Automatica 44 (2008) 552 – 559

www.elsevier.com/locate/automatica

Brief paper
Robust and adaptive variable structure output feedback control of
uncertain systems with input nonlinearity夡
Qinglei Hu a,b,∗ , Guangfu Ma a , Lihua Xie b
a School of Astronautics, Harbin Institute of Technology, Harbin 150001, China
b School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore

Received 8 October 2005; received in revised form 21 December 2006; accepted 16 June 2007
Available online 21 December 2007

Abstract
This brief proposes a robust control algorithm for stabilization of a three-axis stabilized flexible spacecraft in the presence of parametric
uncertainty, external disturbances and control input nonlinearity/dead-zone. The designed controller based on adaptive variable structure output
feedback control satisfies the global reaching condition of sliding mode and ensures that the system state globally converges to the sliding
mode. A modified version of the proposed control law is also designed for adapting the unknown upper bounds of the lumped uncertainties and
perturbations. The stability of the system under the modified control law is established. Numerical simulations show that the precise attitude
pointing and vibration suppression can be accomplished using the derived robust or adaptive controller.
䉷 2007 Elsevier Ltd. All rights reserved.

Keywords: Dead-zone/nonlinearity; Flexible spacecraft; Variable structure control; Vibration suppression

1. Introduction (Bošković, Li, & Mehra, 2001; Chen & Lo, 1993; Hu & Ma,
2005; Iyer & Singh, 1991; Lo & Chen, 1995; Shen, Liu, & Hu,
The fine attitude control of rigid and flexible structures during 2000; Zeng, Araujo, & Singh, 1999). However, in those works,
tracking maneuvers has been an important issue in spacecraft VSC is limited to systems with full state feedback. In practical
applications. The dynamics of large rotational and tracking ma- applications, full measurement of state might be neither possi-
neuvers are time varying and nonlinear and contain bounded ble nor feasible, for example, it is difficult to measure the vari-
parameter uncertainties, and they are affected by various dis- ables describing the flexible motion, the modal position, and
turbances coming from environments. Therefore, the design of the velocity of flexible spacecrafts. Therefore, the direct
effective disturbance rejection control systems that are robust output feedback design in VSC is worth investigating. Sev-
to parametric uncertainty poses a challenging task for space- eral authors have considered VSC with static output feedback
craft designers. (Heck & Ferri, 1989; Yallapragada, Heck, & Finney, 1996;
Variable structure control (VSC) is known to be an ef- Zak & Hui, 1993).
fective control technique applicable to a wide class of non- Another important problem encountered in practice is the
linear systems subject to modeling uncertainty and external control input nonlinearities, such as saturation and dead-zones,
disturbances (Drakunov & Utkin, 1992). It has been applied to originating from actuators in system realization. The phe-
spacecraft attitude control problems in some previous studies nomenon often appears when the reaction wheel is used as
the actuator for a spacecraft (Werta, 1978, Chapter 7). This
kind of input nonlinearity is a source of performance degrada-
夡 This paper was not presented at any IFAC meeting. This paper was
tion or even worse, instability of a system. Consequently, the
recommended for publication in revised form by Associate Editor Xiaohua problem of control design for systems with input nonlinearity
Xia under the direction of Editor M. Araki.
∗ Corresponding author. School of Astronautics, Harbin Institute of Tech- has recently become an area of interests (Hsu, Wang, & Lin,
nology, Harbin 150001, China. 2004; Shyu, Liu, & Hsu, 2003). Hsu et al. (2004) has proposed
E-mail address: [email protected] (Q. Hu). VSC schemes for uncertain dynamic systems with dead-zone
0005-1098/$ - see front matter 䉷 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.automatica.2007.06.024
 Q. Hu et al. / Automatica 44 (2008) 552 – 559 553

