Authors: A. Lj. Juloski, S. Weiland

Abstract: In this paper we develop an algorithm for the identification of piecewise linear output error models for the case where the discrete mode of the underlying hybrid system is not known. The presented algorithm is based on a Bayesian framework, i.e. unknown model parameters are treated as random variables and described with probability density functions. The identification problem is posed as a problem of computing the posterior parameter densities, given the prior densities and the observed data. A suboptimal identification algorithm is derived. Operation of the algorithm is demonstrated on an example.

Keywords: hybrid systems,identification,output error models,Bayesian methods

Identifier: 10.3182/20060329-3-AU-2901.00055

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: R. McVinish, J. H. Braslavsky, K. Mengersen

Abstract: This paper takes a Bayesian-decision theoretic approach to transfer function estimation, nominal model estimation, and quantification of the resulting model error. Consistency of the nonparametric estimate of the transfer function is proved together with a rate of convergence. The required quantities can be computed routinely using reversible jump Markov chain Monte Carlo methods. The proposed methodology has connections with set membership identification which has been extensively studied for this problem.

Keywords: transfer functions,decision theory,non-parametric identification,Monte Carlo calculation,loss minimization

Identifier: 10.3182/20060329-3-AU-2901.00162

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Geordie Z. Zhang, Girish N. Nair, Robin J. Evans, Bjorn Wittenmark

Abstract: This paper addresses the problem of adaptively controlling a plant with unknown parameters using communication-limited feedback. Assuming known dynamics, expressions have recently been obtained for the minimum average feedback data rate required for asymptotic stabilisability. The main purpose of this work is to demonstrate that this minimum rate does not increase if the plant parameters are unknown, and the key elements of a stabilizing, minimum-rate policy are explicitly discussed. By regarding the uncertain plant as a higher-dimensional, nonlinear plant with unknown initial condition, it is shown that this result agrees with the recent concept of local topological feedback entropy. Extensions to the case of uncertain nonlinear plants are discussed.

Keywords: adaptive control,data-rate limited control,entropy

Identifier: 10.3182/20060329-3-AU-2901.00178

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Riccardo Muradore, Fabrizio Bezzo

Abstract: A novel algorithm is presented for the design of inferential estimators for process monitoring and control. The algorithm aims at integrating Partial Least Squares (PLS) techniques and Subspace Identification Methods (SIM) to exploit the main advantages of both methodologies. In particular, the algorithm will retain the PLS computational robustness in dealing with large sets of correlated inputs and outputs, whilst profiting by the SIM dynamic description of the system being investigated.

Keywords: PLS,subspace identification methods

Identifier: 10.3182/20060329-3-AU-2901.00041

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Damian Marelli

Abstract: The subband method allows either identifying or approximating a linear system in the time-frequency domain, with high numerical efficiency. In this paper we propose a functional analysis setting to analyze the subband technique, which yields the following results: (a) We provide an analytical expression to calculate the best subband approximation of a given fullband system. (b) We provide a novel identification strategy which consists in identifying a "low quality" subband model and using it to build the required system model. This identification strategy is computationally more efficient and yields smaller residual errors, when compared with the existing methods.

Keywords: time-frequency representation,system identification

Identifier: 10.3182/20060329-3-AU-2901.00059

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Kenneth Hsu, Tyrone Vincent, Kameshwar Poolla

Abstract: We consider interconnected systems consisting of linear time-invariant systems and static nonlinear maps. Under the assumptions that the linear dynamics are known and the input to the nonlinear maps are measurable, it is shown that the identification problem can be reduced to a least squares problem. An identification algorithm utilizing a kernel-based dispersion function is proposed.

Keywords: nonlinear system identification,structured systems,kernel

Identifier: 10.3182/20060329-3-AU-2901.00193

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Kenneth Hsu, Tyrone Vincent, Kameshwar Poolla

Abstract: In (Hsu et al., 2005c), an algorithm for the identification of structured nonlinear systems was proposed and its computational properties were explored. In this paper, we continue the investigation and formalize notions of identifiability and persistence of excitation. Conditions under which the estimated nonlinearity converges uniformly to the true nonlinearity are developed for a class of kernel based dispersion functions.

