基于重力卫星去混频的卡尔曼滤波融合方法研究

Aliasing error has been one of the most critical problems that restrict the next generation gravity mission. Currently, the most common way to eliminate the aliasing effect is the use of official de-aliasing product, which physically models the high frequent Earth mass flux and thus can be removed beforehand while recovering the monthly gravity fields. In addition, another “self-de-aliasing” approach has been raised recently for eliminate the aliasing effect, of which the main idea is to recover the monthly mean gravity fields and meanwhile the daily low-resolution gravity fields that indeed represent the high frequent mass flux. In this context, our project proposes a data fusion approach based on Kalman filter to combine the a-priori de-aliasing product and the daily low-resolution gravity fields from “self-de-aliasing”. The projects aims to study (1) the combination of current de-aliasing products, (2) realization of “self-de-aliasing” with multiple pairs of high-low tracking satellites, (3) the design of Kalman fusion to fulfill the complementary of a-priori de-aliasing product and the “self-de-aliasing”. Our project shall reduce the aliasing effect so as to improve the accuracy of time-variable gravity fields, and the achievements shall provide a novel solution of de-aliasing problem for next generation gravity mission as well as provide technical supports for Chinese in-house gravity mission.

混频效应已成为制约下一代重力卫星反演精度的关键误差源之一。目前，扣除混频效应通常采用官方去混频产品，其基本思路是建模地球高频时变重力信号并将其从月平均时变重力场解算中扣除。除此之外，最新模拟实验表明采取“自去混频”算法，即同时反演大尺度的高频（每天）时变信号和月平均信号，亦能达到去混频目的。为了充分发挥两种方法的优势，本项目将首次提出针对去混频产品的卡尔曼滤波融合方法，重点研究现有去混频产品的组合策略、多组高低跟踪卫星联合的“自取混频”算法、基于去混频先验信息和“自取混频”提取的大尺度高频信号的卡尔曼滤波融合设计，最终达到实现两种信息互补的目的。本项目的研究将有效削弱重力卫星的混频效应和提高时变重力场的反演精度，研究成果将为解决下一代重力卫星的去混频问题提供全新的思路，为我国自主重力卫星的顺利实施提供技术支撑。

去混频误差是目前和下一代重力卫星的主要误差源，克服或削弱该误差是国内外重力卫星研究的重难点和热点话题之一。为此，基于去混频产品的最新发展态势和下一代重力卫星设计，本项目开展了以削弱卫星重力去混频误差为目标的科学研究，并取得如下成果：(1)分析了星间测距残差和时变重力场之间的关系，建立了去混频模型的质量评价新方法和新理论，解决了背景模型精度难以定量评估的问题；(2)充分研究和对比了当前的主流大气积分算法，在此基础上提出了高精度的三维混合积分理论，并成功将其应用在大气去混频建模中，发布了当前国际最高时空分辨率的大气去混频产品，有效的削弱了去混频产品的插值误差问题；(3)搭建了一套时变重力场模拟和实测数据分析软件，对基于bender-type的自去混频算法进行了实验验证，另外针对卫星姿态确定的扩展卡尔曼滤波进行了理论研究，发布了国内首套GRACE-FO重力卫星姿态产品，该成果可应用于我国自主重力卫星的设计研究。