Design of a class of rigidly structured control systems with adaptive properties. I: Asymptotic properties of the roots of characteristic equations of control systems that are stable under infinitely increasing gains (Q581316)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: Design of a class of rigidly structured control systems with adaptive properties. I: Asymptotic properties of the roots of characteristic equations of control systems that are stable under infinitely increasing gains |
scientific article; zbMATH DE number 4018863
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Design of a class of rigidly structured control systems with adaptive properties. I: Asymptotic properties of the roots of characteristic equations of control systems that are stable under infinitely increasing gains |
scientific article; zbMATH DE number 4018863 |
Statements
Design of a class of rigidly structured control systems with adaptive properties. I: Asymptotic properties of the roots of characteristic equations of control systems that are stable under infinitely increasing gains (English)
0 references
1986
0 references
The following problem, which has an engineering application in designing stable conrollers for plants whose transfer functions have coefficients which are known only by their upper and lower bounds, is considered: Let \[ (1)\quad F_ 0(p)+\sum^{n-1}_{t=1}\sum_ tK_{i_ 1}...K_{i_ t}Q_{tj_ t}(p)+\prod^{n}_{i=1}K_ iF_ n(p)=0 \] where \(\Sigma_ t\) means that for a fixed t, the set of indices \(i_ 1,...,t_ t\) runs through all the \(C^ t_ n\) distinct collections of values belonging to the set (1,...,n), whereas the index \(j_ t\) varies from 1 to \(C^ t_ n\). \(F_ 0(p)\), \(Q_{tj_ t}(p)\) and \(F_ n(p)\) are polynomials in p whose degrees are denoted by (2) deg \(F_ 0(p)=N_ 0\), deg \(F_ n(p)=N_ n\), deg \(Q_{t_ 1}(p)=N_ t\) and satisfy the conditions (3) \(N_ 0>N_ 1>...>N_ n\); \(N_ t\geq \deg Q_{tj_ t}(p)\). The polynomials on the left hand side of (1) are ``interval polynomials'', i.e. their coefficients are specified only by lower and upper bounds. When \(K_ i\to \infty\), \(N_ 0-N_ n\) roots of (1) approach infinity. The problem considered is the roots approach infinity only in the left half complex plane, with as weak as possible constraints on the bounds of the coefficients. The following assumption with regard to \(K_ i\to \infty\), \(i=1,...,n\), is made in the paper: \[ (4)\quad K_ i=C_ iK^{\mu_ i},\quad C_ i>0,\quad \mu_ i\geq \mu_{i+1},\quad K\to \infty. \] For this assumption several necessary conditions with regard to the coefficients bounds are obtained, such that the roots of (1) approach infinity in the left half plane. The proofs are based primarily on Newton's diagrams. The difficulty with ``interval coefficients'' is overcome by using a theorem published in 1979 which expresses stability of an ``interval polynomial'' in terms of stability of four polynomials with fixed coefficients determined by the bounds.
0 references
stable conrollers
0 references
interval polynomials
0 references
Newton's diagrams
0 references
interval coefficients
0 references
time-invariant
0 references