A Study of EM Algorithm as an Imputation Method: A Model-Based Simulation Study with Application to a Synthetic Compositional Data

Compositional data, such as relative information, is a crucial aspect of machine learning and other related fields. It is typically recorded as closed data or sums to a constant, like 100%. The statistical linear model is the most used technique for identifying hidden relationships between underlying random variables of interest. However, data quality is a significant challenge in machine learning, especially when missing data is present. The linear regression model is a commonly used statistical modeling technique used in various applications to find relationships between variables of interest. When estimating linear regression parameters which are useful for things like future prediction and partial effects analysis of independent variables, maximum likelihood estimation (MLE) is the method of choice. However, many datasets contain missing observations, which can lead to costly and time-consuming data recovery. To address this issue, the expectation-maximization (EM) algorithm has been suggested as a solution for situations including missing data. The EM algorithm repeatedly finds the best estimates of parameters in statistical models that depend on variables or data that have not been observed. This is called maximum likelihood or maximum a posteriori (MAP). Using the present estimate as input, the expectation (E) step constructs a log-likelihood function. Finding the parameters that maximize the anticipated log-likelihood, as determined in the E step, is the job of the maximization (M) phase. This study looked at how well the EM algorithm worked on a made-up compositional dataset with missing observations. It used both the robust least square version and ordinary least square regression techniques. The efficacy of the EM algorithm was compared with two alternative imputation techniques, k-Nearest Neighbor (k-NN) and mean imputation (), in terms of Aitchison distances and covariance.

Modified Omega-K algorithm for processing helicopter-borne frequency modulated continuous waveform rotating synthetic aperture radar data

With appropriate geometry configuration, helicopter- borne rotating synthetic aperture radar （ROSAR） can break through the limitations of monostatic synthetic aperture radar （SAR） on forward-looking imaging. With this capability, ROSAR has extensive potential applications, such as self-navigation and self-landing. Moreover, it has many advantages if combined with the frequency modulated continuous wave （FMCW） technology. A novel geometric configuration and an imaging algorithm for helicopter-borne FMCW-ROSAR are proposed. Firstly, by per- forming the equivalent phase center principle, the separated trans- mitting and receiving antenna system is equalized to the case of system configuration with antenna for both transmitting and receiving signals. Based on this, the accurate two-dimensional spectrum is obtained and the Doppler frequency shift effect in- duced by the continuous motion of the platform during the long pulse duration is compensated. Next, the impacts of the velocity approximation error on the imaging algorithm are analyzed in de- tail, and the system parameters selection and resolution analysis are presented. The well-focused SAR image is then obtained by using the improved Omega-K algorithm incorporating the accurate compensation method for the velocity approximation error. FJnally, correctness of the analysis and effectiveness of the proposed al- gorithm are demonstrated through simulation results.

Improved chirp scaling algorithm integrated motion compensation for parallel-track bistatic synthetic aperture radar

To compensate motion errors of images from the parallel-track bistatic synthetic aperture radar（BiSAR）,an improved chirp scaling algorithm（CSA） is proposed.Since velocity vector of the moving aircrafts in the parallel-track BiSAR system can not remain invariant in an aperture,an actual aperture is divided into subapertures so that it is reasonable to assume that the aircrafts move with constant acceleration vector in a subaperture.Based on this model,an improved CSA is derived.The new phase factors incorporate three-dimensional acceleration and velocity.The motion compensation procedure is integrated into the CSA without additional operation required.The simulation results show that the presented algorithm can efficiently resolve motion compensation for parallel-track BiSAR.

Novel method for extraction of ship target with overlaps in SAR image via EM algorithm

The quality of synthetic aperture radar(SAR)image degrades in the case of multiple imaging projection planes(IPPs)and multiple overlapping ship targets,and then the performance of target classification and recognition can be influenced.For addressing this issue,a method for extracting ship targets with overlaps via the expectation maximization(EM)algorithm is pro-posed.First,the scatterers of ship targets are obtained via the target detection technique.Then,the EM algorithm is applied to extract the scatterers of a single ship target with a single IPP.Afterwards,a novel image amplitude estimation approach is pro-posed,with which the radar image of a single target with a sin-gle IPP can be generated.The proposed method can accom-plish IPP selection and targets separation in the image domain,which can improve the image quality and reserve the target information most possibly.Results of simulated and real mea-sured data demonstrate the effectiveness of the proposed method.

