See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224801193

Robust Control

Chapter  in  Lecture Notes in Control and Information Sciences · January 1993

DOI: 10.1007/11375753_4 · Source: DLR

CITATIONS READS

16 287

5 authors, including:

Andrew Bartlett Dieter Kaesbauer

The MathWorks, Inc Munich University of Applied Sciences
24 PUBLICATIONS   732 CITATIONS    32 PUBLICATIONS   880 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Dieter Kaesbauer on 27 January 2015.

The user has requested enhancement of the downloaded file.

 GRADUATE COURSE ON
POLYNOMIAL METHODS FOR
ROBUST CONTROL

Didier HENRION
www.laas.fr/∼henrion
henrion@laas.fr

Laboratoire d’Analyse et d’Architecture des Systèmes

Centre National de la Recherche Scientifique
Toulouse

Ústav Teorie Informace a Automatizace

Akademie Věd České Republiky
Praha

Universidad de los Andes

Mérida, Venezuela
October-November 2001
 Course outline

I Robust stability analysis (Part I)

I.1 General introduction - linear systems, polynomial methods
and robust control
I.2 Single parameter uncertainty - eigenvalue criteria
I.3 Interval uncertainty - Kharitonov’s theorem

II Robust stability analysis (Part II)

II.1 Polytopic uncertainty - edge theorem
II.2 Multilinear and polynomic uncertainty - mapping theorem

III Robust design and convex optimization

III.1 Robust pole placement - approximation of stability region
III.2 Rank-one robust stabilization - Youla-Kučera and
Rantzer-Megretski parametrizations
III.3 Simultaneous stabilization - strong stabilization, Hermite
criterion, open problems

IV New results on robust analysis and design

IV.1 Robust stability analysis - Linear matrix inequalities and
positive polynomial matrices
IV.2 Robust stability design - Numerical examples
 Course Outline

I Robust stability analysis (Part I)

II Robust stability analysis (Part II)
III Robust design and convex optimization
IV New results on robust analysis and design

Scope of the course

Systems Methods
• linear • frequential
• non-linear • state-space
• polynomial

Control scheme Uncertainty

• adaptive • non-parametric
• stochastic • parametric
• robust
 Linear systems

Theory much better developed than for

non-linear systems

Provides background for nonlinearists

Availability of powerful CACSD tools to solve

numerical linear algebra problems

(Old) reference books

• Kailath. Linear systems. Prentice Hall, 1980
• Chen. Linear system theory and design. HRW, 1984

Computer tools used for the course

Matlab 6.0 Polynomial

Toolbox 2.5

www.mathworks.com www.polyx.cz
 Polynomial methods

Based on the algebra of polynomials and

polynomial matrices, typically involve
• linear Diophantine equations
• quadratic spectral factorization

Pioneered in central Europe during the 70s

mainly by Vladimı́r Kučera from the former
Czechoslovak Academy of Sciences

Network funded by the European commission

www.utia.cas.cz/europoly

Polynomial matrices also occur in Jan Willems’

behavorial approach to systems theory

Alternative to state-space methods developed

during the 60s most notably by Rudolf Kalman
in the USA, rather based on
• linear Lyapunov equations
• quadratic Riccati equations
 Ratio of polynomials

A scalar transfer function can be viewed as the

ratio of two polynomials

Example
Consider the mechanical system

k2 u
m
k1

• y displacement • u external force

• k1 viscous friction coeff • k2 spring constant
• m mass

Neglecting static and Coloumb frictions, we

obtain the linear transfer function
y(s) 1
G(s) = =
u(s) ms2 + k1s + k2
 Ratio of polynomial matrices

Similarly, a MIMO transfer function can be

viewed as the ratio of polynomial matrices
−1 −1
G(s) = NR (s)DR (s) = DL (s)NL(s)
the so-called matrix fraction description (MFD)

