Algorithm for Generating Samples of Homogeneous Gaussian Fields

A new algorithm is developed for generating samples of stationary Gaussian random fields. The algorithm is based on a model derived from the spectral representation theorem for weakly stationary random fields. The model consists of a superposition of a random number of waves with random amplitude and frequency, can match the second moment properties of any target random field, and becomes Gaussian as the intensity of two independent Poisson processes, defining the number of waves and their frequencies, increases indefinitely. In contrast to the current Monte Carlo simulation algorithms, the proposed algorithm: (1) does not produce periodic samples; and (2) does not require the discretization of the frequency domain. The proposed Monte Carlo algorithm is applied to generate samples of a stationary Gaussian field defined on 2.     

Statistical methods for detecting activated regions in functional MRI of the brain

Two statistical tests for detecting activated pixels in functional MRI (fMRI) data are presented. The first test (t-test) is the optimal solution to the problem of detecting a known activation signal in Gaussian white noise. The results of this test are shown to be equivalent to the cross-correlation method that is widely used for activation detection in fMRI. The second test (F test) is the optimal solution when the measured data are modeled to consist of an unknown activation signal that lies in a known lower dimensional subspace of the measurement space with added Gaussian white noise. A model for the signal subspace based on a truncated trigonometric Fourier series is proposed for periodic activation-baseline imaging paradigms. The advantage of the second method is that it does not assume any information about the shape or delay of the activation signal, except that it is periodic with the same period as the activation-baseline pattern. The two models are applied to experimental echo-planar fMRI data sets and the results are compared.     

Moment Equation Analysis of Base-Isolated Buildings Subjected to Support Motion

The stationary response of base-isolated buildings subjected to support motion is studied. The isolation device consists of a friction device, while the ground motion is assumed to be a stationary Gaussian white noise random process. The moment equation approach is preferred in this study to characterize the structural response statistically. A non-Gaussian closure is adopted to reveal the degree of non-Gaussianness of the response. The proposed approach is compared with the stochastic equivalent linearization and the numerical simulation, resulting in more accuracy when predicting the moments of the response which is markedly non-Gaussian. The models of cascaded and coupled system for the base isolator and the building are also compared, finding the former to be quite in error.     

Influence of oxygen partial pressure on human and mouse myeloid cell line characteristics

An optimal maximum likelihood (ML) method is described for an unbiased estimation of monoexponential T2 from magnitude spin-echo images. The algorithm is based on a Gaussian assumption of noise distribution. The validity of this assumption was checked by a statistical chi 2 test on spin-echo and fast low-angle shot surface coil images. Monte-Carlo simulations of magnitude data showed that the ML estimate standard deviation is lower than that produced by a weighted least-squares fitting on signal logarithm. Correction schemes are proposed to reduce bias deriving from magnitude reconstruction. The variance of the ML estimate converged rapidly toward the theoretical algebraic expression of the Cramér-Rao lower bound.     

陡脉冲干扰下的心电信号滤波及QRS提取

为消除陡脉冲带来的干扰,分析了陡脉冲干扰的特点,建立了陡脉冲噪声数学模型,提出了基于变分模态分解（Variational mode decomposition, VMD）的心电信号滤波算法,提取叠加在心电信号中陡脉冲干扰分量、识别陡脉冲干扰分量并剔除陡脉冲干扰分量;为减少VMD分解层数、提高实时性并减少内存消耗,提出了心电信号预处理算法;针对医疗环境中的随机噪声伴随陡脉冲出现的情况,分析了VMD后子信号中随机噪声的特点,提出了基于VMD子信号能量估计的阈值去噪算法;利用变分模态分解的带通滤波器组特性,提出了基于变分模态分解子信号重组的QRS波群检测算法,配合滤波算法以提高心电信号特征检测精度。以添加了高斯白噪声和模拟陡脉冲干扰的MIT?BIH数据库心电信号和医疗环境中采集的心电信号为实验对象,分别实现对滤波算法和QRS波群检测算法的定量对比分析。      

Mixed Bayesian networks: a mixture of Gaussian distributions

Mixed Bayesian networks are probabilistic models associated with a graphical representation, where the graph is directed and the random variables are discrete or continuous. We propose a comprehensive method for estimating the density functions of continuous variables, using a graph structure and a set of samples. The principle of the method is to learn the shape of densities from a sample of continuous variables. The densities are approximated by a mixture of Gaussian distributions. The estimation algorithm is a stochastic version of the Expectation Maximization algorithm (Stochastic EM algorithm). The inference algorithm corresponding to our model is a variant of junction three method, adapted to our specific case. The approach is illustrated by a simulated example from the domain of pharmacokinetics. Tests show that the true distributions seem sufficiently fitted for practical application.     

