Stochastic stability of the derivative unscented Kalman filter

This is the second of two consecutive papers focusing on the filtering algorithm for a nonlinear stochastic discretetime system with linear system state equation. The first paper established a derivative unscented Kalman filter(DUKF) to eliminate the redundant computational load of the unscented Kalman filter(UKF) due to the use of unscented transformation(UT) in the prediction process. The present paper studies the error behavior of the DUKF using the boundedness property of stochastic processes. It is proved that the estimation error of the DUKF remains bounded if the system satisfies certain conditions. Furthermore, it is shown that the design of the measurement noise covariance matrix plays an important role in improvement of the algorithm stability. The DUKF can be significantly stabilized by adding small quantities to the measurement noise covariance matrix in the presence of large initial error. Simulation results demonstrate the effectiveness of the proposed technique.     

自适应光学系统变形镜控制电压预测

在校正大气湍流畸变波前相差的自适应光学系统中,利用基于Levenberg-Marquardt学习算法的非线性反向传播神经网络技术（LMBP）对变形镜控制电压进行预测。以对受横向风影响的大气湍流畸变波前斜率数据为研究对象,通过数值仿真方法,研究了基于LMBP算法的自适应光学系统变形镜电压非线性预测控制算法。通过实验发现,预测电压和变形镜实际控制电压拟合效果良好。讨论了回溯帧数对预测效果的影响,并与基于递推最小二乘（RLS）算法的线性预测算法进行比较。对比结果表明,基于LMBP算法的非线性电压预测方法比基于递推最小二乘法的线性电压预测方法能更有效地降低系统由伺服延迟引起的误差。     

基于在线误差修正自适应SVR的非线性不确定分数阶混沌系统滑模控制

提出了一种基于在线误差修正自适应SVR的滑模控制方法, 用于解决一类非线性不确定分数阶混沌系统的控制问题. 分别通过对混沌系统非线性函数的离线SVR估计和基于增量学习的状态跟踪误差在线SVR预测, 解决了不确定分数阶混沌系统模型难以预测的问题. 同时根据Lyapunov稳定性理论设计出SVR权值自适应调整律. 本文以分数阶Arneodo 系统为例进行仿真, 仿真结果表明了, 对于带有外界噪声扰动的非线性不确定分数阶混沌系统, 该方法可以在有限时间内将系统稳定至期望状态, 提高对非线性函数的预测精度, 改善控制性能.     

A design for a general digital filter for state estimation of an arbitrary stochastic sound system

This article describes a new attempt at the design of a general digital filter for the state estimation of a nonstationary nonlinear stochastic sound system. A recursive algorithm for estimating the higher-order statistics of arbitrary-function type, mean, and variance is obtained by introducing a new expansion form of Bayes' theorem. Further, the state probability density function (PDF) can also be estimated in a unified form of orthogonal or nonorthogonal series expansions by using these estimates. This method is widely applicable for cases where the random-noise fluctuation is non-Gaussian. The estimation algorithm proposed in this article agrees completely with a well-known Kalman filtering theory [J. Basic Eng. 82, 35-45 (1960); Kalman and Buchy, J. Basic Eng. 83, 95-108 (1961)], as a simplified special case when the stochastic system is of linear type with Gaussian random excitation. The validity and effectiveness of the proposed theory were confirmed experimentally by applying it to actually observed room acoustic data and road-traffic noise data.     

