Design of a fixed-interval smoother using covariance information based on the innovations approach in linear discrete-time stochastic systems

This paper describes a design for a recursive least-squares Wiener fixed-interval smoother using the covariance information in linear discrete-time stochastic systems. The estimators require information from the observation matrix, the system matrix for the state variable, related to the signal, the variance of the state variable, the cross-variance function of the state variable with the observed value and the variance of the white observation noise. It is assumed that the signal is observed with additive white noise.     

Design of recursive least-squares fixed-lag smoother using covariance information in linear continuous stochastic systems

This paper newly designs the recursive least-squares (RLS) fixed-lag smoother and filter using the covariance information in linear continuous-time stochastic systems. It is assumed that the signal is observed with additive white observation noise and the signal is uncorrelated with the observation noise. The estimators require the covariance information of the signal and the variance of the observation noise. The auto-covariance function of the signal is expressed in the semi-degenerate kernel form.     

Design of RLS fixed-lag smoother using covariance information in linear discrete stochastic systems

This paper newly designs the recursive least-squares fixed-lag smoother using the covariance information in linear discrete-time stochastic systems. It is assumed that the signal is observed with additive white observation noise and the signal is uncorrelated with the observation noise. The fixed-lag smoother uses the covariance function of the signal in the semi-degenerate kernel form and the variance of the observation noise. The proposed fixed-lag smoother is suitable for the estimations of stationary or non-stationary stochastic signals generally.     

RLS fixed-lag smoother using covariance information in linear continuous stochastic systems

This paper newly designs the recursive least-squares fixed-lag smoother using the covariance information in linear continuous-time stochastic systems. It is assumed that the signal is observed with additive white observation noise and the signal is uncorrelated with the observation noise. The fixed-lag smoother uses the covariance function of the signal in the semi-degenerate kernel form and the variance of the observation noise. The proposed fixed-lag smoother is appropriate for the estimations of stationary or non-stationary stochastic signals generally.     

Design of RLS Wiener estimators from randomly delayed observations in linear discrete-time stochastic systems

This paper presents the design of a new recursive least-squares (RLS) Wiener filter and fixed-point smoother based on randomly delayed observed values by one sampling time in linear discrete-time wide-sense stationary stochastic systems. The mixed observed value y(k) consists of the past observed value by one sampling time with the probability p(k) and of the current observed value at time k with the probability 1 − p(k). It is assumed that the delayed measurements are characterized by Bernoulli random variables. The observation is given as the sum of the signal z(k) and the white observation noise v(k). The RLS Wiener estimators explicitly require the following information: (a) the system matrix for the state vector; (b) the observation matrix; (c) the variance of the state vector; (d) the delayed probability p(k); (e) the variance of white observation noise v(k).     

Linear Kalman–Bucy Filter with Autoregressive Signal and Noise

In the Kalman—Bucy filter problem, the observed process consists of the sum of a signal and a noise. The filtration begins at the same moment as the observation process and it is necessary to estimate the signal. As a rule, this problem is studied for the scalar and vector Markovian processes. In this paper, the scalar linear problem is considered for the system in which the signal and noise are not Markovian processes. The signal and noise are independent stationary autoregressive processes with orders of magnitude higher than 1. The recurrent equations for the filter process, its error, and its conditional cross correlations are derived. These recurrent equations use previously found estimates and some last observed data. The optimal definition of the initial data is proposed. The algebraic equations for the limit values of the filter error (the variance) and cross correlations are found. The roots of these equations make possible the conclusions concerning the criterion of the filter process convergence. Some examples in which the filter process converges and does not converge are given. The Monte Carlo method is used to control the theoretical formulas for the filter and its error.     

Blind SNR estimation for limited time series

A blind SNR estimation method for discrete data is presented. The original noise-free data is assumed to follow a known signal model with an unknown attenuation. The SNR in noisy data is estimated using a polynomial fit obtained from the correlative characteristics between SNR and variance fractal dimension values. The bias and the “standard error” (root mean square error) of the estimator are used as performance measures. The simulated performance of the estimator for a specific signal model with real additive white Gaussian noise assumption is compared to that of a published decision-aided (nonblind) SNR estimator.     

