Upload
yong-he
View
212
Download
0
Embed Size (px)
Citation preview
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROLInt. J. Robust Nonlinear Control 2005; 15:923–933Published online 4 November 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rnc.1039
Augmented Lyapunov functional and delay-dependentstability criteria for neutral systems
Yong He1,2, Qing-Guo Wang1,n,y, Chong Lin1 and Min Wu2
1Department of Electrical and Computer Engineering, National University of Singapore,
10 Kent Ridge Crescent, Singapore 119260, Singapore2School of Information Science and Engineering, Central South University, Changsha 410083, China
SUMMARY
In this paper, an augmented Lyapunov functional is proposed to investigate the asymptotic stability ofneutral systems. Two methods with or without decoupling the Lyapunov matrices and system matrices aredeveloped and shown to be equivalent to each other. The resulting delay-dependent stability criteria areless conservative than the existing ones owing to the augmented Lyapunov functional and the introductionof free-weighting matrices. The delay-independent criteria are obtained as an easy corollary. Numericalexamples are given to demonstrate the effectiveness and less conservativeness of the proposed methods.Copyright # 2005 John Wiley & Sons, Ltd.
KEY WORDS: neutral systems; delay-dependent stability; linear matrix inequality (LMI); Lyapunovfunctional
1. INTRODUCTION
Neutral systems are frequently encountered in various engineering systems, includingpopulation ecology, heat exchangers and repetitive control [1–3]. The existing stability criteriafor neutral systems can be classified into two types: delay-dependent ones which includeinformation on the size of delays [4–17], and delay-independent ones which are applicable todelays of arbitrary size [18–20]. Delay-dependent results are usually less conservative than delay-independent ones, especially when the size of the delay is small.
In the past two decades, the fixed model transformation and parameterized modeltransformation are usually employed to obtain the delay-dependent conditions for neutralsystem. Four basic fixed model transformations are addressed in Reference [12]. Among them,the descriptor model transformation method combined with Park’s or Moon et al.’s inequalities[21, 22] is the most efficient [11, 12, 14]. Recently, in order to reduce the conservatism, a free-weighting matrix method is proposed in References [16, 23–25], in which the boundingtechniques on some cross-product terms are not involved. Especially, Wu et al. [17] combined
Received 2 August 2005Revised 13 September 2005
Accepted 26 September 2005Copyright # 2005 John Wiley & Sons, Ltd.
yE-mail: [email protected]
nCorrespondence to: Q.-G. Wang, Department of Electrical and Computer Engineering, National University ofSingapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore.
the parameterized model transformation into the Lyapunov functional and employed the free-weighting matrix method to obtain less conservative delay-dependent stability criteria forneutral systems. We notice that it is hard to further reduce the conservatism by using the sametypes of Lyapunovs as in the above works.
In this paper, we propose an augmented Lyapunov functional, which takes into account thedelay term. Owing to the augmented Lyapunov functional, improved delay-dependent stabilitycriteria for neutral systems are derived by using the free-weighting matrix method. We first developa delay-dependent condition by introducing some free-weighting matrices with the aid of theLeibniz–Newton formula. Then, we derive an alternative delay-dependent criterion by retainingthe derivative of the state and introducing additional free-weighting matrices with the aid of thesystem equation. The relationship between two methods is clearly established. Finally, numericalexamples are given to demonstrate the effectiveness and the merit of the proposed method.
2. PROBLEM FORMULATION
Consider the following neutral system S:(’xðtÞ � C ’xðt� tÞ ¼ AxðtÞ þ Adxðt� tÞ; t > 0
xðtÞ ¼ fðtÞ; t 2 ½�t; 0�ð1Þ
where xðtÞ 2 Rn is the state vector; t > 0 is a constant time delay; and A;Ad and C are constantmatrices with appropriate dimensions. fðtÞ denotes an initial condition which is continuousvector-valued initial function of t 2 ½�t; 0�: Define operator D: Cð½�t; 0�;RnÞ ! Rn as
Dxt ¼ xðtÞ � Cxðt� tÞ
Definition 1 (Hale and Verduyn Lund [2])Operator D is said to be stable if the zero solution of the homogeneous difference equationDxt ¼ 0; t50; x0 ¼ c 2 ff 2 Cð½�t; 0� : Df ¼ 0g is uniformly asymptotically stable.
