基于PCA和核FDA的人脸识别研究

基于主分量分析的特征脸识别是人脸识别中重要的识别方法,具有简单、实用等特点。Fisher判别分析是统计分析一种常用的降维方法,多类Fisher判别分析在模式识别领域得到广泛应用。核方法技术是设计全局非线性算法最流行的工具之一,应用核方法技术使得低维空间线性不可分的样本在高维空间线性可分。先对ORL人脸数据库中的图像应用主分量分析提取主分量,再应用核Fisher判别方法把特征向量做隐式变换,最后把得到的数据采用k-紧邻分类器进行分类识别,并对实验结果做了比较分析。     

基于子空间的线性约束最小二乘恒模算法

线性约束最小二乘恒模算法能够有效克服恒模算法中存在的干扰捕获问题,然而,在实际的系统中,由于受到来自噪声子空间分量的影响,从而导致性能下降。针对这一问题,该文提出了一种基于子空间方法的线性约束最小二乘恒模(SLSCM)多用户检测算法并对其性能进行了理论分析。该算法收敛速度快,信道跟踪能力强,有较好的输出信干噪比和误码率性能。理论分析和仿真实验验证了该文算法的有效性和优越性。     

基于核邻域保持投影的人脸识别

提出了一种有效的非线性子空间学习方法:核邻域保持投影.其主要思想是通过引入线性变换矩阵来近似经典的局部线性嵌入(LLE),然后通过核方法的技巧在高维空间里求解.经过推导,实际的子空间的计算可归结为标准的特征值分解问题而非推广的特征值分解问题.在AR人脸数据库上的试验表明该方法是有效的.     

基于GA-LS指数衰减正弦信号参数估计算法

本文提出基于GA-LS指数衰减正弦信号的参数估计方法,根据系统模型特性将待估计参数分成两类,非线性类和线性类,将遗传算法和最小二乘法估计结合起来对参数进行估计。在观察与分析信号的频域特征后,确定各非线性部分的参数范围,采用遗传算法在参数空间里搜索,再用当下最优的非线性类参数,通过最小线性二乘法确定线性类参数。通过缩小了遗传算法的搜索空间维数,考虑了系统的线性部分参数间的约束关系,达到提高参数的搜索效率,提高参数的估计精度。大量仿真实验表明算法收敛速度优于标准遗传算法,精度高。用于估计冲击实验的采集信号满足实际需要。     

非线性主动噪声控制的随机傅里叶特征-核滤波算法

传统的线性主动噪声控制算法在噪声信号或噪声通道呈现非线性特性的情况下控制效果不佳。核-滤波最小均方误差算法(Kernel Filtered x Least Mean Square,KFxLMS)通过将输入噪声信号映射到高维再生核希尔伯特空间,再用线性方法在高维空间中进行处理。然而,随着新噪声信号的输入,KFxLMS算法递增的核函数运算需要较高的成本。为进一步改进KFxLMS算法,本文提出了随机傅里叶特征核滤波最小均方误差算法(Random Fourier Feature - Kernel Filtered x Least Mean Square,RFF-KFxLMS)。在仿真实验部分讨论了算法的参数选择,给出了算法的计算耗时,并验证了提出的RFF-KFxLMS算法在非线性噪声通道情况下,针对不同频率分量的正弦噪声都能够达到理想的性能。      

基于LS-SVM的天线指向模型研究

提出了一种新的基于最小二乘支持向量机的天线指向模型(LSSVM-PM).不同于线性指向模型(Linear-PM)对天线指向偏差的分项进行代数相加,LSSVM-PM是对影响天线指向的因素进行考察,将因素变量作为支持向量机的输入,指向偏差作为输出,模型的求解是将原空间非线性地映射到一个高维的特征空间,然后在此空间中引入最小二乘法进行优化求解.因此,LSSVM-PM模型可解决天线指向偏差的非线性问题,同时将影响指向的外部因素考虑周全.实验数据模拟结果表明,该模型与线性模型相比,指向预测精度提高了17.36%,能更有效地提高天线的指向精度.最后,对LSSVM-PM模型的参数选择、学习样本数量及分布对模型精度的影响进行了分析.     