Analysis of Class of Nonlinear System under Deterministic and Stochastic Excitations

In this paper, some analysis techniques of nonlinear dynamics are applied to physical systems which may be modeled by the Duffing nonlinear differential equation. The response of the Duffing oscillator to both deterministic sinusoidal and stochastic loadings is investigated and distinct regimes of the response motion are discerned and discussed. The stochastic input to the system is low-pass Gaussian white noise. The efficacy of studying the variation in time of the probability density of one or more of the system output states to determine the type of motion of the system is examined. Attractors in phase space are defined via Poincaré mapping and bounds on motion which serve as signatures for particular types of motion (e.g., chaotic, periodic) are identified by a hypervolume measurement technique. An accepted method for adapting one measured output state into a higher dimensional space by using time-delayed coordinates is used in conjunction with correlation dimension calculation to supply a lower-bound estimate of the fractal dimension and insight into the character of the motion of a nonlinear dynamic system.     

Modal Identification Algorithm with Unmeasured Input

This paper investigates the possibility of generating a time‐domain modal identification algorithm that does not need measurements of the input signals. Such a technique is useful when complete input data acquisition cannot be performed. The approach is based on the Yule‐Walker equations, extended to the case where the input signal is not white noise. The algorithm is written recursively both to minimize data acquisition and to be flexible enough when time‐varying modal parameters are tracked. Natural frequencies and damping ratios are extracted with an error magnitude inferior to impulse response techniques but superior to methods using the input time history. From a computational aspect, the algorithm only introduces scalar inversions, which presents an important gain of time and stability. Examples and comparisons with other techniques are presented. The case where a change in the modal parameter values occur is also highlighted. A normalized error, taking into account the quality and the swiftness of the new estimation, is introduced.     

Identification of a Distribution of Stiffness Reduction in Reinforced Concrete Slab Bridges Subjected to Moving Loads

The method for identifying arbitrary stiffness reduction in damaged reinforced concrete slab bridges under moving loads is proposed and dynamic signals measured at several points are used as response data to reflect the properties of the moving loads sensitivity. In particular, the change in stiffness in each element before and after damage, based on the system identification method, is described and discussed by using a modified bivariate Gaussian distribution function. The proposed method in this work is more feasible than the conventional element-based damage detection method from the computational efficiency because the procedure of finite-element analysis coupled with microgenetic algorithm using six unknown parameters irrespective of the number of elements are considered. The validity of the technique is numerically verified using a set of dynamic data obtained from a simulation of the actual bridge modeled with a three-dimensional solid element. The numerical calculations show that the proposed technique is a feasible and practical method that can prove the exact location of a damaged region as well as inspect the complex distribution of deteriorated stiffness, although there is a modeling error between actual bridge results and numerical model results as well as a measurement error like uncertain noise in the response data.     

联合改进作用距离和Dijkstra算法的复杂结构声发射定位方法及应用

声发射定位对含空区等复杂结构的连续动态安全监测具有重要意义。针对直线路径定位方法不适用于复杂结构,而常规Dijkstra搜索算法常出现局部最优路径等问题,提出了一种联合改进作用距离和Dijkstra算法的声发射定位方法,实现复杂结构下声发射的高精度定位。理论测试显示：基于改进作用距离的Dijkstra算法得到的P波传播路径长度小于等于常规作用距离的Dijkstra算法,即P波走时更为准确。理论测试和断铅试验定位测试表明：本文提出的定位方法在复杂结构声发射定位时的整体误差在0.50 cm范围之内,断铅事件平均定位误差由常规作用距离Dijkstra算法的0.95 cm下降至本文的0.54 cm。改进的声发射定位方法在复杂结构声发射定位方面具有很好的应用前景。     