Power-series approximations for the optimal feedback control of noisy nonlinear systems

Power-series methods are developed for designing approximately optimal state regulators for a nonlinear system subject to white Gaussian random disturbances. The performance index of the control is an ensemble average of a quadratic form. A perfect observation of the system state is assumed. When the system nonlinearity is small and it is characterized by a polynomial function of the state, a definite method is presented to construct a suboptimal feedback control of a power-series form in a small nonlinearity parameter. If the variance of noise is small, an alternative method is also applicable which yields a suboptimal control in a power series with respect to a variance parameter. A simple one-dimensional problem is examined to make comparison between controls of the two different forms.     

Seemingly unrelated regression models

The cross-covariance matrix of observation vectors in two linear statistical models need not be zero matrix. In such a case the problem is to find explicit expressions for the best linear unbiased estimators of both model parameters and estimators of variance components in the simplest structure of the covariance matrix. Univariate and multivariate forms of linear models are dealt with.     

Stabilized hard thresholding for an unknown noise level

The effect wa ys of estimating noise variance on the statistical characteristics of the stabilized hard thresholding of signal wavelet coefficients is studied. An unbiased estimator of the mean-square risk is analyzed. It is shown that under certain conditions, the estimator distribution tends to a normal law with variance that depends on the type of noise variance estimate.     

对左截断数据估计平稳序列

设平稳信号过程$\{X_t\}$被白噪声序列$\{Y_t\}$干扰. 只有$X_t>Y_t$时可以观测到信号过程$X_t$, 否则只能观测到白噪声$Y_t$. 这种数据模型被称为左截断数据模型. 本文在左截断数据模型下估计平稳信号过程的$\{X_t\}$均值, 自协方差函数, 和自相关系数. 证明所给的估计量是强相合估计. 当$X_t$是自回归序列时, 本文给出自回归模型的强相合的参数估计.     

Limit Theorems for Risk Estimate in Models with Non-Gaussian Noise

The problem of constructing an estimate of a signal function from noisy observations, assuming that this function is uniformly Lipschitz regular, is considered. The thresholding of empirical wavelet coefficients is used to reduce the noise. As a rule, it is assumed that the noise distribution is Gaussian and the optimal parameters of thresholding are known for various classes of signal functions. In this paper a model of additive noise whose distribution belongs to a fairly wide class, is considered. The mean-square risk estimate of thresholding is analyzed. It is shown that under certain conditions, this estimate is strongly consistent and asymptotically normal.     

The Asymptotic Distribution of REML Estimators

Restricted maximum likelihood (REML) estimation is a method employed to estimate variance-covariance parameters from data that follow a Gaussian linear model. In applications, it has either been conjectured or assumed that REML estimators are asymptotically Gaussian with zero mean and variance matrix equal to the inverse of the restricted information matrix. In this article, we give conditions under which the conjecture is true and apply our results to variance-components models. An important application of variance components is to census undercount; a simulation is carried out to verify REML′s properties for a typical census undercount model.     

Kalman滤波的自适应算法

1 引 言 本文,我们讨论时不变线性随机系统 这里A、Γ和C分别是已知的n×n,n×p和q×n阶常数矩阵,1≤p,q≤n,且{ξ_k}{η_k}是均 值为零的高斯白噪声序列,有     

Robustesse de la solution des problemes de filtrage avec bruit blanc independant †

We suppose that the observation of a signal process is corrupted by an independent white noise. In a case which is more general than the classical case of semimartingales, we construct the optimal estimate of the signal as a continuous function on the space of observation trajectories. We show that moreover, it is continuous in a certain sense with respect to the a priori law of the signal     

