Variance reduction of GNSS ambiguity in (inverse) paired Cholesky decorrelation transformation

It has been discovered that (a) the variance of all entries of the ambiguity vector transformed by a (inverse) paired Cholesky integer transformation is reduced relative to that of the corresponding entries of the original ambiguity vector; (b) the higher the dimension of the ambiguity vector, the more significantly the transformed variance will be decreased. The property of variance reduction is explained theoretically in detail. In order to better measure the property of variance reduction, an efficiency factor on variance reduction of ambiguities is defined. Since the (inverse) paired Cholesky integer transformation is generally performed many times for the GNSS high-dimensional ambiguity vector, the computation formula of the efficiency factor on the multi-time (inverse) paired Cholesky integer transformation is deduced. The computation results in the example show that (a) the (inverse) paired Cholesky integer transformation has a very good property of variance reduction, especially for the GNSS high-dimensional ambiguity vector; (b) this property of variance reduction can obviously improve the success rate of the transformed ambiguity vector.     

GNSS模糊度降相关算法及其评价指标研究

针对Gauss、LDL和LLL算法构造整数阵存在的实数阵元素计算、实数至整数阵转换的排序问题,分别研究了相应的元素升序降相关算法和整逆型(先求逆后取整)降相关算法。分析了谱条件数、降相关系数和平均相关系数等降相关算法评价指标的优缺点,提出了等效相关系数评价指标。研究结果表明,等效相关系数较其他3种指标能更有效地评价不同维数方差阵,尤其是高维情况的降相关算法效果;逆整型优于整逆型降相关算法,升序(逆整型)降相关算法更佳,且优劣顺序为升序LDL、升序Gauss和升序LLL算法。     

下三角Cholesky分解的整数高斯变换算法

针对全球导航卫星系统(GNSS)载波相位测量中,基于整数最小二乘估计准则解算整周模糊度问题。目前以LAMBDA降相关算法和Lenstra-Lenstra-Lovász(LLL)为代表的规约算法应用最为广泛。由于不同算法采用的模糊度方差-协方差阵的分解方式不同,导致难以合理地进行不同算法性能的比较。该文通过分析LAMBDA算法的降相关特点,从理论上推出基于下三角Cholesky分解多维情形下的整数高斯变换的降相关条件及相应公式,并与分解方式不同的LAMBDA和LLL算法作了对比。实验结果表明,降相关采用的分解方式将会直接影响计算复杂度和解算性能,因此该文推导的整数高斯变换算法便于今后基于下三角Cholesky分解的降相关算法间的合理比较。     

GPS降相关算法的比较研究

利用矩阵分解理论分别对整数高斯法、联合去相关法、基于矩阵乔里斯基分解的迭代法、逆整数乔里斯基法和LLL法等降相关算法进行了分类和比较。仿真计算表明：逆整数乔里斯基分解法优于联合去相关法,联合去相关法优于LLL法。     

Random simulation and GPS decorrelation

 (i) A random simulation approach is proposed, which is at the centre of a numerical comparison of the performances of different GPS decorrelation methods. The most significant advantage of the approach is that it does not depend on nor favour any particular satellite–receiver geometry and weighting system. (ii) An inverse integer Cholesky decorrelation method is proposed, which will be shown to out-perform the integer Gaussian decorrelation and the Lenstra, Lenstra and Lovász (LLL) algorithm, and thus indicates that the integer Gaussian decorrelation is not the best decorrelation technique and that further improvement is possible. (iii) The performance study of the LLL algorithm is the first of its kind and the results have shown that the algorithm can indeed be used for decorrelation, but that it performs worse than the integer Gaussian decorrelation and the inverse integer Cholesky decorrelation. (iv) Simulations have also shown that no decorrelation techniques available to date can guarantee a smaller condition number, especially in the case of high dimension, although reducing the condition number is the goal of decorrelation. Received: 26 April 2000 / Accepted: 5 March 2001     

利用两种z变换算法的PS-DInSAR相位解缠与等价性证明

在介绍PS-DInSAR相位解缠函数模型的基础上,给出了应用LAMBDA方法求解模糊度和形变参数的过程,并将两种改进的z变换降相关算法——逆整乔列斯基和LLL应用于PS-DInSAR相位解缠。以z变换过程的迭代次数、z变换后的模糊度向量间的平均相关系数和协因数阵的谱条件数为准则,对两种算法进行仿真模拟和分析,结果表明逆整乔列斯基算法和LLL算法等价。最后从理论上对两种降相关算法的一致性进行了解释。     

