Search

Multi-input multi-output (MIMO) process identification is studied, where the purpose of identification is control system design. An identification procedure is presented by which one can estimate not only a nominal parametric process model, but also an upper bound of the model errors in the frequency domain. The basic steps of this method consists of high-order model estimation and subsequent model reduction. In this framework, fundamental problems such as input design and model structure selection can easily be solved. The method is also numerically simple and reliable. A simulation example is given to illustrate the method. >

For the time being, the notion of ‘intelligent measurements’ is still rather vague (and, according to these authors, often misused). It is clear, however, that the implicit or explicit use of models in such measurements is quite pronounced. For that reason it is appropriate, perhaps even necessary, to explore the relations with system identification (ie, model building + parameter estimation). That is the subject of this paper. Also discussed are the actual and potential contributions of identification, the availability of appropriate software, and the desirabilities with respect to further contributions to intelligent measurements.

Using cross-validation ideas, two procedures are proposed for making a choice between different model structures used for (approximate) modelling of multivariable systems. The procedures are derived under fairly general conditions: the ‘true’ system does not need to be contained in the model set; model structures do not need to be nested and different criteria may be used for model estimation and validation. The proposed structure selection rules are invariant to parameter scaling. Under certain conditions (essentially requiring that the system belongs to the model set and that the maximum likelihood method is used for parameter estimation) they are shown to be asymptotically equivalent to the (generalized) Akaike structure selection criteria

In the development of the interaction health care/ engineering the development of models of physiological systems has received much attention. In spite of some scepsis in the past with respect to the relevance of such models for medical care there is a definite trend towards the possible application of quantitative models in diagnosis and curative treatment. In this paper some of these trends are discussed. In this way a (potential) impact of (control) engineering techniques on health care is illustrated.

An account is given of useful and effective applications of information theory in identification and some new problems are indicated, which can be solved within an information-theoretic framework. In the first part of the paper a brief survey is presented of information measures and their mutual relations. Restriction has been made to probabilistic information measures. Non-probabilistic measures did not find any application in identification so far. The second part of the paper deals with the practical applications of information theory in identification. Special attention is paid to problems, the solution of which in the information-theoretic framework seems to be the most effective or perhaps even the only possible one*).

In electrical and control engineering education the stochastic signal theory is gradually evolving to a middle-undergraduate level. During pre- and early-university training the student has been exposed predominantly to deterministic thinking. This implies that for numerous students the mastering of stochastic concepts is quite a difficult assignment. In order to provide an effective aid for this learning process, a series of lecture demonstrations and laboratory experiments has been developed and built. By means of these devices the student may strengthen his knowledge and insight into aspects of stochastic processes such as: probability density functions including conditional ones auto- and cross-correlation coefficients as well as -functions; the influence on accuracy of a measurement interval of finite length; measurement of dynamic characteristics by crosscorrelation techniques; etc. An outline is given of the underlying philosophy, of the demonstrations and the experiments. An indication is given of several inexpensive instruments that have been designed with the above-mentioned goals in mind. A more detailed technical description of the instruments is given separately (Gerretsen and Eykhoff, 1988; in the sequel this paper is referred to as 'op cit.').

It is well known that the (naive) least squares approach to the estimation problem in functional relationships leads to inconsistent results if the information matrix is not known exactly. In this paper an estimation method, denoted by “Least-Squares-Like”, is developed. This method can be used if two sets of observations of the explaining variables, both containing disturbances, are available. Its statistical properties are derived. The method is discussed from the point of view of system parameter estimation using input and output measurements contaminated by noise. The properties of this estimator are compared to those of the maximum likelihood one. The instrumental variable method can be considered as a special case of this least-squares-like approach.

Using cross-validation ideas, two procedures are proposed for making a choice between different model structures used for (approximate) modelling of multivariable systems. The procedures are derived under fairly general conditions: the ‘true’ system does not need to be contained in the model set; model structures do not need to be nested and different criteria may be used for model estimation and validation. The proposed structure selection rules are shown to be invariant to parameter scaling. Under certain conditions (essentially requiring that the system belongs to the model set and that the maximum likelihood method is used for parameter estimation) they are shown to be asymptotically equivalent to the (generalized) Akaike structure selection criteria.

To many engineers and scientists, "experts" as well as non-experts in identification, this field presents itself still as "a bag of tricks" due to the great variety of methods advocated by numerous authors. Yet a multitude of methods can be presented as elements in a coherent picture, particularly for model-representations that are linear-in-the-parameters. Such methods can be recognized as belonging to one of three related classes, with specific properties. This leads to increased insight and to new, practical estimation schemes, adaptable for a wide variety of situations.

