应用预测最小二乘准则的AR模型判阶方法

本文通过多种ＡＲ模型的判阶准则的比较提出了应用最小二乘预测误差的滑窗预测最小二乘（Ｓｌｉｄｉｎｇ Ｗｉｎｄｏｗ Ｐｒｅｄｉｃｔｉｖｅ Ｌｅａｓｔ Ｓｑｕａｒｅｓ，ＳＷＰＬＳ）判阶准则。采用这种准则的主要优点除了准确的判阶性能外，对于时变ＡＲ模型具有良好的跟踪特性，同时算法容易在线实时处理。文中主要对时变模型参数和时变模型阶数的多种情况进行了判阶模拟，验证了文中提出的滑窗最小二乘预测判阶准则的有效性     

LMD结合时变AR模型在瞬时频率测量中的应用

时变AR模型是一种新的时频分析方法,对简单的非平稳信号瞬时频率的测量具有很高的精度,但却不适用于复杂信号。为使时变AR模型同样适合于测量复杂信号的瞬时频率,提出LMD和时变AR模型相结合的方法,该方法使时变AR模型极大地扩展了适用对象。分别采用最小二乘和卡尔曼平滑算法求解信号参数模型,并求得信号瞬时频率,仿真对比分析表明,与直接用时变AR模型测量瞬时频率相比,该改进方法能大大提升瞬时频率测量的准确度。     

混沌时间序列局域偏最小二乘回归多步预测模型

针对混沌序列局域一阶多步预测问题,提出了基于偏最小二乘回归的混沌时间序列局域直接多步预测模型,偏最小二乘用于混沌时序重构相空间中演化轨迹前后相点信息间的建模。该模型克服了以往一阶局域单步预测模型进行多步预测时存在的误差累积,而且能抑制重构相空间中多重共线性的影响,提高了预测精度。试验中使用交叉验证方法将偏最小二乘的提取成分数。通过对Chen’s混沌序列和Mackey-Glass混沌序列的多步预测试验,验证了该模型在混沌时序预测方面具有很好的效果。     

递归实现固定阶的偏最小二乘问题——几种新的快速算法

本文提出了递归实现固定阶的偏最小二乘问题的三种新算法——无窗、预窗和滑动窗形式的超归一化快速算法,以满足不同条件下的实际应用。对于每个数据点,迭代这些算法所需的计算量正比于模型的阶数。     

基于混沌压缩感知的稀疏时变信号在线估计

混沌压缩感知是一种利用混沌系统实现非线性测量的压缩感知理论。针对稀疏时变信号的混沌压缩感知,该文提出稀疏时变信号的在线估计架构,构建一种递归最小二乘准则下的稀疏约束目标函数;通过利用迭代加权非线性最小二乘算法求解目标函数最小化问题,实现稀疏时变信号的参数估计。以Henon混沌系统为例仿真分析了频域时变稀疏信号的估计性能,数值模拟证明了该方法的有效性。     

一类Wiener非线性时变系统的迭代学习辨识

针对Wiener非线性时变系统的参数辨识问题,该文提出一种基于重复轴的迭代学习算法来实现对时变甚至突变参数的估计.文中将维纳系统输出非线性部分的反函数进行多项式展开,进而构造了回归模型,未知参数及中间变量用其估计替代,分别给出了采用迭代学习梯度算法和迭代学习最小二乘算法实现时变参数辨识的方法.仿真结果表明,与带遗忘因子的递推算法和迭代学习梯度算法相比,迭代学习最小二乘算法更具有参数估计收敛速度快,辨识精度高,系统输出误差小等优势,验证了所提学习算法的有效性.     