The stability of operator D is necessary for the stability of neutral system S; which is alwayssatisfied when jjCjj51:
3. THE RESULTS
By introducing an augmented Lyapunov functional, we obtain a delay-dependent stabilitycriterion for system S as follows.
Theorem 1Given a scalar t50; the neutral system S is asymptotically stable if the operator D is stable andthere exist matrices
L ¼
L11 L12 L13
LT12 L22 L23
LT13 LT
23 L33
2664
377550 with L11 > 0
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
Y. HE ET AL.924
Q ¼Q11 Q12
QT12 Q22
" #50
Z ¼Z11 Z12
ZT12 Z22
" #50 and Mi; i ¼ 1; 2; 3
such that the following LMI holds:
G ¼
G11 G12 G13 G14 �tM1 ATS
GT12 G22 G23 G24 �tM2 AT
d S
GT13 GT
23 G33 G34 �tM3 CTS
GT14 GT
24 GT34 �tZ11 �tZ12 0
�tMT1 �tMT
2 �tMT3 �tZT
12 �tZ22 0
SA SAd SC 0 0 �S
26666666666664
3777777777777550 ð2Þ
where
G11 ¼ GAþ ATGT þ L13 þ LT13 þQ11 þ tZ11 þM1 þMT
1
G12 ¼ GBþ ATL12 þ LT23 � L13 þMT
2 �M1
G13 ¼ GC þ L12 þMT3
G14 ¼ tðL33 þ ATL13Þ
G22 ¼ ATd L12 þ LT
12Ad � L23 � LT23 �Q11 �M2 �MT
2
G23 ¼ LT12C þ L22 �Q12 �MT
3
G24 ¼ tð�L33 þ ATd L13Þ
G33 ¼ �Q22
G34 ¼ tðL23 þ CTL13Þ
S ¼ Q22 þ tZ22
G ¼ L11 þQ12 þ tZ12
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
AUGMENTED LYAPUNOV FUNCTIONAL AND DELAY-DEPENDENT STABILITY CRITERIA 925
ProofChoose a Lyapunov functional candidate to be
VðxtÞ :¼ zT1 ðtÞLz1ðtÞ þZ t
t�tzT2 ðsÞQz2ðsÞ dsþ
Z 0
�t
Z t
tþyzT2 ðsÞZz2ðsÞ ds dy ð3Þ
where
L ¼
L11 L12 L13
LT12 L22 L23
LT13 LT
23 L33
2664
377550 with L11 > 0
Q ¼Q11 Q12
QT12 Q22
" #50 and Z ¼
Z11 Z12
ZT12 Z22
" #50
are to be determined and
z1ðtÞ ¼
xðtÞ
xðt� tÞR tt�t xðsÞ ds
2664
3775; z2ðtÞ ¼
xðtÞ
’xðtÞ
" #
From the Leibniz–Newton formula, the following equation is true for any matrices Mi withappropriate dimensions, i ¼ 1; 2; 3:
2½xTðtÞM1 þ xTðt� tÞM2 þ ’xTðt� tÞM3� xðtÞ �Z t
t�t’xðsÞ ds� xðt� tÞ
� �¼ 0 ð4Þ
Calculating the derivative of VðxtÞ along the solution of S yields
’VðxtÞ ¼ 2zT1 ðtÞL’z1ðtÞ þ zT2 ðtÞQz2ðtÞ � zT2 ðt� tÞQz2ðt� tÞ
þ tzT2 ðtÞZz2ðtÞ �Z t
t�tzT2 ðsÞZz2ðsÞ ds
¼ 2zT1 ðtÞL
’xðtÞ
’xðt� tÞ
xðtÞ � xðt� tÞ
2664
3775þ zT2 ðtÞQz2ðtÞ � zT2 ðt� tÞQz2ðt� tÞ
þ tzT2 ðtÞZz2ðtÞ �Z t
t�tzT2 ðsÞZz2ðsÞ ds
þ 2½xTðtÞM1 þ xTðt� tÞM2 þ ’xTðt� tÞM3�
� xðtÞ �Z t
t�t’xðsÞ ds� xðt� tÞ
� �
:¼1
t
Z t
t�tZT1 ðt; sÞ #GZ1ðt; sÞ ds ð5Þ
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
Y. HE ET AL.926
where
Z1ðt; sÞ ¼ ½xTðtÞ xTðt� tÞ ’xTðt� tÞ xTðsÞ ’xTðsÞ�T
#G ¼
G11 þ ATSA G12 þ ATSAd G13 þ ATSC G14 �tM1