投影近似子空间跟踪算法及其在盲线性多用户检测中的应用

本文针对CDMA系统中的多用户检测,探讨信号子空间作为一种无约束最小化问题的解这一新的解释,在此基础上分析并研究了一种新的跟踪信号子空间的方法PASTd算法。分析表明,通过适当的投影近似,最小化任务简化为众所周知的指数加权最小二乘问题,从而可以运用递归最小二乘(RLS)方法来有效地跟踪信号子空间。该算法的计算复杂度为O(nr)。计算机仿真结果证明了PASTd算法的有效性,从而该算法对于基于子空间的盲线性多用户检测器的实现具有研究和参考意义。     

基于等距映射和空间信息的光谱解混算法

由于多重反射和散射,高光谱图像中的混合像元实际上是非线性光谱混合。传统的光谱解混算法是以线性光谱混合模型为基础,因此解混精度不高。本文在光谱非线性混合模型的基础上,提出一种将等距映射与空间信息结合的非线性光谱解混算法。该算法通过等距映射算法将原始高光谱数据非线性降维到低维空间,并结合空间信息实现端元提取。得到的端元采用全约束的最小二乘法计算相应丰度。真实高光谱遥感数据实验结果表明,采用该算法得到的结果优于N-FINDR算法和基于测地线距离的最大单形体体积(GSVM)算法。     

基于线性最小二乘方法的主动段目标初值估计

主动段目标初值估计是导弹目标战术参数估计的重要内容,是监视系统进行实时跟踪的前提条件。在天基观测背景下,该问题是非线性最小二乘求解问题。该文建立了主动段8态重力转弯模型和天基测量模型,通过用常加加速度线性模型来近似目标运动学参数,并对非线性测量进行伪线性化处理,将非线性最小二乘转化为线性最小二乘求解问题。特别地,给出了天基伪线性化测量及其统计特征的详尽推导,并进一步考察了测量伪线性化方法的适用范围。仿真实验中,通过与CRLB和传统的Gauss-Newton迭代方法比较,证明了该文方法在估计精度和运算效率上的优势。     

MC-CDMA系统中一种线性共轭MOE多用户检测算法

该文提出MC-CDMA系统下一种基于递归最小二乘(Recursive Least-Squares, RLS)的最小输出能量(Minimum Output Energy, MOE)噪声抑制线性共轭多用户检测算法.该算法定义了一种新的基于MOE准则的代价函数,同时将噪声子空间作为MOE代价函数的约束条件,设计了一种噪声抑制的线性共轭检测器,并采用RLS算法自适应得到权向量.所提算法将权向量和噪声子空间正交,消除了权向量中的噪声分量,并且利用了伪自相关矩阵的信息,从而提高了系统的性能.仿真结果证明了本文算法的有效性和优越性.     

基于约束最小二乘的近空间雷达网定位算法

该文研究了一种基于归一化约束最小二乘的近空间雷达网定位算法。首先将距离与角度信息的非线性方程转换为线性方程,通过一阶Taylor展开分析了噪声对线性方程的影响,然后将定位问题转化为归一化约束总体最小二乘问题,并通过Lagrange函数将其转换为无约束的优化问题,根据定位均方误差最小原则选取加权因子得到定位解,最后进行了定位误差分析,仿真结果表明了该算法的有效性。     

Low-Cost 3D Bluetooth Indoor Positioning with Least Square

This paper discusses low-cost 3D indoor positioning with Bluetooth smart device and least square methods. 3D indoor location has become more and more attractive and it hasn’t been well resolved. Almost each smart phone has a Bluetooth component and it can be used for indoor positioning and navigation in the nature of things. Least square algorithms are the powerful tools for linear and nonlinear parameters estimation. Various linear and nonlinear least square methods and their theoretical basics and application performance for indoor positioning have been studied. Simulation and hardware experiments results prove that nonlinear least square method is suitable for parameters estimation of Bluetooth signal propagation, and generalized least square method has better performance than total least square methods. Simulation and hardware experiments results also show that proposed method has the advantages of low cost, lost power consumption, perfect availability and high location accuracy.     

State-dependent vector hybrid linear and nonlinear ARMA modeling: Theory

A new model is proposed to represent a general vector nonstationary and nonlinear process by setting up a state-dependent vector hybrid linear and nonlinear autoregressive moving average (SVH-ARMA) model. The linear part of the process is represented by a vector ARMA model, the nonlinear part is represented by a vector nonlinear ARMA model employing a multilayer feedforward neural network, and the nonstationary characteristics are captured with a hidden Markov chain. Based on a unifiedQ-likelihood function, an expectation-maximization algorithm for model identification is derived, and the model parameters are estimated by applying a state-dependent training and nonlinear optimization technique iteratively, which finally yields maximum likelihood estimation of model parameters. This model can track the nonstationary varying of a vector linear and/or nonlinear process adaptively and represent a vector linear and/or nonlinear system with low order. Moreover, it is able to characterize and track the long-range, second-order correlation features of many time series and thus can be used for reliable multiple step ahead prediction. Some impressive applications of the SVH-ARMA model are being presented in the companion paper by Zheng et al., pp. 575–597, this issue.     