Zero-correlation transformation and threshold for efficient GNSS carrier phase ambiguity resolution

The key point of accurate and precise applications of Global Navigation Satellite Systems lies in knowing how to efficiently obtain correct integer ambiguity. One of the methods in solving the ambiguity resolution problem is applying the ambiguity searching technique coupled with an ambiguity decorrelation technique. Traditionally, an integer-valued limitation of the transformation matrix ensures that the integer characteristic of candidates exists after the inverse transformation, but this also makes the decorrelation imperfect. In this research, the float transformation matrix will be considered. To ensure both the integer characteristic and perfect decorrelation can be reached, the float transformation is used indirectly. To solve the ambiguity resolution problem, the problem is transformed by integer and float transformation matrices. The objective of integer transformation is reducing the number of candidates. The target of float transformation is validating these reduced candidates. A zero correlation domain or a near complete diagonalization covariance matrix can be obtained via the float transformation. A space in this domain will be used as the threshold; hence the zero correlation domain is called the threshold domain. The number of ambiguity candidates based on integer transformation can be reduced once again through the proposed method. The experiments in this paper prove that the method can make the ambiguity resolution become more efficient without any drop in the accuracy.     

一种单频单历元BDS／GPS组合整周模糊度解算方法

针对单频单历元组合载波相位差分技术(RTK)定位过程中存在的秩亏及模糊度解算病态等问题,提出了一种模糊度降相关的新方法。该方法引入伪距观测值进行辅助解算。首先采用经验分权法对伪距与载波相位观测值分配权重,并通过加权最小二乘法获得整周模糊度浮点解及协方差。然后通过对整周模糊度浮点解的方差-协方差矩阵进行降序排列和剔除病态模糊度。最后利用修正后的浮点解迭代搜索模糊度的整数解。试验结果表明而且可以起到良好的模糊度降相关的效果定位。      

A new method for three-carrier GNSS ambiguity resolution

A new method for resolving the carrier-phase integer ambiguity in Global Navigation Satellite Systems (GNSS) is presented: the MOdified Cholesky factorization for Ambiguity (MOCA) resolution. The characteristics and features of this method are described and results obtained using a software simulator and an emulator are presented to validate its efficiency. The results are then compared to those obtained using another existing method and good performance of the MOCA method in new GNSS systems is shown. Furthermore, the proposed method yields accurate results even when short time spans are used or when there are poor estimations of measurement error, making it immune to non-ideal conditions and ultimately a practical solution for real applications.     

A comparative study of algorithms for reducing the fill-in during Cholesky factorization

In this paper several ordering algorithms for the unknowns in geodetic least squares systems are compared. The comparison is restricted to the case of the well known Cholesky factorization of the normal matrixA into a lower triangular factorL. p ]The algorithms which were investigated are: minimum degree, minimum deficiency, nested dissection, reverse Cuthill-McKee, King's-, Snay's-, and Levy's-banker's and Gibbs-King. p ]Also some strategies are presented to reduce the time needed to compute the ordering using a priori information about the way the unknowns are connected to each other. p ]The algorithms are applied to normal matrices of the least squares adjustment of 2D geodetic terrestrial networks, photogrammetric bundle-block adjustments, and a photogrammetric adjustment using independent models. p ]The results show that ordering the unknowns yields a considerable decrease of the cpu time for computing the Cholesky factor, and that in general the minimum degree and Snay's banker's ordering perform best. Furtheron they show that a priori information about the connection structure of the unknowns speeds up the computation of the ordering substantially.Supported by the Netherlands Organization for Scientific Research (NWO)     

差分GPS载波相位整周模糊度快速解算方法

本文提出了一种整周模糊度的快速求解方法,将差分GPS的测量值分配到主要测量值集合和次要测量值集合中,用主要集合中的相位测量值限定简约搜索空间,而次要集合中的相位测量值用来验证候选集合。利用已知的基线长度的约束条件,对搜索空间进行了简约,提高了求解整周模糊度的速度,同时,通过Cholesky分解提高搜索效率。     

On lattice reduction algorithms for solving weighted integer least squares problems: comparative study