Keywords: nonlinear system identification,kernel,identifiability,persistence of excitation,convergence

Identifier: 10.3182/20060329-3-AU-2901.00194

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Sunil L. Kukreja, Johan Lofberg, Martin J. Brenner

Abstract: Identification of parametric nonlinear models involves estimating unknown parameters and detecting its underlying structure. Structure computation is concerned with selecting a subset of parameters to give a parsimonious description of the system which may afford greater insight into the functionality of the system or a simpler controller design. In this study, a least absolute shrinkage and selection operator (LASSO) technique is investigated for computing efficient model descriptions of nonlinear systems. The LASSO minimises the residual sum of squares by the addition of a l1 penalty term on the parameter vector of the traditional l2 minimisation problem. Its use for structure detection is a natural extension of this constrained minimisation approach to pseudolinear regression problems which produces some model parameters that are exactly zero and, therefore, yields a parsimonious system description. The performance of this LASSO structure detection method was evaluated by using it to estimate the structure of a nonlinear polynomial model. Applicability of the method to more complex systems such as those encountered in aerospace applications was shown by identifying a parsimonious system description of the F/A-18 Active Aeroelastic Wing using flight test data.

Keywords: system identification,nonlinear systems,structure detection,aeroelasticity

Identifier: 10.3182/20060329-3-AU-2901.00128

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Erik Etien, Damien Halber, Gerard Champenois, Regis Ouvrard

Abstract: Jiles Atherton model is described in discrete form. Least square identification is improved using normalization of sensibility functions. Experimental tries validate the proposed method.

Keywords: magnetic hysteresis,Jiles-Atherton model,least square identification,convergence improvement

Identifier: 10.3182/20060329-3-AU-2901.00018

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Pierre Vacher, Alain Bucharles

Abstract: Flutter flight tests is a crucial and feared phase of the flight test program of a new aircraft. The specific operational context of flutter surveillance implies the development of automatic and reliable tools operating in real-time. At ONERA, we recently developed a toolbox dedicated to data processing for flutter tests. It was used for the latest Airbus aircraft from the A340-600 up to the A380. In this article, we present the main procedure of the toolbox : the identification routine together with the graphical interfaces that were designed to help the operator to get a rapid knowledge of the identification results.

Keywords: identification,algorithms,polynomials,transfer functions,frequency domain

Identifier: 10.3182/20060329-3-AU-2901.00098

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Hao Ye, Rui He, Heng Liu, Guizeng Wang

Abstract: A new approach for Fault Detection (FD) of Networked Control Systems (NCSs) based on Parity Relation (PR) and Principal Component Analysis (PCA) is proposed in this paper. The method can realize a good decoupling from unknown and random network-induced delay, and further ensure the robustness to traditional unknown inputs under an optimization sense. The approach is compared with the traditional PR based FD method for systems without network-induced delay and the existing PR based method for FD of NCSs.

Keywords: fault detection,networks,delay,decoupling,robustness

Identifier: 10.3182/20060329-3-AU-2901.00101

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Lennart Svensson, Joakim Gunnarsson

Abstract: Accurate vehicle motion models are of essential importance for modern car safety systems, as they enable more precise tracking and prediction of the traffic. During normal driving, the driver controls the vehicle almost completely, yet, standard models, such as the constant acceleration kinematic model, fail to acknowledge the impact of the driver. In this paper, we propose a modified version of the above mentioned model, in which the effect of the driver is introduced as an additional acceleration. To calculate this acceleration, we approximate the driver with an optimal regulator, and derive an optimal trajectory through which the acceleration can be found. Our definition of an optimal trajectory, is such that resulting path should be both comfortable, safe, fast and legal. Simulations indicate that the new model can lead to significant gains in both tracking and prediction performance.

Keywords: motion models,active safety,tracking,prediction,driver behavior,Kalman filtering

Identifier: 10.3182/20060329-3-AU-2901.00223

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Koichi Onodera, Genichi Emoto, S. Joe Qin

Abstract: In this paper we present a new subspace identification method applicable to closed-loop data. First, Kalman predictor Markov parameters in a framework of ARX modeling with high order are obtained. These parameters are made available for subspace identification to help the estimation of the Hankel matrix which consists of the estimated predictor Markov parameters. To estimate the observer matrices, eigensystem realization algorithm (ERA) with weightings related to canonical correlation analysis (CCA) is applied to the Hankel matrix. System matrices are easily derived from the estimated observer matrices. We then demonstrate the effectiveness of the proposed algorithm via simulated and industrial closed loop data.

Keywords: subspace identification,Kalman predictor Markov parameters,eigensystem realization algorithm (ERA),canonical correlation analysis (CCA),observer/Kalman filter identification (OKID)

Identifier: 10.3182/20060329-3-AU-2901.00169

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Marion Gilson, Hugues Garnier, Peter Young, Paul Van den Hof

Abstract: This paper describes an optimal instrumental variable method for identifying discrete-time transfer function models of the Box-Jenkins transfer function form in the closed-loop situation. This method is based on the Refined Instrumental Variable (RIV) algorithm which, because of an appropriate choice of particular design variables, achieves minimum variance estimation of the model parameters. The Box-Jenkins model is the most natural since it does not constrain the process and the noise models to have common polynomials. The performance of the proposed approach is evaluated by Monte Carlo analysis in comparison with other alternative closed loop estimation methods.