IMAGING ANALYSIS OF FREQUENCY MODULATION CONTINUOUS WAVE SYNTHETIC APERTURE RADAR IN NEAR RANGE MODE

Frequency-Modulation Continuous-Wave Synthetic Aperture Radar(FMCW SAR)has shown great potential in the applications of civil and military fields because of its easy deployment and low cost.However,most of these work and analysis are concentrated on airborne FMCW SAR,where the characteristics of the imaging geometry and signal are much similar to that of traditional pulsed-SAR.As a result,a series of test campaigns of automobile-based FMCW SAR were sponsored by Institute of Electronics,Chinese Academy of Sciences(IECAS)in the autumn of 2012.In this paper,we analyze the imaging issues of FMCW SAR in automobile mode(named as near range mode),where a vehicle is used as moving platform and a large looking angle is configured.The imaging geometry and signal properties are analyzed in detail.We emphasize the difference of the near range mode from the traditional airborne SAR mode.Based on the analysis,a focusing approach is proposed in the paper to handle the data focusing in the case.Simulation experiment and real data of automobile FMCW SAR are used to validate the analysis.

The study on aperture configuration of optical synthetic aperture imaging system

A model based on Fourier domain consideration on aperture configuration of optical synthetic aperture imaging system is introduced in this paper. The derivation of the model is directly linked to the restoration error of the original object from the recorded image. The aperture configuration is a function of the maximum frequency of interest, and takes into account the diameter of the aperture. The simulative results of genetic algorithms illustrate the usefulness of this model lot designing a synthetic aperture optical imaging system, and the aperture configuration of a good synthetic aperture optical imaging system should be non-redundant.

Web-Based Synthetic Optimization Design System of Micro-Components

In order to meet the requirement of network synthesis optimization design for a micro component, a three-level information frame and functional module based on web was proposed. Firstly, the finite element method （FEM） was used to analyze the dynamic property of coupled-energy-domain of virtual prototype instances and to obtain some optimal information data. Secondly, the rough set theory （RST） and the genetic algorithm （GA） were used to work out the reduction of attributes and the acquisition of principle of optimality and to confirm key variable and restriction condition in the synthesis optimization design. Finally, the regression analysis （RA） and GA were used to establish the synthesis optimization design model and carry on the optimization design. A corresponding prototype system was also developed and the synthesis optimization design of a thermal actuated micro-pump was carded out as a demonstration in this paper.

Retrieval of Sea Surface Wind Speed by One-Dimensional Synthetic Aperture Microwave Radiometer

One-dimensional synthetic aperture microwave radiometers have higher spatial resolution and record measurements at multiple incidence angles.In this paper,we propose a multiple linear regression method to retrieve sea surface wind speed at an incidence angle between 0°65°.We assume that a one-dimensional synthetic aperture microwave radiometer operates at frequencies of 6.9,10.65,18.7,23.8 and 36.5 GHz.Then,the microwave radiative transfer forward model is used to simulate the measured brightness temperatures.The sensitivity of the brightness temperatures at 0°65°to the sea surface wind speed is calculated.Then,vertical polarization channels(VR),horizontal polarization channels(HR)and all channels(AR)are used to retrieve the sea surface wind speed via a multiple linear regression algorithm at 0°65°,and the relationship between the retrieval error and incidence angle is obtained.The results are as follows:(1)The sensitivity of the vertical polarization brightness temperature to the sea surface wind speed is smaller than that of the horizontal polarization.(2)The retrieval error increases with Gaussian noise.The retrieval error of VR first increases and then decreases with increasing incidence angle,the retrieval error of HR gradually decreases with increasing incidence angle,and the retrieval error of AR first decreases and then increases with increasing incidence angle.(3)The retrieval error of AR is the lowest and it is necessary to retrieve the sea surface wind speed at a larger incidence angle for AR.

合成孔径雷达快速后向投影算法综述

后向投影(BP)算法是合成孔径雷达成像算法发展的重要方向之一。然而,由于BP算法具有较大的计算量,阻碍了其在工程应用上的发展。因此,近年来如何有效地提高BP算法的运算效率受到了广泛的重视。该文讨论了基于多种成像面坐标系的快速BP算法,包括距离-方位平面坐标系、地平面坐标系和非欧氏坐标系。该文首先简要介绍了原始BP算法的原理和不同坐标系对加速BP算法的影响,并对BP算法的发展历程进行梳理。然后讨论了基于不同成像面坐标系的快速BP算法的研究进展,并重点介绍了作者所在研究团队近年来在快速BP成像方面完成的研究工作。最后介绍了快速BP算法在工程上的应用,并展望了未来快速BP成像算法的研究发展趋势。

