Advanced ensemble modeling method for space object state prediction accounting for uncertainty in atmospheric density

For objects in the low Earth orbit region, uncertainty in atmospheric density estimation is an important source of orbit prediction error, which is critical for space traffic management activities such as the satellite conjunction analysis. This paper investigates the evolution of orbit error distribution in the presence of atmospheric density uncertainties, which are modeled using probabilistic machine learning techniques. The recently proposed “HASDM-ML,” “CHAMP-ML,” and “MSIS-UQ” machine learning models for density estimation (Licata and Mehta, 2022b; Licata et al., 2022b) are used in this work. The investigation is convoluted because of the spatial and temporal correlation of the atmospheric density values. We develop several Monte Carlo methods, each capturing a different spatiotemporal density correlation, to study the effects of density uncertainty on orbit uncertainty propagation. However, Monte Carlo analysis is computationally expensive, so a faster method based on the Kalman filtering technique for orbit uncertainty propagation is also explored. It is difficult to translate the uncertainty in atmospheric density to the uncertainty in orbital states under a standard extended Kalman filter or unscented Kalman filter framework. This work uses the so-called “consider covariance sigma point (CCSP)” filter that can account for the density uncertainties during orbit propagation. As a test-bed for validation purposes, a comparison between CCSP and Monte Carlo methods of orbit uncertainty propagation is carried out. Finally, using the HASDM-ML, CHAMP-ML, and MSIS-UQ density models, we propose an ensemble approach for orbit uncertainty quantification for four different space weather conditions.     

Accurate estimation of relative atmospheric density error on the example of uncertain geomagnetic activity information

Errors in neutral atmospheric density are the dominant contributor to unrealistic orbital state-vector covariances in low Earth orbits (LEO). Density uncertainty is caused by model-uncertainty at spatial scales below and within the model resolution, as well as input-uncertainty of the environmental parameters supplied to the semi-empirical atmospheric model.The paper at hand provides multiple contributions. First, analytic equations are derived to estimate the relative density error due to an input parameter uncertainty in any of the environmental parameters supplied to the model. Second, it is shown on the example of uncertain geomagnetic activity information, how to compute the required inputs to facilitate the accurate estimation of the relative density error.A clamped cubic splining approach for the conversion from geomagnetic amplitude (ap) to the kp index is postulated to perform this uncertainty propagation, as other algorithms were found unsuitable for this task. Results of numerical simulations with three popular semi-empirical models are provided to validate the set of derived equations. It is found that geomagnetic input uncertainty is especially important to consider in case of low global geomagnetic activity. The findings seamlessly integrate with prior work by the authors to perform density-uncertainty considering orbit estimation.     

Storm-time atmospheric density modeling using neural networks and its application in orbit propagation

Upper atmospheric densities during geomagnetic storms are usually poorly estimated due to a lack of clear understanding of coupling mechanisms between the thermosphere and magnetosphere. Consequently, the orbit determination and propagation for low-Earth-orbit objects during geomagnetic storms have large uncertainties. Artificial neural networks are often used to identify nonlinear systems in the absence of rigorous theory. In the present study, an attempt has been made to model the storm-time atmospheric density using neural networks. Considering the debate over the representative of geomagnetic storm effect, i.e. the geomagnetic indices ap and Dst, three neural network models (NNM) are developed with ap, Dst and a combination of ap and Dst respectively. The density data used for training the NNMs are derived from the measurements of the satellites CHAMP and GRACE. The NNMs are evaluated by looking at: (a) the mean residuals and the standard deviations with respect to the density data that are not used in training process, and (b) the accuracy of reconstructing the orbits of selected objects during storms employing each model. This empirical modeling technique and the comparisons with the models NRLMSIS-00 and Jacchia-Bowman 2008 reveal (1) the capability of neural networks to model the relationship between solar and geomagnetic activities, and density variations; and (2) the merits and demerits of ap and Dst when it comes to characterizing density variations during storms.     