Keywords: system identification,closed-loop identification,optimal instrumental variable

Identifier: 10.3182/20060329-3-AU-2901.00143

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Mats Ekman, Mei Hong, Torsten Soderstrom

Abstract: It is well-known that the least-squares identification method generally gives biased parameter estimates when the observed input-output data are corrupted with noise. Previously, an extended version of compensated least-squares (ECLS), based on an overdetermined linear system of equations, was proposed as a method for handling problems where the input and output data are corrupted by white noise. This paper considers the problem where the noise is colored and, thus, extends previous results of the ECLS method. By considering the ECLS problem as a separable nonlinear LS problem, it is shown that the parameters, associated with the noise terms, can be obtained from solving a variable projection minimization problem. The accuracy of the parameter estimates is investigated, and it is also shown that the estimates, under some general assumptions, are consistent.

Keywords: parameter estimation,least-squares method,noisy input-output systems,separable nonlinear least squares

Identifier: 10.3182/20060329-3-AU-2901.00022

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Thomas Hanefeld Sejeroe, Niels Kjolstad Poulsen, Ole Ravn

Abstract: In this paper we present a simulation platform to evaluate methods for simultaneous location and mapping. The platform is based on the Kalmtool 3 toolbox which is a set of MATLAB tools for state estimation of nonlinear systems. The toolbox contains functions for extended Kalman filtering as well as for two new filters called the DD1 filter and the DD2 filter. It also contains function for Uncented Kalman filters as well as three versions of particle filters. The toolbox requires MATLAB ver. 6, but no additional toolboxes are required.

Keywords: simulation,toolbox,state estimation,Kalman filtering,nonlinear systems,autonomous robots

Identifier: 10.3182/20060329-3-AU-2901.00149

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Sette Diop

Abstract: In this Communication we show that for any dynamics given by a rational state vector equation dx/dt = f(u,x) there always is a scalar linear observation y = α1x1 + α2x2 + ... + αnxn which makes the state x observable provided that the coefficients α1, α2, ... , αn are allowed to be nonconstants. Moreover, any such scalar observation makes the system observable if, and only if, the coefficients are linearly independent over constants in a difierential algebraic sense.

Keywords: observability,identifiability,parameter estimation,nonlinear control systems,algebraic systems theory

Identifier: 10.3182/20060329-3-AU-2901.00131

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Keith R. Godfrey, Ai Hui Tan, H. Anthony Barker, W. Dhammika Widanage

Abstract: Three computer programs, all freely available on the World Wide Web, for designing and generating different classes of perturbation signals for system identification are described. Two of the programs are for designing pseudorandom signals, which have fixed power spectra. The first program is for designing signals based on five known classes of pseudorandom binary or near-binary sequences, and the second program is for designing pseudorandom signals based on maximum-length sequences, both binary and multilevel. The third program is for designing multilevel multiharmonic signals, in which the user can specify the harmonic pattern required, and a computer optimization routine is then used to meet the specification as closely as possible.

Keywords: binary signals,frequency responses,input signals,multilevel codes,pseudorandom sequences,system identification,time-domain responses

Identifier: 10.3182/20060329-3-AU-2901.00108

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Jian-Xun Peng, Kang Li, Er-Wei Bai

Abstract: A two-stage algorithm is proposed for fast identification of optimal linear-in-the-parameters models for nonlinear dynamic systems. In the first stage, an initial model is selected from a significant number of candidates, using a stepwise forward procedure. The significance of each selected model term is reviewed iteratively at the second stage using a fast review procedure and insignificant terms are then replaced, resulting in a locally optimised compact model. The contribution is that both the forward and backward model selection is performed within a well-defined regression context, leading to significantly reduced computational complexity. The computational complexity analysis confirms the arithmetic efficiency and the simulation results demonstrate the effectiveness.

Keywords: system identification,nonlinear system,linear-in-the-parameters model,model structure selection,computational complexity

Identifier: 10.3182/20060329-3-AU-2901.00087

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Mei Hong, Torsten Soderstrom, Wei Xing Zheng

Abstract: The bias-eliminating least squares (BELS) method is one of the consistent estimators for identifying dynamic errors-in-variables systems. The attraction of the BELS method lies in its good accuracy and its modest computational cost. In this paper, we investigate the accuracy properties of the BELS estimates. It is shown that the estimated system parameters and the estimated noise variances are asymptotically Gaussian distributed. An explicit expression for the normalized covariance matrix of the estimated parameters is derived and supported by some numerical examples.

