Scribd
Upload
31 views

Week10-State-feedback Control-Part2

This document discusses robust state-feedback control systems. It presents theorems and methods for designing robust LQR and H-infinity state-feedback controllers to stabilize systems with parameter uncertainties and disturbances. Linear matrix inequalities are used to solve for controller gains.

Uploaded by

Reedus
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
31 views

Week10-State-feedback Control-Part2

This document discusses robust state-feedback control systems. It presents theorems and methods for designing robust LQR and H-infinity state-feedback controllers to stabilize systems with parameter uncertainties and disturbances. Linear matrix inequalities are used to solve for controller gains.

Uploaded by

Reedus
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

ECSL

State-feedback control systems

University of Ulsan
Prof. KIM

c
°2023-2024 All rights reserved
ECSL

4 Robust state-feedback control


• The considered system model is as follows:
¡ ¢ ¡ ¢
ẋ(t) = A + ∆A x(t) + B + ∆B u(t). (21)

• The parameter uncertainties ∆A ∈ Rn×n and ∆B ∈ Rn×m are


described as, respectively,

∆A = EΥH1 , ∆B = EΥH2 (22)

where E ∈ Rn×p , H1 ∈ Rq×n , and H2 ∈ Rq×m are known real


constant matrices; and Υ ∈ Rp×q is an unknown matrix but satisfies

ΥT Υ ≤ I. (23)
ECSL

• Let us employ the following state-feedback control law:

u(t) = Kx(t) (24)

where K ∈ Rm×n denotes the control gain to be designed later.


• As a result, the closed-loop control system is described as follows:
¡ ¢ ¡ ¢
ẋ(t) = A + EΥH1 x(t) + B + EΥH2 Kx(t)
¡ ¢
= A + BK + EΥ(H1 + H2 K) x(t)
¡ ¢
= Ā + EΥH̄ x(t) (25)

by defining

Ā = A + BK, H̄ = H1 + H2 K.
ECSL

Theorem 3. System (21) is robustly stable at the origin if there


exist matrices P̄ = P̄ T ∈ Rn×n , K̄ ∈ Rm×n , and a scalar ² > 0
satisfying

P̄ > 0 (26)
 © ª 
He AP̄ + B K̄ + ²EE T ?
  < 0. (27)
H1 P̄ + H2 K̄ −²I

Furthermore, the control gain can be reconstructed as follows:

K = K̄ P̄ −1 .
ECSL

Example 4.1. Obtain the linear matrix inequalities (LMI) needed


to design a robust LQR state-feedback controller for system (21).
ECSL

5 H∞ state-feedback control
• This state-feedback control is designed with the aim of minimizing
the effect of disturbances on the closed-loop system while using
feedback of the state variables.
• To begin with, let us consider the following linear system with
external disturbances belonging to the L2 space:

ẋ(t) = Ax(t) + Bu(t) + N w(t)
(34)
z(t) = Cx(t) + Du(t).
ECSL

• Let us employ the following state-feedback control law:

u(t) = Kx(t)

where K ∈ Rm×n is the control gain to be designed later.


• As a result, the closed-loop control system is described as follows:

ẋ(t) = Āx(t) + N w(t)
(35)
z(t) = C̄x(t)

where

Ā = A + BK, C̄ = C + DK.
ECSL

Theorem 4. System (34) is stable and has an H∞ disturbance


attenuation level γ, if there exist symmetric matrices P̄ ∈ Rn×n and
K̄ ∈ Rm×n , and a scalar γ > 0 satisfying

P̄ > 0 (36)
 © 
He AP̄ + B K̄} ? ?
 
 T 2
0 
 N −γ I  < 0. (37)
C P̄ + DK̄ 0 −I

Furthermore, the control gain can be reconstructed as follows:

K = K̄ P̄ −1 .

You might also like