Automatica 46 (2010) 1860–1869
Contents lists available at ScienceDirect
Automatica
journal homepage: www.elsevier.com/locate/automatica
Brief paper
Insensitive H
∞
filter design for continuous-time systems with respect to filter
coefficient variations
✩
Guang-Hong Yang
∗
, Xiang-Gui Guo
College of Information Science and Engineering, Northeastern University, Shenyang, 110004, PR China
Key Laboratory of Integrated Automation of Process Industry (Ministry of Education), Northeastern University, Shenyang, 110004, PR China
a r t i c l e i n f o
Article history:
Received 19 September 2009
Received in revised form
19 April 2010
Accepted 25 June 2010
Available online 27 July 2010
Keywords:
Sensitivity
H
∞
filtering
Filter coefficient variations
Linear matrix inequality (LMI)
a b s t r a c t
This paper is concerned with the problem of designing insensitive H
∞
filters for linear continuous-time
systems. Coefficient sensitivity functions of transfer functions with respect to filter additive/multiplicative
coefficient variations are defined, and the H
∞
norms of the sensitivity functions are used to measure the
sensitivity of the transfer functions with respect to additive filter coefficient variations. In addition, in
order to deal with the filter design problem for the multiplicative filter coefficient variation case, new
measures based on the average of the sensitivity functions are also defined. Consequently, the insensitive
H
∞
filter design problem is reduced to a multi-objective filter design problem, which minimizes the
coefficient’s sensitivity and meets the prescribed H
∞
norm constraint simultaneously. First, a novel
method for designing insensitive H
∞
filters subjected to additive filter coefficient variations is given
in terms of the linear matrix inequality (LMI) optimization techniques. Furthermore, based on the new
sensitivity measures, the obtained results are extended to the multiplicative coefficient variation case.
In addition, an indirect method for solving the multiplicative variations is also proposed. Finally, two
numerical examples are provided to demonstrate the effectiveness of the proposed method.
© 2010 Elsevier Ltd. All rights reserved.
1. Introduction
An important topic in filter design is filter coefficient sensitiv-
ity (Wong & Ng, 2001) because very small perturbations in the
coefficient of the designed filter may result in a serious deterio-
ration of the system’s performance, including instability. Such per-
turbations may be due to, among other things, roundoff errors in
numerical computation during the filter or controller implemen-
tation and the need to provide practicing engineers with safe-
tuning margins (Yang & Che, 2008). Therefore, the filter should be
designed to be insensitive to some amount of error with respect
to its coefficients. On the other hand, sensitivity analysis allows
the analyst to assess the effects of changes in the parameter val-
ues (Castillo, Gutiérrez, & Hadi, 1997). Hence, it is very useful to
✩
This work was supported in part by the Funds for Creative Research Groups of
China (No. 60821063), National 973 Program of China (Grant No. 2009CB320604),
the Funds of National Science of China (Grant No. 60974043), and the 111
Project (B08015), the Fundamental Research Funds for the Central Universities
(No. N090604001, N090604002). The material in this paper was not presented at
any conference. This paper was recommended for publication in revised form by
Associate Editor Tongwen Chen under the direction of Editor Ian R. Petersen.
∗
Corresponding author at: College of Information Science and Engineering,
Northeastern University, Shenyang, 110004, PR China. Tel.: +86 13504182968;
fax: +86 24 83681939.
E-mail addresses: yangguanghong@ise.neu.edu.cn, yang_guanghong@163.com
(G.-H. Yang), guoxianggui@163.com, guoxianggui@foxmail.com (X.-G. Guo).
understand how changes in the parameter values influence the de-
sign (Castillo, Mínguez, & Castillo, 2008). After the hard work of
many researchers in more than one decade, fundamental results
have been obtained for the study of sensitivity analysis and perfor-
mance limitations in automatic control systems (see, e.g., Chevrel
& Yagoubi, 2004; Li, 1997; Whidborne, Istepanian, & Wu, 2001;
Wu, Chen, Li, Istepanian, & Chu, 2001; Yamaki, Abe, & Kawamata,
2008; Yan & Moore, 1992 and the references therein), and many
different definitions of sensitivity have been used for sensitivity
analysis. One of the effective synthesis methods is the coefficient
sensitivity method, which describes the variations in performance
due to variations in the parameters that affect the system’s dynam-
ics, given in Chevrel and Yagoubi (2004), Hilaire, Chevrel, and Trin-
quet (2005), Li (1998, 2004), Lutz and Hakimi (1988), Tavsanoğlu
and Thiele (1984), Whidborne et al. (2001) and Wong and Ng
(2001). Chevrel and Yagoubi (2004) proposed a design approach
to reduce the sensitivity of the performance with regard to the
system’s parametric variations. The problem of designing opti-
mal realizations of controllers via a reduction of the coefficient’s
sensitivity within the implicit state-space framework was inves-
tigated in Hilaire et al. (2005), Li (1998), Whidborne et al. (2001)
and Wu et al. (2001), while the optimal realization of state-space
digital filters which minimized the coefficient’s sensitivity was in-
vestigated in Li (1997), Lutz and Hakimi (1988), Tavsanoğlu and
Thiele (1984) and Wong and Ng (2001). Noting that in the above
mentioned works on the sensitivity analysis, though a considerable
amount of literature is available on the optimal realization of state-
space digital filters, to the best of the authors’ knowledge, there
0005-1098/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.automatica.2010.07.004