Keywords: system identification,errors-in-variables,bias-eliminating least squares,accuracy analysis

Identifier: 10.3182/20060329-3-AU-2901.00024

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Ole Ravn

Abstract: Scilab/Scicos are open source alternatives for Matlab/Simulink. The Adaptive Toolbox (TAT) has been designed to mirror The Adaptive Blockset (TAB) for Simulink. Design considerations and implementational aspects of the Adaptive Toolbox for Scilab/Scicos are presented. The basics of indirect adaptive controllers are summarized. The concept behind the Adaptive Toolbox for Scilab/Scicos is to bridge the gap between simulation and prototype controller implementation. This is done using the code generation capabilities of CodeGen and RTAI in combination with C function blocks for adaptive control in Scicos. In the paper the design of each group of blocks normally fund in adaptive controllers is outlined. The block types are, identification, controller design, controller and state variable filter. The use of the Adaptive Toolbox is demonstrated using a laboratory setup. Both the use of the toolbox for system identification and adaptive control are shown. Furthermore a comparison is made with The Adaptive Blockset (TAB) for Simulink.

Keywords: system identification,adaptive control,simulation,CACSD,controller implementation

Identifier: 10.3182/20060329-3-AU-2901.00145

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Domenico Bellomo, Robert Babuska, David Naso

Abstract: Adaptive fuzzy control (AFC) has been an active research area over the last decade and several stable adaptive fuzzy controllers have been proposed in the literature. Such controllers are generally based on feedback linearization and their parameters are updated by tracking error-based adaptive laws, designed by Lyapunov synthesis. However, most of AFC schemes have only been evaluated on relatively simple simulations examples. In this paper, we study and compare different indirect adaptive schemes (linear and fuzzy, with standard and with composite adaptive laws), by means of an experimental benchmark consisting of two coupled servo drives. Parametric and structural changes are introduced to the controlled plant, in order to emphasize the advantages and limitations of the considered adaptive controllers. Experimental results demonstrate that AFC achieves significantly better tracking performance than the linear adaptive controller and that the composite adaptive laws provide a further improvement over the standard adaptive laws.

Keywords: adaptive control,fuzzy systems,model-reference control,servo-drives,feedback linearization,Lyapunov stability

Identifier: 10.3182/20060329-3-AU-2901.00175

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Oyvind N. Smogeli, Asgeir J. Sorensen

Abstract: Recently, new control designs and performance monitoring schemes for electrically driven marine propellers have been developed. The new control designs have been shown to increase the thrust capability, reduce mechanical wear and tear, and limit power transients. To achieve optimum performance, these designs rely on knowledge of the friction in the propeller drive system. Since this information not necessarily is available, a nonlinear adaptive observer for estimating the propeller load torque and the coefficient of friction is developed. Experiments are provided to verify the observer performance.

Keywords: adaption,observers,propulsion control,thrust losses,marine vessels

Identifier: 10.3182/20060329-3-AU-2901.00047

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: Jorn Op Den Buijs, Lizette Warner, Nicolas W. Chbat, Hal H. Ottesen

Abstract: In this paper, we propose a novel method for the real-time identification and simultaneous order estimation of a class of piecewise linear systems, namely jump-linear systems. The parameters, orders and time instances of switching are identified, based on a recursive weighted least-squares scheme and pole-zero cancellations. Several examples demonstrate the effectiveness of the algorithm. The proposed real-time technique is promising for use in patient monitoring and feedback control systems.

Keywords: order,piecewise,recursive,system identification

Identifier: 10.3182/20060329-3-AU-2901.00053

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006

Authors: C. E. Kinney, R. A. de Callafon

Abstract: This paper details the design of an adaptive internal model-based controller to attenuate periodic disturbances in the presence of random noise. The internal model-based controller is designed in two steps, making use of the separation principle to design a controller with the desired properties that stabilizes the closed-loop system. Periodic disturbances appear as harmonics of a fundamental frequency. The fundamental frequency of the disturbance is estimated with a magnitude/phase-locked loop and is used to adjust the parameters of the controller, specifically the natural frequency of the internal model and the feedback gain. An active noise control study shows the ability of the proposed method to cancel time varying disturbances in an acoustic system.

Keywords: feedback control,adaptive control,disturbance rejection,active noise control,frequency estimation

Identifier: 10.3182/20060329-3-AU-2901.00038

Conference: 14th IFAC Symposium on System Identification, 2006

Location: , Australia

Start Date: Wed Mar 29 2006 - End Date: Fri Mar 31 2006