Decorrelation or reduction theory deals with identifying appropriate lattice bases that aid in accelerating integer search to find the optimal integer solution of the weighted integer least squares problem. Orthogonality defect has been widely used to measure the degree of orthogonality of the reduced lattice bases for many years. This contribution presents an upper bound for the number of integer candidates in the integer search process. This upper bound is shown to be a product of three factors: (1) the orthogonality defect, (2) the absolute value of the determinant of the inverse of the generator matrix of the lattice, and (3) the radius of the search space raised to the power of the dimension of the integer ambiguity vector. Four well-known decorrelation algorithms, namely LLL, LAMBDA, MLAMBDA, and Seysen, are compared. Many simulated data with varying condition numbers and dimensions as well as real GPS data show that the Seysen reduction algorithm reduces the condition number much better than the other algorithms. Also, the number of integer candidates, before and after the reduction process, is counted for all algorithms. Comparing the number of integer candidates, condition numbers, and orthogonality defect reveals that reducing the condition number and the orthogonality defect may not necessarily result in decreasing the number of integer candidates in the search process. Therefore, contrary to the common belief, reducing the orthogonality defect and condition number do not always result in faster integer least squares estimation. The results indicate that LAMBDA and MLAMBDA perform much better in reducing the number of integer candidates than the other two algorithms, despite having a larger orthogonality defect and condition number in some cases. Therefore, these two algorithms can speed up the integer least squares estimation problem in general and the integer ambiguity resolution problem in particular.     

回代解算的LAMBDA方法及其搜索空间

基于回代解算的序贯条件最小二乘技术和上三角的Cholesky分解算法,提出求解载波相位模糊度的一种新算法--回代解算的LAMBDA方法.该方法同Teunissen提出的LAMBDA方法相比,有着不同的目标函数和不同的搜索空间.实例试算表明这两种方法有着不同的执行过程,但有着相同的模糊度整数解和相当的搜索效率.     

The least-squares ambiguity decorrelation adjustment: a method for fast GPS integer ambiguity estimation

The GPS double difference carrier phase measurements are ambiguous by an unknown integer number of cycles. High precision relative GPS positioning based on short observational timespan data, is possible, when reliable estimates of the integer double difference ambiguities can be determined in an efficient manner. In this contribution a new method is introduced that enables very fast integer least-squares estimation of the ambiguities. The method makes use of an ambiguity transformation that allows one to reformulate the original ambiguity estimation problem as a new problem that is much easier to solve. The transformation aims at decorrelating the least-squares ambiguities and is based on an integer approximation of the conditional least-squares transformation. This least-squares ambiguity decorrelation approach, flattens the typical discontinuity in the GPS-spectrum of ambiguity conditional variances and returns new ambiguities that show a dramatic improvement in correlation and precision. As a result, the search for the transformed integer least-squares ambiguities can be performed in a highly efficient manner.     

Computed success rates of various carrier phase integer estimation solutions and their comparison with statistical success rates

The success rate of carrier phase ambiguity resolution (AR) is the probability that the ambiguities are successfully fixed to their correct integer values. In existing works, an exact success rate formula for integer bootstrapping estimator has been used as a sharp lower bound for the integer least squares (ILS) success rate. Rigorous computation of success rate for the more general ILS solutions has been considered difficult, because of complexity of the ILS ambiguity pull-in region and computational load of the integration of the multivariate probability density function. Contributions of this work are twofold. First, the pull-in region mathematically expressed as the vertices of a polyhedron is represented by a multi-dimensional grid, at which the cumulative probability can be integrated with the multivariate normal cumulative density function (mvncdf) available in Matlab. The bivariate case is studied where the pull-region is usually defined as a hexagon and the probability is easily obtained using mvncdf at all the grid points within the convex polygon. Second, the paper compares the computed integer rounding and integer bootstrapping success rates, lower and upper bounds of the ILS success rates to the actual ILS AR success rates obtained from a 24 h GPS data set for a 21 km baseline. The results demonstrate that the upper bound probability of the ILS AR probability given in the existing literatures agrees with the actual ILS success rate well, although the success rate computed with integer bootstrapping method is a quite sharp approximation to the actual ILS success rate. The results also show that variations or uncertainty of the unit–weight variance estimates from epoch to epoch will affect the computed success rates from different methods significantly, thus deserving more attentions in order to obtain useful success probability predictions.     

