GNSS非模型化误差处理理论与方法

正由于系统误差的时空复杂特性及认知有限,即使通过了差分和线性组合、经验模型改正及传统参数化吸收等方式,在GNSS观测值中仍然会有残余系统误差的存在,这些残余系统误差可称为非模型化误差。非模型化误差会对GNSS精密定位与导航产生负面作用。为突破现有GNSS应用的精度和可靠性,本文研究了GNSS非模型化误差的处理理论与方法,包括不同应用模式和观测环境下的处理策略,并建立随外界条件变化而自适应调整的非模型化误差处理流程。     

GPS中长基线观测值随机特性分析

探讨了中长基线GPS观测值随机特性的评估方法,先采用历元间高次差分的高通滤波算法有效地消除低频系统误差,只保留高频噪声;然后建立评估中长基线观测值精度、交叉相关性和时间相关性的数学模型。实测数据计算分析了GPS观测值的精度与高度角的关系、不同类型观测值的交叉相关性以及各类观测值的时间相关性。     

立体坐标量测仪系统误差的测定与改正

本文探讨了立体坐标量测仪系统误差改正方法,及系统误差对观测值的影响。指出仪器准系统误差、归心安置定向误差是引起系统误差的主要原因。通过实验验证仿射变换改正公式对消除系统变形趋势有较好作用;同时高次多项式改正能提高更多的观测精度。利用最小二乘配置又可最后确定仪器的噪声影响,为评定仪器精度提供了依据。     

误差与先验信息对导航卫星自主定轨的影响研究

介绍了利用星间距离观测值和由预报星历提供的先验信息采用最小二乘配置法进行导航卫星自主定轨的方法、模型及软件;采用自编软件进行了一系列的模拟计算,分别探讨了观测值随机噪声、残余系统误差和先验信息的好坏对自主定轨精度的影响。结果表明,随机噪声对自主定轨精度的影响不明显,而系统误差和先验信息的影响则较为重要,值得关注。     

三角网水平角观测系统误差的检验和中误差的计算

本文主要讨论三角测量水平角观测误差的检验问题,认为应该用多种统计检验的方法来检验水平角观测值中是否有系统误差。检验结果表明,在一般情况下,Ⅱ等全面网的水平角观测误差主要是偶然误差,而系统误差不是显著的。同时,本文还对几种计算三角测量方向(或角度)中误差的方法进行了讨论。     

一种Partial EIV半参数模型的系统误差处理方法

在使用总体最小二乘求解参数时,若观测值中包含系统误差,此时得到的参数估值则会受到系统误差的影响,从而得到不可靠的解,因此必须削弱系统误差对参数估计的影响,以获得相对可靠的解。本文提出在partial errors-in-variables （Partial EIV）模型的基础上给观测值增加非参数部分（系统误差）,从而构建Partial EIV半参数模型;基于补偿最小二乘准则进行公式推导,并分别通过选取适当的正则化矩阵及通过L曲线法确定平滑因子。通过算例结果分析表明,与传统方法相比,本文的方法在一定程度上能够削弱系统误差的影响,得到更为可靠的参数解,从而验证了该方法的有效性和可行性。     

手持式GPS定位精度研究

在绝对定位模式下系统地探讨手持式GPS定位的稳定性,合适的定位观测时间,观测值中所含误差的性质.通过对比研究发现手持式GPS定位中系统误差占有绝对主导地位,认为改正系统误差是提高手持式GPS定位精度的重要途径.此外,还对手持式GPS同步和异步距离测量方法进行对比研究.     

用拟准检定法和补偿最小二乘估计解算污染半参数模型

现代自动化测绘技术采集的测量数据中系统误差和粗差都在所难免,需要同时考虑系统误差和粗差共存的平差问题.针对污染半参数模型中含有系统误差和观测值受到污染而不严格服从正态分布的情形,文章探讨了用拟准检定法抵御粗差,用补偿最小二乘法分离系统误差.算例表明:用拟准检定法对粗差进行定位定值并对观测误差进行修正后,再利用补偿最小二乘法解算半参数模型,可以得到较好的结果.     

长期验潮站潮位观测误差的综合探测及修复方法研究

在分析潮位观测值中粗差和系统误差特点及现有检测和修复方法的基础上,提出了一种基于潮汐调和分析、概率论与数理统计相结合的潮位误差综合探测和修复方法。在长期潮位站实现了粗差和系统误差的准确探测和修复。     

含有粗差或系统误差的方差分量估计

结合Helmert方差分量估计,讨论当观测值中含有粗差或系统误差时对方差分量估计的影响,并推导顾及粗差或系统误差时的方差分量估计公式.最后,针对实用情况给出一些建议.     