雷达成像近似二维模型及其超分辨算法

现有的雷达成像超分辨算法是基于目标回波信号的二维正弦信号模型,所以模型误差,特别是距离走动误差,将使算法性能严重下降或失败。为此,本文采用距离走动误差下的一阶近似雷达成像二维信号模型,提出了一种基于非线性最小二乘准则的参数化超分辨算法。在算法中,距离走动误差补偿与目标参量估计联合进行,文中同时给出了算法估计性能的Ｃｒａｍｅｒ－Ｒａｏ界及仿真结果。     

遗传规划和最小二乘法在数据拟合中的应用

为了分析和预测数据,需要预先确定其拟合函数模型.最小二乘法是一种常用的对测量数据进行数据拟合的方法,但当拟合函数模型未知时,该方法即失效.提出了一种联合遗传规划和最小二乘法寻求拟合模型的方法.利用遗传规划方法只需给出数据点及允许误差即可得到匹配的拟合函数式,并可对复杂函数式合理地简化.以此结果作为最小二乘法的拟合函数模型,进一步估计其中的参数,实现了对测量数据的更精确拟合.文中给出了应用实例,说明了本方法的有效性.     

基于自适应滑窗递归稀疏主成分分析的工业过程故障监测

本文提出一种自适应滑窗递归稀疏主成分分析方法,用于时变工业过程的在线故障监测.首先,通过滑窗提取正常过程数据空间的特征信息,并对当前窗口数据块矩阵进行稀疏主成分分析,构建稀疏主成分分析故障监测模型;然后,根据相邻窗口的相似度实时调整遗忘因子以自适应更新滑窗大小,使得所建立的稀疏主成分故障监测模型可以有效追踪复杂的时变过程;最后,通过递归更新滑窗稀疏载荷矩阵来动态更新故障监测模型.非线性数值仿真系统与田纳西-伊斯曼过程的故障监测结果表明,所提方法可以有效提高故障检测的准确率,适应于长流程时变工业过程在线故障监测.     

基于动态加权LS-SVM的网络流量混沌预测

网络流量是具有复杂非线性、不确定时变性的混沌时间序列.为提高标准最小二乘支持向量机的预测精度与自适应性,提出一种基于动态加权最小二乘支持向量机的网络流量混沌预测方法.该方法在标准LS-SVM回归机的训练样本误差设置时间权,增强对非线性样本的逼近能力.然后结合滚动窗与迭代求逆法实现模型动态在线校正,进而克服网络变化时的累积误差.仿真实验结果表明,相对常规LS-SVM,该模型能降低预测误差、减少计算时间,实现高精度实时混沌流量估计.     

Real-frequency stability criteria for linear time-varying systems

Bounds on the time-varying parameters that are sufficient to insure stability in linear time-varying systems are developed. The important and significant property of the stability criterion developed is that the bounds on the time-varying parameters can be established from the real-frequency characteristics of the time-invariant part of the system. Therefore, the stability criterion is applied easily in high-order systems. Techniques which permit the determination of decaying exponential bounds on the system signals are also presented. The application of the results obtained to parametric devices, satellite attitude control and adaptive systems is illustrated via examples.     

Robust Output-Feedback Control for Linear Continuous Uncertain State Delayed Systems with Unknown Time Delay

The state-delayed time often is unknown and independent of other variables in most real physical systems. In this paper, a new stability criterion for uncertain systems with a state time-varying delay is proposed. Then, a robust observer-based control law based on this criterion is constructed via the sequential quadratic programming method. We also develop a separation property so that the state feedback control law and observer can be independently designed and maintain closed-loop system stability. An example illustrates the availability of the proposed design method.     

Popov-like stability criterion for a class of time-varying continuous systems

Using functional-analysis methods, a Popov-like stability criterion is derived for use in the analysis of a class of time-varying control systems, consisting of a time-varying gain followed by a causal time-invariant linear convolution operator. The stability criterion provides a simple graphical means of obtaining stability information which complements the existing information for this class of systems.     

区间变时滞不确定系统鲁棒稳定性分析

本文针对一类线性区间变时滞不确定系统的鲁棒稳定性分析问题进行了研究.基于时滞中点法和凸组合技术,借助于构造一个包含四重积分项的新Lyapunov-Krasovskii（L-K）泛函,并利用积分不等式方法给出了LMI（Linear Matrix Inequality）形式的时滞相关稳定性新判据.与已有文献相比,该判据能大大降低理论推导和计算上的复杂性.最后通过三个具有代表性的数值例子对比验证了本文所提出方法在降低结论保守性方面的优越性.     

