偏最小二乘回归模型内涵分析方法研究

偏最小二乘回归是一种新型的多元分析方法。它可以在自变量多重相关的条件下,有效地构造出对系统解释性最强的子空间,进行发建模,使模型的精度和可靠性得到很大的提高。本文提出采用因素分析方法,对偏最小二乘回归的最优子空间进行正交变换。这种变换方法对偏最小二乘回归的模型结果没有任何影响,却可以使最优子空间的实际含义得到更好的解释。案例研究表明,经过正交变换后,原始变量被分为若干变量组,每个变量组分别对应于最     

1000兆赫≤f≤9400兆赫太阳射电流量与黑子数的关系

本文给出了9400、3750、2800、2000和1000兆赫太阳射电流量与黑子数的统计相关关系式,采用了最小二乘法、百分最小二乘法和复合最小二乘法三种处理方法.结果表明:百分最小二乘法得出的三次方程为最佳拟合关系式.由此得出的宁静太阳射电频谱与用射电干涉仪实测结果符合.      

基于半光程差的天线反射面型面精度检测

型面拟合最小量的选取是天线面板型面检测中的关键技术之一.天线表面各点半光程差的均方根是衡量天线型面精度的指标.根据圆锥曲面的几何定义及其光学性质,提出了圆锥曲面天线面板半光程差的直接计算方法,给出了基于半光程差的非线性最小二乘曲面拟合算法.以旋转抛物面天线面板为例,通过模拟和实际测量,对比分析了分别选取轴向误差和半光程差作为最小量拟合计算结果.试验验证了选取半光程差作为最小量拟合的算法具有更高的数值计算精度及稳定性.      

A multi-spectral spatial convolution approach of rainfall forecasting using weather satellite imagery

Flood forecasting has long been a major topic of hydrologic research. Recent events and studies indicate that the success of flood forecasting in Taiwan depends heavily on the accuracy of real-time rainfall forecasting. In this study, we demonstrate a multi-spectral spatial convolution approach for real-time rainfall forecasting using geostationary weather satellite images. The approach incorporates cloud-top temperatures of three infrared channels in a spatial convolution context. It not only characterizes the input–output relationship between cloud-top temperature and rainfall at the ground level, but also is more consistent with physical and remote sensing principles than single-pixel matches. Point rainfall measurements at raingauge sites are up-scaled to pixel-average-rainfall by block kriging, then related to multi-spectral cloud-top temperatures derived from Geostationary Meteorological Satellite images by spatial convolution. The kernel function of the multispectral spatial convolution equation is solved by the least squares method. Through a cross-validation procedure, we demonstrate that the proposed approach is capable of achieving high accuracy for 1- to 3-h-lead pixel-average-rainfall forecasting.     

Accuracy assessment of single and double difference models for the single epoch GPS compass

The single epoch GPS compass is an important field of study, since it is a valuable technique for the orientation estimation of vehicles and it can guarantee a total independence from carrier phase slips in practical applications. To achieve highly accurate angular estimates, the unknown integer ambiguities of the carrier phase observables need to be resolved. Past researches focus on the ambiguity resolution for single epoch; however, accuracy is another significant problem for many challenging applications. In this contribution, the accuracy is evaluated for the non-common clock scheme of the receivers and the common clock scheme of the receivers, respectively. We focus on three scenarios for either scheme: single difference model vs. double difference model, single frequency model vs. multiple frequency model and optimal linear combinations vs. traditional triple-frequency least squares. We deduce the short baseline precision for a number of different available models and analyze the difference in accuracy for those models. Compared with the single or double difference model of the non-common clock scheme, the single difference model of the common clock scheme can greatly reduce the vertical component error of baseline vector, which results in higher elevation accuracy. The least squares estimator can also reduce the error of fixed baseline vector with the aid of the multi-frequency observation, thereby improving the attitude accuracy. In essence, the “accuracy improvement” is attributed to the difference in accuracy for different models, not a real improvement for any specific model. If all noise levels of GPS triple frequency carrier phase are assumed the same in unit of cycles, it can be proved that the optimal linear combination approach is equivalent to the traditional triple-frequency least squares, no matter which scheme is utilized. Both simulations and actual experiments have been performed to verify the correctness of theoretical analysis.     

