利用静止气象卫星监测初生对流的研究进展

本文主要对近年来利用静止气象卫星监测、分析和研究对流初生的国内外若干研究结果和进展给予了简要综述。主要涉及对流的判识、对流的追踪技术、初生对流的多光谱云顶特征、初生对流的判据和问题与展望等方面的内容。总体而言,利用静止气象卫星监测初生对流,以及发展的对流,是目前国内外研究和探索的热点之一。研究表明,通过精心挑选对对流敏感的光谱通道、通道组合和时间演变判据,可以对初生对流进行监测,可比地面雷达更早地预警降雨系统的发展,显示出在临近预报中的应用潜力。监测对流初生的算法流程大体上分为目标判识、目标追踪和初生对流判识三个部分。国际上主要的一些与对流初生相关的成熟算法各具特色,如RDT(Rapidly Developing Thunderstorms)算法在对流判识中强调了垂直形态的峰值检测,追踪对流时利用了速度外推并检测重叠面积。ForTraCC(Forecasting and Tracking the Evolution of Cloud Clusters)算法关注活跃对流,一定程度上考虑了对流合并和分裂的情况,具有外推预报功能。GOES-R(The Geostationary Operational Environmental Satellite R-Series Program)算法特点是多光谱的使用,利用多光谱判识技术进行对流云顶检测,是面向下一代静止气象卫星探测能力的算法。UWCI(The University of Wisconsin Convective Initiation)构思巧妙,所提出的box-average技术简单易行,适合软件工程化。随着卫星传感器和探测能力的发展,以及计算机技术的快速发展,更多卫星资料将用于联合判识对流。同时,更新的计算机视觉、计算机图像处理和模式识别技术也将用来解决追踪中的复杂问题,进一步改进初生对流的判识准确率。     

高轨空间碎片的观测模拟

对高轨空间碎片进行了观测模拟。在观测模拟中，分别选取南京和云南两个观测站；采用2种跟踪方式：跟踪恒星和跟踪卫星；对于同一地理经度上的同步卫星进行了观测；提供了发现新目标的方法；并给出了新目标的初轨。这为我国空间碎片的数据库的建立作了有益的尝试，对于卫星和航天器的飞行安全具有特别重要的意义。     

Intel MKL在Bernese GNSS数据处理中的应用

为提高Bernese GNSS software数据处理效率，将英特尔数学核心函数库(math kernel library，MKL)应用于Bernese 精密定轨数据处理，对比分析多个MKL矩阵求逆函数与Bernese 原有程序的计算效率。使用2019-03全球200个测站北斗/GNSS数据进行实验分析，结果表明，采用参数预消除策略时，参数预消除步骤消耗时间明显大于矩阵求逆，使用MKL处理数据效率提升不明显；而未采用参数预消除策略时，使用MKL矩阵求逆函数可显著提高矩阵求逆效率，其中dpotri函数矩阵求逆计算效率最高，消耗时间平均值为133 s，相比Bernese 原有程序计算速度可提高13倍。     

基于全球MGEX数据的北斗导航星座精密轨道确定

利用全球分布的MGEX站观测数据,使用"两步法"以双差方式对北斗卫星进行精密定轨。主要研究北斗卫星轨道确定的处理策略,重点分析轨道确定的流程,并通过实验评价了3种不同类型轨道的精度。结果显示,MEO和IGSO卫星的定轨内符合精度优于0.2m,其中径向优于10cm,外符合精度优于30cm;GEO(以C05号为例)卫星外符合精度优于0.7m,且径向优于10cm。     

Automated design of gravity-assist trajectories to Mars and the outer planets

In this paper, a new approach to planetary mission design is described which automates the search for gravity-assist trajectories. This method finds all conic solutions given a range of launch dates, a range of launch energies and a set of target planets. The new design tool is applied to the problems of finding multiple encounter trajectories to the outer planets and Venus gravity-assist trajectories to Mars. The last four-planet grand tour opportunity (until the year 2153) is identified. It requires an Earth launch in 1996 and encounters Jupiter, Uranus, Neptune, and Pluto. Venus gravity-assist trajectories to Mars for the 30 year period 1995–2024 are examined. It is shown that in many cases these trajectories require less launch energy to reach Mars than direct ballistic trajectories.Assistant Professor, School of Aeronautics and AstronauticsGraduate Student, School of Aeronautics and Astronautics     

