力学过程噪声随机模型构建；北斗卫星非平稳运动状态下动力学模型补偿方法；研制北斗卫星实时滤波精密定轨处理软件，实现基于MGEX、iGMAS、北斗地基增强系统等实时数据流的业务处理与验证。研究成果有望丰富实时滤波定轨方法，对提升北斗二号、三号卫星实时定轨精度与稳定具有实用价值。

北斗导航卫星存在动力学模型精度不如GPS系统、运动状态异常变化（姿态切换、地影、轨道机动）较为频繁等问题，常用的轨道预报方法难以获得连续可靠的实时高精度轨道。实时滤波定轨方法可以更为及时快速的调节观测信息与动力学信息的权重，对提升北斗实时精密轨道处理性能具有重要意义，项目为此开展了北斗实时滤波精密定轨方法研究与软件研制。基于长期观测数据分析了北斗卫星动力学噪声随机特性，建立了适用于北斗不同类型卫星的动力学噪声模型，实现了卫星平稳运行状态下滤波定轨的连续稳定运行；提出了一种适用于实时滤波定轨处理的非差模糊度固定方法，无需进行独立基线选取或整网独立模糊度搜索，其处理效率和解算精度要优于传统双差模糊度固定方法；提出了基于相位三差残差线性回归分析的北斗卫星轨道机动探测方法，实现北斗卫星轨道机动时刻的实时精确探测，探测精度优于5分钟，有效提升了机动探测时间准确度；针对北斗卫星运动状态异常变化下的实时滤波定轨问题，建立了附加钟差模型约束的自适应滤波定轨方法，有效提升了机动期间和机动后卫星轨道的精度和可用性；研制了北斗/GNSS实时滤波精密定轨处理软件，可业务化提供北斗/GPS/GLONASS/Galileo四大系统五分钟实时更新的精密轨道产品。.项目研究成果发表SCI论文5篇，EI论文2篇，其中6篇为第一标注，申请/授权发明专利2项，获批软件著作权1项，获2022年卫星导航科技进步特等奖（项目负责人排名第二）。