Nonlinear estimation applying an unscented transformation in systems with correlated uncertain observations

An approximation to the least squares filter is proposed for discrete signals whose evolution is governed by nonlinear functions, when the estimation is based on nonlinear observations with additive noise which can consist only of random noise; this uncertainty in the observation process is modelled by Bernoulli random variables which are correlated at consecutive time instants and are otherwise independent. The proposed recursive approximation is based on the unscented principle; successive applications of the unscented transformation to a suitable augmented state vector enable us to approximate the one-stage state and observation predictors from the state filter at the previous time instant. The performance of the proposed algorithm is compared with that of an extended algorithm in a numerical simulation example.     

Data-driven recursive least squares methods for non-affined nonlinear discrete-time systems

Aiming at identifying nonlinear systems, one of the most challenging problems in system identification, a class of data-driven recursive least squares algorithms are presented in this work. First, a full form dynamic linearization based linear data model for nonlinear systems is derived. Consequently, a full form dynamic linearization-based data-driven recursive least squares identification method for estimating the unknown parameter of the obtained linear data model is proposed along with convergence analysis and prediction of the outputs subject to stochastic noises. Furthermore, a partial form dynamic linearization-based data-driven recursive least squares identification algorithm is also developed as a special case of the full form dynamic linearization based algorithm. The proposed two identification algorithms for the nonlinear nonaffine discrete-time systems are flexible in applications without relying on any explicit mechanism model information of the systems. Additionally, the number of the parameters in the obtained linear data model can be tuned flexibly to reduce computation complexity. The validity of the two identification algorithms is verified by rigorous theoretical analysis and simulation studies.     

An efficient algorithm for discrete modelling of tracer kinetics

Effective study of certain health care problems and biomedical systems requires development and utilization of duly validated models for characterization of tracer kinetics in compartmental systems. This report presents an efficient algorithm for evaluation of discrete-time models in this context, starting from real patient data. The procedure evolved is systematic and involves parameter identification, model order determination and ascertaining the validity of the model: mathematical techniques proposed for this three-tier approach are robust as well as simple. A method for deriving state-space representations for multicompartmental systems directly from observations vector is also outlined. An illustrative example is given which demonstrates the effectiveness of the algorithm when applied to tracer analysis of the hepatobiliary system. The methodology proposed is recursive in nature such that it can be implemented very conveniently utilizing a microcomputer or even a programmable pocket calculator.     

A least squares identification algorithm for a state space model with multi-state delays

A parameter estimator is presented for a state space model with time delay based on the given input–output data. The basic idea is to expand the state equations and to eliminate some state variables, and to substitute the state equation into the output equation to obtain the identification model which contains the information vector and parameter vector. A least squares algorithm is developed to estimate the system parameter vectors. Finally, an illustrative example is provided to verify the effectiveness of the proposed algorithm.     

On-line identification of language measure parameters for discrete-event supervisory control

The discrete-event dynamic behavior of physical plants is often represented by regular languages that can be realized as deterministic finite state automata (DFSA). The concept and construction of signed real measures of regular languages have been recently reported in literature. Major applications of the language measure are: quantitative evaluation of the discrete-event dynamic behavior of unsupervised and supervised plants; and analysis and synthesis of optimal supervisory control algorithms in the discrete-event setting. This paper formulates and experimentally validates an on-line procedure for identification of the language measure parameters based on a DFSA model of the physical plant. The recursive algorithm of this identification procedure relies on observed simulation and/or experimental data. Efficacy of the parameter identification procedure is demonstrated on the test bed of a mobile robotic system, whose dynamic behavior is modelled as a DFSA for discrete-event supervisory control.     

Composite Quantile Estimation in Partial Functional Linear Regression Model Based on Polynomial Spline

In this paper, we consider composite quantile regression for partial functional linear regression model with polynomial spline approximation. Under some mild conditions, the convergence rates of the estimators and mean squared prediction error, and asymptotic normality of parameter vector are obtained. Simulation studies demonstrate that the proposed new estimation method is robust and works much better than the least-squares based method when there are outliers in the dataset or the random error follows heavy-tailed distributions. Finally, we apply the proposed methodology to a spectroscopic data sets to illustrate its usefulness in practice.     