nonlinearity for solving regulation problems. However, the where

information of upper bound of uncertainties is required. In ⎧ umax
⎪
⎪ if ui − u0 > umax ,
general, this bound is difficult to measure in practical applica- ⎪ ui − u 0
tions. An adaptive approach may offer a simple and effective ⎨ 
⎪
if ui − u0 umax for ui u0 ,
(ui − u0 ) = 1 (5)
tool to solve this problem (Yoo & Chung, 1992). ⎪
⎪ if u i + u 0  − u max for ui  − u0 ,
⎪
⎪
The contribution of this brief is the design of a nonlinear ⎩ −umax if u + u < − u
i 0 max
controller to achieve the attitude maneuver of a three-axis sta- ui − u 0
bilized flexible spacecraft while actively suppressing the vibra-
and such that 0 < i (ui − u0 ) 1.
tions of the flexible appendages under parametric uncertainty,
external disturbances and control input nonlinearity/dead-zone.
3. Robust adaptive variable structure output feedback
The proposed controller ensures the global reaching condition
control design
of the sliding mode of the uncertain system in the presence of
input dead-zone nonlinearity. In addition, the developed adap-
A linear switching surface is defined as
tive variable structure output feedback controller is also ex-
tended to a modified case without the limitation of knowing S(t) = Gy = GCx(t), (6)
the bounds of the uncertainties and perturbations. The stability
of the modified control system is established. Numerical sim- where S(t) = [s1 (t), s2 (t), . . . , sm (t)]T is a vector of sliding
ulations performed on a three-axis stabilized spacecraft with surfaces and G = [g1 , . . . , gm ]T ∈ R m×p is a specified constant
flexible appendages during rest-to-rest maneuver demonstrate matrix such that matrix GCB is nonsingular.
the effectiveness and feasibility of the method. Let S̃(t) = [s̃1 (t), s̃2 (t), . . . , s̃m (t)]T = B T C T GT S(t), and
then the following VSC law with nonlinearity/dead-zone is
2. Uncertain systems proposed:
⎧ s̃i (t)
A general description of nonlinear systems containing ⎪
⎪
⎪ −kˆ(t) + u0 if s̃i (t) < 0,
nonlinearity/dead-zone in the input function is given in the form ⎨ S̃(t)
ui = 0 if s̃i (t) = 0, (7)
ẋ(t) = (A + A(t))x(t) + B(u) + f (t), ⎪
⎪
⎪
⎩ −kˆ(t)
s̃i (t)
− u0 if s̃i (t) > 0,
y = Cx, (1)
S̃(t)
where x(t) ∈ Rnis the state, u ∈ the control input, y ∈ R p
Rm
where k = max(GCA, GCB),  > 1,  > 0, which will be
the output, and f (t) ∈ R the external disturbance. A(t)
m
defined later, and ˆ (t) is generated as the solution of the linear
denotes time-varying bounded uncertainty and (u) the input
differential equation
nonlinearity.
Throughout this paper, the following assumptions are made. ˙ˆ (t) = kˆ3 (t)B T C T GT S(t) (8)
Assumption 1. The triplet (A, B, C) is controllable and ob- with ˆ (0) = 0 . Note that 0 is the bounded positive initial value
servable. of ˆ (t).
Assumption 2. The uncertainty matrix A(t) and disturbance Lemma 1. For all the input nonlinearities (u) satisfying As-
f satisfy the following matching conditions: sumption 3, the control law (7) satisfies that
A(t) = BH(t), H (t) 1 , (2a)
S̃ T (t)(u)  − kˆ(t)S̃ T (t), (9)
f (t) = Bd(t), d(t) < 2 , (2b)
where 0 <  i (ui − u0 ) 1.
where H and d are with appropriate dimensions.
Proof. From Eqs. (4) and (7), ui > ui0 implies that s̃i (t) < 0
Assumption 3. The nonlinear input function (u) satisfied the and thus
saturation and dead-zone nonlinearity. All three components of
s̃i
the control torque u(t) are constrained by a saturation value (ui − ui0 )(ui ) = − kˆ(t) (ui )
umax and dead-zone value u0 , expressed by S̃
ui (t) − u0 umax for ui (t)u0 , (3a) s̃i2
(ui − ui0 )2 =  [kˆ(t)]2 , (10a)
S̃2
ui (t) + u0  − umax for ui (t) − u0 . (3b)
whereas for ui < − ui0 , s̃i (t) > 0 and thus
Note that a conservative common saturation limit umax and
u0 is adopted here to streamline the analysis. Eq. (3) can be s̃i
expressed as (ui + ui0 )(ui ) = − kˆ(t) (ui )
S̃
(ui ) = (ui − u0 ) · (ui − u0 ) for ui u0 , (4a) s̃i2
(ui + ui0 )2 =  [kˆ(t)]2 . (10b)
(ui ) = (ui + u0 ) · (ui + u0 ) for ui  − u0 , (4b) S̃2
554 Q. Hu et al. / Automatica 44 (2008) 552 – 559

