Authors: Sboui, Amine; Farza, Mondher; Cherrier, Estelle; MSaad, Mohammed

Abstract: In this paper, one proposes adaptive observers for a class of uniformly observable nonlinear systems with general nonlinear parameterizations and in the presence of one (or more) variable and known delay(s). The proposed approach consists of a systematic procedure for the high gain observers where the state and the unknown parameters of the considered systems are supposed to lie in bounded domains whose size can be arbitrarily large. The exponential convergence of the observer relies on the resolution of an algebraic Lyapunov equation and leads to an explicit expression of the observer gain. The performances of the proposed observer are evaluated through a numerical example at the end of the paper.

Keywords: Nonlinear System Identification

Identifier: 10.3182/20090706-3-FR-2004.00025

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Soemers, Marcus; Alsmeyer, Frank

Abstract: For the long-term storage of measured data from production processes, process information management systems (PIMS) have been established in the last years. The use of these measurement data offers optimization potential if the relevant process information can be extracted. This contribution gives an overview of the innovative algorithms the software platform ALANDA provides for online and offline analysis of process data. On the basis of configuration-free algorithms, the effort for data analysis and model building can be reduced significantly. An introduction to the methods for PIMS configuration, basic preprocessing, and trend detection is given. These methods, which are predominantly based on wavelet analysis are used for the identification of a soft sensor in an industrial application. Finally, we present a tutorial demonstration of ALANDA in terms of a trend detection in a separation process.

Keywords: Others; Identification for Control; Filtering and Smoothing

Identifier: 10.3182/20090706-3-FR-2004.00083

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Markovsky, Ivan

Abstract: Closed-loop data-driven simulation refers to the problem of constructing trajectories of a closed-loop system directly from data of the plant and a representation of the controller. Conditions under which the problem has a solution are given and an algorithm for computing the solution is presented. The problem formulation and its solution are in the spirit of the deterministic identification algorithms, i.e., in the theoretical analysis of the method, the data is assumed exact (noise free).

Keywords: Subspace Methods; Identifiability; Identification for Control

Identifier: 10.3182/20090706-3-FR-2004.00018

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Ugryumova, Diana; Lau, Katrina; Braslavsky, Julio H.; Meinsma, Gjerrit

Abstract: Magnetotelluric (MT) surveying is an Electromagnetic (EM) surveying technique used in geophysics and mineral exploration. The main problem in MT surveying is the estimation of the impedance of the ground, which is obtained as the ratio between the natural environmental electric and magnetic fields measured on the surface of the target area. Because these measurements are inherently corrupted by noise, the impedance estimate may be biased (Errors-In-Variables (EIV)), which is a difficulty well-known in MT literature and typically overcome by the use of additional independent measurements. This paper formulates the MT problem as a standard system identification problem, and uses output-error model structures to obtain scalar (SISO) and vector (TISO) ground impedance estimates. Estimation bias is minimised by selecting segments of data with high Signal-to-Noise Ratio (SNR). The SISO and TISO modelling approaches are discussed and compared on results obtained from experimental MT data.

Keywords: Other; Errors in Variables Identification; Continuous Time System Estimation

Identifier: 10.3182/20090706-3-FR-2004.00161

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Kachour, Maher; Yao, Jiang-Feng

Abstract: We introduce a new class of autoregressive models for integer-valued time series using the rounding operator. Compared to classical INAR models based on the thinning operator, the new models have several advantages: simple innovation structure; autoregressive coefficients with arbitrary signs; possible negative values for time series; possible negative values for the autocorrelation function. Focused on the first order RINAR(1) model, we give conditions for its ergodicity and stationarity. For parameter estimation, a least squares estimator is introduced and we prove its consistency under suitable identifiability condition. An analysis of real data set is carried out to access the performance of the model.

Keywords: Time Series; Model Validation; Identifiability

Identifier: 10.3182/20090706-3-FR-2004.00240

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Ravariu, Cristian; Bondarciuc, Ala; Wilhelm, Kappel; Bondarciuc, Vlad; Ravariu, Florina

Abstract: The objectives of this study are the particularities of the laser close infrared spectrum (850nm) in vivo interaction with the human tissues, using some statistical tools. The main statistical variable is the average reflection coefficient, ARC, in the intact tissues and in the pathologically modified tissues (edemas, hematoma). The determined ARC coefficient in the intact tissues is an index of the health-state, having a range between 55,7 to 68 mW + 2,1 mW, very stable in time. The determined average reflection coefficient in the pathologically modified tissues constantly decreases from 58 up to 42 mW + 3,4 mW and varies in time accordingly with the evolution of the pathological process. However, the overlap for the ARC ranges occurs among diseases that induce uncertain diagnosis. Therefore a statistical approach is favorable. The laser bio-photometry provides the experimental tables that are processed to compute the statistical parameters of the probabilistic variable. The developed Identification function, Id, allows the pathological processes finding, besides to the evaluation and monitoring of their evolution.