Stationarity conditions for linear models

The problem of building a linear stationary model for a process given by evenly spaced discrete or continuous observations is considered. Criteria are proposed for the existence of such a model governed by a vector difference or differential equation. Various model representations in discrete and continuous forms are studied and numerical methods for their identification are developed. This gives the order and dynamics of a model in the canonical form. To include processes in noisy environment, a moving average of observations is introduced into deterministic identification algorithms. Different integral forms of the moving average of continuous observations are proposed for identification of models governed by a system of linear stationary differential equations. Discussion of some experimental and computational results is presented.     

在不确定观测下离散状态时滞系统的最优滤波

研究了在不确定观测下离散状态时滞系统的最优滤波问题,观测值的不确定性则通过一个满足Bernoulli分布且统计特性已知的随机变量来描述. 一般采用状态增广方法将时滞系统转换为无时滞随机系统, 再利用Kalman滤波器的设计方法解决最优状态估计问题, 但是当系统时滞较大时,转换后的系统状态维数很高, 这样增加了计算负担. 为此,基于最小方差估计准则, 利用射影性质和递归射影公式得到了一个新的滤波器设计方法, 而且保证了滤波器的维数与原系统相同.最后, 给出一个仿真例子说明所提方法的有效性.     

递推阻尼最小二乘法的收敛性与稳定性

递推最小二乘法是参数辨识中最常用的方法,但容易产生参数爆发现象.因此对一种更稳定的辨识方法——递推阻尼最小二乘法进行了收敛特性的分析.在使用算法之前先归一化测量向量,结果表明,参数化距离收敛于一个零均值随机变量,并且在持续激励条件下,适应增益矩阵的条件数有界.参数化距离的方差有界.     

测量误差模型的自适应LASSO变量选择方法研究

本文研究测量误差模型的自适应LASSO(least absolute shrinkage and selection operator)变量选择和系数估计问题.首先分别给出协变量有测量误差时的线性模型和部分线性模型自适应LASSO参数估计量,在一些正则条件下研究估计量的渐近性质,并且证明选择合适的调整参数,自适应LASSO参数估计量具有oracle性质.其次讨论估计的实现算法及惩罚参数和光滑参数的选择问题.最后通过模拟和一个实际数据分析研究了自适应LASSO变量选择方法的表现,结果表明,变量选择和参数估计效果良好.     

A conservative bound on the estimation error covariance matrix in the presence of correlated driving noise and correlated discrete measurement noise

In an earlier paper, the conservative and minimal bound to the crosscorrelation terms between estimation error and a random forcing function was presented. That bound was found to be a particular linear combination of the estimation error covariance and the forcing function covariance involving a free scalar parameter. The bound was then substituted for the cross-correlation terms in the differential equation for the estimation error covariance matrix in order to approximate its behavior between discrete measurement times. The time history of the free parameter which minimized a linear combination of the elements of the estimated covariance matrix at the next measurement time was found as the noniterative solution to an optimal control problem with a matrix state.In this paper, necessary and sufficient conditions are presented for the problem of minimizing a linear combination of the elements of the approximated estimation error covariance at the end of an interval in which are linearly incorporated a finite number of discrete vector measurements corrupted by white and/or correlated measurement noise. Although the determination of the optimal trajectory in general requires iteration, a particularly simple algorithm is presented. Numerical results are presented for the case of a satellite in a highly elliptic orbit about a model Earth.     

Linear system identification via an asymptotically stable observer

This paper presents a formulation for identification of linear multivariable systems from single or multiple sets of input-output data. The system input-output relationship is expressed in terms of an observer, which is made asymptotically stable by an embedded eigenvalue assignment procedure. The prescribed eigenvalues for the observer may be real, complex, mixed real and complex, or zero corresponding to a deadbeat observer. In this formulation, the Markov parameters of the observer are first identified from input-output data. The Markov parameters of the actual system are then recovered from those of the observer and used to realize a state space model of the system. The basic mathematical formulation is derived, and numerical examples are presented to illustrate the proposed method.     

Parameter and state estimation algorithm for single-input single-output linear systems using the canonical state space models

This paper presents a new parameter and state estimation algorithm for single-input single-output systems based on canonical state space models from the given input–output data. Difficulties of identification for state space models lie in that there exist unknown noise terms in the formation vector and unknown state variables. By means of the hierarchical identification principle, those noise terms in the information vector are replaced with the estimated residuals and a new least squares algorithm is proposed for parameter estimation and the system states are computed by using the estimated parameters. Finally, an example is provided.     