Then 4. Modified adaptive variable structure output feedback

control design
s̃i2
s̃i i (ui ) − kˆ(t) . (11)
S̃ In Section 3, we have presented a robust adaptive variable
Therefore, the following inequality holds: structure output feedback control for uncertain systems with
 saturation nonlinearity and dead-zone where the bounds of the
m m
s̃i2 (t) uncertainties and disturbances are known. In this section, we
S̃ (t)(u) =
T
s̃i (t)(ui )  −kˆ(t) shall address a modified adaptive variable structure output feed-
i=1 i=1
S̃(t)
back when these bounds are unknown.
= − kˆ(t)S̃ T (t). (12) We now propose the following adaptive variable structure
Note that a nonsingular transformation of switching surface will output feedback control:
not change the sliding mode dynamics. If a particular switching ⎧ s̃i (t)
surface S1 = G1 y is chosen, it can be transformed to S = Gy, ⎪
⎪ −k ˆ
ˆ (t) + u0 if s̃i (t) < 0,
⎪
⎨ S̃(t)
where G=(G1 CB)−1 G1 . Therefore, without loss of generality, ui = 0 if s̃i (t) = 0, (20)
assume that GCB = I in the following.  ⎪
⎪
⎪
⎩ −k ˆ
ˆ (t)
s̃i (t)
− u0 if s̃i (t) > 0,
S̃(t)
Theorem 1. Consider the nonlinear system (1) subjected to
Assumptions 1–3. If the input u(t) is given as that indicated by with ˆ (t) given as
Eq. (7) with the adaptive control Eq. (8), then the system trajec-
tories asymptotically converge to the sliding manifold S(y)=0. ˙ˆ (t) = k ˆ
ˆ 3 (t)B T C T GT S(t). (21)

Proof. Consider a Lyapunov function candidate Note that (t)

ˆ is the estimate of , and ˆ can be obtained from
the following dynamics:
V (t) = 21 S T S + 21 ˜ 2 , (13)
˙ˆ = q̄ −1 kB T C T GT S(t),
(t) (0)
ˆ = 0 , (22)
where ˜ (t) = ˆ −1 (t) − .
Then, the time derivative of V (t) has where q̄ > 0 and 0 is positive and bounded initial value of (t).
ˆ
Let (t)
˜ = (t)
ˆ −  denote the adaptation error. Because 
V̇ (t) = S T GC[Ax + B(u) + BHx + Bd] is assumed to be constant, then the following expression keeps
+ ˜ (t)˙˜ (t)kS(x + H  + d) valid:
+ S T GCB(u) + ˜ (t)˙˜ (t). (14)
(t) ˙ˆ
˙˜ = (t). (23)
Now define a constant  which always exists, satisfying
Based on the above parameter adaptive controller, we obtain
∞ > (x + H  + d) (15) the following theorem.
for all ∞ > t 0, and apply Lemma 1 and the assumption
GCB = I , we have Theorem 2. Consider the nonlinear system as described in Eq.
(1). If the modified law is designed as Eq. (20) with Eqs. (21)
V̇ (t)kS − kˆ(t)S T (t) + ˜ (t)˙˜ (t) and (22), and Assumptions 1–3 hold, then the trajectory of the
= kS − (kˆ(t) + ˜ (t)kˆ(t))S(t) nonlinear system converges to the sliding manifold S(y) = 0.
= (1 − )kS 0. (16)
Proof. Consider the Lyapunov function candidate as V̄˙ (t) =
If w̄(t)=(−1)kS is defined, integrating the above equation V (t) + (q̄/2)˜ 2 , q̄ > 0. Using Eq. (20) and the same manipu-
from zero to t yields lations as Eqs. (14)–(16), the time derivative of V̄ (t) becomes

