A link between H control of a discrete-time nonlinear system and its linearization (Q4259098)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1326183
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | A link between H control of a discrete-time nonlinear system and its linearization |
scientific article; zbMATH DE number 1326183 |
Statements
A link between H control of a discrete-time nonlinear system and its linearization (English)
0 references
4 April 2000
0 references
nonlinear \(H_\infty\) control
0 references
discrete time nonlinear systems
0 references
linearization
0 references
0.870398998260498
0 references
This paper studies relations between nonlinear \(H_\infty\) control of discrete time nonlinear systems and \(H_\infty\) control of the linearized system. The nonlinear system under consideration is described by \[ x_{k+1}= A(x_k)x_k+ B_1(x_k) w_k+ B_2(x_k) u_k,\;x_0=0, \] \[ z_k= C_1(x_k)x_k+ D_{11}(x_k) w_k+D_{12}(x_k) u_k,\tag{1} \] \[ y_k= C_2(x_k)x_k+ D_{21}(x_k) u_k, \] where \(w_k\in \mathbb{R}^r\) is the disturbance input which belongs to \(l_2[0,\infty)\), (i.e. \(\sum^\infty_{k=0}\|w_k\|^2<\infty\)), \(z_k\in \mathbb{R}^p\) is the regulated output and \(y_k\in \mathbb{R}^q\) is the measured output. All matrix functions in (1) are assumed to be continuous. Given a scalar \(\gamma>0\), the local problem of nonlinear \(H_\infty\) control is to design a nonlinear controller \(u={\mathcal K}(y)\), where \({\mathcal K}\) can be a dynamic controller, such that the resulting closed-loop system of (1) is internally stable on a neighbourhood \({\mathcal W}\) of the origin and \[ \|z\|_2< \gamma\|w\|_2\tag{2} \] for any nonzero \(w\in l_2[0,\infty)\) (i.e. \(\sum^\infty_{k= 0}\|w_k\|^2<\infty\)) for which the state trajectory of the closed-loop system does not leave \({\mathcal W}\). Associated with the system (1), we introduce a linearized system \[ \overline x_{k+1}=\overline A\overline x+\overline B_1 w_k+\overline B_2 u_k, \] \[ \overline z_k= \overline C_1\overline x_k+\overline D_{11}w_k+\overline D_{12}u_k,\tag{3} \] \[ \overline y_k=\overline C_2\overline x_k+\overline D_{21} w_k, \] where \[ \overline A= A(0),\quad \overline B_1= B_1(0),\quad \overline B_2= B_2(0),\quad\overline C_1= C_1(0),\quad\overline C_2= C_2(0), \] \[ \overline D_{11}= D_{11}(0),\quad\overline D_{12}= D_{12}(0),\quad\overline D_{21}= D_{21}(0). \] It is shown that an input-output discrete-time nonlinear system has a local \(L_2\) gain \(\gamma>0\), with internal stability, if its linearized system is asymptotically stable and has an \(L_2\) gain strictly less than \(\gamma\). Using this fundamental result, it is further proved that the existence of a local solution to the problem of disturbance attenuation with internal stability for nonlinear systems is guaranteed by the solution to the corresponding \(H_\infty\) control problem for the linearized systems. All the results are established directly via the simple and intuitive linearization technique which provides, in contrast with the existing approach presented by \textit{W. Lin} and \textit{C. I. Byrnes} [IEEE Trans. Autom. Control 41, No. 4, 494-510 (1996; Zbl 0857.93035)], a new insight and an appealing alternative to the discrete-time nonlinear \(H_\infty\) control problem.
0 references