Stability analysis of a noise-induced Hopf bifurcation

We study analytically and numerically the noise-induced transition between an absorbing and an oscillatory state in a Duffing oscillator subject to multiplicative, Gaussian white noise. We show in a non-perturbative manner that a stochastic bifurcation occurs when the Lyapunov exponent of the linearised system becomes positive. We deduce from a simple formula for the Lyapunov exponent the phase diagram of the stochastic Duffing oscillator. The behaviour of physical observables, such as the oscillators mean energy, is studied both close to and far from the bifurcation.Received: 8 August 2003, Published online: 19 November 2003PACS: 05.40.-a Fluctuation phenomena, random processes, noise, and Brownian motion - 05.10.Gg Stochastic analysis methods (Fokker-Planck, Langevin, etc.) - 05.45.-a Nonlinear dynamics and nonlinear dynamical systems     

利用粒子滤波从雷达回波实时跟踪反演大气波导

粒子滤波(particle filter,PF)是利用蒙特卡洛仿真方法处理递推估计问题的非线性滤波算法,这种方法不受模型线性和高斯假设的约束,是处理非线性非高斯动态系统状态估计的有效算法,适用于雷达回波反演大气波导(RFC)这类非线性非高斯问题.文中分别介绍了PF的基本思想和具体算法实现步骤,最后导出PF反演算法的迭代求解格式.数值试验结果表明,与扩展卡尔曼滤波(extended kalman filter,EKF)和不敏卡尔曼滤波(unscented kalman filter,UKF)相比,PF更适用于RFC这类高度非线性反演问题,可有效提高反演结果的稳定性和精度.     

Filtering skill for turbulent signals for a suite of nonlinear and linear extended Kalman filters

The filtering skill for turbulent signals from nature is often limited by errors due to utilizing an imperfect forecast model. In particular, real-time filtering and prediction when very limited or no a posteriori analysis is possible (e.g. spread of pollutants, storm surges, tsunami detection, etc.) introduces a number of additional challenges to the problem. Here, a suite of filters implementing stochastic parameter estimation for mitigating model error through additive and multiplicative bias correction is examined on a nonlinear, exactly solvable, stochastic test model mimicking turbulent signals in regimes ranging from configurations with strongly intermittent, transient instabilities associated with positive finite-time Lyapunov exponents to laminar behavior. Stochastic Parameterization Extended Kalman Filter (SPEKF), used as a benchmark here, involves exact formulas for propagating the mean and covariance of the augmented forecast model including the unresolved parameters. The remaining filters use the same nonlinear forecast model but they introduce model error through different moment closure approximations and/or linear tangent approximation used for computing the second-order statistics of the augmented stochastic forecast model. A comprehensive study of filter performance is carried out in the presence of various moment closure errors which are enhanced by additional model errors due to incorrect parameters inducing additive and multiplicative stochastic biases. The estimation skill of the unresolved stochastic parameters is also discussed and it is shown that the linear tangent filter, despite its popularity, is completely unreliable in many turbulent regimes for both parameter estimation and filtering; moreover, regimes of filter divergence for the linear tangent filter are identified. The results presented here provide useful guidelines for filtering turbulent, high-dimensional, spatially extended systems with more general model errors, as well as for designing more skillful methods for superparameterization of unresolved intermittent processes in complex multi-scale models. They also provide unambiguous benchmarks for the capabilities of linear and nonlinear extended Kalman filters using incorrect statistics on an exactly solvable test bed with rich and realistic dynamics.     

Adaptive step-size modified fractional least mean square algorithm for chaotic time series prediction

This paper presents an adaptive step-size modified fractional least mean square （AMFLMS） algorithm to deal with a nonlinear time series prediction. Here we incorporate adaptive gain parameters in the weight adaptation equation of the original MFLMS algorithm and also introduce a mechanism to adjust the order of the fractional derivative adaptively through a gradient-based approach. This approach permits an interesting achievement towards the performance of the filter in terms of handling nonlinear problems and it achieves less computational burden by avoiding the manual selection of adjustable parameters. We call this new algorithm the AMFLMS algorithm. The predictive performance for the nonlinear chaotic Mackey Glass and Lorenz time series was observed and evaluated using the classical LMS, Kernel LMS, MFLMS, and the AMFLMS filters. The simulation results for the Mackey glass time series, both without and with noise, confirm an improvement in terms of mean square error for the proposed algorithm. Its performance is also validated through the prediction of complex Lorenz series.     