用于低维混沌时间序列预测的一种非线性自适应预测滤波器

在二阶Volterra滤波器基础上,提出了一种用于低维混沌时间自适应预测的非线性自适应预测器。基于最小均方误差准则导出了一种NLMS类型的自适应算法来实时调整这种非线性滤波预测器的系数,仿真实验结果表明：这种线性化的非线性自适应滤波预测器能够有效地预测低维混时间序列,且它的模块化特征更易于VLSI电路实现,具有广泛的工程应用价值。     

基于噪声子空间跟踪的OFDM盲信道估计

基于子空间分解的OFDM信道估计算法利用信号子空间和噪声子空间的正交性可以对信道参数进行盲估计,但是子空间分解的运算量大,使盲信道估计算法的实用化受到限制。本文在噪声子空间自适应跟踪的基础上进行OFDM盲信道估计,显著降低了运算量。仿真结果表明,在适当选择学习因子后可以实现对噪声子空间的快速跟踪,在噪声子空间跟踪的基础上得到的信道估计性能接近于子空间分解法。     

Nonlinear Bayesian Filtering Using the Unscented Linear Fractional Transformation Model

For nonlinear state space model involving random variables with arbitrary probability distributions, the state estimation given a sequence of observations is based on an appropriate criterion such as the minimum mean square error (MMSE). This leads to linear approximation in the state space of the extended Kalman filter (EKF) and the unscented Kalman filter (UKF), which work reasonably well only for mildly nonlinear systems. We propose a Bayesian filtering technique based on the MMSE criterion in the framework of the virtual linear fractional transformation (LFT) model, which is characterized by a linear part and a simple nonlinear structure in the feedback loop. LFT is an exact representation for any differentiable nonlinear mapping, so the virtual LFT model is amenable to a wide range of nonlinear systems. Simulation results demonstrate that the proposed filtering technique gives better approximation and tracking performance than standard methods like the UKF. Furthermore, for highly nonlinear systems where UKF diverges, the LFT model estimates the conditional mean with reasonable accuracy.      

基于线性噪声特征值的信源数估计

为了提高信源数目估计的准确性,本文通过对采样协方差矩阵特征值的分析,得出噪声子空间对应的特征值近似线性而信号子空间对应的特征值不在该直线上的性质,根据此性质提出一种简单有效的基于最小二乘直线拟合的信源数估计算法,可以在低信噪比的情况下准确估计信号源数目。     

Adaptive model-based motion estimation

A general discrete-time, adaptive, multidimensional framework is introduced for estimating the motion of one or several object features from their successive nonlinear projections on an image plane. The motion model consists of a set of linear difference equations with parameters estimated recursively from a nonlinear observation equation. The model dimensionality corresponds to that of the original, nonprojected motion space, thus allowing to compensate for variable projection characteristics such as paning and zooming of the camera. Extended recursive least-squares and linear-quadratic tracking algorithms are used to adaptively adjust the model parameters and minimize the errors of either smoothing, filtering or prediction of the object trajectories in the projection plane. Both algorithms are derived using a second order approximation of the projection nonlinearities. All the results presented here use a generalized vectorial notation suitable for motion estimation of any finite number of object features and various approximations of the nonlinear projection. The application of the model-based motion estimator for temporal decimation/interpolation in digital video sequence compression systems is presented.     

A linear CMOS transconductance element of an adaptively biased source-coupled differential pair using a quadritail cell

A novel circuit design technique for realizing a linear CMOS transconductance element, consisting of an adaptively biased source-coupled differential pair using a quadritail cell, is proposed. In the circuitry, the quadritail cell, which provides an output current proportional to the square of a differential input voltage, cancels a nonlinear term of the source-coupled differential pair. The circuit have a superior linearity and a wide linear input voltage range compared with the conventional linear CMOS transconductance elements because the transconductance characteristic is theoretically linear over wide input voltage range when all the MOS field-effect transistors (MOSFETs) are operating in the saturation region and the MOSFETs' behaviors are according to the relation based on the square-law characteristic. The proposed adaptively biased source-coupled differential pair was verified by using transistor-arrays and discrete components on a breadboard.