GT12 þ AT
d SA G22 þ ATd SAd G23 þ AT
d SC G24 �tM2
GT13 þ CTSA GT
23 þ CTSAd G33 þ CTSC G34 �tM3
GT14 GT
24 GT34 �tZ11 �tZ12
�tMT1 �tMT
2 �tMT3 �tZT
12 �tZ22
26666666664
37777777775
S ¼Q22 þ tZ22
and Gij ; i ¼ 1; 2; 3; i4j43 are defined as in Theorem 1. By Schur complement, the inequalityG50 is equivalent to #G50; which gives ’VðxtÞ5� ejjxðtÞjj2 for a sufficiently small e > 0: Thisproves the asymptotic stability of system S: &
It is seen that Theorem 1 is established by adopting the Leibniz–Newton formula whichprovides free-weighting matrices Mi; i ¼ 1; 2; 3 as in (4). In the following, we derive analternative delay-dependent criterion by taking into account the system equation S; whichprovides another set of free-weighting matrices.
Theorem 2Given a scalar t50; the neutral system S is asymptotically stable if operator D is stable andthere exist matrices
L ¼
L11 L12 L13
LT12 L22 L23
LT13 LT
23 L33
2664
377550 with L11 > 0
Q ¼Q11 Q12
QT12 Q22
" #50
Z ¼Z11 Z12
ZT12 Z22
" #50; U; Mi and Tj ; i ¼ 1; 2; 3; j ¼ 1; . . . ; 6
such that the following LMI holds:
F ¼
F11 F12 F13 F14 F15 F16
FT12 F22 F23 F24 F25 F26
FT13 FT
23 F33 F34 F35 F36
FT14 FT
24 FT34 F44 F45 F46
FT15 FT
25 FT35 FT
45 �tZ11 �tZ12
FT16 FT
26 FT36 FT
46 �tZT12 �tZ22
266666666664
37777777777550 ð6Þ
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
AUGMENTED LYAPUNOV FUNCTIONAL AND DELAY-DEPENDENT STABILITY CRITERIA 927
whereF11 ¼ L13 þ LT
13 þQ11 þ tZ11 þM1 þMT1 � T1A� ATTT
1
F12 ¼ L11 þQ12 þ tZ12 þUT þ T1 � ATTT2
F13 ¼ LT23 � L13 þMT
2 �M1 � T1Ad � ATTT3
F14 ¼ L12 þMT3 � T1C � ATTT
4
F15 ¼ tL33 � ATTT5
F16 ¼ �tM1 � ATTT6
F22 ¼ Q22 þ tZ22 þ T2 þ TT2
F23 ¼ L12 �U þ TT3 � T2Ad
F24 ¼ �T2C þ TT4
F25 ¼ tL13 þ TT5
F26 ¼ �tU þ TT6
F33 ¼ �L23 � LT23 �Q11 �M2 �MT
2 � T3Ad � ATd T
T3
F34 ¼ L22 �Q12 �MT3 � T3C � AT
d TT4
F35 ¼ �tL33 � ATd T
T5
F36 ¼ �tM2 � ATd T
T6
F44 ¼ �Q22 � T4C � CTTT4
F45 ¼ tL23 � CTTT5
F46 ¼ �tM3 � CTTT6
ProofChoose the same augmented Lyapunov functional as in (3). It is clear that the followingequation holds by system equation S:
2
Z t
t�tZT2 ðt; sÞT ½ ’xðtÞ � C ’xðt� tÞ � AxðtÞ � Adxðt� tÞ� ds ¼ 0 ð7Þ
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
Y. HE ET AL.928
where
T ¼ ½TT1 TT
2 TT3 TT
4 TT5 TT
6 �T
Z2ðt; sÞ ¼ ½xTðtÞ ’xTðtÞ xTðt� tÞ ’xTðt� tÞ xTðsÞ ’xTðsÞ�T
On the other hand, ’xðtÞ in ’VðxtÞ is retained (contrasting with the proof of Theorem 1, in which