t
t
V1 (0) V1 (t) + w̄() d w̄() d. (17) V̇ (t) k S
ˆ − k ˆˆ (t)S T (t)
+ ˜ (t)˙˜ (t) + ˜ (1
˙˜
0 0
− )k S
ˆ 0. (24)
Taking the limit as t → ∞ on both side of Eq. (17) gives

t For the moment, using the result of Theorem 1, we can con-
∞ > V1 (0) lim w̄() d. (18) clude that S(y) → 0 as t → ∞. Hence the proof is achieved
t→∞ 0 completely. 
Thus according to Barbalat lemma (Popov, 1973), we obtain
Remark. In this section, the modified adaptive version of the
lim w̄(t) = lim ( − 1)kS = 0. (19) proposed adaptive variable structure output feedback control
t→∞ t→∞
law has been also designed for adapting the unknown upper
Since  > 0 for all t > 0,  > 1, k > 0 and GCB = 0, Eq. (19) bounds of the lumped uncertainties and perturbations so that
implies S(y) → 0 as t → ∞. Hence the proof is achieved the limitation of knowing the bound of the uncertainties and
completely.  perturbations in advance is released.
 Q. Hu et al. / Automatica 44 (2008) 552 – 559 555

5. Application to a flexible spacecraft 5.2. Simulation results

5.1. Mathematical model In this section, application of the proposed control schemes
to the attitude control of a flexible spacecraft is presented using
The dynamics of spacecraft with flexible appendages can be MATLAB/SIMULINK software. The nominal moments of in-
obtained from the Lagrangian approach (Gennaro, 1998). With ertia parameters, coupling matrices, and the natural frequencies
the assumption of a small Euler angle rotations, the dynamic are taken from Gennaro (1998).
model can be approximated as Suppose the initial values of the three attitude angles are
⎡ ⎤ ⎡ ⎤ (0) = 6◦ ,
(0) = 4◦ , and (0) = −4◦ . The control objec-
 ¨    ˙ tive is to drive each of the attitude angles to zero from their
(JT − JR )  ⎢
¨ ⎥
T
0 0 ⎢
˙ ⎥ initial value within tf = 150 s, and each of the vibration
⎣¨⎦+ ⎣˙⎦
 I  0 D  mode coordinates satisfies i = ˙ i = i (tf ) = ˙ i (tf ) = 0,
¨ ˙ i = 1, 2, 3, 4. Suppose that the matched uncertainty and distur-
⎡ ⎤
    bance are of the forms: As = Be(1+sin xs ) and fs = B[sin 5t],
0 0 ⎢
⎥ I
+ ⎣ ⎦= ( (·) − us ), (25) respectively.
0 K  0 In the numerical simulation, for comparison, three cases are
 conducted: (1) attitude maneuver control using the proposed
robust adaptive variable structure output feedback controller
where ,
and  are roll, pitch and yaw attitude angles, and
(7); (2) attitude maneuver control using the traditional vari-
(·) is a function of the moments of inertia of the flexible
able structure output feedback controller given in Yallapragada
spacecraft, the coupling matrix between attitude and vibrations
et al. (1996) as follows:
modes, etc.
Introducing a new variable S
u = Ny − (GCB)−1 (29)
S + 
z = [
 ] T
(26)
and (3) attitude maneuver control using the modified adaptive
and setting variable structure output feedback controllers (20). The attitude
response and vibration suppression are analyzed to study the
   