Direct Generation of Non-Gaussian Weighted Integrals

A simple numerical algorithm is proposed for directly generating realizations of weighted integrals (or, similarly, local averages) of non-Gaussian random fields for use in simulation-based stochastic finite-element analyses. The method uses a Gaussian quadrature integration rule to numerically evaluate an individual weighted integral (or local average), thus reducing the computation of the integral to a summation of a small number of properly weighted non-Gaussian random variables. Consequently, the need to generate actual realizations of the non-Gaussian random field is eliminated. The vector of non-Gaussian random variables is obtained from a nonlinear mapping of a vector of properly correlated Gaussian random variables, which in turn is obtained from a vector of uncorrelated Gaussian random variables using modal decomposition. The “proper” correlation structure of the Gaussian random variables is established a priori from the correlation structure of the non-Gaussian random variables, which itself is established a priori from the known or desired correlation structure of the non-Gaussian random field. Numerical results are provided to demonstrate the statistical equivalence of weighted integrals (or local averages) generated using the proposed approach to those computed using conventional numerical integration of actual realizations of the non-Gaussian random field.     

Nonparametric modeling of oxygen-converter processes

Preliminary data analysis in the identification of multidimensional discrete–continuous processes is considered. A method is proposed for generating a working sample from an initial training sample consisting of normal operating data. The method somewhat resembles the bootstrap process. In the present case, the process begins with a training sample that reflects the properties of the object to be identified. By means of the proposed method, the unknown stochastic dependence at the limit of definition of the corresponding input–output variables for the object may be automatically derived. The identification of the oxygen-converter process in converter shop 2 at OAO EVRAZ ZSMK is considered in the case with insufficient available information and gaps in the observation sample. The model is based on a new working sample containing both the measurements and data generated by the proposed method. By using the working sample as a training sample, the precision of identification is doubled.     

Sensitivity analysis of kernel estimates: implications in nonlinear physiological system identification

Many techniques have been developed for the estimation of the Volterra/Wiener kernels of nonlinear systems, and have found extensive application in the study of various physiological systems. To date, however, we are not aware of methods for estimating the reliability of these kernels from single data records. In this study, we develop a formal analysis of variance for least-squares based nonlinear system identification algorithms. Expressions are developed for the variance of the estimated kernel coefficients and are used to place confidence bounds around both kernel estimates and output predictions. Specific bounds are developed for two such identification algorithms: Korenberg's fast orthogonal algorithm and the Laguerre expansion technique. Simulations, employing a model representative of the peripheral auditory system, are used to validate the theoretical derivations, and to explore their sensitivity to assumptions regarding the system and data. The simulations show excellent agreement between the variances of kernel coefficients and output predictions as estimated from the results of a single trial compared to the same quantities computed from an ensemble of 1000 Monte Carlo runs. These techniques were validated with white and nonwhite Gaussian inputs and with white Gaussian and nonwhite non-Gaussian measurement noise on the output, provided that the output noise source was independent of the test input.     

A GNSS Signal Blind-decoding Algorithm at Low SNR

The pseudo-random noise (PRN) code modulated in satellite navigation signals impacts the system positioning performance directly, and the code monitoring is one of the key technologies. However, the received signal is often buried in noise, and the ranging codes can not visible in time domain. Considering local clock bias, the signal model in transmis-sion link is derived in this paper, and a PRN code blind-decoding method is proposed also. It calculates the signal' s cyclic spectrum by using fast Fourier transform accumulation method (FAM) , and estimates the code rate and Doppler frequency making use of the noise eliminating characteristic in non-zero cycle frequency cross-section. Wiped off the Doppler shift,the navigation message or secondary code bits are determined and removed by slide-correlating a small slice of itself with the whole data. The start of the code is determined by stacking multiple periods of the whole data into a code period, and then the whole data is shifted to the start of the PRN code, and is restacked. Then the individual period of PRN code is esti-mated. An experiment for the proposed algorithm is performed by simulated vector signal analyzer (VSA) collected data.The results indicate that the algorithm is effective and reliable.     