多子阵SAS方位空变运动补偿子孔径算法

为解决方位空变的侧摆和偏航误差存在情形下合成孔径声纳的快速运动补偿与成像问题,提出一种多子阵合成孔径声纳方位空变运动补偿子孔径算法。首先,建立运动误差存在情形下的双程距离历程模型,并利用泰勒级数展开对双根号形式距离历程进行近似;然后,利用子孔径运动补偿和单基等效处理,将含有方位空变的侧摆和偏航误差的多子阵回波数据转换为理想的单阵回波数据;最后,利用经典的单阵频域逐线成像算法,实现快速运动补偿和高分辨成像。仿真实验与实测数据成像结果均验证了所提算法的有效性。

基于SiamFC的海上舰船跟踪实验设计

为了提高海上舰船跟踪准确率,设计了一种基于SAR图像的海上舰船跟踪实验方案。首先利用三组连续的卫星SAR图像构建数据集,然后根据海上舰船尺寸小、全卷积孪生网络(SiamFC)算法网络层数少的特点,提出将SiamFC算法应用于海上舰船跟踪过程。通过调整模型相关参数,并构建数据集,对模型进行了训练及测试。测试结果表明,该模型具有可行性。为了达到更好的跟踪效果,引入伽马算子对SAR图像进行增强,减少了SAR图像的相干斑点噪声。最后,对改进后的模型进行测试的结果表明,该方案的舰船跟踪准确率大大提高。

多子阵合成孔径声呐6阶方位空变运动补偿快速算法

为解决6自由度运动误差存在情形下的多子阵合成孔径声呐快速运动补偿与成像问题,本文提出一种多子阵合成孔径声呐6阶方位空变运动补偿快速算法。在考虑6自由度运动误差随方位时间6阶空变特性的基础上,对双根号形式距离历程进行泰勒级数展开并保留至6阶项,进而推导得到严格解析的点目标响应二维谱,通过对传统四阶模型的距离多普勒算法进行改进,同时实现了6自由度运动误差的快速补偿和图像重建。仿真分析和实测数据成像结果表明了所提算法的有效性和高效率,可为大场景宽测绘带多子阵合成孔径声呐快速运动补偿和高分辨成像提供理论依据。

基于相位补偿核的GNSS-SAR改进后向投影成像算法

在利用全球导航卫星系统反射信号(GNSS-R)的无源合成孔径雷达体制中,目前常采用后向投影算法进行成像处理,但该算法逐点精确聚焦的处理过程使得算法的运算量大。针对上述问题,提出了一种数据分块和构造相位补偿核的改进后向投影算法。改进算法对回波矩阵进行分块,将目标区域成像网格点的最长延迟距离与最短延迟距离等间隔划分,建立相位补偿核,通过距离向搜索操作得到各网格点的相位补偿结果,在方位向相干叠加形成子图,递归合并得到最终图像。定量分析得出,改进算法可有效降低运算量,同时节省内存。仿真结果验证了所提算法的有效性。

基于最大对比度的联合相位误差校正方法

针对高分辨率逆合成孔径雷达(ISAR)成像过程中,因目标转动所引入的转动相位误差无法被忽略,以及由于转动相位误差使得传统平动相位误差校正方法性能严重下降的问题,提出了一种基于最大对比度的联合相位误差校正的方法,可同时校正残余的平动相位误差和转动相位误差,进一步提高成像的质量。首先利用传统的运动补偿方法对回波信号进行处理,其次根据最大对比度准则,将相位误差校正问题转化为最大对比度优化问题,最后利用改进的拟牛顿算法进行求解。该方法能使高分辨率ISAR进行精准成像,且所得图像聚焦效果良好。通过对仿真数据和实测数据的处理,验证了文中方法的有效性。

一种基于随机森林的OFDM系统自适应算法

针对动态变化的信道环境,自适应正交频分复用(Orthogonal Frequency Division Multiplexing,OFDM)系统可以对子载波间隔和循环前缀长度进行调整,以最大化系统的吞吐量。为了能够快速准确地找到OFDM系统在不同信道环境中的最优子载波间隔和循环前缀长度取值,本文提出了基于随机森林的OFDM系统自适应算法。随机森林算法基于集成的思想,能够有效处理高维度数据,并且具有高效率、高准确率和强泛化能力等优势,可以在复杂的数据场景下进行有效的分类。通过提取通信过程中信噪比、用户移动速度、最大多普勒频率和均方根时延扩展等信道特征与OFDM系统的子载波间隔和循环前缀长度组成训练样本,利用随机森林算法创建了OFDM系统参数多分类模型。所提模型可以根据输入的信道特征,实现OFDM系统子载波间隔和循环前缀长度的自适应分配。同时,针对训练样本主要集中在少数几个系统参数类别的情况,利用合成少数类过采样技术对较少样本数的类别进行扩充,满足了随机森林算法对训练样本类别平衡化的需求,进一步提高了算法的分类准确率。相比传统的自适应算法,所提算法具有更高的分类准确率和模型泛化能力。分析和仿真结果表明,与子载波间隔和循环前缀长度固定的OFDM系统相比,本文所提出的自适应算法能够准确选择出最优的系统参数,可以有效地减轻信道中符号间干扰和子载波间干扰的影响,从而在整个信噪比范围上提供最大的平均频谱效率。基于随机森林的OFDM系统自适应算法能够动态地分配子载波间隔和循环前缀长度,增强OFDM系统的通信质量和抗干扰能力,实现在不同信道环境下的可靠传输。