Recursive computational formulas of the least squares criterion functions for scalar system identification

The paper discusses recursive computation problems of the criterion functions of several least squares type parameter estimation methods for linear regression models, including the well-known recursive least squares (RLS) algorithm, the weighted RLS algorithm, the forgetting factor RLS algorithm and the finite-data-window RLS algorithm without or with a forgetting factor. The recursive computation formulas of the criterion functions are derived by using the recursive parameter estimation equations. The proposed recursive computation formulas can be extended to the estimation algorithms of the pseudo-linear regression models for equation error systems and output error systems. Finally, the simulation example is provided.     

Associative memory approach to the identification of structural and mechanical systems

This paper presents a new method for identification of parameters in nonlinear structural and mechanical systems in which the initial guesses of the unknown parameter vectors may be far from their true values. The method uses notions from the field of artificial neural nets and, using an initial set of training parameter vectors, generates in an adaptive fashion other relevant training vectors to yield identification of the parameter vector in a recursive fashion. The simplicity and power of the technique are illustrated by considering three highly nonlinear systems. It is shown that the technique presented here yields excellent estimates with only a limited amount of response data, even when each element of the set comprising the initial training parameter vectors is far from its true value—in fact, sufficiently far that the usual recursive identification schemes fail to converge.     

Analysis of equation submodels

In bond graph models, the atomic submodels are described by sets of equations. Because of the physical justification of the bond graph formalism, it provides extensive possibilities for verification of the model at the graphical level. The equation formulation on the other hand is founded in the mathematical domain, so the need for a check against physical criteria is both more needed and more difficult.

Causality assignment is the meeting point of the graphical level and the equation level. In bond graph modeling, causality assignment is a vital step in analysis and simulation. The assignment process in the graph is based on the causality restrictions of the atomic submodels. In this article, an automatic procedure for the derivation of causality restrictions of atomic submodels is presented. This process not only generates the correct set of causality restrictions, but also provides a detailed verification of the correctness of the submodel     

On the combination of kernel principal component analysis and neural networks for process indirect control

ABSTRACT

A new adaptive kernel principal component analysis (KPCA) for non-linear discrete system control is proposed. The proposed approach can be treated as a new proposition for data pre-processing techniques. Indeed, the input vector of neural network controller is pre-processed by the KPCA method. Then, the obtained reduced neural network controller is applied in the indirect adaptive control. The influence of the input data pre-processing on the accuracy of neural network controller results is discussed by using numerical examples of the cases of time-varying parameters of single-input single-output non-linear discrete system and multi-input multi-output system. It is concluded that, using the KPCA method, a significant reduction in the control error and the identification error is obtained. The lowest mean squared error and mean absolute error are shown that the KPCA neural network with the sigmoid kernel function is the best.     

A procedure for one-parametric linear programming

A procedure is proposed for the parametric linear programming problem where all the coefficients are linear or polynomial functions of a scalar parameter. The solution vector and the optimum value are determined explicitly as rational functions of the parameter. In addition to standard linear programming technique, only the determination of eigenvalues is required. The procedure is compared to those by Dinkelbach and Zsigmond, and a numerical example is given.     

Auxiliary model based parameter estimation for dual-rate output error systems with colored noise

The dual-rate sampled-data systems can offer better quality of control than the systems with single sampling rate in practice. However, the conventional identification methods run in the single-rate scheme. This paper focuses on the parameter estimation problems of the dual-rate output error systems with colored noises. Based on the dual-rate sampled and noise-contaminated data, two direct estimation algorithms are addressed: the auxiliary model based recursive extended least squares algorithm and the recursive prediction error method. The auxiliary model is employed to estimate the noise-free system output. An example is given to test and illustrate the proposed algorithms.     

A guaranteed estimate of the parameter reconstruction error in a linear difference system

We consider the problemon reconstructing parameters of a linear autonomous difference discrete-time system from a finite set of approximate observations of the system state. We impose minimal assumptions on the observation error. Namely, we assume that the absolute value of the difference between the state vector and the corresponding observation is componentwise bounded from above by some given constant. Under these assumptions, we propose a theorem on the minimal guaranteed estimate of the parameter reconstruction error and describe the corresponding reconstruction algorithm.