【作者】 王彬；

【导师】 刘经南；

【作者基本信息】 武汉大学 ， 大地测量学与测量工程， 2016， 博士

【摘要】 我国从2004年开始启动北斗卫星导航系统建设,经过8年研发,顺利完成区域导航卫星星座组网,并于2012年12月27日对外宣布提供区域导航、定位、授时服务。按照三步走计划,我国将在2020年前后建成覆盖全球的卫星导航系统。GNSS系统性能与卫星钟性能以及系统对卫星钟特征参数的估计和预报能力密切相关。卫星钟作为GNSS系统的关键设备,是获取高精度时频信息的前提和条件。对北斗在轨卫星钟特征进行分析,并开展相应的卫星钟差预报研究,是一项具有理论意义和实用价值的研究课题,对于理解北斗系统服务现状,讨论北斗系统服务性能优化或改进方法具有重要意义。与地面原子钟不同,导航卫星原子钟特征分析和预报具有自己的特点：首先,卫星原子钟特征分析数据不能直接获取,卫星原子钟特征分析通常采用由星地时间比对算法估计的卫星钟差数据,该数据不仅包含卫星钟的影响,还受到时间比对算法的影响；其次,由于运行环境的不同,导航卫星原子钟受卫星轨道动力学环境的影响。星地时间比对算法包括单向和双向两种方式,常用的单向时间比对技术为ODTS技术,双向时间比对技术为TWTT技术。BDS系统采用TWTT技术进行卫星时间与地面系统时间的同步,具有创新性和独特性。TWTT技术与ODTS技术的不同主要体现在观测数据类型、数据处理策略以及算法时间基准等方面,两种技术估计的卫星钟钟差数据具有各自不同的特点：与ODTS钟差相比,TWTT钟差受卫星轨道误差的影响较小。两种钟差数据的对比分析,对于识别BDS卫星轨道及卫星钟差估计误差,进而提高其估计精度具有重要意义。目前,国内外对BDS卫星钟的特征分析主要集中在系统特征和随机特征两个方面,采用的数据量普遍较少,分析时间段普遍较短,尚未开展BDS卫星钟周期变化特征及其原因的分析工作。TWTT技术是BDS系统的创新,目前利用TWTT钟差与ODTS钟差进行BDS卫星钟特征分析的工作仍然较少。国内在卫星钟差预报方面的研究主要集中钟差预报函数模型组合、钟差随机分量的数学方法建模等方面,在卫星钟随机特征对卫星钟差预报模型参数估计误差的影响方面以及卫星钟预报性能分析方面的研究工作相对较少。本文以BDS卫星钟在轨性能评估为主要目标,以在轨特征分析为主要内容,以提高BDS卫星钟差预报精度为主要目的,利用ODTS及TWTT钟差数据,对BDS卫星钟的系统特征、周期变化特征、随机特征以及可预报性能进行分析,结合BDS卫星钟差预报现状,讨论BDS钟差预报精度优化及改进的相关方法,对BDS系统性能优化及改进具有一定的创新性和应用价值。本文主要研究内容和成果包括：1)总结了导航卫星钟差估计、特征分析及预报理论与方法研究现状,分析讨论了ODTS及TWTT钟差在BDS卫星钟特征分析、建模的作用,明确了BDS卫星钟特征分析、建模及卫星钟差预报误差特性分析对BDS系统服务性能优化和提升的重要性和必要性；2)研究了精确时间信息在GNSS系统中的生成与传递；分析了当前常用的卫星钟差估计算法及其数据处理策略对卫星钟差估值产生的影响；从卫星钟特征分析、建模及预报的角度出发,确定了本文的特征分析方案及与卫星钟差预报有关的研究内容;3)讨论了ODTS定轨策略对卫星钟差估计的影响；分析了WHU-BDS精密钟差的内符合精度及与其他分析中心精密钟差的符合程度；基于ODTS钟差数据,分析了BDS卫星钟的系统特征,讨论了GEO/IGSO卫星轨道长半轴变化对BDS卫星钟频漂的相对论影响；分析了BDS卫星钟周期变化特征,研究了卫星钟差周期变化特征与卫星动力学环境的相关性,讨论了BDS卫星钟周期变化的原因；分析了BDS卫星钟亚天频率稳定性及其时变特征,幂律噪声类型以及噪声水平等随机特征,从频率稳定性角度出发,给出了BDS卫星钟的稳定性分析参数；4)分析了TWTT钟差的内符合精度,讨论了TWTT钟差与ODTS钟差差异的周期变化与卫星轨道之间的相关性,得出了ODTS钟差受卫星轨道误差影响的结论；基于TWTT钟差数据,分析了BDS卫星钟的系统特征、周期变化特征及随机性特征；对比分析了TWTT中差与ODTS钟差数据特征及其相应的BDS卫星钟特征分析结果：5)基于卫星钟特征分析结果,定义了BDS卫星钟系统的随机微分方程,推导了该微分方程的Kalman滤波解,该系统能够顾及多种随机噪声以及卫星钟周期变化的影响；分别以ODTS钟差和TWTT钟差为观测量,利用Kalman滤波算法估计卫星钟状态,取得了良好结果；利用Kalman滤波算法,实现了ODTS钟差和TWTT钟差数据的融合；6)评价了BDS广播星历钟差精度,从理论角度分析了广播星历钟差预报精度改进的相关方法;进行了BDS卫星钟差滑动拟合、预报实验,讨论了多项式模型参数估计误差对BDS卫星钟差预报误差的影响；在此基础上设计了BDS卫星钟差混合区间拟合预报策略；以周期项对拟合、预报结果的影响为评价标准,讨论了周期项参数在BDS卫星钟拟合、预报数学模型中的显著性。更多还原

【Abstract】 BeiDou Satellite navigation system (BDS) has been constructed by China from 2004, after eight years of development, the regional navigation satellite constellation is completed. The regional navigation, positioning, and also timing services of BDS were declared on December 27,2012. In accordance with the three-step plan, global navigation satellite system will be built in 2020. GNSS performance is closely related to the satellite clock performance and also the ability to estimate and predict satellite clock. Satellite clock is a key device for GNSS, and is also the premise for obtaining the accurate time and frequency information. Characteristic analysis of BDS satellite clock in orbit, and research of BDS satellite clock prediction is important for understanding the current situation and discussing the optimization or improvement of BDS system performance, and is also a research topic with theoretical and practical value.Different from the characteristics analysis and prediction methods of ground atomic clock, satellite clock has its own feature:firstly, satellite clock data can not be directly acquired, while typically obtained by the time comparison performed between ground station and satellite, which includes not only the impact of