The increasing applicability of system identification, as noticed in purely-engineering problems, can also be recognized in bio-medical tasks. This development of actual applications takes place in spite of the well-known specific limitations and problems related to such bio-systems: -restricted amount of a priori qualitative Knowledge; -very limited (if any) quantitative knowledge; -nonlinearities; -variability in time of the objects; -variability between objects; -poor measurability/observability; -difficult experimental conditions, as well as limitations due to ethical aspects In this workshop the attention is focussed on the continuing development of (potential) successful applications, with an emphasis on the more recent contributions. Ways of application include situations like: research, diagnosis, medication, surgery, rehabilitation, medical training This paper summarizes a number of pertinent general considerations. The fields of application discussed in this workshop include: - metabolic and endocrine systems - respiratory and cardiovascular systems - neuromuscular systems - sensory systems and neuronal systems One has to realize, of course, that these fields do not constitute an exhaustive list. They are, however, certainly representative for the challenges, feasibilities and limitations in this fascinating domain of human endeavour

Two criteria for choosing between different model-structures are proposed. Their derivation is within a natural cross-validatory assessment context and is fairly assumption-free. In particular, the two criteria can be used for discriminating between non-nested model structures and, more importantly, the ‘true’ system is not required to belong to the considered set of models. Should the true system belong to the model set, the two proposed criteria will asymptotically reduce to some well-known structure selection criteria. This is believed to be a desirable feature of our proposals. On the other hand, it provides a nice cross-validation interpretation of some well-known model structure selection rules. Also, the cross-validation interpretation helps one to choose which criteria to use in a given application The paper also has a second purpose, somewhat decoupled from that mentioned above. It contains an extensive survey of the literature that is useful in its own right.

Forced by university curriculum reform (5-→ 4-year curriculum duration) the introductory control engineering course has been shifted from the 3rd year to the 2nd year and the total teaching program has been reconstructed. Therefore the idea originated to concentrate the laboratory experiments around one central process, on which sequentially various aspects can be studied, such as measurements, modelling, identification, stochastic characteristics, analog control, digital control, etc. This central process consists of three cascaded water reservoirs. The students are required to study this process over a period of five afternoons. They have to develop their own models, make the measurements, identify the system, estimate the parameters and verify the results under rather unprescribed circumstances. They themselves will conclude that, due to very big time constants, simulation is necessary. In these simulations comparisons of analog and digital simulations are made and finally an optimal controller has to be developed. In addition, special attention has been paid to the various tasks that can be done using the available small digital computers, e.g. acquisition, storage and displaying of data, aid for computations such as s → z-transformation, simulation, digital control, threshold checking etc. The process and some preliminary experiences will be reported on.

In this paper the problem of input signal design for system identification is considered. The input signal is amplitude-constrained and the case of a finite-length Input sequence Is Investigated. Various optiraality criteria are used to derive the optimal identifying input sequence. In particular: Rissanen's Minimum Description length (MHL) criterion, the determinant of the information matrix, the information matrix determinant ratio, and the trace of the parameter error covariance matrix are considered as the criteria for optimality of identification. These input design criteria can also be used in structure selection (multivariate systems) or in model order (univariate systems) estimation procedures. The emphasis is pur on autoregressive models, and the input signal sequences are derived using a dynamic programming approach. A comparative study based on analytical results as well as on simulations is presented.

Based on the notions of template functions a multitude of system parameter estimation methods is presented as a coherent picture. This leads to increased insight and to new, practical estimation schemes, adaptable for a wide variety of situations. The goal of parameter estimation is the derivation of quantitative knowledge on unknown characteristics of dynamical systems. Consequently it is of basic interest to determine whether, fundamentally, the particular situation at hand offers enough information, i.e. whether the best estimation algorithm available could indeed provide the knowledge required. The this question is about the efficiency of the various estimation algorithms, i.e. the part of the potentially available information that becomes artually present. In this paper partial answers to these questions are given. This holds for recursive estimation as well as for “one shot” estimation in cases of lin-earity in the parameters.

Occasionally, system identification is called the measuring of ‘unmeasurable’ quantities. It is characterised by a marked ambiguity. On the one hand, a number of its theoretical aspects clearly belong to the realm of science ; on the other hand, when applied to practical problems, identification shows traits that refer to it being an art as well. It is this duality that is partly responsible for the wide variety and the large number of publications, differing in scope, in depth, in practicability and in appeal. In the present paper, a few aspects of this duality of character are being discussed through: on the one hand, indicating some coherence of the many parameter estimation schemes that have been proposed and advocated in literature; this provides an example of a part of the identification field that is clearly developing towards a ‘scientific’ framework; on the other hand, outlining essential features of identification procedures, related to practical applications as well as limitations, that can still be characterised as belonging to the ‘art’ of the experimenter/model builder.

In this paper the problem of input signal design for parameter and order estimation in linear dynamical systems is considered. The input signal is assumed to be amplitude constrained. The accuracy criterion of the parameter estimates is given by the determinant ratio of the approximated covariance matrix. A model order test is proposed using the instrumental variable approach. The influence of input signal design on the discrimination power of the order test is discussed.