基于广义正则化最小二乘的天基测向初定轨

传统天基测向初定轨的不足,主要是由于观测数据存在系统误差和观测方程组的系数矩阵病态或不可逆。文章建立观测方程的半参数回归模型,提出基于补偿最小二乘估计和岭估计的广义正则化最小二乘估计,推导了估计公式,并证明了相关统计性质。引入选主列Givens-QR分解算法,提高观测方程求解效率和数值计算稳定性。仿真结果表明:该方法应用于天基测向初定轨可行,可以提高定轨精度和解算成功率。     

基于贝叶斯网的业务端编程方法

针对现有的业务端编程方法缺乏对于没有软件开发背景的最终用户的指导机制,提出了一种基于贝叶斯网的业务端编程方法.首先提出了一种基于贝叶斯网的最终用户编程基本框架,然后基于贝叶斯网结构学习方法和过程日志建立了用来支撑业务端编程的领域专家系统,在此基础上提出了以领域专家系统为核心的启发式业务端编程方法.该方法在编程过程中依赖领域专家系统对最终用户推荐业务活动,提供实时指导,帮助最终用户逐步完成应用系统的构造.最后对该方法的有效性进行了定性的评估和分析,结果表明,基于贝叶斯网的业务端编程方法可以在一定程度上提高最终用户编程的效率.     

Regional gravity field model in Pakistan area from the combination of CHAMP,GRACE and ground data using least squares collocation: A case study

This study describes a methodology of recovery of the Earth’s gravity field from CHAMP and GRACE satellites data in Pakistan using least squares collocation (LSC) based downward continuation technique. The CHAMP height anomalies and GRACE gravity disturbances derived from the observed satellite data have been used in combination solution using LSC with observed gravity values at the Earth surface. The combined covariance functions of height anomalies and/or gravity disturbances at satellite altitudes and observed gravity anomalies at Earth surface have been used as the basis for combination and downward continuation solution. The variance of predicted gravity anomalies from GRACE gravity disturbances is relatively lower than the corresponding results of gravity anomalies from CHAMP height anomalies. This fact may be attributed partly to the amplification of noise and partly to the unstable inverse transformation process of height anomalies to gravity anomalies. The impact of data error variance has been studied in the context of smoothing and noise reduction in the final solution of downward continuation using least squares collocation. The raising of data error suppresses the noise and as a result a smooth final solution is obtained. The prediction results appear to be dependent on the quality of data and goodness of combined covariance function, which are fairly comparable for the CHAMP and GRACE data. The recovered gravity field from satellite data appears to contribute mainly to medium and long wavelength parts of total gravity field spectrum. Due to flexibility of data handling in least squares collocation, this procedure is applicable to any observable of gravity field being at different altitudes and with different data spacing.     

脉冲星导航的整周模糊度解算方法研究

脉冲相位是X射线脉冲星导航方法的基本观测量,在测量的脉冲相位与预报相位比对的过程中存在整周期模糊度问题．借鉴GPS载波相位模糊度的解算方法,提出了单差搜索法、最小二乘搜索法和模糊度函数法3种空间搜索方法用以解算脉冲星导航系统的整周模糊度,并针对地球卫星完全迷失的情况进行了搜索仿真验证．结果表明,这3种方法都能够准确地解算出整周模糊度,均可用于脉冲星导航系统的在轨解算．其中最小二乘搜索法的性能稳定,搜索速度快;模糊度函数法搜索速度慢,但适用于装配单X射线探测器的脉冲星导航系统．     

严格回归轨道的管道导航方法研究

分析了作为参考轨道的严格回归轨道与卫星在轨运行状态的相对运动关系,提出近地遥感卫星的管道导航方法。由于参考轨道的设计只考虑高精度的地球非球形摄动,与在轨卫星的动力学环境存在差别,这导致两者之间存在切航向漂移。基于高精度的轨道动力学模型和位置确定方法,设计了卫星与参考轨道采样点的沿航向对齐算法,从而获取了卫星相对参考轨道采样点的相位时间偏差和卫星在参考轨道编队坐标系切航向平面内的相对运动轨迹,进而引入椭圆的“最小二乘适配法”获取相对运动轨迹的特征量。所研究的管道导航方法可应用于基于GNSS测量数据的卫星自主轨迹保持。     

基于最小偏差LS-SVR的模电系统多故障诊断

为解决支持向量分类机多分类存在的样本重复训练、训练模型过多的问题,保证模拟电子系统在整体和局部多故障模式上的诊断正确率,提出基于最小偏差的最小二乘支持向量回归机多故障诊断方法.通过引进样本各维度拟合误差相对于平均拟合误差的偏差平方项,最小化维度间的拟合误差间距,得到能够输出多维变量及具有高分辨率的最小二乘支持向量回归机模型.将模型多维输出值与预设的各个多故障模式值相乘,所得结果集中最大值所对应的多故障模式即为最终诊断结果.仿真结果表明:提出的方法在简化训练过程的同时,能够保持良好的整体和局部多故障诊断正确率.     