(JT − JR ) T 0 0 performances of the controllers.
M̄ = , D̄ = ,
 I 0 D
5.2.1. Robust adaptive variable structure output feedback
 
0 0 control (RAVSOFC)
K̄ = , B̄ = [I 0]T , Fig. 1 shows the results of the proposed robust adaptive vari-
0 K
able structure output feedback controller in the presence of un-
the following equation can be obtained: certainty and disturbance. From Fig. 1(a), it can be seen that
each of the three-attitude angle responses (solid line) almost
M̄ z̈ + D̄ ż + K̄z = B̄[ (·) − us ]. (27) approaches zero around the time of 100 s, and the steady error
of each angle is less than 0.002◦ . Fig. 1(b) shows the responses
Define the state variable xs = [
  ˙
˙ ˙ ˙ ]T ∈ (solid line) of angular velocity of the flexible spacecraft and
2( n̄+3)
R , the state-space spacecraft model describing the dy- the steady errors are no more than 0.002◦ /s. Responses of the
namics behavior of the flexible spacecraft can be described by vibrations modes are illustrated in Fig. 1(d). The maximum in-
duced vibration coordinate of the first mode is effectively sup-
ẋs (t) = (As + As )xs (t) + Bs s (us ) + fs (t),
pressed. It is noted that all the vibration modal coordinates and
ys = Cs xs , (28) vibration rates approach zero at time 10 s.
where
5.2.2. Traditional variable structure output feedback control
   
0 I 0 (TVSOFC)
As = , Bs = ,
−M̄ −1 K̄ −M̄ −1 D̄ M̄ −1 B̄ For the purpose of comparison, the system employed by TV-
SOFC law in Eq. (32) proposed by Yallapragada et al. (1996)
 
I3×3 03×n̄ 03×3 03×n̄ is also considered in this section. The same simulation case is
f (t) = Bs (·), Cs = . repeated with the TVSOFC replacing the proposed RAVSOFC
03×3 03×n̄ I3×3 03×n̄
for a fair complexion and the results of simulation were also
Suppose that As is the parametric uncertainty matrix of As shown in Fig. 1 (dotted line). For this case, a significant amount
and n̄ is the number of elastic modes considered. Here the of the oscillations of the attitude angle and rate occurred during
nonlinearities in the input s (us ) is taken as saturation and the maneuvering as demonstrated in Figs. 1(a) and (b) (dot-
dead-zone whose saturated value and dead-band value are 10 ted line) with steady errors more than 0.05◦ and 0.05◦ /s, re-
and 0.5 N m, respectively. In addition, the system satisfies all spectively. Moreover, the oscillations do not settle within 300 s.
of the assumptions given earlier. The disturbance rejection is markedly worse than that achieved
556 Q. Hu et al. / Automatica 44 (2008) 552 – 559

Fig. 1. Attitude maneuver control using proposed RAVSOFC and TVSOFC. (a) Responses of attitude angle. (b) Responses of attitude angle rate. (c) Coordinate
responses of first–fourth vibration modes.
 Q. Hu et al. / Automatica 44 (2008) 552 – 559 557

Fig. 2. Attitude maneuver control using modified AVSOFC. (a) Responses of attitude angle. (b) Responses of attitude angle rate. (c) Coordinate responses of
first–fourth vibration modes.