基于自适应GA-RF的用户流失预测研究

针对电信用户流失问题,文章提出一种自适应遗传算法优化随机森林的预测模型。首先对Kaggle平台提供的电信数据进行数据清洗、特征提取及无量纲化处理,然后运用SMOTE过采样以解决数据不平衡问题,对决策树、随机森林等模型预测的召回率、F1和AUC值进行对比。最后提出一种自适应遗传算法优化随机森林的电信用户流失预测模型。结果表明,自适应遗传算法优化的随机森林模型的预测性能优于单一分类模型。

基于遗传算法的无人机载穿墙三维SAR航迹规划方法

传统手持或车载穿墙雷达由于架设高度受限,无法对城市高层建筑内部目标进行透视成像,无人机载穿墙雷达具有灵活机动、高效便捷、无高度限制等优势,可对城市高层楼宇进行大范围三维穿透探测。三维层析合成孔径雷达(SAR)成像广泛采用多基线扫描模式,以获得高度向高分辨能力,但存在高度向空域欠采样导致的栅瓣问题。对此,该文提出一种基于遗传算法的无人机载穿墙三维SAR航迹规划方法,通过非均匀化飞行航迹,削弱周期性的雷达回波能量叠加,从而抑制栅瓣、实现更优的成像质量。该算法结合飞行距离与无人机载穿墙雷达成像质量的内在关系,建立了无人机航迹规划代价函数;利用遗传算法对3种典型的无人机飞行轨迹关键控制点进行基因编码,并进行基因杂交、变异等以优化种群与个体;最终通过最小化代价函数,分别筛选出3种典型飞行模式下的最优飞行航迹。仿真和实测数据的三维成像结果表明:相较于传统等间距多基线飞行模式,所提方法显著抑制了成像目标的栅瓣效应;此外,无人机斜线飞行的航迹长度明显缩短,提高了无人机载穿墙SAR成像效率。

结合偏航校正的合成孔径声呐大测绘带运动补偿方法

传统基于频域补偿相位的运动补偿方法难以校正姿态误差造成的波束照射方向偏转,因此提出了一种基于偏航校正的合成孔径声呐大测绘带运动补偿方法。首先,利用时延的距离空变特性校正时延估计野值,提高估计精度和稳健性;其次,通过在后向投影算法的成像过程中对实孔径成像网格进行旋转来校正偏航造成的波束照射方向偏转;最后,通过分段二维轨迹估计和分段成像克服运动误差的距离空变特性,实现大测绘带运动补偿。仿真结果表明,所提方法能够有效抑制偏航误差导致的点目标图像主瓣展宽、旁瓣升高和模糊现象。实验结果表明,相比于参考方法,所提方法在小偏航角和大偏航角条件下均显著提高了成像质量,有效校正了偏航造成的运动误差,实现了大测绘带精确成像。

合成孔径声呐时延估计误差校正改进算法

由相位解缠错误造成的时延估计误差会严重降低合成孔径声呐时延估计的精度。现有的时延估计误差校正算法采用二次函数作为时延的拟合模型,该模型在距离向近端和远端处不符合时延空变规律,拟合误差较大且无法估计载体的横荡和升沉运动。针对该问题,文章提出一种改进的时延估计误差校正算法,利用时延的距离空变函数代替二次函数作为拟合模型。数值仿真和实验结果表明,相较于参考算法,改进算法校正时延估计误差的效果更好、速度更快,同时还能准确地估计载体的横荡和升沉运动。运动补偿结果显示了改进算法能较好地提升成像质量。

舰船目标的合成孔径雷达成像研究

合成孔径雷达(SAR)成像技术已得到广泛应用。传统距离-多普勒(RD)算法在SAR成像中无法满足精度较高、运算率较小的需求,而线频调变标(CS)算法基于RD算法引入线性调频参考信号,可有效避免距离徙动校正中的插值计算,从而提高成像算法的计算效率,解决距离迁移问题。文中基于CS算法,从点目标拓展到简单立方体模型,最后对复杂舰船目标进行仿真成像,得到较为清晰的SAR仿真图像,验证了CS算法的有效性,可为复杂目标的辨别监测、反演研究,以及判断舰船航行方向及状态等提供一定的参考。