地基Headwall高谱影像大气校正方法

遥感影像大气校正是定量遥感研究的前提与难点之一.大气校正有多种方法和模型.本文研究了目前常用的大气校正方法,包括FLAASH模型、6S模型、ELC方法和QUAC方法,并应用这几种方法对Headwall野外高光谱成像仪获取的高光谱影像进行大气校正.通过对4种典型地物的表观反射率、校正后反射率以及实测反射率进行对比分析,发现4种模型均能有效去除大气的影响,能够较好地恢复地物光谱的典型特征.相对于其他3种大气校正方法,经验线性法对地基Headwall高光谱影像校正效果最好.      

大气阻力参数修正对低轨空间目标轨道预报精度的改进

对于低轨空间目标, 大气阻力是影响轨道预报精度的主要摄动力. 本文提出了一种 基于空间环境数据和神经网络模型的空间目标大气阻力参数修正方法, 基于目 标的历史两行元根数, 通过模拟得到外推一天轨道预报中预报结果与观测数据 符合最好的阻力调制系数, 分析表明其与太阳F_10.7指数和地磁Ap指数具有很好的相关性. 根据已有数据, 构建神经网络模型, 实现对阻力调制系数 的补偿计算, 从而改进低轨目标外推一天的轨道预报. 结果表明, 神经网络模 型相比两行元根数能够更及时地对空间环境变化进行响应. 将该方案应用于天 宫一号和国际空间站的外推一天轨道预报, 验证了方案的正确性和普适性, 对 地磁扰动引起的较大预报误差改进效果更好, 误差能够降低50%~60%; 平均而言, 预报精度可以提高约30%, 改进成功率达到80%左右.      

高精度在轨实时轨道机动决策

为保证在轨机动实时性和高精度的要求,提出了一种基于机器学习的在轨实时机动决策方法。通过优化算法离线获得摄动下的精确解,减去二体解得到速度增量差,将其投影到轨道坐标系获得速度增量摄动修正项,以此作为神经网络输出,设计网络参数并训练得到摄动修正网络、组合应用摄动修正网络和二体解实现高精度的在轨实时轨道机动决策。仿真结果表明:卫星按照该决策机动完成后的终端位置偏差与按照优化算法给出的决策机动完成后终端位置偏差精度一致,且前者决策耗时仅为后者决策耗时的0.01%左右。所提轨道机动决策方法兼顾了精度与实时性,适用于星上决策。      

Measuring atmospheric density using GPS–LEO tracking data

We present a method to estimate the total neutral atmospheric density from precise orbit determination of Low Earth Orbit (LEO) satellites. We derive the total atmospheric density by determining the drag force acting on the LEOs through centimeter-level reduced-dynamic precise orbit determination (POD) using onboard Global Positioning System (GPS) tracking data. The precision of the estimated drag accelerations is assessed using various metrics, including differences between estimated along-track accelerations from consecutive 30-h POD solutions which overlap by 6 h, comparison of the resulting accelerations with accelerometer measurements, and comparison against an existing atmospheric density model, DTM-2000. We apply the method to GPS tracking data from CHAMP, GRACE, SAC-C, Jason-2, TerraSAR-X and COSMIC satellites, spanning 12 years (2001–2012) and covering orbital heights from 400 km to 1300 km. Errors in the estimates, including those introduced by deficiencies in other modeled forces (such as solar radiation pressure and Earth radiation pressure), are evaluated and the signal and noise levels for each satellite are analyzed. The estimated density data from CHAMP, GRACE, SAC-C and TerraSAR-X are identified as having high signal and low noise levels. These data all have high correlations with anominal atmospheric density model and show common features in relative residuals with respect to the nominal model in related parameter space. On the contrary, the estimated density data from COSMIC and Jason-2 show errors larger than the actual signal at corresponding altitudes thus having little practical value for this study. The results demonstrate that this method is applicable to data from a variety of missions and can provide useful total neutral density measurements for atmospheric study up to altitude as high as 715 km, with precision and resolution between those derived from traditional special orbital perturbation analysis and those obtained from onboard accelerometers.     