Keywords: Biological Systems; Identifiability; Model Validation

Identifier: 10.3182/20090706-3-FR-2004.00167

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Kenne, Godpromesse; Ahmed-Ali, Tarek; Lamnabhi-Lagarrigue, Françoise; Arzande, Amir; Vannier, Jean-Claude

Abstract: In this paper, an online rotor resistance estimator for induction motor adaptive control is presented. The proposed algorithm is an improvement of the previous work Kenne et al.. The rotor resistance scheme uses the rotor speed, the stator current, voltage provided by the controller, time-derivatives of the stator current and voltage. The time derivatives of the stator current and voltage are estimated by using 2nd order sliding mode observer. The convergence of the estimated rotor resistance to the nominal value in finite time is achieved when the classical persistent excitation condition is satisfied by the input signal. Experimental results with 100% variation of the stator resistance and online variation of the rotor resistance show that the proposed algorithm gives better performance compared to the previous work.

Keywords: Identification for Control; Nonlinear System Identification; Closed Loop Identification

Identifier: 10.3182/20090706-3-FR-2004.00050

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Bolea, Yolanda; Martinez-Gonzalez, Ruben; Grau, Antoni; Martinez-Garcia, Herminio

Abstract: In this paper an LPV non-integer order control model of an irrigation canal is derived from system identification experiments. This model is experimentally obtained by using the described LPV fractional identification procedure. This procedure consists on the identification of a non-linear order model in each operation point of the canal. The global LPV model is obtained by the polynomial interpolation of the parameters of the local models. The validation results show that fractional models are more accurate than integer models. Therefore the fractional models have an important role to play in management and efficient use of water resources.

Keywords: Process Control

Identifier: 10.3182/20090706-3-FR-2004.00297

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Tóth, Roland; Heuberger, Peter S.C.; Van den Hof, Paul M.J.

Abstract: Describing nonlinear dynamic systems by Linear Parameter-Varying (LPV) models has become an attractive tool for control of complicated systems with regime-dependent (linear) behavior. For the identification of LPV models from experimental data a number of methods has been presented in the literature but a full picture of the underlying identification problem is still missing. In this contribution a solid system theoretic basis for the description of model structures for LPV systems is presented, together with a general approach to the LPV identification problem. Use is made of a series-expansion approach, employing orthogonal basis functions.

Keywords: Nonlinear System Identification; Basis Functions

Identifier: 10.3182/20090706-3-FR-2004.00221

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Kantas, Nikolas; Doucet, Arnaud; Singh, Sumeetpal S.; Maciejowski, Jan

Abstract: Nonlinear non-Gaussian state-space models arise in numerous applications in control and signal processing. Sequential Monte Carlo (SMC) methods, also known as Particle Filters, provide very good numerical approximations to the associated optimal state estimation problems. However, in many scenarios, the state-space model of interest also depends on unknown static parameters that need to be estimated from the data. In this context, standard SMC methods fail and it is necessary to rely on more sophisticated algorithms. The aim of this paper is to present a comprehensive overview of SMC methods which have been proposed to perform static parameter estimation in general state-space models. We discuss the advantages and limitations of these methods.

Keywords: Particle Filtering/Monte Carlo Methods; Maximum Likelihood Methods; Bayesian Methods

Identifier: 10.3182/20090706-3-FR-2004.00129

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Tanaka, Hideyuki

Abstract: This paper studies a realization algorithm for modeling dynamical systems subject to a bounded error. An iterative algorithm is developed based on ``indefinite realization'', which is a generalization of stochastic realization. A Riccati difference equation for indefinite realization is derived, and it is shown that the Riccati recursion converges to the stabilizing solution to the steady state Riccati equation under certain assumptions. Numerical simulation results are also included.

Keywords: Subspace Methods; Multivariable System Identification; Bounded Error Identification

Identifier: 10.3182/20090706-3-FR-2004.00017

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Schoukens, Johan; Barbé, Kurt; Vanbeylen, Laurent; Pintelon, Rik

Abstract: This paper analyses the variance of linear models that are identified in the presence of nonlinear distortions. The best linear approximation GBLA to a nonlinear system, driven by a Gaussian distributed random input, is identified in least squares sense. Even in the absence of disturbing noise, the estimated model G^BLA will vary from one realization of the excitation signal to the other. In this paper it will be shown that the variance of the nonparametric frequency response function (FRF) G^BLA(k) can still be obtained from the experimental data using the classical variance formulas. However, for parametric G^BLA(q,theta) estimates , an increased variance is observed. The contributing terms to this additional variance are studied in this paper.