基于小波变换的语音净化新方法

语音信号由于其时变特性,传统的小波算法虽然能够衰减语音中含有的噪声,但易造成语音的失真。丈中提出了基于小波变换的语音净化新方法,改进了阈值的选择和小波系数量化算法。仿真实验结果表明,在去除噪声和提高信噪比方面,本文方法是一种有效语音净化方法。     

Signal selection in communication and radar systems

A criterion for selecting a finite set of transmitter signals for a continuous communication channel is proposed. The "optimum" signal sets using this criterion are selected to maximize the minimum divergence between hypothesis pairs being tested at the receiver. The resulting signal sets have the property that the error probability using these signals is less than the error probability for any other choice of signals for some {em a priori} message statistics. The signal selection procedure may be applied without a knowledge of the {em a priori} message statistics and does not require an evaluation of error probabilities. Four examples of signal selection are included to illustrate the procedure.     

A New Robust Stabilization Analysis Result for Uncertain Systems with Time-Varying Delay

Time delay and uncertainty are frequently encountered in various engineering systems. The robust stabilization problem for uncertain systems with time-delay has been paid more attentions by many researchers[1-8]. Most results of the robust stabilization are based on the systems matrices norm approach. The uncertainties of the systems discussed there must satisfy the so-called matching-conditions, which is more strict for the practical systems. The memoryless state feedback control gain matrice…     

Flexible decision-aided maximum likelihood phase estimation in coherent optical phase-shift-keying systems

Decision-aided maximum likelihood (DA-ML) phase estimation has been applied in coherent optical communication systems due to its high computational efficiency. However, conventional DA-ML scheme only assumes constant phase noise within each observation block, thus causing block length effect (BLE) which degrades system performance. In this paper, we take into account the time-varying laser phase noise and propose a flexible DA-ML phase estimation method for carrier phase recovery in coherent optical phase-shift-keying systems so as to eliminate BLE. Weighted coefficients based on ML criterion are introduced to strengthen the estimation accuracy. The statistical property of phase estimation error is derived, and the bit error rate (BER) performance is also evaluated. Numerical simulation results show that our flexible DA-ML scheme is very robust against time-varying phase noise. Compared with conventional DA-ML receiver, it can significantly reduce the phase estimation variance, improve the BER performance and increase the laser linewidth tolerance. By adopting the flexible DA-ML method with a relatively larger block length, BLE can be effectively eliminated. Thus, the BER performance can be significantly improved without carefully finding out the optimum block length or the optimum forgotten factor.     

Relevance feedback using generalized Bayesian framework with region-based optimization learning.

This paper presents a generalized Bayesian framework for relevance feedback in content-based image retrieval. The proposed feedback technique is based on the Bayesian learning method and incorporates a time-varying user model into the formulation. We define the user model with two terms: a target query and a user conception. The target query is aimed to learn the common features from relevant images so as to specify the user's ideal query. The user conception is aimed to learn a parameter set to determine the time-varying matching criterion. Therefore, at each feedback step, the learning process updates not only the target distribution, but also the target query and the matching criterion. In addition, another objective of this paper is to conduct the relevance feedback on images represented in region level. We formulate the matching criterion using a weighting scheme and proposed a region clustering technique to determine the region correspondence between relevant images. With the proposed region clustering technique, we derive a representation in region level to characterize the target query. Experiments demonstrate that the proposed method combined with time-varying user model indeed achieves satisfactory results and our proposed region-based techniques further improve the retrieval accuracy.     

Vector Autoregressive Model-Order Selection From Finite Samples Using Kullback's Symmetric Divergence

In this paper, a new small-sample model selection criterion for vector autoregressive (VAR) models is developed. The proposed criterion is named Kullback information criterion ($ KIC_ vc$), where the notation$ vc$stands for vector correction, and it can be considered as an extension of the KIC, for VAR models.$ KIC_ vc$adjusts KIC to be an unbiased estimator for the variant of the Kullback symmetric divergence, assuming that the true model is correctly specified or overfitted. Furthermore,$ KIC_ vc$provides better VAR model-order choices than KIC in small samples. Simulation results show that the proposed criterion selects the model order more accurately than other asymptotically efficient methods when applied to VAR model selection in small samples. As a result,$ KIC_ vc$serves as an effective tool for selecting a VAR model of appropriate order. A theoretical justification of the proposed criterion is presented.