’xðtÞ is replaced with the system equation S) and (4) is slightly modified to
2½xTðtÞM1 þ ’xTðtÞU þ xTðt� tÞM2 þ ’xTðt� tÞM3�
� xðtÞ �Z t
t�t’xðsÞ ds� xðt� tÞ
� �¼ 0 ð8Þ
Similar to the proof of Theorem 1 with (7) and (8) being added into the expression of ’VðxtÞ; theresult follows immediately. &
Remark 1Theorems 1 and 2 are based on a newly proposed augmented Lyapunov functional of form (3),which contains a structure more general than the traditional ones as those in References[8, 16, 17]. For instance, the first term of (3) involves the delay term xðt� tÞ; and thus is moregeneral than the simply case of operator D: This new type of Lyapunov functional enables us toestablish less conservative results. For theoretical comparison, let us take Reference [17] forexample. It can be verified that the result in Reference [17] is recovered by setting L11 ¼ P11;L12 ¼ �P11C; L13 ¼ P12; L22 ¼ CTP11C; L23 ¼ �CTP12; L33 ¼ P22; Q12 ¼ Z12 ¼ 0; Z11 ¼W ;Z22 ¼ Z; Q11 ¼ Q; Q22 ¼ R; M1 ¼ N1; U ¼ N2; M2 ¼ N3; M3 ¼ N4; T5 ¼ 0 and T6 ¼ 0:
In fact, Theorems 1 and 2 are equivalent to each other, which is shown as follows. Setting
J ¼
I AT 0 0 0 0
0 I 0 0 0 0
0 ATd I 0 0 0
0 CT 0 I 0 0
0 0 0 0 I 0
0 0 0 0 0 I
2666666664
3777777775
ð9Þ
then it follows that
P ¼ JFJT ¼
P11 P12 P13 P14 G14 P16
PT12 F22 P23 P24 F25 F26
PT13 PT
23 P33 P34 G24 P36
PT14 PT
24 PT34 P44 G34 P46
GT14 FT
25 GT24 GT
34 �tZ11 �tZ12
PT16 FT
26 PT36 PT
46 �tZT12 �tZ22
266666666664
377777777775
ð10Þ
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
AUGMENTED LYAPUNOV FUNCTIONAL AND DELAY-DEPENDENT STABILITY CRITERIA 929
where
P11 ¼ G11 þUTAþ ATU þ ATSA
P12 ¼ Gþ T1 þUT þ ATS þ ATT2
P13 ¼ G12 � ATU þUTAd þ ATSAd
P14 ¼ G13 þUTC þ ATSC
P16 ¼ �tðM1 þ ATUÞ
P23 ¼ L12 �U þ TT2 Ad þ SAd þ TT
3
P24 ¼ SC þ TT2 C þ TT
4
P33 ¼ G22 �UTAd � ATdU þ AT
d SAd
P34 ¼ G23 �UTC þ ATd SC
P36 ¼ �tðM2 þ ATdUÞ
P44 ¼ G33 þ CTSC
P46 ¼ �tðM3 þ CTUÞ
and S;G and Gij ; i ¼ 1; 2; 3; i4j43 are defined as in Theorem 1. From (10), it remains to showthat P50 is equivalent to #G50: To this end, if P50 holds, then #G50 holds since P reduces to#G by removing the second row and the second column in P and replacing variables M1;M2 andM3 in #G by M1 þ ATU; M2 þ AT
dU and M3 þ CTU; respectively. On the other hand, if #G50;P50 is feasible by setting T1 ¼ �G ¼ �ðL11 þQ12 þ tZ12Þ; T2 ¼ �S ¼ �ðQ22 þ tZ22Þ;T3 ¼ �L12; T4 ¼ 0; T5 ¼ �tL13; T6 ¼ 0 and U ¼ 0: Therefore, this shows the equivalenceof Theorems 1 and 2.