by the proposed approach. In addition, from the dotted line as 5.2.3. Adaptive variable structure output feedback control
shown in Fig. 1(c), TVSOFC also introduces much more vi- For the modified adaptive control case in Eq. (20) with adap-
brations to the first four modes than proposed RAVSOFC and tive laws (21) and (22), the same tests are also repeated with the
the maximum induced vibration coordinate of the first mode is same control parameters and the simulation results are shown
nearly close to 0.5 kg1/2 m. in Fig. 2. In this case, the attitude angles and rates response of
558 Q. Hu et al. / Automatica 44 (2008) 552 – 559

each axis are shown in Figs. 2(a) and (b), respectively. As ex- Iyer, A., & Singh, S. N. (1991). Variable structure slewing control and
pected, the responses of both the angle displacement and the vibration damping of flexible spacecraft. Acta Astronautica, 25(1), 1–9.
angular velocity almost also converge to zero in 100 s. The Lo, S. C., & Chen, Y. P. (1995). Smooth sliding mode control for spacecraft
attitude tracking maneuvers. Journal of Guidance, Control and Dynamics,
maximum induced vibration coordinate of the first mode is ef- 18(6), 1345–1349.
fectively suppressed to below 0.1 kg1/2 m as shown in Fig. 2(c). Popov, V. M. (1973). Hyperstability of control system. Berlin: Spinger.
It is seen that the adaptive control law, as expected, is essen- Shen, Y., Liu, C., & Hu, H. (2000). Output feedback variable structure control
tially capable of suppressing the vibrations while maintaining for uncertain systems with input nonlinearities. Journal of Guidance,
the attitude maneuvering capability of flexible spacecraft even Control and Dynamics, 23(4), 762–764.
Shyu, K. K., Liu, W. J., & Hsu, K. C. (2003). Decentralized variable
the information of upper bound of the perturbations and uncer- structure control of uncertain large-scale systems containing a dead-zone.
tainties is not required in advance. IEE Proceedings of Control Theory and Applications, 150(5), 467–475.
Werta, J. R. (1978). Spacecraft attitude determination and control (pp.
6. Conclusions 270–272). Dordrecht, Holland: D. Reidel Publishing Company.
Yallapragada, S. V., Heck, B. S., & Finney, J. D. (1996). Reaching conditions
for variable structure control with output feedback. Journal of Guidance,
In this paper, the problem of vibration suppression of three- Control and Dynamics, 19(4), 848–853.
axis stabilized flexible spacecraft during attitude maneuver in Yoo, D. S., & Chung, M. J. (1992). A variable structure control with simple
the presence of bounded model uncertainty, external distur- adaptation laws for upper bounds on the norm of the uncertainties. IEEE
bances and control input nonlinearity/dead-zone has been in- Transactions on Automatic Control, AC-37(6), 860–864.
Zak, S. H., & Hui, S. (1993). On variable structure output feedback controllers
vestigated. A robust adaptive variable structure output feedback for uncertain dynamic systems. IEEE Transactions on Automatic Control,
controller to stabilize uncertain dynamics system with explic- AC-38(10), 1509–1512.
itly considering nonlinearity/dead-zone control in control in- Zeng, Y., Araujo, A. D., & Singh, S. N. (1999). Output feedback variable
put has been proposed. It is shown that the designed controller structure adaptive control of a flexible spacecraft. Acta Astronautica, 44(1),
guarantees the global reaching condition of the sliding mode 11–22.
for the uncertain system. In addition, we have also proposed
a modified adaptive version of this algorithm, which removes Qinglei Hu obtained his B.Eng. degree in 2001
from the Department of Electrical and Elec-
the constraint of known parametric uncertainty bound and dis- tronic Engineering at the Zhengzhou University,
turbance bound. Simulations have verified the effectiveness of PR China, and M.Eng. and Ph.D. degree from
the proposed designs. the Department of Control Science and Engi-
neering at Harbin Institute of Technology with
specialization in controls, PR China, in 2003
Acknowledgments and 2006, respectively.
Since 2003, he has been with the Department
of Control Science and Engineering at Harbin
This present work was supported by National Natural Science Institute of Technology and was promoted as