Psychophysical evidence for fast region-based segmentation processes in motion and color

Theories of image segmentation suggest that the human visual system may use two distinct processes to segregate figure from background: a local process that uses local feature contrasts to mark borders of coherent regions and a global process that groups similar features over a larger spatial scale. We performed psychophysical experiments to determine whether and to what extent the global similarity process contributes to image segmentation by motion and color. Our results show that for color, as well as for motion, segmentation occurs first by an integrative process on a coarse spatial scale, demonstrating that for both modalities the global process is faster than one based on local feature contrasts. Segmentation by motion builds up over time, whereas segmentation by color does not, indicating a fundamental difference between the modalities. Our data suggest that segmentation by motion proceeds first via a cooperative linking over space of local motion signals, generating almost immediate perceptual coherence even of physically incoherent signals. This global segmentation process occurs faster than the detection of absolute motion, providing further evidence for the existence of two motion processes with distinct dynamic properties.     

Bayesian Modal Updating using Complete Input and Incomplete Response Noisy Measurements

The problem of identification of the modal parameters of a structural model using complete input and incomplete response time histories is addressed. It is assumed that there exist both input error (due to input measurement noise) and output error (due to output measurement noise and modeling error). These errors are modeled by independent white noise processes, and contribute towards uncertainty in the identification of the modal parameters of the model. To explicitly treat these uncertainties, a Bayesian framework is adopted and a Bayesian time-domain methodology for modal updating based on an approximate conditional probability expansion is presented. The methodology allows one to obtain not only the optimal (most probable) values of the updated modal parameters but also their uncertainties, calculated from their joint probability distribution. Calculation of the uncertainties of the identified modal parameters is very important if one plans to proceed with the updating of a theoretical finite-element model based on these modal estimates. The proposed approach requires only one set of excitation and corresponding response data. It is found that the updated probability density function (PDF) can be well approximated by a Gaussian distribution centered at the optimal parameters at which the posterior PDF is maximized. Numerical examples using noisy simulated data are presented to illustrate the proposed method.     

铜冶炼分析X射线荧光仪工作标准样的研制

通过对铜冶炼分析的X射线荧光工作标准样品的制备、化学分析法定值、压片法制样、建立标准曲线、用基本参数法和经验系数法进行校正、最后进行精密度和准确度验证的全部过程。分析数据表明,这套用于铜冶炼控制分析的X射线荧光工作标样在制备和应用上都比较满意。     

基于局部特征和稀疏表示的鲁棒人耳识别方法

作为图像局部特征区域的有效描述方法,局部二值模式是目前对二维图像最有效的纹理分析特征之一.本文提出了基于局部二值模式特征的稀疏表示人耳识别方法.该识别算法首先提取训练人耳图像的局部二值模式特征描述子作为稀疏表示的字典,然后将测试样本的局部二值模式特征描述子表示为字典中所有局部二值模式原子的稀疏线性组合,最后通过求解稀疏表示模型得到稀疏编码系数,根据测试人耳图像的重建误差进行识别.在UND-J2人耳库和USTB人耳库上的实验结果表明,基于局部二值模式特征的稀疏表示人耳识别方法对人耳图像光照变化、姿态变化以及人耳遮挡具有更好的鲁棒性,实现了更高的识别率.      

Probabilistic Models and Simulation of Irregular Masonry Walls

A method is proposed for generating samples of irregular masonry walls that capture the essential statistics of a given population. The method first entails characterizing the geometry of scaled star-like inclusions by means of a non-Gaussian random field model and second packing these inclusions together to form a virtual material specimen. The model used in the first step is a nonlinear memoryless mapping of a sum of harmonic functions with Gaussian coefficients while in the second step the model proposed transforms Poisson fields into a domain of inclusions with a sieving curve that matches the sample specimen. The two random field models are used to develop Monte Carlo algorithms which produce virtual material specimens that include two levels of probabilistic characterization, a first level that is correlated to the inclusion geometry, and a second that is dictated by the global morphology of the sample material specimen.