基于补偿最小二乘的航天器轨道确定方法研究

航天器动力学模型误差是降低轨道确定精度的关键因素之一,文章将航天器动力学模型中的模型误差转化为测量模型的系统误差,建立了测量方程的部分线性模型;并根据补偿最小二乘原理,推导了部分线性模型参数估计方法,并证明了相关性质。仿真试验表明:部分线性模型更接近测量数据的真实表现;基于补偿最小二乘的航天器轨道改进方法能有效逼近航天器状态真值,定轨精度显著优于传统最小二乘估计。     

时钟偏差辅助的GPS完整性监测算法

对伪距残差最小二乘的接收机自主完整性监测(RAIM)算法进行了分析.在此基础上,通过对接收机时钟偏差的建模,将模型预测的时钟偏差引入伪距残差最小二乘的RAIM算法中,保证了在可见星仅为4颗的情况下,仍能利用χ2检验法对全球定位系统(GPS)进行完整性监测,从而达到提高完整性监测算法有效性的目的.计算机仿真的结果显示,这种辅助方式不仅计算简单,而且有效可行.      

Computerized ionospheric tomography with the IRI model

Computerized ionospheric tomography (CIT) is a method to estimate ionospheric electron density distribution by using the global positioning system (GPS) signals recorded by the GPS receivers. Ionospheric electron density is a function of latitude, longitude, height and time. A general approach in CIT is to represent the ionosphere as a linear combination of basis functions. In this study, the model of the ionosphere is obtained from the IRI in latitude and height only. The goal is to determine the best representing basis function from the set of Squeezed Legendre polynomials, truncated Legendre polynomials, Haar Wavelets and singular value decomposition (SVD). The reconstruction algorithms used in this study can be listed as total least squares (TLS), regularized least squares, algebraic reconstruction technique (ART) and a hybrid algorithm where the reconstruction from the TLS algorithm is used as the initial estimate for the ART. The error performance of the reconstruction algorithms are compared with respect to the electron density generated by the IRI-2001 model. In the investigated scenario, the measurements are obtained from the IRI-2001 as the line integral of the electron density profiles, imitating the total electron content estimated from GPS measurements. It has been observed that the minimum error between the reconstructed and model ionospheres depends on both the reconstruction algorithm and the basis functions where the best results have been obtained for the basis functions from the model itself through SVD.     

一类有色噪声干扰随机系统最小二乘递推辨识

提出一类有色噪声干扰随机系统的辅助模型最小二乘递推辨识算法.此算法结合辅助模型辨识思想和递推增广最小二乘理论,用辅助模型的输出代替辨识模型信息向量中未知中间变量,用估计残差代替信息向量中不可测噪声项,从而可以运用递推辨识思想来估计系统所有参数,包括噪声模型参数.仿真例子说明提出算法的有效性.      

经验Bayes岭估计在GINS车载试验工具误差辨识中的应用

在利用车载试验进行GINS工具误差辨识过程中,由于输入加速度很小使得系统存在严重的复共线性。应用传统的最小二乘方法会出现增大最小二乘估计量的方差、参数估计值不稳定、产生弃真错误等问题。本文引入经验Bayes岭估计方法来进行GINS车载试验工具误差辨识工作。仿真结果表明,和传统最小二乘方法相比,经验Bayes岭估计的辨识精度有所提高,并可克服系统存在的复共线性的影响。     

基于连续小波变换的飞行器结构模态参数辨识

给出了一种基于连续小波变换的多输入多输出MIMO(Multiple-Input Multiple-Output)飞行器结构模态参数的辨识方法.对结构离散运动方程进行连续小波变换建立了小波域内的系统AR(Auto Regression)模型,AR模型的系数矩阵决定着系统的动力学特性,可以通过最小二乘法求得.模态参数可以通过求解由AR系数矩阵构成的特征矩阵的特征值来获得.与已有基于小波变换的模态参数辨识方法相比,该方法应用了连续小波的时移共变性和小波变换的滤波能力来确保辨识的效率.在辨识过程中,采用优化算法提高了辨识的精度和稳定性.算例仿真结果表明该方法具有较高的计算精度和稳定性,能用于飞行器结构模态参数的辨识.      

基于三维最小二乘方法的空间直线度误差评定

空间直线度误差是评定机械产品精度的一项重要指标,实际工程中对空间直线度误差评定算法的精度要求越来越高.为了准确评定空间直线度误差,参照国家标准(GB/T 11336—2004),采用三维最小二乘方法建立了空间直线拟合的数学模型,并给出了该数学模型的精确解.基于最小二乘拟合中线,采用空间投影、坐标变换和格点法求得最小二乘中线包容圆柱面直径.采用数值算例验证了新方法的有效性.提出的空间直线度误差评定方法精度高、鲁棒性好且易于编程实现.