Remark 2Theorem 2 is suitable for dealing with uncertain systems with polytopic-type uncertainties byusing parameter-dependent Lyapunov functionals as in Reference [23]. This is because the LMIcondition of Theorem 2 does not involve the product of system matrices and Lyapunovmatrices. As for norm-bounded uncertain case, Theorems 1 and 2 are applicable for the robuststability analysis following a similar line to References [17, 24].
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
Y. HE ET AL.930
Before concluding this section, we would like to point out that our method can be used toestablish delay-independent criteria. To see this, setting Z ¼ eI ; where e > 0 is a sufficiently smallpositive scalar, L13 ¼ L23 ¼ L33 ¼ 0; T5 ¼ T6 ¼ 0 and U ¼M1 ¼M2 ¼M3 ¼ 0 in conditionsof Theorem 2, the following delay-independent result is straightforward.
Corollary 1For any scalar t50; the neutral system S is asymptotically stable if operator D is stable andthere exist matrices
L ¼L11 L12
LT12 L22
" #50 with L11 > 0; Q ¼
Q11 Q12
QT12 Q22
" #50 and Tj ; j ¼ 1; . . . ; 4
such that the following LMI holds:
C11 C12 C13 C14
CT12 C22 C23 C24
CT13 CT
23 C33 C34
CT14 CT
24 CT34 C44
26666664
3777777550 ð11Þ
where
C11 ¼ Q11 � T1A� ATTT1
C12 ¼ L11 þQ12 þ T1 � ATTT2
C13 ¼ �T1Ad � ATTT3
C14 ¼ L12 � T1C � ATTT4
C22 ¼ Q22 þ T2 þ TT2
C23 ¼ L12 � T2Ad
C24 ¼ �T2C þ TT4
C33 ¼ �Q11 � T3Ad � ATd T
T3
C34 ¼ L22 �Q12 � T3C � ATd T
T4
C44 ¼ �Q22 � T4C � CTTT4
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
AUGMENTED LYAPUNOV FUNCTIONAL AND DELAY-DEPENDENT STABILITY CRITERIA 931
With a similar treatment, an alternative delay-independent result, which is equivalent toCorollary 1, can be obtained as a direct corollary of Theorem 1. For brevity, we omit the details.
4. NUMERICAL EXAMPLES
In this section, we use two numerical examples to show the less conservativeness of our results.
Example 1Consider the neutral system S with
A ¼�2 0
0 �0:9
" #; Ad ¼
�1 0
�1 �1
" #; C ¼
c 0
0 c
" #; 04c51
Table I lists the maximum upper bound of t for this system in case of different c’s. It is clear thatthe method in this paper produces better results than those in References [8, 12, 17].
Example 2Consider the neutral system S with
A ¼�1:7073 0:6856
0:2279 �0:6368
" #; Ad ¼
�2:5026 �1:0540
�0:1856 �1:5715
" #; C ¼
0:0558 0:0360
0:2747 �0:1084
" #
The upper bound of delay obtained in References [16, 17, 25] are 0.5735, 0.6054 and 0.5937,respectively. By Theorems 1 and 2, the computed value is 0.6189, which is larger than those inReferences [16, 25]. In addition, setting C ¼ 0; this system reduces to a linear nominal systemwith time-invariant delay. The upper bound of delay obtained in References [11, 12, 14, 23, 24]are all 0.6903, and 0.7163 in Reference [17], while the computed value is 0.7918 by Theorems 1and 2, which is larger than those in References [11, 12, 14, 17, 23, 24].
5. CONCLUSIONS
In this paper, an augmented Lyapunov functional is constructed to study the delay-dependentstability problems for neutral systems. The resulting criteria are shown less conservative thansome existing ones. Numerical examples given have convinced the less conservativeness.
Table I. Maximum upper bound of t:
c 0 0.1 0.3 0.5 0.7 0.9
Fridman et al. [12] 4.47 3.49 2.06 1.14 0.54 0.13Han [8] 4.35 4.33 4.10 3.62 2.73 0.99Wu et al. [17] 4.47 4.35 4.13 3.67 2.87 1.41Theorems 1 and 2 4.47 4.42 4.17 3.69 2.87 1.41
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
Y. HE ET AL.932
REFERENCES
1. Nakano M, Hara S. In Microprocessor-based Repetitive Control, in Microprocessor-Based Control Systems, SinhaNK (ed.). D. Reidel Publishing Company: Dordrecht, 1986.
