摘要:

The computation of bounds on the parameters, rather than point estimates and covarianccs, is considered for autoregressive-moving-average-exogenous (ARMAX) models. The bounds derive from observations affected by bounded noise. The influence of noise structure on the performance of the best-known parameter-bounding algorithm is examined. A problem analogous to the bias in least-squares estimates when regressors and noise are mutually correlated is also found to arise in parameter-bounding. An explanation is offered, and the possibility of ARMAX parameter bounds being non-convex is pointed out.

ISSN:0020-7179    

DOI:10.1080/00207178708933738

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

This paper is concerned with the robustness of LQ state feedback regulators in the presence of system uncertainty. Assuming that the perturbations satisfy the matching condition, several robust stability results are explicitly derived for linear, non-linear, time-varying, and dynamical perturbations in terms of bounds of perturbations and the weighting matrices in the performance index. Some examples are included to show the applicability and the limitation of the present results.

ISSN:0020-7179    

DOI:10.1080/00207178708933739

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

This paper presents a method for designing a type-one servomechanism for a conlinuous-time linear system subject to a previewable desired output and an unmeasurable step disturbance. By applying the linear quadratic integral technique, a state-feedback controller with integral and preview actions is derived. We show that the closed-loop system is asymptotically stable and hence achieves a complete regulation under the stabilizability and detectability assumption. We also show that the preview servomechanism problem is the dual of a smoothing problem for a linear stochastic system with a brownian-motion input. An example is included to illustrate the transient response and frequency characteristics of the closed-loop system.

ISSN:0020-7179    

DOI:10.1080/00207178708933740

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

Harris and Mellichamp (1980) demonstrated the successful use of multifrequency binary sequences in process identification. In the present work it is shown how the open-loop frequency response of the process generated by using multifrequency binary sequences can be exploited to identify the model order and obtain initial parameter estimates. This not only avoids the necessity of knowing the model order a priori, but also yields faster convergence to the optimal parameter values. The proposed technique, when extended to model reduction, is shown to yield equivalent results with only eight frequencies as against the hundred used by Luus (1980(. This can result in a considerable saving of computer time.

ISSN:0020-7179    

DOI:10.1080/00207178708933741

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

Given a general non-linear plant, a method is presented for designing model reference feedback controllers for the plant. It is shown that the feedback controllers produced will guarantee prespecified closed-loop properties if the error between the plant and model lies within computable bounds. In particular, an upper bound on the tracking error is established and conditions are given that ensure closed-loop stability.

ISSN:0020-7179    

DOI:10.1080/00207178708933742

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

Methods of identifying bilinear systems from recorded input-output data are discussed in this article. A short survey of the existing literature on the topic is given. ‘Standard’ methods from linear systems identification, such as least squares, extended least squares, recursive prediction error and instrumental variable methods are transferred to bilinear, input-output model structures and tested in simulation. Special attention is paid to problems of stabilizing the model predictor, and it is shown how a time-varying Kalman filter and associated parameter estimation algorithm can deal with this problem.

ISSN:0020-7179    

DOI:10.1080/00207178708933743

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

An (L2-)stability criterion is derived for non-linear time-varying (NLTV) systems with a separable gain(k(t).f(.))in the feedback-path, wheref(.)is monotone, using a simple R-L multiplier (s + β)/(s + α)(α > β) and involving the first and second derivatives ofk(t). The criterion seems to be very much relaxed for systems with a slowly time-varying gaink(t).

ISSN:0020-7179    

DOI:10.1080/00207178708933744

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

Following the work of Huntet at. (1986), we derive necessary and sufficient conditions for equivalence of time-varying non-linear systems with time-varying linear ones.

ISSN:0020-7179    

DOI:10.1080/00207178708933745

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

In association with robust control-system design and analysis, the Hurwitz property of interval matrices and interval polynomials has recently been actively investigated. However, its discrete counterpart, the convergence property, has seemingly not been much discussed. In this paper, this property is studied in comparison with the Hurwitz counterpart. Some conditions under which interval matrices or interval polynomials are convergent are derived.

ISSN:0020-7179    

DOI:10.1080/00207178708933746

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor

摘要:

A self-tuning algorithm is proposed, which minimizes a certain quadratic criterion with closed-loop error transfer-function pole and zero placement. The necessary and sufficient conditions for simultaneous error transfer-function pole and zero assignment have been given. Simulated examples illustrate the usefulness of the suggested method in solving the model-following problem.

ISSN:0020-7179    

DOI:10.1080/00207178708933747

出版商:Taylor & Francis Group    

年代:1987

数据来源: Taylor