the satellite clock, but also the impact of the time comparison algorithms; secondly, satellite clock is affected by the orbital dynamics environment. There are two methods to realize the time comparison between satellite and ground station, one-way method and two-way method. The one-way time comparison technology is commonly referred to ODTS (Orbit Determination and Time Synchronization) technology, and the two-way time comparison technology is commonly referred to TWTT (Two-Way radio Time Transfer) technology.It’s innovative and unique for BDS system to adopt TWTT technology for the time synchronization between satellite and ground station. The differences between TWTT and ODTS are mainly reflected in the following aspects:observation data, data processing strategy and reference time, etc., satellite clock time offset data obtained by these two algorithms have their own different characteristics:compared with the satellite clock time offset data obtained by ODTS algorithm, the influences of satellite orbit error are smaller for the satellite clock time offset data obtained TWTT algorithm. Comparison of these two type of satellite clock time offset data is of great significance for identifying the error and improving precision of satellite orbit determination and time synchronization.At present, the characteristic analysis of BDS satellite clock in orbit are mainly concentrated in the systematic and stochastic characteristic. The analyzed data is small, and the analyzed time is short, and the periodic variation in the BDS satellite clock has not yet been analyzed. TWTT is one of the innovations for BDS system, while contrast analysis of BDS satellite clock characteristic analysis results derived from the analysis of TWTT and ODTS satellite clock time offsets is still less. Domestic research of satellite clock time offset prediction focus on functional model combinations and mathematical modeling of random component of satellite clock time offset. But the researches about the effect of satellite clock stochastic characteristic on the estimation error of satellite clock state, and also the prediction performance analysis of satellite clock are relatively small.In this paper, the main objective is the performance evaluation of BDS satellite clocks in-orbit, and the main content is the characteristic analysis, and the main purpose is improvement of the prediction accuracy of BDS satellite clock time offset. ODTS and TWTT satellite clock time offset data are used to analyze the systematic, periodic, and stochastic characteristics, and also the prediction performance the of BDS satellite clocks. Based on the present situation of BDS satellite clock time offset prediction, discussion of methods related to the optimization and improvement of prediction accuracy of satellite clock time offset is of innovation and application value for the optimization and improvement of BDS system performance. In this paper, the main research contents and achievements include:1) The role of ODTS and TWTT satellite clock data in the characteristic analysis and modelling of BDS satellite clock is discussed on the basis of the research status about the estimation, characteristic analysis and the prediction theory of satellite clock data. Then the importance and necessity of analysis, modelling and prediction research of BDS satellite clock in orbit