同步卫星观测误差的分布规律及处理

本文利用经典平差方法对鑫诺卫星的观测数据进行处理,采用偏度、峰度正态假设检验其残差值是否服从偶然分布,进而判断参数估计结果可靠与否。检验结果表明,残差时间序列的频率直方图不服从正态分布,最小二乘平差方法在处理此观测数据所得到的估计结果是不可靠的。鑫诺卫星观测数据不仅含有偶然误差,还存在系统误差等。本文采用数字滤波方法模型化系统误差及迭代平差的方法,达到提高数据处理精度的目的。     

A procedure for the significance testing of unmodeled errors in GNSS observations

It is a crucial task to establish a precise mathematical model for global navigation satellite system (GNSS) observations in precise positioning. Due to the spatiotemporal complexity of, and limited knowledge on, systematic errors in GNSS observations, some residual systematic errors would inevitably remain even after corrected with empirical model and parameterization. These residual systematic errors are referred to as unmodeled errors. However, most of the existing studies mainly focus on handling the systematic errors that can be properly modeled and then simply ignore the unmodeled errors that may actually exist. To further improve the accuracy and reliability of GNSS applications, such unmodeled errors must be handled especially when they are significant. Therefore, a very first question is how to statistically validate the significance of unmodeled errors. In this research, we will propose a procedure to examine the significance of these unmodeled errors by the combined use of the hypothesis tests. With this testing procedure, three components of unmodeled errors, i.e., the nonstationary signal, stationary signal and white noise, are identified. The procedure is tested by using simulated data and real BeiDou datasets with varying error sources. The results show that the unmodeled errors can be discriminated by our procedure with approximately 90% confidence. The efficiency of the proposed procedure is further reassured by applying the time-domain Allan variance analysis and frequency-domain fast Fourier transform. In summary, the spatiotemporally correlated unmodeled errors are commonly existent in GNSS observations and mainly governed by the residual atmospheric biases and multipath. Their patterns may also be impacted by the receiver.     

Potentialities of multifrequency ionospheric correction in Global Navigation Satellite Systems

The first-order ionospheric error is reduced in the dual-frequency Global Navigation Satellite Systems (GNSS). In this paper, the possibility of eliminating ionospheric higher-order errors in the multifrequency GNSS is explored. Since the second-order error associated with the geomagnetic field effect on the refractive index can be eliminated in dual-frequency measurements, we explore the possibility of eliminating third-order errors in triple-frequency GNSS in view of phase scintillations. A connection between the possibility of improving the multifrequency GNSS accuracy and diffraction effects in radio signal propagation through the randomly inhomogeneous ionosphere is shown. The numerical simulation has revealed that the systematic, residual ionospheric error is considerably reduced when we pass on from dual-frequency to triple-frequency measurements. The change in the residual error variance during such a transition depends however on the relationship between the inner scale of the turbulent spectrum of ionospheric irregularities and the Fresnel radius. Given the inner scale larger than the Fresnel radius, not only the systematic error, but also the standard deviation reduces when we pass on from dual-frequency to triple-frequency measurements. Otherwise, when the Fresnel radius exceeds the inner scale, the variance increases with increasing number of frequencies in use.     

地面三维激光扫描仪系统误差标定

介绍了地面三维激光扫描仪工作原理,采用自检校法对仪器系统误差进行标定并详细推导了自检校系统误差模型。利用徕卡HDS3000获取的数据进行分析计算,得到该仪器的系统误差值,同时对获取的数据进行改正,并通过检验点验证系统误差改正效果。实验表明,经过系统误差改正后,HDS3000扫描仪的测量准确度可以提高一倍,效果良好。     