Keywords: Error Quantification; Nonlinear System Identification; Model Validation

Identifier: 10.3182/20090706-3-FR-2004.00105

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Mendelson, Alexander; Tenno, Robert

Abstract: In this paper the Zakai filtering method is applied for identification of nonlinear physical models describing electrode kinetics of electrochemical systems. The filtering equations for both stirred and non-agitated solutions are derived. The model parameters are estimated and convergence of the estimates is analyzed on several models introduced in order of increasing complexity.

Keywords: Nonlinear System Identification; Other; Recursive Identification

Identifier: 10.3182/20090706-3-FR-2004.00237

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Markovsky, Ivan

Abstract: A number of problems in system theory, signal processing, and computer algebra fit into a generic structured low-rank approximation problem. Several problems of this type are reviewed and efficient local optimization methods for solving them are outlined.

Keywords: Errors in Variables Identification; Time Series; Maximum Likelihood Methods

Identifier: 10.3182/20090706-3-FR-2004.00186

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Soumelidis, Alexandros; Bokor, Jozsef; Schipp, Ferenc

Abstract: This paper is devoted to the construction of wavelet-type transforms with the purpose to represent functions belonging to the Hardy-space H². The concept of the affine wavelet-transform defined in the space L² is extended to the Hardy space H² on the basis of the Blaschke group. A discrete hyperbolic wavelet scheme is also constructed that results in computable forms. The wavelet construction obtained possesses good localization properties with respect to functions in H², hence forms an adequate tool for signal and system identification purposes.

Keywords: Basis Functions; Frequency Domain Identification; Nonparametric Methods

Identifier: 10.3182/20090706-3-FR-2004.00222

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Poulimenos, Aggelos; Fassois, Spilios D.

Abstract: Functional Series Time-dependent AutoRegressive Moving Average (FS-TARMA) models constitute an important class of models for non-stationary signal modeling. However, the asymptotic properties of “general” (that is not necessarily periodically evolving) FS-TARMA estimators are not, yet, well understood. In the present study, a general framework for the asymptotic analysis of “general” FS-TARMA estimators is developed and applied to the case of “general” Weighted Least Squares, Maximum Likelihood and Multi Stage estimators. The validity of the asymptotic analysis results is confirmed through Monte Carlo experiments.

Keywords: Time Series

Identifier: 10.3182/20090706-3-FR-2004.00242

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Pronzato, Luc

Abstract: We present new conditions for the strong consistency and asymptotic normality of the least squares estimator in nonlinear stochastic models when the design variables vary in a finite set. The application to self-tuning optimisation is considered, with a simple adaptive strategy that guarantees simultaneously the convergence to the optimum and the strong consistency of the estimates of the model parameters. An illustrative example is presented.

Keywords: Input and Excitation Design; Identification for Control

Identifier: 10.3182/20090706-3-FR-2004.00026

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Barbé, Kurt; Pintelon, Rik; Schoukens, Johan

Abstract: It is well known that under very general assumptions the discrete Fourier coefficients of filtered noise is asymptotically independent circular complex Gaussian distributed, based on a generalized central limit theorem (CLT). The standard results on the consistency and the asymptotic uncertainty of the frequency domain Errors-in-Variables (EIV) estimator are derived under the assumption that the Fourier coefficients are circular complex Gaussian distributed and independent over the different frequency bins. In this paper, we shall study the influence of this assumption on the consistency and the efficiency of the frequency domain EIV-estimator. We show that a slightly stronger form of the CLT is needed to preserve the classically obtained uncertainty bounds if independent complex Gaussian Fourier coefficients are not assumed. Our analysis reveals that the classical derived asymptotic uncertainty bounds are valid for a very wide class of distributions.

Keywords: Errors in Variables Identification; Frequency Domain Identification; Maximum Likelihood Methods

Identifier: 10.3182/20090706-3-FR-2004.00189

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: None

Abstract:

Keywords:

Identifier: 10.3182/20090706-3-FR-2004.90004

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Du, Dajun; Li, Kang; Fei, Minrui; Irwin, George W.

Abstract: This paper proposes a new automatic forward model selection method based on leave-one-out (LOO) cross-validation and a fast forward recursive algorithm. It can automatically select a sparse model by incrementally minimizing a LOO test error from a pool of candidate model terms, without the need to specify an additional termination criterion. By defining a proper regression context, the computational efficiency is ensured. A numerical example demonstrates the effectiveness of the proposed method.

Keywords: Nonlinear System Identification; Recursive Identification; Subspace Methods

Identifier: 10.3182/20090706-3-FR-2004.00145

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Peeters, Ralf; Olivi, Martine; Hanzon, Bernard

Abstract: Lossless systems have many applications in system and control theory, including signal processing, filter design, system identification, system approximation, and the parameterization of classes of linear systems. In this survey paper we address the issue of parameterization of the space of rational lossless matrix functions by successfully combining two approaches. The first approach proceeds in state-space from balanced realizations and triangular pivot structures of reachability matrices. The second approach concerns interpolation theory with linear fractional transformations and the tangential Schur algorithm. We construct balanced realizations (and canonical forms) in terms of the Schur parameters encountered in the tangential Schur algorithm, and conversely, we interpret balanced realizations in discrete-time and in continuous-time in terms of Schur parameters. A number of application areas are discussed to illustrate the importance of this theory for a variety of topics.

