摘要:

Quite recently, examples of non-linear adaptive controllers have been found that adaptively stabilize any scalar first-order linear system. In this paper we describe two general classes of such ‘universal’ adaptive stabilizers that include the previously proposed controllers as special cases, Different types of perturbations of these controllers are studied and we report on extensive simulation experiments. Our results demonstrate considerable qualitative differences in the dynamical behaviour of various adaptive universal stabilizers. While the previously proposed controllers all exhibit the bursting phenomena, we give an explicit example of a stabilizing controller that does not show bursting behaviour.

ISSN:0020-7179    

DOI:10.1080/00207178808906241

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

The problem of stabilizing a non-linear recursive system ∑ using a closed-loop configuration with compensation and feedback is considered. Most attention is devoted to the design of the dynamical behaviour of the stabilized closed-loop configuration. It is shown that, except for some obvious restrictions, any stable dynamics can be assigned to the closed-loop system. Compensators that yield the desired dynamics are explicitly constructed in implementable recursive form, and their formulae are expressed in terms of quantities derived directly from the given descriptions of the system ∑ and the desired dynamics. It is assumed that the state of the system ∑ is accessible. The resulting closed-loop configurations are internally stable.

ISSN:0020-7179    

DOI:10.1080/00207178808906242

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

A recent parameterization of the class of linear output-feedback controllers that assign a set of desired self-conjugate eigenvalues to the closed-loop system is used to formulate and solve a fundamental response insensitivity problem. It is established to what extent output-feedback control can be used to render the closed-loop system response insensitive to possibly many not necessarily small parameter variations in the open-loop state space model. A non-conservative sequential design procedure is developed for making as many of the closed-loop system eigenmodes as possible totally insensitive while retaining arbitrary assignment of the maximum number of closed-loop eigenvalues. The main result is a class of desensitizing fixed-gain output-feedback controllers explicitly specified by a set of free parameters which may be chosen to satisfy additional design requirements. Conditions are given for total modal decoupling insensitivity to possibly many not necessarily small parameter variations in the open-loop state model. The mechanism of modal decoupling insensitivity for a given mode is interpreted in terms of the insensitivity of the corresponding left eigenmode. A parametric approach to output feedback design for modal decoupling insensitivity is discussed.

ISSN:0020-7179    

DOI:10.1080/00207178808906243

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

The problem of designing stable, robust closed-loop systems is considered in this paper. A new bound of robustness measure is introduced. Free design parameters are exploited in the achievement of desirable performance. Designs of minimal and reduced-order dynamic output controllers are carried out using special algorithms and an illustrative example of the design method is included.

ISSN:0020-7179    

DOI:10.1080/00207178808906244

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

Properties of discrete time systems obtained by sampling continuous time systems are described. By introducing prefilters, we can treat different ways of sampling within one framework. Results on the convergence of poles and zeros of transfer functions and noise filters as the sampling interval tends to zero are given. These results are generalizations of the results of Åström, Hagander and Sternby (1984) on the convergence of poles and zeros for zero-order hold sampled transfer functions. Sampled noise models are also analysed. Knowledge of these properties is very important in, for example, discrete time simulations of continuous time systems, and identification of continuous time models based on discrete time measurements.

ISSN:0020-7179    

DOI:10.1080/00207178808906245

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

A complete solution is given for the inverse problem of LQ optimal regulators for the discrete-time multi-input case. The Kalman equation for discrete-time systems is derived to characterize the optimal regulators. It is then shown that this equation holds for some quadratic weighting if and only if the feedback satisfies a condition corresponding to the circle criterion for continuous-time systems. The optimality criterion given here consists of this condition and an additional condition to ensure detectability. An example is given to illustrate the results.

ISSN:0020-7179    

DOI:10.1080/00207178808906246

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

A constructive algorithm is presented for solving a two-dimensional (2-D) polynomial matrix equation that plays an important role in the output feedback stabilization of multivariable 2-D linear systems. It is shown that this algorithm can have some computational advantages over the previous known ones based on the Gröbner-basis approach. As an application of the algorithm, the class of all stabilizing strictly causal 2-D compensators is parametrized for a given stabilizable causal (not necessarily strictly causal) 2-D plant.

ISSN:0020-7179    

DOI:10.1080/00207178808906247

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

This note presents a method for designing the linear feedback of a single-input single-output (SISO) bilinear system. By using the concept of matrix measure guaranteed estimations are obtained for the main characteristics of the control system. These results allow a quite simple selection of the feedback gain according to the design specifications.

ISSN:0020-7179    

DOI:10.1080/00207178808906248

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

This paper introduces a new synthesis method, based on a computation method for solving H∞-optimization problem given by Glover (1986) and a technique ofjω-axis shifting, for designing an H∞-optimal robust controller for the MIMO system. With this method we can optimize the excess stability margin under the requirement that all the poles of the nominal closed-loop system (Cl-system) be located left of thejω-axis with a required distancedand ensure asymptotic regulation and disturbance rejection for a class of system input and disturbance if the plant perturbation does not exceed the excess stability margin. Another advantage is that there are not any constraints on the poles and zeros of the plant when using this method

ISSN:0020-7179    

DOI:10.1080/00207178808906249

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor

摘要:

Network realzability theory provides the basis for a unified approach to the stability of a polynomial or a family of polynomials. In this paper conditions are given, in terms of certain decompositions of a given polynomial, that are necessary and sufficient for the given polynomial to be Hurwitz. These conditions facilitate the construction of stability domains for a family of polynomials through the use of linear inequalities. This approach provides a simple interpretation of recent results for polynomials with real coefficients and also leads to the formulation of corresponding results for the case of polynomials with complex coefficients.

ISSN:0020-7179    

DOI:10.1080/00207178808906250

出版商:Taylor & Francis Group    

年代:1988

数据来源: Taylor