Foundation of China (Project Number: 60674101) and Research an Associate Professor in 2006. He is currently working as a Postdoctoral Re-
search fellow in the School of Electrical and Electronic Engineering, Nanyang
Fund for the Doctoral Program of Higher Education of China Technological University. His research interests include variable structure
(Project Number: 20050213010). The authors would like to control and application, robust control and application, active vibration con-
thank the reviewers and the Editor for many suggestions that trol, spacecraft attitude control, and fuzzy control and application. He was
an Associate Editor for the Journal of the Franklin Institute.
helped improve the paper.
Guangfu Ma is currently a Professor in the
References Department of Control Science and Engineer-
ing, Harbin Institute of Technology. He received
his Ph.D. and M.S. in electrical engineering
Bošković, J. D., Li, S. M., & Mehra, R. K. (2001). Robust adaptive variable from Harbin Institute of Technology in 1993
structure control of spacecraft under control input saturation. Journal of and 1987, respectively. In 1992, he became an
Guidance, Control and Dynamics, 24(1), 14–22. Associate Professor and in 1997, Professor at
Chen, Y., & Lo, S. (1993). Sliding mode controller design for spacecraft Harbin Institute of Technology, where he cur-
attitude tracking maneuvers. IEEE Transaction on Aerospace and rently teaches and performs research in the fields
Electronics, 1328–1333. of spacecraft attitude control and aerospace con-
trol systems.
Drakunov, S. V., & Utkin, V. I. (1992). Sliding mode control in dynamic
systems. International Journal of Control, 55, 1029–1037.
Lihua Xie received the B.E. and M.E. degrees
Gennaro, S. D. (1998). Adaptive robust tracking for flexible spacecraft in in electrical engineering from Nanjing Univer-
presence of disturbance. Journal of Optimization Theory and Applications, sity of Science and Technology in 1983 and
98(3), 545–568. 1986, respectively, and the Ph.D. degree in elec-
Heck, B. S., & Ferri, A. A. (1989). Application output feedback for variable trical engineering from the University of New-
structure systems. Journal of Guidance, Control and Dynamics, 12(6), castle, Australia, in 1992. Since 1992, he has
932–935. been with the School of Electrical and Elec-
Hsu, K. C., Wang, W. Y., & Lin, P. Z. (2004). Sliding mode control tronic Engineering, Nanyang Technological Uni-
for uncertain nonlinear systems with multiple inputs containing sector versity, Singapore, where he is a Professor. He
held teaching appointments in the Department
nonlinearities and dead-zone. IEEE Transactions on Systems, Man, and
of Automatic Control, Nanjing University of
Cybernetics, Part B: Cybernetics, 34(1), 374–380. Science and Technology from 1986 to 1989.
Hu, Q. L., & Ma, G. F. (2005). Variable structure control and active vibration He has also held visiting appointments with the University of Newcastle,
suppression of flexible spacecraft during attitude maneuver. Aerospace the University of Melbourne, Hong Kong Polytechnic University, and South
Science and Technology, 9, 307–317. China University of Technology.
 Q. Hu et al. / Automatica 44 (2008) 552 – 559 559

Dr. Xie’s research interests include robust control and estimation, net- Transactions on Circuits and Systems-II. He is also a member of the Editorial
worked control systems, time delay systems, control of disk drive systems Board of IET Proceedings on Control Theory and Applications. He served
and sensor networks. In these areas, he has published a number of pa- as an Associate Editor of International Journal of Control, Automation and
pers and co-authored two patents and the books H-infinity Control and Systems from 2004 to 2006 and an Associate Editor of the Conference
Filtering of Two-dimensional Systems (with C. Du); Optimal and Robust Editorial Board, IEEE Control Systems Society from 2000 to 2005. He was
Estimation (with F.L. Lewis and D. Popa) and Control and Estimation of the General Chairman of the ninth International Conference on Control,
Systems with Input/Output Delays (with H. Zhang). He is an Associate Automation, Robotics and Vision. Dr. Xie is a Fellow of IEEE and a Fellow
Editor of IEEE Transactions on Automatic Control, Automatica and IEEE of Institution of Engineers, Singapore.