摘要:

In this paper, we study a general class of linear time invariant discrete time distributed systems. We consider both single-input-single-output (SISO) and multi-input-multi-output (MIMO) systems, and study design procedures. We develop a commutative algebra of transfer function, b(p0), for a general class of SISO discrete time convolution systems, which covers sampled distributed systems and, of course, lumped systems as a special case. Each element of b(p0) is formulated as a ratio of two elements in an algebra l1−(p0) of causal p0-stable transfer functions. We demonstrate that l1−(p0) indeed a euclidean ring, give necessary and sufficient conditions for coprimeness between elements in l1−(p0) and characterize poles and zeros for elements in b(p0). In contrast to the algebra l1the algebra b(p0) includes both stable and unstable systems; furthermore since p0<1 this formulation allows us to study the dominant poles inside the unit disc of the complex plane. We study next MIMO systems whose transfer functions are matrices with elements in b(p0). We establish the matrix fraction representation theory and use it to develop : the dynamic interpretation of poles and transmission zeros, the feedback interconnection of such MIMO systems, and the problem of controller design to achieve stabilization (analogous to arbitrary closed-loop eigenvalue assignment), asymptotic tracking and disturbance rejection ; finally, for the case of stable square plants, we show how to achieve complete decoupling with detailed pole assignment and finite settling time, subject to, of course, the limitations imposed by the plant transmission zeros outside the open unit disc.

ISSN:0020-7179    

DOI:10.1080/00207178208932903

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

The parsimony principle is frequently used in system identification on more or less heuristical grounds. Recently, a formal proof of this principle has been given. Here the assumptions used in that proof are examined to some extent. It will be shown by means of some counterexamples that these assumptions cannot be relaxed unless additional restrictions are introduced. A result showing how the parsimony principle should be used when estimating the parameters of regression models by the least squares method is also presented.

ISSN:0020-7179    

DOI:10.1080/00207178208932904

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

The closed loop pole positions of a linear, multivariable system with unity rank output feedback are examined as a scalar feedback gain varies. In particular, consideration of the terminal locations of the poles as the gain increases indefinitely leads to a minor modification of a well known pole placement algorithm, resulting in a dyadic output feedback algorithm for zero assignment. This allows the designer to manipulate the root loci directly, and a consequence of the unity rank feedback is that the root loci of the multivariable system can be analysed in terms of a classical, scalar root locus diagram. The technique can be used to provide an improved ‘ point of departure ’ for the design of gain injecting compensators.

ISSN:0020-7179    

DOI:10.1080/00207178208932905

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

In this paper we present a simple procedure for obtaining stabilization of a class of continuous and discrete linear systems subjected to control structure constraints, giving more attention to the so-called decentralized control. Contrary to the continuous case where much work has been done, the solution presented here for the discrete case is one of the first to be presented. Some examples are treated and, for the continuous case, some comparisons are made with other methods found in the literature.

ISSN:0020-7179    

DOI:10.1080/00207178208932906

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

Linear discrete stochastic control systems containing unknown multiple time delays, plant parameters and noise variances are considered. An algorithm is established which uses the maximum-likelihood technique to identify the unknown parameters. An estimated likelihood function is evaluated based on the previous parameter estimates, which in turn generates a new descent direction vector to update the unknown parameters. The delays and plant parameters are identified in their respective parameter spaces. An example of a second-order stochastic system has been implemented by digital simulation to demonstrate the applicability of the algorithm.

ISSN:0020-7179    

DOI:10.1080/00207178208932907

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

This paper discusses the structure estimation problem in the identification of vector difference equations (VDE). The structural properties of VDEs derived from canonical state space and matrix fraction descriptions of multivariable linear systems are studied. Based on the rules established for the structure of canonical VDE representation, a residual error type structure seeking procedure is developed. The effectiveness of the proposed method is illustrated by simulation examples.

ISSN:0020-7179    

DOI:10.1080/00207178208932908

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

A method of identifying the structure and the corresponding parameters in a general single-input-single-output lumped continuous linear system from a knowledge of the input-output data is presented in this paper. In the identification of structure indices, such as system order etc., the present technique is found to be superior to the methods which are based on evaluation of rank of certain information matrices. The proposed method is successfully tested with illustrative examples with deterministic and noisy data.

ISSN:0020-7179    

DOI:10.1080/00207178208932909

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

This paper describes a general solution to the system identification problem. The technique is referred to as ‘ intermediate domain processing ’ and can result in faster and more efficient computations. The paper begins by outlining a general theory of translation invariant linear systems. By making specific choices within this general framework, different special cases can be derived. A method is developed which facilitates the conversion between data vectors associated with each of these special cases. This conversion technique is used in deriving a generalized solution to the system identification problem. A practical example is considered in which the unknown system is assumed to be time invariant linear (TIL), whereas computations are carried out in the domain of an equivalent and computationally advantageous dyadic invariant linear (DIL) system.

ISSN:0020-7179    

DOI:10.1080/00207178208932910

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

The necessary and sufficient conditions for the existence of an observer for linear multivariable system with immeasurable, arbitrary disturbances are formulated. The theorem is a generalization of the observability conditions for linear systems. Based on the proof of the theorem, the algorithm of an observer can be constructed. A numerical example is given.

ISSN:0020-7179    

DOI:10.1080/00207178208932911

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor

摘要:

Criteria are derived to determine the distribution of the characteristic roots of a Byatem with respect to parabolas and hyperbolas in the complex plane, which are related to certain physical properties of the system. The criteria are given in terms of a classical half-plane algorithm, e.g. a Routh array, carried out on a transformed polynomial of double the degree of the original characteristic polynomial. For parabolas, left, right or shifted, the algorithm provides the complete root distribution including criteria for root clustering and root exclusion as Bpecial cases. For hyperbolas, the algorithm provides in all cases criteria for root clustering inside the (open or closed) left or right branch of the hyperbola, and for characteristic polynomials which are Hurwitz, the algorithm provides the complete root distribution with respect to the left branch of the hyperbola. Also, by an additional classical test (Sturm), the algorithm provides in all cases the exact number of roots on the left boundary of the hyperbola and on the right boundary of the hyperbola. By a limiting process the results on hyperbolas are extended to linear sectors, and by another limiting process the results are extended to symmetrical intervals on the real axis.

ISSN:0020-7179    

DOI:10.1080/00207178208932912

出版商:Taylor & Francis Group    

年代:1982

数据来源: Taylor