整周模糊度整数变换前后LAMBDA方法的执行结果比较

首先介绍了求取模糊度整数解的整数最小二乘方法的基本原理和LAMBDA方法，然后讨论了降相关的可容许整数变换对于LAMBDA方法求取双差模糊度整数解的影响。通过一个短基线的实例计算发现：对原始的双差模糊度进行降相关的可容许整数变换，不仅可提高模糊度整数解的准确性，而且还能提高模糊度的求取速度。     

改进的GPS模糊度降相关LLL算法

模糊度降相关技术可以有效提高模糊度求解的效率及成功率,LLL(A.K.Lenstra,H.W.Lenstra,L.Lovasz)算法是新出现的模糊度降相关方法。详细分析LLL算法,针对该算法中存在的缺陷,提出逆整数乔勒斯基、整数高斯算法和升序调整矩阵辅助的改进LLL算法。利用谱条件数及平均相关系数为准则,以300个随机模拟的对称正定矩阵作为模糊度方差-协方差矩阵,对LLL算法和改进的LLL算法进行仿真计算。比较与分析结果表明,改进LLL算法模糊度降相关处理更加彻底,能有效地加速整周模糊度搜索及成功解算。     

Integer estimation in the presence of biases

 Carrier phase ambiguity resolution is the key to fast and high-precision GNSS (Global Navigation Satellite System) kinematic positioning. Critical in the application of ambiguity resolution is the quality of the computed integer ambiguities. Unsuccessful ambiguity resolution, when passed unnoticed, will too often lead to unacceptable errors in the positioning results. Very high success rates are therefore required for ambiguity resolution to be reliable. Biases which are unaccounted for will lower the success rate and thus increase the chance of unsuccessful ambiguity resolution. The performance of integer ambiguity estimation in the presence of such biases is studied. Particular attention is given to integer rounding, integer bootstrapping and integer least squares. Lower and upper bounds, as well as an exact and easy-to-compute formula for the bias-affected success rate, are presented. These results will enable the evaluation of the bias robustness of ambiguity resolution. Received: 28 September 2000 / Accepted: 29 March 2001     

Ambiguity validation with combined ratio test and ellipsoidal integer aperture estimator

To ensure reliable ambiguity resolution, ambiguity validation is an indispensable step. It has been a challenge for many years and is far from being resolved. Over the past years, various ambiguity validation methods have been proposed, such as F-ratio test, R-ratio test, difference test, projector test, ellipsoidal integer aperture (EIA) estimator and penalized integer aperture (PIA) estimator. In this paper, through analysis and testing, we find that, when the aperture region of EIA is not allowed to be overlapped, the efficiency of ambiguity resolution with EIA is low and it is not applicable to fast static positioning or real-time kinematic (RTK) applications. Then EIA with overlapped aperture regions is recommended and the resulted fail-rate becomes upper bound of the actual one. After that, it is suggested that combined use of overlapped EIA and R-ratio test can increase the reliability of ambiguity resolution. Finally, numerical tests are carried out based on practical buoy data and simulated Galileo data.     

On the importance of intra-frame and inter-frame covariances in frame transformation theory

The coordinate frame transformation (CFT) problem in geodesy is typically solved by a stepwise approach which entails both inverse and forward treatment of the available data. The unknown transformation parameters are first estimated on the basis of common points given in both frames, and subsequently they are used for transforming the coordinates of other (new) points from their initial frame to the desired target frame. Such an approach, despite its rational reasoning, does not provide the optimal accuracy for the transformed coordinates as it overlooks the stochastic correlation (which often exists) between the common and the new points in the initial frame. In this paper we present a single-step least squares approach for the rigorous solution of the CFT problem that takes into account both the intra-frame and inter-frame coordinate covariances in the available data. The optimal estimators for the transformed coordinates are derived in closed form and they involve appropriate corrections to the standard estimators of the stepwise approach. Their practical significance is evaluated through numerical experiments with the 3D Helmert transformation model and real coordinate sets obtained from weekly combined solutions of the EUREF Permanent Network. Our results show that the difference between the standard approach and the optimal approach can become significant since the magnitude of the aforementioned corrections remains well above the statistical accuracy of the transformation results that are obtained by the standard (stepwise) solution.