for the optimization and promotion of BDS system performance is emphasized.2) Generation and transfer of precise time information in GNSS system is studied. The impact of estimation algorithm and data processing strategies on the satellite clock time offset estimation is analyzed. From the perspective of analysis, modelling and prediction research of satellite clock in orbit, the analysis scheme and research content related to the satellite clock prediction is determined.3) The influence of orbit determination strategy on the estimation of satellite clock time offset is discussed. The internal consistency of WHU-BDS precise clock and the accordance with the BDS satellite clock data of other analytical center is analyzed. Systematic characteristic of BDS satellite clock is analyzed using the ODTS satellite clock data, and the frequency drift caused by linear variation of semi-major axis of GEO/IGSO satellite orbit is also discussed. Periodic signals have been detected in BDS satellite clock. In order to identify the underlying cause of the BDS clock harmonics, the relationship between the periodic signals and the orbital dynamics is studied. BDS sub-daily clock stabilities and its time variation is discussed, and also the power law noises and their spectral density. Performance indicators of BDS satellite clock is obtained from the perspective of frequency stability.4) The internal consistency of TWTT satellite clock data is analyzed. The conclusion that satellite clock data estimated by ODTS algorithm are affacted by the satellite orbit determination error is confirmed through the correlation analysis between the difference of ODTS and TWTT satellite clock data and the orbital dynamics. Systematic, periodic, and also the stochastic characteristic of BDS satellite clock is analyzed based on the TWTT satellite clock data. Comparison of BDS satellite clock analysis result derived from ODTS and TWTT satellite clock data respectively are also discussed.5) Stochastic differential equation of BDS satellite clock system is defined on the basis of the analysis results of BDS satellite clock in orbit, and the Kalman filter is used to solve the differential equation. This differential equation system takes the impact of a variety of random noises into account, and can be used for satellite clock state estimation that takes different kinds of clock observables as input data; ODTS and TWTT satellite clock data are used for the validation of Kalman filter in this thesis, and good results are achieved. Kalman filter is also used to realize the data fusion of ODTS and TWTT satellite clock data.6) Clock prediction accuracy of the BDS broadcast ephemeris is evaluated. Methods for the improvement of broadcast ephemeris clock prediction accuracy is theoretically analyzed. The impact of parameter estimation error in the polynomial model on the BDS satellite clock prediction accuracy is discussed on the basis of sliding window polynomial fitting and predicting experiment. And the hybrid fitting intervals method for the BDS satellite clock predicting is designed on the basis of the discussion result. The significance of periodic variation for BDS satellite clock fitting model and prediction model is discussed on the basis of analysis result about the influences of periodic variations on the fitting and predicting of satellite clock data.更多还原