【作者】 郭海荣；

【导师】 杨元喜；

【作者基本信息】 解放军信息工程大学 ， 大地测量学与测量工程， 2006， 博士

【摘要】 在卫星导航定位中，精确位置测量实际上是精确时间的测量。原子钟作为导航系统测距的星上时间基准，是卫星导航系统有效载荷的核心部分，其性能直接决定用户的导航定位精度。因此，对导航卫星原子钟时频特性进行分析具有重要意义。本文主要研究了导航卫星原子钟时频特性分析的理论与方法，主要内容和创新点概括如下： 1、分析了原子钟数据特征，主要包括系统性变化和随机性变化两部分。其系统性变化可用时差、频差和频漂以及周期变化分量来表示，而随机性变化可用七种独立的能量谱噪声来描述，并分析了这几种噪声的产生机制及其时频域特性。在分析原子钟数据特征的基础上，总结了原子频标时频特性主要技术参数的测试分析方法。 2、针对原子钟数据预处理，指出了数据绘图表示的重要性，给出了粗差探测方法，特别强调剔除粗差前要结合原子钟运行条件进行分析，断定异常点是粗差还是频标噪声。给出了相位跳变、频率跳变以及数据间断处理方法，详细分析了不同噪声情况下的频差、频漂估计方法。 3、对实验室真空状态三台Rb钟约60天的频率数据以及GPS Rb钟和Cs钟651天的五分钟精密钟差数据进行了预处理，结果表明：真空状态Rb钟的日漂移率和频率偏差比GPS Block IIR Rb钟约低两个数量级，此外，所有钟还受到明显的周期变化影响。 4、对原子钟时域频率稳定性分析方法进行了系统研究。详细分析了时域频率稳定性分析的各种方差，重点研究了其等效自由度，得出一些有意义的结论：重叠估计和相应的总方法估计对高频噪声等效自由度的提高优于低频噪声，而非重叠估计与之相反。从频域角度揭示了时域频率稳定性分析方法的本质，推导了各种方差的传递函数，并在此基础上分析了各种方差与能量谱噪声的关系，进一步明确各种方差的适用范围。提出了一种分段拟合的噪声水平估计方法。 5、对原子钟频域稳定性分析方法进行了系统研究。比较分析了几种常用的功率谱密度模型，提出了能量谱分析法，该方法能明确地确定原子钟受周期性环境因素影响情况。推导了时频域稳定性分析的严格转换公式，给出由时域稳定性分析计算噪声水平系数的精确公式。 6、重点研究了影响时域稳定性的主要因素，给出了相应的改正方法，并用模拟数据验证了方法的有效性。这主要包括能量谱噪声、频率漂移、相位跳变和频率峰值、无数据段、测量分辨率噪声、周期性环境因素影响以及能量渗漏误差和截断误差等。 7、对实验室真空状态的三台Rb钟和GPS Rb钟、Cs钟进行了时频域稳定性分析，结果表明：所有钟都受到明显的周期变化影响，但不同类型原子钟对环境变化敏感程度不同；Rb钟都有明显的频漂，且受到甚低频噪声影响，Block IIR Rb钟还受到测量噪声引起的调相噪声影响；Rb钟应采用哈达玛系列方差分析其频率稳定性，Cs钟应采用阿仑系列方更多还原

【Abstract】 Measuring of precise distances and positions is really measuring of precise time in a satellite navigation system. The atomic clock is the on-board time datum for measuring the distance in satellite navigation. And it is also one of the key navigation payloads of the satellite navigation system. The performance of the on-board clock affects the navigation and position directly. Therefore, It is very important to study the time and frequency characterization of the on-board atomic clocks. This dissertation mainly focuses on the analysis theories and algorithms of the time and frequency characterization for the atomic clocks of navigation satellites. The main works and contributions are summarized as follows:1. It is shown that the time deviation (or time error) and frequency-offset of atomic clock can be modeled by a deterministic part and a random part by analyzing its data. The deterministic part is quantified by the addition of a constant time offset, constant frequency offset, constant linear frequency drift and periodic components induced by environmental changes. The random part is characterized by seven independent power-law noises whose characterizations and noise mechanisms are also analyzed. Furthermore, the methods of testing and evaluating the time and frequency characterization of frequency standards are summarized and analyzed.2. Data plotting is one of the most important steps in preprocessing the measurement data for atomic clocks. A simple and effective technique for recognizing and removing outliers is presented. It is important to explain all outliers with the help of running environment of the frequency standard, thereby determining whether they are noise or the abnormal data caused by measurement process. Methods for dealing with phase steps, frequency steps and gaps are also developed. Additionally, the frequency offset and frequency drift are estimated and compared in the case of different kinds of noises.3. The data series of 6-months frequency offsets of three Rb clocks in the vacuum sampled with 100-seconds and the 651-day time errors of GPS Rb clocks and Cs clocks sampled with 5-minutes are processed. It is shown that the daily frequency drifts and the frequency offsets of the Rb clocks in the vacuum are more than 2 orders in magnitude than those of GPS Block IIR Rb clocks and all clocks are obviously affected by the periodic components.4. The methods for time-domain stability analysis of atomic clocks are discussed and their equivalent degrees of freedom (EDF) are studied in detail. The interesting results show that the improvement on the EDF of high frequency noise is more than that of low frequency noise for the overlapped estimators and the total estimators of stability variances, but the inverse更多还原