The conditioning of least‐squares problems in variational data assimilation

In variational data assimilation a least‐squares objective function is minimised to obtain the most likely state of a dynamical system. This objective function combines observation and prior (or background) data weighted by their respective error statistics. In numerical weather prediction, data assimilation is used to estimate the current atmospheric state, which then serves as an initial condition for a forecast. New developments in the treatment of observation uncertainties have recently been shown to cause convergence problems for this least‐squares minimisation. This is important for operational numerical weather prediction centres due to the time constraints of producing regular forecasts. The condition number of the Hessian of the objective function can be used as a proxy to investigate the speed of convergence of the least‐squares minimisation. In this paper we develop novel theoretical bounds on the condition number of the Hessian. These new bounds depend on the minimum eigenvalue of the observation error covariance matrix and the ratio of background error variance to observation error variance. Numerical tests in a linear setting show that the location of observation measurements has an important effect on the condition number of the Hessian. We identify that the conditioning of the problem is related to the complex interactions between observation error covariance and background error covariance matrices. Increased understanding of the role of each constituent matrix in the conditioning of the Hessian will prove useful for informing the choice of correlated observation error covariance matrix and observation location, particularly for practical applications.     

Blind Channel Identification Based on Noisy Observation by Stochastic Approximation Method

A stochastic approximation algorithm for estimating multichannel coefficients is proposed, and the estimate is proved to converge to the true parameters a.s. up-to a constant scaling factor. The estimate is updated after receiving each new observation, so the output data need not be collected in advance. The input signal is allowed to be dependent and the observation is allowed to be corrupted by noise, but no noise statistics are used in the estimation algorithm.     

Optimality and robustness of combinations of moving averages

A combination of moving averages has been shown previously to be more accurate than simple moving averages, under certain conditions, and to be more robust to non-optimal parameter specification. However, the use of the method depends on specification of three parameters: length of greater moving average, length of shorter moving average, and the weighting given to the former. In this paper, expressions are derived for the optimal values of the three parameters, under the conditions of a steady state model. These expressions reduce a three-parameter search to a single-parameter search. An expression is given for the variance of the sampling error of the optimal combination of moving averages and this is shown to be marginally greater than that for exponentially weighted moving averages (EWMA). Similar expressions for optimal parameters and the resultant variance are derived for equally weighted combinations. The sampling variance of the mean of such combinations is shown to be almost identical to the optimal general combination, thus simplifying the use of combinations further. It is demonstrated that equal weight combinations are more robust than EWMA to noise to signal ratios lower than expected, but less robust to noise to signal ratios higher than expected.     

Removal of blur in images based on least squares solutions

We propose an image restoration method. The method generalizes image restoration algorithms that are based on the Moore–Penrose solution of certain matrix equations that define the linear motion blur. Our approach is based on the usage of least squares solutions of these matrix equations, wherein an arbitrary matrix of appropriate dimensions is included besides the Moore–Penrose inverse. In addition, the method is a useful tool for improving results obtained by other image restoration methods. Toward that direction, we investigate the case where the arbitrary matrix is replaced by the matrix obtained by the Haar basis reconstructed image. The method has been tested by reconstructing an image after the removal of blur caused by the uniform linear motion and filtering the noise that is corrupted with the image pixels. The quality of the restoration is observable by a human eye. Benefits of using the method are illustrated by the values of the improvement in signal‐to‐noise ratio and in the values of peak signal‐to‐noise ratio. Copyright © 2013 John Wiley & Sons, Ltd.     

Monitoring an Exponential Smoothing Forecasting System

The first part of this paper is concerned with the variance of the smoothed error when the forecasting system being used is exponential smoothing. The expression derived for this variance involves the variance of the noise, the smoothing constant and a sum of squared binomial coefficients. It is also shown that the variance of the sum of errors equals the variance of the smoothed error for one less degree of freedom divided by the square of the smoothing constant.The second part of the paper considers the practical application of the above result and also gives values for the tracking signal limits, obtained by simulation, which could be used in the automatic monitoring of a forecasting system.