Lightly damped structures such as oil derricks,

regional power models, earthquakes models,
mechanical multi-body systems, damped gyro-
scopic systems are most naturally represented
by second order polynomial MFDs
(D0 + D1s + D2s2)y(s) = N0u(s)
Example
The (simplified) oscillations of a wing in an air stream
is captured by properties of the quadratic polynomial
matrix [Lancaster 1966]
   
121 18.9 15.9 7.66 2.45 2.1
D(s) = 0 2.7 0.145 + 0.23 1.04 0.223 s+
11.9 3.64 15.5 0.6 0.756 0.658
 
17.6 1.28 2.89
1.28 0.824 0.413 s2
2.89 0.413 0.725
 First-order polynomial MFD

Example
RCL network

R
L
C
u y1 y2

• y1 voltage through inductor

• y2 current through inductor
• u voltage

Applying Kirchoff’s laws and Laplace transform we get

    
1 −Ls y1 (s) 0
= u(s)
Cs 1 + RCs y2 (s) Cs
and thus the first-order left system MFD
 −1  
1 −Ls 0
G(s) = .
Cs 1 + RCs Cs
 Second-order polynomial MFD

Example
mass-spring system

Vibration of system governed by 2nd-order differential

equation M ẍ + C ẋ + Kx = 0 where e.g. n = 250, mi =
1, κi = 5, τi = 10 except κ1 = κn = 10 and τ1 = τn = 20

Quadratic matrix polynomial

D(s) = M s2 + Cs + K
with
M =I
C = tridiag(−10, 30, −10)
K = tridiag(−5, 15, −5).
 Another second-order polynomial MFD

Example
Inverted pendulum on a cart

Linearization around the upper vertical

position yields the left polynomial MFD
(M + m)s2 + bs lms2
h ih i h i
x(s) 1
= f (s)
lms2 (J + l2 m)s2 + ks − lmg φ(s) 0

With J = mL2/12, l = L/2 and g = 9.8, M =

2, m = 0.35, l = 0.7, b = 4, k = 1, we obtain
the denominator polynomial matrix
" #
5s + 3s2 0.35s2
D(s) =
0.35s2 −3.4 + s + 0.16s2
 More examples of polynomial MFDs

Higher degree polynomial matrices can also

be found in aero-acoustics (3rd degree) or in
the study of the spatial stability of the Orr-
Sommerfeld equation for plane Poiseuille flow
in fluid mechanics (4rd degree)

Pseudospectra of Orr-Sommerfeld equation

For more info see Nick Higham’s homepage at

www.ma.man.ac.uk/∼higham
 Uncertainty

When modeling systems we face several sources

of uncertainty, including

• non-parametric (unstructured) uncertainty

• unmodeled dynamics
• truncated high frequency modes
• non-linearities
• effects of linearization, time-variation..
• parametric (structured) uncertainty
• physical parameters vary within given bounds
• interval uncertainty (l∞)
• ellipsoidal uncertainty (l2)
• l1 uncertainty

How can we overcome uncertainty ?

• model predictive control
• adaptive control
• robust control
 Robustness

We seek a control law valid over the whole

range of admissible uncertainty

• off-line
• simple
• cheap
• secure

Basically we will study two classes of problems

Robust stability analysis
Parts I, II & IV of this course

Robust controller design

Parts III & IV of this course
 Course material

Information on the course can be found at

www.laas.fr/∼henrion/courses/polyrobust.html

Most of the material of the first two parts of

the course (robust stability analysis) is taken
from the textbooks
• J. Ackermann. Robust control: systems with uncertain physical
parameters. Springer, 1993
• B. R. Barmish. New tools for robustness of linear systems.
MacMillan, 1994
• S. P. Bhattacharyya, H. Chapellat, L. H. Keel. Robust control -
The parametric approach. Prentice Hall, 1995

The third part (robust design) describes recent

results published from 1992 to 1999 in
• IEEE Transactions on Automatic Control
• IFAC Automatica
• System and Control Letters
• International Journal of Control
• SIAM Journal on Control and Optimization

The last part of the course contains mostly

new, previously unpublished material
View publica