Klobuchar电离层模型误差分析及预测

电离层延迟误差是全球导航卫星系统(global navigation satellite system,GNSS)中的重要误差源之一.目前在电离层延迟改正模型中,应用最广泛的是Klobuchar参数模型,但是该模型的改正率仅能达到60％左右,无法满足日益增长的精度需求.将国际GNSS监测评估系统(internation...     

基于天基雷达的空间目标轨道预测

根据天基雷达获取的空间目标位置和速度参数, 研究了计算空间目标轨道根数的方法, 以实现对空间目标的初定轨. 分析了雷达坐标系下对目标的观测误差给协议天球坐标系下的目标参数估计带来的影响. 提出了空间二体相遇问题的一种解决方案. 利用已知的空间站轨道, 仿真分析了空间目标在一周之内和空间站的相遇情况, 同时给出了目标轨道预测的误差分析.      

基于摄动轨道的卫星自主天文导航仿真研究

针对星光折射间接敏感地平的卫星自主天文导航方法 ,利用推广的卡尔曼滤波方法进行仿真研究。为了准确建立运动模型 ,在系统方程中引入了非球形地球引力中的二阶带谐项 ;在考虑具有指数密度的球状分层大气的基础上 ,建立了以星光视高度为观测量的量测方程。在建立了推广的卡尔曼滤波方程后 ,文章进行了计算机仿真 ,并对仿真结果进行了详细的误差分析 ,结果表明基于摄动轨道的星光折射间接敏感地平的卫星自主天文导航方法能取得较高的导航精度     

非均匀螺距的双绞线串扰统计特性分析

双绞线的螺距误差会直接影响其抗干扰水平,螺距误差的随机性会造成双绞线串扰预测的不确定性.通过双绞线制作原理的分析,利用制作参数可以较准确得到任意双绞线螺距误差的概率密度.采取等弧度离散法和广义多端口网络的概念,快速建立非均匀螺距双绞线的频域串扰模型,得到感性耦合和螺距误差的关系.基于螺距误差的概率密度函数,可以有效分析任意双绞线串扰的统计特性.仿真结果表明双绞线制作参数转轴角速度的均值和方差是影响双绞线感性耦合大小的关键因素,在双绞线制作过程中必须予以考虑,以保证双绞线的抗干扰能力.      

遗传算法在星座多星接近轨道优化中的应用

以三颗非共轨的Walker星座卫星为研究对象, 对航天器无需变轨与其接近的可能性进行研究. 将Lambert方法得到的航天器轨道作为初始轨道, 利用遗传算法对初始轨道进行优化. 对初始轨道在参考时刻位置和速度的改变量进行编码,形成对应的种群. 以航天器与星座卫星之间的最近距离为适应度函数, 通过种群的繁殖得到优化结果. 结合仿真算例, 分析了最小二乘算法和遗传算法在轨道优化中的优劣以及接近过程中轨道摄动的影响. 结果表明, 遗传算法适用于所提出的轨道改进问题. 研究结果可为单航天器无需变轨对星座多星接近问题提供理论依据.      

基于线性协方差方法的交会对接误差分析

将线性协方差分析方法和蒙特卡罗仿真相结合,按交会任务和飞行特征把交会过程分为变轨飞行、自由飞行和中途速度修正三种特征段,研究了状态误差的传播规律和交会过程中各种误差对交会对接精度的影响。在变轨飞行段,分析了追踪航天器的姿态误差、控制系统性能状态估计误差,以及目标航天器轨道摄动对状态误差传播的影响。在自由飞行段,分析了追踪航天器估计状态误差的先验值和测轨误差对状态误差传播的影响。在中途速度修正段,分析了追踪航天器姿态误差和控制系统性能误差对状态误差传播的影响。仿真结果表明,误差分析方法设计合理,可以指导交会对接的轨道设计工作,能对已经设计好的交会策略进行误差分析和设计验证。     