Orbit Determination of Binary Asteroids

In addition to the detection of an asteroid moon or a binary asteroid, the knowledge of the satellite’s true orbit is of high importance to derive fundamental physical parameters of the binary system such as its mass and to shed light on its possible formation history and dynamical evolution (prograde/retrograde orbit, large/small eccentricity or inclination, etc.). A new methodology for preliminary orbit determination of binary asteroids – and visual binaries in general – is proposed. It is based on Thiele–Innes method combined with a ‘trial and error’ Monte-Carlo technique. This method provides the full set of solutions (bundle of orbits, with the 7 orbital elements) even for a reduced number of observations. The mass is a direct by-product of this orbit determination, from which one can next infer the bulk-density and porosity. In addition to the bundle of orbits, the method provides the marginal probability densities of the foreseen parameters. Such error analysis – since it avoids linear approximation – can be of importance for the prediction of the satellite’s position in the plane-of-sky during future stellar occultations or subsequent observations, but also for the analysis of the orbit’s secular evolution. After briefly describing the method, we present the algorithm and its application to some practical cases, with particular emphasis on asteroids binaries and applications on orbital evolution.     

月球重力场对绕月低轨卫星的长期影响

从解析形式出发,利用月球重力场模型JGL165P1,分析了月球重力场(带谐项)对绕月低轨卫星的长期影响。为了减少计算误差,保证计算精度,在分析解中使用循环公式来计算倾角函数。结果指出对于一个高度为100km的极月轨道卫星,冻结轨道存在的可能性不大,但是当轨道倾角在i=90°附近或者高度再高一些,则有可能存在冻结轨道;对于100km高的初始圆轨道,卫星在无控的情况下半年内将会坠落到月球表面,如果高度增加到200km,则不进行轨道控制也不会坠落到月面上。利用仿真软件GEODYN解算出来的结果证实了上述结论。     

Homoclinic and heteroclinic orbits in the photogravitational restricted three-body problem

We study numerically the asymptotic homoclinic and heteroclinic orbits around the hyperbolic Lyapunov periodic orbits which emanate from Euler's critical points L ₁ and L ₂, in the photogravitational restricted plane circular three-body problem. The invariant stable-unstable manifolds associated to these Lyapunov orbits, are also presented. Poincaré surface of sections of these manifolds on appropriate planes and several homoclinic and heteroclinic orbits for the gravitational case as well as for varying radiation factor q ₁, are displayed. Homoclinic-homoclinic and homoclinic-heteroclinic-homoclinic chains which link the interior with the exterior Hill's regions, are illustrated. We adopt the Sun-Jupiter system and assume that only the larger primary radiates. It is found that for small deviations of its value from the gravitational case (q ₁ = 1), the radiation pressure exerts a significant impact on the Hill's regions and on these asymptotic orbits.     

GLONASS卫星位置的动力学改进算法

为了减弱由于广播星历拟合引起的误差,本文提出了一种基于动力学轨道改进的GLONASS卫星位置算法。该算法仅采用GLONASS广播星历信息就能够提高GLONASS卫星位置坐标的计算精度。算例表明,该方法能够有效减弱GLONASS卫星广播星历的拟合误差,卫星位置的精度损失从0.5m降到mm级;外推2-3h,精度损失也仅在cm级水平。     

高轨卫星轨道预报中神经网络模型优化设计

高轨卫星是我国卫星导航系统的重要组成部分。提升该类卫星的轨道预报精度有利于用户定位精度的提高。提出了一种改进高轨卫星轨道预报精度的新方法。该方法避开了精化动力学模型的困难,尝试从轨道预报误差的规律中寻找突破。利用神经网络作为建立预报模型的工具,将某历史时刻的轨道预报误差作为训练样本,利用训练好的神经网络模型补偿当前时刻的预报轨道以提高轨道预报精度。对影响神经网络模型补偿效果的各因素进行了详细分析,制定了适应于高轨卫星短期、中期和长期预报的神经网络最优模型。利用实测数据进行了试验分析,结果表明:预报8,15及30 d应选择的训练步长分别为10,20及25 min;轨道预报8~30 d时,训练噪声均选取0.01。神经网络模型有效地改进了高轨卫星的轨道预报精度,预报4~30 d,轨道精度提高幅度为34.67%~82.37%不等。