Adaptive modeling, adaptive data assimilation and adaptive sampling

For efficient progress, model properties and measurement needs can adapt to oceanic events and interactions as they occur. The combination of models and data via data assimilation can also be adaptive. These adaptive concepts are discussed and exemplified within the context of comprehensive real-time ocean observing and prediction systems. Novel adaptive modeling approaches based on simplified maximum likelihood principles are developed and applied to physical and physical–biogeochemical dynamics. In the regional examples shown, they allow the joint calibration of parameter values and model structures. Adaptable components of the Error Subspace Statistical Estimation (ESSE) system are reviewed and illustrated. Results indicate that error estimates, ensemble sizes, error subspace ranks, covariance tapering parameters and stochastic error models can be calibrated by such quantitative adaptation. New adaptive sampling approaches and schemes are outlined. Illustrations suggest that these adaptive schemes can be used in real time with the potential for most efficient sampling.     

Estimating parameters in stochastic systems: A variational Bayesian approach

This work is concerned with approximate inference in dynamical systems, from a variational Bayesian perspective. When modelling real world dynamical systems, stochastic differential equations appear as a natural choice, mainly because of their ability to model the noise of the system by adding a variation of some stochastic process to the deterministic dynamics. Hence, inference in such processes has drawn much attention. Here a new extended framework is derived that is based on a local polynomial approximation of a recently proposed variational Bayesian algorithm. The paper begins by showing that the new extension of this variational algorithm can be used for state estimation (smoothing) and converges to the original algorithm. However, the main focus is on estimating the (hyper-) parameters of these systems (i.e. drift parameters and diffusion coefficients). The new approach is validated on a range of different systems which vary in dimensionality and non-linearity. These are the Ornstein-Uhlenbeck process, the exact likelihood of which can be computed analytically, the univariate and highly non-linear, stochastic double well and the multivariate chaotic stochastic Lorenz ’63 (3D model). As a special case the algorithm is also applied to the 40 dimensional stochastic Lorenz ’96 system. In our investigation we compare this new approach with a variety of other well known methods, such as the hybrid Monte Carlo, dual unscented Kalman filter, full weak-constraint 4D-Var algorithm and analyse empirically their asymptotic behaviour as a function of observation density or length of time window increases. In particular we show that we are able to estimate parameters in both the drift (deterministic) and the diffusion (stochastic) part of the model evolution equations using our new methods.     

Constrained adaptive lifting and the CAL4 metric for helicopter transmission diagnostics

This paper presents a methodology for detecting and diagnosing gear faults in the planetary stage of a helicopter transmission. This diagnostic technique is based on the constrained adaptive lifting (CAL) algorithm, an adaptive manifestation of the lifting scheme. Lifting is a time domain, prediction-error realization of the wavelet transform that allows for greater flexibility in the construction of wavelet bases. Adaptivity is desirable for gear diagnostics as it allows the technique to tailor itself to a specific transmission by selecting a set of wavelets that best represent vibration signals obtained while the gearbox is operating under healthy-state conditions. However, constraints on certain basis characteristics are necessary to enhance the detection of local wave-form changes caused by certain types of gear damage. The proposed methodology analyzes individual tooth-mesh waveforms from a healthy-state gearbox vibration signal that was generated using the vibration separation synchronous signal-averaging algorithm. Each waveform is separated into analysis domains using zeros of its slope and curvature. The bases selected in each analysis domain are chosen to minimize the prediction error, and constrained to have approximately the same-sign local slope and curvature as the original signal. The resulting set of bases is used to analyze future-state vibration signals and the lifting prediction error is inspected. The constraints allow the transform to effectively adapt to global amplitude changes, yielding small prediction errors. However, local waveform changes associated with certain types of gear damage are poorly adapted, causing a significant change in the prediction error. A diagnostic metric based on the lifting prediction error vector termed CAL4 is developed. The CAL diagnostic algorithm is validated using data collected from the University of Maryland Transmission Test Rig and the CAL4 metric is compared with the classic metric FM4.     

