2480 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 54, NO. 10, OCTOBER 2009
Fig. 1. Reduction of deformable mirror model. solid—original model,
dashed—Hankel model reduction, dashed—optimization based method
(QCO).
have close zero frequency response. The model is truncated to 20 states
by means of the described quasi-convex optimization technique (QCO
method), Hankel model reduction.
We implement QCO method on the frequency grid with 84 samples
with tolerance in bisection procedure
10
0
6
. The optimization together
with calculating frequency samples took 74 seconds and the resulting
approximation error is
2
:
9
1
10
0
5
. Hankel model reduction took around
20 minutes providing the error
7
:
98
1
10
0
5
. Results, see in the Fig. 1.
For the given frequency interval QCO provided a better model than
Hankel reduction. However, in general we do not expect QCO approx-
imations to be better than Hankel reduction approximations. This ex-
ample shows, that for large/medium scale systems we win sufficiently
in time and do not really lose in approximation quality.
VII. C
ONCLUSION
In this technical note we have discussed multi-input-multi-output ex-
tension of [5], where convex optimization is used to search for low
order models. We have shown that the same approximation gap bound
for MIMO extension stands as in SISO methods. The method can be
very useful for large scale systems, since it is rational fit algorithm with
stability guarantee and relatively low computational complexity.
A
CKNOWLEDGMENT
The authors would like to thank P. Hagander for valuable comments,
and P. Giselsson for providing the model of a deformable mirror.
REFERENCES
[1] K. Zhou, J. C. Doyle, and K. Glover, Robust and Optimal Control.
Upper Saddle River, NJ: Prentice-Hall, 1996.
[2] G. Obinata and B. Anderson, Model Reduction for Control System De-
sign. London, U.K.: Springer-Verlag, 2001.
[3] R. W. Freund, “Model reduction methods based on krylov subspaces,”
Acta Numerica, vol. 12, p. 2003, 2003.
[4] K. C. Sou, A. Megretski, and L. Daniel, “A quasi-convex optimization
approach to parameterized model order reduction,” in Proc. IEEE/ACM
Design Autom. Conf., Jun. 2005, pp. 933–938.
[5] A. Megretski, “H-infinity model reduction with guaranteed subopti-
mality bound,” in Proc. Amer. Control Conf., Jun. 2006, pp. 448–453.
[6] T. Kailath, Linear Systems. Englewood Cliffs, NJ: Prentice-Hall,
1980.
[7] A. H. Sayed and T. Kailath, “A survey of spectral factorization
methods,” Numerical Linear Algebra With Applications, vol. 8, no.
6-7, pp. 467–496, 2001.
[8] J. Sturm, “Using SeDuMi 1.02, a MATLAB toolbox for optimiza-
tion over symmetric cones,” Optim. Methods Softw., vol. 11-12, pp.
625–653, 1999.
[9] J. Löfberg, “Yalmip: A toolbox for modeling and optimization in
MATLAB,” in Proc. CACSD Conf., Taipei, Taiwan, 2004, pp. 284–289.
[10] A. Rantzer, “On the Kalman-Yakubovich-Popov lemma,” Syst. Control
Lett, vol. 28, no. 1, pp. 7–10, 1996.
[11] P. Giselsson, “Modeling and Control of a 1.45 m Deformable Mirrors”
Master’s thesis, Dept. Autom. Control, Lund Univ., Lund, Sweden,
2006.
Set-Membership Filtering for Discrete-Time Systems
With Nonlinear Equality Constraints
Fuwen Yang, Senior Member, IEEE, and
Yongmin Li, Senior Member, IEEE
Abstract—In this technical note, the problem of set-membership filtering
is considered for discrete-time systems with nonlinear equality constraint
between their state variables. The nonlinear equality constraint is first lin-
earized and transformed into a state linear equality constraint with two
uncertain quantities related to linearizing truncation error and base point
error. S-procedure method is then applied to merge all inequalities into one
inequality and the solution to the unconstrained set-membership filtering
problem is provided. The set-membership filter with state constraint is fi-
nally derived from projecting the unconstrained set-membership filter onto
the constrained surface by using Finsler’s Lemma. A time-varying linear
matrix inequality optimization based approach is proposed to design the
set-membership filter with nonlinear equality constraint. A recursive algo-
rithm is developed for computing the state estimate ellipsoid that guaran-
tees to contain the true state. An illustrative example is provided to demon-
strate the effectiveness of the proposed set-membership filtering with non-
linear equality constraint.
Index Terms—Nonlinear equality constraint, state constraint, state es-
timate ellipsoid, set-membership filtering, time-varying linear matrix in-
equality (LMI) optimization based approach.
I. INTRODUCTION
Filtering technique has been playing an important role in target
tracking, image processing, signal processing and control engineering
[2]. Most filtering approaches require the system noises including
process noise and measurement noise in a stochastic framework and
then provide a probabilistic state estimation [31], [33]–[35]. The
probabilistic nature of the estimates leads to the use of mean and
variance to describe the state spreads (distributions). These spreads
Manuscript received March 04, 2009; revised May 29, 2009. First published
September 22, 2009; current version published October 07, 2009. This work was
supported in part by the Engineering and Physical Sciences Research Council
(EPSRC) of the U.K. under Grant EP/C007654/1, the National Nature Science
Foundation of China under Grants 60874059 and 60604027. Recommended by
Associate Editor T. Zhou.
F. Yang is with the School of Information Science and Engineering, East
China University of Science and Technology, Shanghai 200237, China and also
with the Department of Information Systems and Computing, Brunel Univer-
sity, Uxbridge, Middlesex UB8 3PH, U.K. (e-mail: fwyang@ecust.edu.cn).
Y. Li is with the Department of Information Systems and Computing, Brunel
University, Uxbridge, Middlesex UB8 3PH, U.K..
Color versions of one or more of the figures in this technical note are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TAC.2009.2029403
0018-9286/$26.00 © 2009 IEEE
Authorized licensed use limited to: Fuzhou University. Downloaded on December 20, 2009 at 09:38 from IEEE Xplore. Restrictions apply.