基于经验模态分解的GPS基线解算模型

非建模系统误差是影响高精度GPS基线解算精度的一个重要因素,文章给出基于经验模态分解的GPS基线解算模型,有效消弱系统误差对基线解的影响。在现有经验模态分解理论的基础上,定义经验模态分解的多尺度分解与重构结构,并由此给出基于经验模态的系统趋势分离模型,依据累积标准化模量的均值随尺度的变化确定系统误差与噪声分离尺度的选择标准。给出基于经验模态分解的GPS基线解算的技术路线,首先计算GPS相位双差观测方程的浮点解残差序列,分离出系统误差并用于修正GPS双差观测值,重新计算双差浮点解,采用Lambda算法固定整周模糊度,计算固定基线解,从而消弱系统误差对基线解算的影响,提高基线固定解的可靠性。并采用实测GPS数据验证模型,F-ratio指数与W-ratio指数表明系统误差消弱后,基线固定解可靠性得到明显提高,重新计算的残差序列表明系统误差得到很好的消弱。     

顾及有色噪声的卡尔曼滤波在多路径误差削弱中的应用

多路径误差为一时空环境效应,难以构建准确数学模型消除其影响,且该误差在基线两端不具有空间相关性,运用现有差分技术也无法很好消除,是高精度短基线测量中主要误差之一．为进一步削弱多路径误差,本文以监测站坐标时间序列中多路径误差为研究对象,根据多路径误差在历元间的时变特性,建立多路径误差状态空间模型,采用标准卡尔曼滤波和顾及有色噪声的卡尔曼滤波从监测站第一天双差固定解坐标残差序列中估计多路径误差改正序列,并根据多路径误差的周日重复特性,利用第一天得到的多路径误差改正序列对之后各天坐标序列进行改正．最后通过实验分析,得出顾及有色噪声的卡尔曼滤波估计方法优于标准卡尔曼滤波的结论．研究方法对提高GNSS定位精度具有重要实用价值．      

航空重力向下延拓的误差影响分析

本文从分析航空重力向下延拓过程中偶然误差和系统误差的变化特性入手,进而提出处理办法。首先,利用试验说明移去恢复法局限性,同时表明需处理系统误差和偶然误差的必要性。然后,采用理论推演和数值模拟计算分别估计了系统误差和偶然误差影响,试验结果发现:系统误差影响和偶然误差影响均与数据格网间隔、向下延拓高度呈线性关系,当格网化间隔较小和延拓高度较高时系统误差影响和偶然误差影响较大。最后,提出使用半参数模型和正则化算法的两步法估计系统误差和减弱偶然误差影响,试验结果说明两步法处理向下延拓各类误差影响优于仅用半参数模型或正则化算法的结果,在试验数据的偶然误差标准差为2×10~(-5) m/s~2、恒值系统误差3×10~(-5) m/s~2和变值系统误差标准差约1.3×10~(-5) m/s~2时,以及向下延拓高度6.3 km和格网间隔6′的条件下,两步法向下延拓结果的精度可达2.3×10~(-5) m/s~2。     

关于平差残差和单位权中误差的统计分析

从残差的定量关系出发评述了检测粗差的数据探测法和抗差估计的选权迭代法 ,指出了其中用残差代替真误差所产生的问题及解决办法 ,并且导出了单位权方差估值、单位权中误差估值的概率分布 ,给出了后者的数字特征 ,对平差问题的精度体系作了统计分析。     

空间前方交会中的系统误差处理

空间前方交会测量中,基线长及基线方向值的误差对所有观测点的坐标影响是系统性的。本文提出一种方法,利用空间交会测量中的多余观测,可以精确计算出这些系统误差,同时在交会点的坐标中也消去了基线误差的影响。     

Reliable estimation of orbit errors in spaceborne SAR interferometry

An approach to improve orbital state vectors by orbit error estimates derived from residual phase patterns in synthetic aperture radar interferograms is presented. For individual interferograms, an error representation by two parameters is motivated: the baseline error in cross-range and the rate of change of the baseline error in range. For their estimation, two alternatives are proposed: a least squares approach that requires prior unwrapping and a less reliable gridsearch method handling the wrapped phase. In both cases, reliability is enhanced by mutual control of error estimates in an overdetermined network of linearly dependent interferometric combinations of images. Thus, systematic biases, e.g., due to unwrapping errors, can be detected and iteratively eliminated. Regularising the solution by a minimum-norm condition results in quasi-absolute orbit errors that refer to particular images. For the 31 images of a sample ENVISAT dataset, orbit corrections with a mutual consistency on the millimetre level have been inferred from 163 interferograms. The method itself qualifies by reliability and rigorous geometric modelling of the orbital error signal but does not consider interfering large scale deformation effects. However, a separation may be feasible in a combined processing with persistent scatterer approaches or by temporal filtering of the estimates.