Analysis of statistical characteristics of adaptive antenna array with independent reference-signal generation

The steady-state statistical characteristics of an adaptive antenna array whose weight-factor vector is adjusted by means of a least-mean-square algorithm are studied using a nonlinear inertialess transformation of the array output as a reference signal. The case of phase-shift keying of the desired signal with Gaussian interference and noise is examined. An optimal nonlinear reference-signal transformation is found for which mean-square error of signal estimation is minimized and the output signal-to-noise ratio is maximized. The theoretical results are confirmed by computer modeling.     

The fluctuating phase error analysis in the digital grating phase-shifting profilometry

The nonlinear response of the experimental system and the saturation of fringe patterns can induce the fluctuating phase error in the projection grating phase-shifting profilometry. Two major factors of the fluctuating phase error are discussed by simulation. The fluctuating phase error caused by the nonlinear response of the system is four times the frequency of the fringe pattern when the conventional four-frame phase extracting algorithm is used. However, such error can be decreased by five-frame algorithm. On the other hand, the fluctuating phase error caused by the fringe saturation is five times the frequency of the fringe pattern by using conventional five-frame phase extracting algorithm. A novel phase recovering algorithm is used to decrease the phase error caused by the saturation. Furthermore, the applicability range of the proposed phase recovering algorithm is analyzed by simulation and experiments with different saturation degree of the fringe pattern and nonlinearity of the measurement system.     

一种时间序列的弱非线性检验方法

时间序列的非线性是判定该时间序列具有混沌特性的必要条件.提出一种基于线性和非线性AR模型归一化多步预测误差比值的非线性检验量δ_NAR,采用替代数据法来检测时间序列中的弱非线性.以Lorenz时间序列为例,分析了估计非线性检验量δ_NAR时各相关参数对弱非线性检测性能的影响.通过混沌时间序列非线性检测试验,对4种混沌时间序列中的3种,非线性检验量δ_NAR都表现出比基于AIC模型选择准则的非线性检验量相似文献   

基于时频差的正交容积卡尔曼滤波跟踪算法

针对基于时频差测量的无源跟踪中面临的非线性估计问题, 提出一种正交容积卡尔曼滤波跟踪算法. 该算法在容积卡尔曼滤波算法的基础上, 通过引入特定正交矩阵改进容积采样方法, 在高维状态估计下减小因采样产生的误差, 在没有增加计算量的前提下, 有效提高收敛速度及跟踪精度. 仿真结果表明, 在基于到达时差和到达频差的联合无源跟踪问题中, 与扩展卡尔曼滤波及容积卡尔曼滤波算法相比, 本文所提算法在跟踪性能上有明显提升.     

基于P范数的核最小对数绝对差自适应滤波算法

为了进一步提高在a稳定分布噪声背景下非线性自适应滤波算法的收敛速度,本文提出了一种新的基于p范数的核最小对数绝对差自适应滤波算法(kernel least logarithm absolute difference algorithm based on p-norm, P-KLLAD).该算法结合核最小对数绝对差算法和p范数,一方面利用最小对数绝对差准则保证了算法在a稳定分布噪声环境下良好的鲁棒性,另一方面在误差的绝对值上添加p范数,通过p范数和一个正常数a来控制算法的陡峭程度,从而提高该算法的收敛速度.在非线性系统辨识和Mackey-Glass混沌时间序列预测的仿真结果表明,本文算法在保证鲁棒性能的同时提高了收敛速度,并且在收敛速度和鲁棒性方面优于核最小均方误差算法、核分式低次幂算法、核最小对数绝对差算法和核最小平均p范数算法.     