Determination of visible coordinates of the low-orbit space objects and their photometry by the CCD camera with the analogue output. Initial image processing

The number of artificial space objects in the low Earth orbit has been continuously increasing. That raises the requirements for the accuracy of measurement of their coordinates and for the precision of the prediction of their motion. The accuracy of the prediction can be improved if the actual current orientation of the non-spherical satellite is taken into account. In so doing, it becomes possible to directly determine the atmospheric density along the orbit. The problem solution is to regularly conduct the photometric surveillances of a large number of satellites and monitor the parameters of their rotation around the centre of mass. To do that, it is necessary to get and promptly process large video arrays, containing pictures of a satellite against the background stars.     

Uncertainty-aware Cube algorithm for medium-term collision risk assessment

The number of Earth orbiting objects is constantly growing, and some orbital regions are becoming risky environments for space assets of interest, which are increasingly threatened by accidental collisions with other objects, especially in Low-Earth Orbit (LEO). Collision risk assessment is performed by various methods, both covariance and non-covariance based. The Cube algorithm is a non-covariance-based method used to estimate the collision rates between space objects, whose concept consists in dividing the space in cubes of fixed dimension and, at each time instant, checking if two or more objects share the same cube. Up to now its application has been limited to the long-term scenarios of orbital debris evolutionary models, where considering the uncertainties is not necessary and impractical. Within operative contexts, instead, medium-term collision risk analysis may be an important task, in which the propagation-related uncertainties play a prominent role, but the timescale poses challenges for the application of standard covariance-based conjunction analysis techniques. In this framework, this paper presents an approach for the evaluation of the medium-term collision frequency for objects in LEO, called Uncertainty-aware Cube method. It is a modified version of the Cube, able to take the possible errors in the space objects’ position into account for the detection of the conjunctions. As an object’s orbit is propagated, the along-track position error grows more and more, and each object could potentially be in a different position with respect to the one determined by numerical propagation and, thus, in a different cube. Considering the uncertainties, at each time instant the algorithm associates more than one cube to each object and checks if they share at least one cube. If so, a conjunction is detected and a degree of confidence is evaluated. The performance of the method is assessed in different LEO scenarios and compared to the original Cube method.     

基于地月合影图像的“嫦娥5T”指向校正算法

以"嫦娥5T"地月合影图像作为基本数据,提出一种基于信息融合的相机指向校正方法。首先利用形态分析从合影图像中提取天体的形心坐标,然后构建目标函数,通过优化算法估计相机的安装误差矩阵。该方法的典型优势是只需要利用单个相机对天体所成的单幅图像,便可快速确定相机指向。仿真数据和实测数据得到相机指向误差估计值的偏差保持在1%度量级,证明了算法的有效性。算法可以为实际工程任务和应用提供参考。     

High accuracy satellite drag model (HASDM)

The dominant error source in force models used to predict low-perigee satellite trajectories is atmospheric drag. Errors in operational thermospheric density models cause significant errors in predicted satellite positions, since these models do not account for dynamic changes in atmospheric drag for orbit predictions. The Air Force Space Battlelab’s High Accuracy Satellite Drag Model (HASDM) estimates and predicts (out three days) a dynamically varying global density field. HASDM includes the Dynamic Calibration Atmosphere (DCA) algorithm that solves for the phases and amplitudes of the diurnal and semidiurnal variations of thermospheric density near real-time from the observed drag effects on a set of Low Earth Orbit (LEO) calibration satellites. The density correction is expressed as a function of latitude, local solar time and altitude. In HASDM, a time series prediction filter relates the extreme ultraviolet (EUV) energy index E_10.7 and the geomagnetic storm index a_p, to the DCA density correction parameters. The E_10.7 index is generated by the SOLAR2000 model, the first full spectrum model of solar irradiance. The estimated and predicted density fields will be used operationally to significantly improve the accuracy of predicted trajectories for all low-perigee satellites.