In electrical and control engineering education the stochastic signal theory is gradually evolving to a middle-undergraduate level. During pre- and early-university training the student has been exposed predominantly to deterministic thinking. This implies that for numerous students the mastering of stochastic concepts is quite a difficult assignment. In order to provide an effective aid for this learning process a series of lecture demonstrations and laboratory experiments has been developed and built. By means of these devices the student may strengthen his knowledge and insight into aspects of stochastic processes like: probability density functions including conditional ones; auto- and crosscorrelation coefficients as well as -functions; the influence on accuracy of a finite measurement interval; measurement of dynamic characteristics by crosscorrelation techniques; etc. Separately (Eykhoff and Gerretsen, 1988; in the sequel indicated by 'op.cit.') a general outline as well as a survey is given of the philosophy of the demonstrations and the experiments. In this paper a more detailed technical description is given of several inexpensive instruments, that have been designed with the above mentioned goals in mind.

The meaning of the terms process identification, adaptive and optimalizing control is indicated. The basic difference between these types of control and the ordinary ones is the need for a "learning process". Using multiplication plus integration as the simplest possible type of "learning", simple schemes are obtained in which sinusoidals or random fluctuations may be used as test signals. A more detailed study and an analog simulation have been made of such an optimalizing system. This one is free from the defects of formerly proposed systems, i.e.: - the detrimental effects of system dynamics on the optimalizing control action, - the loop gain changes and even instability of the optimalizing control loop when the process dynamics change with time. This has been achieved by introducing a model of the process dynamics, the parameter(s) of which are controlled by an identifying-control loop. A typical value for the convergence speed towards the optimum of the optimalizing control can be given as follows: using a sinusoidal test signal, the transient for a step-disturbance is approximately of the form \epsilon^{-t/\tau} where the value of τ corresponds with three periods of the test signal used. Although the system is nonlinear and time varying, some analytical results have been obtained which check the computer results. The ideas presented can be extended to more-dimensional optimalization and more-parameter identification.

The paper presents a coherent picture of the parameter-estimation problem. Starting from the theory of minimum risk- or Bayes estimation the paper shows how other statistical estimation techniques can be interpreted as special cases (viz. maximum likelihood-, Markov-, and least squares estimation), The most important properties of these estimates are summarized. The engineering approaches based on these statistical techniques can be divided into two classes, viz. ''using explicit mathematical relations'' and ''using adjustment of a model''. Each of these classes is discussed briefly. The majority of parameter estimation techniques can be embodied in this framework. A very brief discussion is given on the problem of process state estimation which is related to parameter estimation. A few examples are used to illustrate the notions presented and to indicate some engineering considerations.

This paper describes the construction of an instrument for measuring the impedance using a physical model. The model has to be adjusted in such a way that its differential equation corresponds with that of the unknown impedance. For the implementation of model and adjustment mechanism, standard hybrid computer components are used. Also the testsignal is easily implemented and is composed in such a way that it is possible to measure three parameters of the impedance simultaneously with little dependence between the parameter adjustments. Some attention is paid to the stability-problem.

The purpose of this paper is to give a coherent picture of the process-parameter estimation. It offers a philosophy by introducing a "generalized model." From this, two classes of schemes can be derived. In each class there is quite some freedom with respect to the choice of elements of the scheme, the choice of the error criterion, and the choice of signal representation (continuous, sampled, quantized, etc.). These choices account for a whole range of estimating systems. The dynamic properties of many of these systems can be studied by the approach presented in this paper. The extension to estimation of nonlinear processes is also discussed.

In electrical and control engineering education the stochastic signal theory is gradually evolving to a middle-undergraduate level. During pre- and early-university training the student has been exposed predominantly to deterministic thinking. This implies that for numerous students the mastering of stochastic concepts is quite a difficult assignment.

This paper attempts to summarize some recent essential developments in the field of system identification and parameter estimation. The topics discussed include: the identification protocol, model representations, estimation methods, identification tools, identification applications, expectations on the further developments. It is recognized that identification in an engineering sense has both a ‘science’ and an ‘arts’ aspect.

In electrical and control engineering education the stochastic signal theory is gradually evolving to a middle-undergraduate level. During pre- and early-university training the student has been exposed predominantly to deterministic thinking. This implies that for numerous students the mastering of stochastic concepts is quite a difficult assignment. In order to provide an effective aid for this learning process, a series of lecture demonstrations and laboratory experiments has been developed and built. By means of these devices the student may strengthen his knowledge and insight into aspects of stochastic processes such as: probability density functions including conditional ones auto- and cross-correlation coefficients as well as -functions; the influence on accuracy of a measurement interval of finite length; measurement of dynamic characteristics by crosscorrelation techniques; etc. An outline is given of the underlying philosophy, of the demonstrations and the experiments. An indication is given of several inexpensive instruments that have been designed with the above-mentioned goals in mind. A more detailed technical description of the instruments is given separately (Gerretsen and Eykhoff, 1988; in the sequel this paper is referred to as ‘op cit.’).