A subpixel motion estimation algorithm based on digital correlation for illumination variant and noise image sequences

One of the challenges in practical subpixel motion estimation is how to obtain high accuracy with sufficient robustness to both illumination variations and additive noise. Motivated by the fact that the normalized spatial cross-correlation is invariant to illumination, we introduce a gradient-based subpixel registration method by maximizing the digital correlation (DC) function between the reference and target frames. Such DC function is remodeled with the presence of image noise, yielding that the correlation coefficient is only sensitive to noise standard variance. To fairly suppress the noise corruption, not only the target frame but also the reference one is reformulated into Taylor gradient expression with half but opposite motion vector. The final solution to motion estimates can be approximated into a closed form by reserving first-order coefficient terms of unregistered motion variables. The error trend of approximated solution is discussed. Computer simulations and actual experiments’ results demonstrate the superiority of the proposed method to the LMSE-based method and ordinary DC method when illumination variations and noise exist. Among the experiments, the influences of real subpixel translation value and noise variance degree on accuracy are studied; correspondingly, an optimized iterative idea for big translations and the recommended noise level adaptive to our method are introduced.     

Calibration of a stochastic health evolution model using NHIS data

This paper presents and calibrates an individual’s stochastic health evolution model. In this health evolution model, the uncertainty of health incidents is described by a stochastic process with a finite number of possible outcomes. We construct a comprehensive health status index (HSI) to describe an individual’s health status, as well as a health risk factor system (RFS) to classify individuals into different risk groups. Based on the maximum likelihood estimation (MLE) method and the method of nonlinear least squares fitting, model calibration is formulated in terms of two mixed-integer nonlinear optimization problems. Using the National Health Interview Survey (NHIS) data, the model is calibrated for specific risk groups. Longitudinal data from the Health and Retirement Study (HRS) is used to validate the calibrated model, which displays good validation properties. The end goal of this paper is to provide a model and methodology, whose output can serve as a crucial component of decision support for strategic planning of health related financing and risk management.     

粒子区别性稀疏表征的小弱运动目标跟踪算法

评估每个粒子的重要性是确保粒子滤波法跟踪目标准确性的重要因素。针对背景杂波和噪声干扰形成的大量虚警导致小弱目标跟踪识别的随机性和不确定性问题, 提出了一种基于粒子区别性稀疏表征的小弱目标跟踪方法。该方法根据红外图像信号自适应构建分类超完备字典, 即反映目标信号特征的目标字典和表示背景杂波的背景字典, 有利于突出目标粒子和背景粒子在联合分类字典的稀疏表征差异程度;建立基于目标粒子和背景粒子稀疏重构残差差异性的粒子滤波观测模型, 采用随机估计法对字典子空间进行在线更新, 实现对目标状态估计与跟踪。理论分析和试验结果表明, 该方法增强了随机粒子的状态估计能力, 提升了粒子稀疏表征对小弱运动目标的适应能力和跟踪识别准确度。     

非高斯冲激干扰下基于Softplus函数的核自适应滤波算法

针对非高斯环境下一般自适应滤波算法性能严重下降问题,本文提出了一种基于Softplus函数的核分式低次幂自适应滤波算法(kernel fractional lower algorithm based on Softplus function,SP-KFLP),该算法将Softplus函数与核分式低次幂准则相结合,利用输出误差的非线性饱和特性通过随机梯度下降法更新权重.一方面利用Softplus函数的特点在保证了SP-KFLP算法具有良好的抗脉冲干扰性能的同时提高了其收敛速度;另一方面将低次幂误差的倒数作为权重向量更新公式的系数,利用误差突增使得权重向量不更新的方法来抵制冲激噪声,并对其均方收敛性进行了分析.在系统辨识环境下的仿真表明,该算法很好地兼顾了收敛速度和跟踪性能稳定误差的矛盾,在收敛速度和抗脉冲干扰鲁棒性方面优于核最小均方误差算法、核分式低次幂算法和S型核分式低次幂自适应滤波算法.