2. Hale JK, Verduyn Lunel SM. Introduction to Functional Differential Equations, Applied Mathematical Sciences,vol. 99. Springer: New York, 1993.
3. Kuang Y. Delay Differential Equations with Applications in Population Dynamics. Academic Press: Boston, 1993.4. Lien CH, Yu KW, Hsieh JG. Stability conditions for a class of neutral systems with multiple delays. Journal of
Mathematical Analysis and Applications 2000; 245:20–27.5. Chen JD, Lien CH, Fan KK, Chou JH. Criteria for asymptotic stability of a class of neutral systems via a LMI
approach. IEE Proceedings D, Control Theory and Applications 2001; 148:442–447.6. Niculescu SI. On delay-dependent stability under model transformations of some neutral linear systems.
International Journal of Control 2001; 74:609–617.7. Fridman E. New Lyapunov-Krasovskii functionals for stability of linear retarded and neutral type systems. Systems
and Control Letters 2001; 43:309–319.8. Han QL. Robust stability of uncertain delay-differential systems of neutral type. Automatica 2002; 38:719–723.9. Han QL. Stability criteria for a class of linear neutral systems with time-varying discrete and distributed delays. IMA
Journal of Mathematical Control and Information 2003; 20:371–386.10. Ivanescu D, Niculescu SI, Dugard L, Dion JM, Verriest EI. On delay-dependent stability for linear neutral systems.
Automatica 2003; 39:255–261.11. Fridman E, Shaked U. A descriptor system approach toH1 control of linear time-delay systems. IEEE Transactions
on Automatic Control 2002; 47:253–270.12. Fridman E, Shaked U. Delay-dependent stability and H1 control: constant and time-varying delays. International
Journal of Control 2003; 76:48–60.13. Han QL. On robust stability of neutral systems with time-varying discrete delay and norm-bounded uncertainty.
Automatica 2004; 40:1087–1092.14. Han QL. A descriptor system approach to robust stability of uncertain neutral systems with discrete and distributed
delays. Automatica 2004; 40:1791–1796.15. Han QL. A new delay-dependent stability criterion for linear neutral systems with norm-bounded uncertainties in all
system matrices. International Journal of Systems Science 2005; 36:469–475.16. He Y, Wu M, She JH, Liu GP. Delay-dependent robust stability criteria for uncertain neutral systems with mixed
delays. Systems and Control Letters 2004; 51:57–65.17. Wu M, He Y, She JH. New delay-dependent stability criteria and stabilizing method for neutral systems. IEEE
Transactions on Automatic Control 2004; 49:2266–2271.18. Hu GD, Hu GD. Some simple stability criteria of neutral delay-differential systems. Applied Mathematics and
Computation 1996; 80:257–271.19. Hui GD, Hu GD. Simple criteria for stability of neutral systems with multiple delays. Internatinoal Journal of
Systems Science 1997; 28:1325–1328.20. Mahmoud MS. Robust H1 control of linear neutral systems. Automatica 2000; 36:757–764.21. Park P. A delay-dependent stability criterion for systems with uncertain time-invariant delays. IEEE Transactions on
Automatic Control 1999; 44:876–877.22. Moon YS, Park P, Kwon WH, Lee YS. Delay-dependent robust stabilization of uncertain state-delayed systems.
International Journal of Control 2001; 74:1447–1455.23. He Y, Wu M, She JH, Liu GP. Parameter-dependent Lyapunov functional for stability of time-delay systems with
polytopic-type uncertainties. IEEE Transactions on Automatic Control 2004; 49:828–832.24. Wu M, He Y, She JH, Liu GP. Delay-dependent criteria for robust stability of time-varying delay systems.
Automatica 2004; 40:1435–1439.25. Xu S, Lam J, Zou Y. Further results on delay-dependent robust stability conditions of uncertain neutral systems.
International Journal of Robust and Nonlinear Control 2005; 15:233–246.
Copyright # 2005 John Wiley & Sons, Ltd. Int. J. Robust Nonlinear Control 2005; 15:923–933
AUGMENTED LYAPUNOV FUNCTIONAL AND DELAY-DEPENDENT STABILITY CRITERIA 933