Keywords:

Identifier: 10.3182/20090706-3-FR-2004.00045

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Villemonteix, Julien; Vazquez, Emmanuel; Walter, Eric

Abstract: Bayesian optimization uses a probabilistic model of the objective function to guide the search for the optimum. It is particularly interesting for the optimization of expensive-to-evaluate functions. For the last decade, it has been increasingly used for industrial optimization problems and especially for numerical design involving complex computer simulations. We feel that Bayesian optimization should be considered with attention by anyone who has to identify the parameters of a model based on a very limited number of model simulations because of model complexity. In this paper, we wish to describe, as simply as possible, how Bayesian optimization can be used in parameter identification and to present a new application. We concentrate on two algorithms, namely EGO (for Efficient Global Optimization) and IAGO (for Informational Approach to Global Optimization), and describe how they can be used for parameter identification when the budget for evaluating the cost function is severely limited. Some open questions that must be addressed for theoretical and practical reasons are indicated.

Keywords: Nonparametric Methods; Machine Learning and Data Mining; Mechanical and Aerospace

Identifier: 10.3182/20090706-3-FR-2004.00266

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Truong, Nguyen-Vu; Wang, Liuping

Abstract: This paper addresses the identification of a Wiener-Hammerstein benchmark nonlinear system using wavelet based State Dependent Parameter(WSDP) models. The major attractive feature of the proposed technique lies on the systematic approach to nonlinear model structure determination using the Predicted Residual Sums of Squares(PRESS) statistic and forward regression. This procedure detects parametrically efficient structures for the wavelet based parameterization of SDP models, thus further enhances the parsimony of the model and helps to avoid over-parameterization. The obtained simulation results demonstrate the merit of the approach in identifying this benchmark nonlinear system.

Keywords: Nonlinear System Identification; Identification for Control

Identifier: 10.3182/20090706-3-FR-2004.00140

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Vanbeylen, Laurent; Pintelon, Rik; de Groen, Pieter

Abstract: This paper is concerned with the maximum likelihood identification of discrete-time Wiener systems from noisy output measurements only (blind identification). Prior work has been devoted to the blind identification of Wiener and Hammerstein systems in a noiseless situation. Applying these methods to output-noise corrupted data unavoidably results in biased estimates. Fortunately, the bias could be proven to be small for high signal-to-noise ratios. Nevertheless, it is clearly desirable to have a method which is consistent at any noise level. Therefore, this paper extends the existing method, by assuming a second (independent) white Gaussian noise source added to the output before measurement. Due to the presence of an extremely high dimensional integral in the expression of the likelihood function, the problem is very hard in practice. The 'curse of dimensionality' is avoided by approximating this integral by Laplace's method for integrals. The paper includes the illustration of the method on a simulation example, showing that the bias is possibly lower than in the method that ignores the presence of noise.

Keywords: Nonlinear System Identification; Maximum Likelihood Methods; Blind Estimation

Identifier: 10.3182/20090706-3-FR-2004.00280

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009

Authors: Vandersteen, Gerd; Barbé, Kurt; Pintelon, Rik; Schoukens, Johan

Abstract: The identification of linear dynamic system in a frequency domain Errors-In-Variables (EIV) framework often assumes the knowledge of a nonparametric noise model prior to parametric identification. This noise model can be extracted for periodic excitations using various sample mean/sample variance techniques starting from at least P=4 measured periods. The use of overlap and circular-overlap methods were recently introduced to increase the efficiency of the sample maximum likelihood (SML) estimator. This enabled halving the number of required periods while maintaining the efficiency loss of 1.46dB for P=2 periods with respect to Maximum Likelihood (ML) estimates with known covariance matrix. In this paper, the overlap and circular-overlap methods are applied to the sample BTLS method. A Monte Carlo analysis on a typical low-pass system shows that the efficiency loss with respect to the ML is only 0.12dB when P=2 periods are available. In addition, the example shows that the efficiency loss for P=1.2 periods is smaller than the loss introduced by the circular-overlap SML with P=4 periods.

Keywords: Errors in Variables Identification; Frequency Domain Identification; Nonparametric Methods

Identifier: 10.3182/20090706-3-FR-2004.00260

Conference: 15th IFAC Symposium on System Identification, 2009

Location: Saint-Malo Convention Center, Saint-Malo, France

Start Date: Mon Jul 06 2009 - End Date: Wed Jul 08 2009