Professional Documents
Culture Documents
Dual-Quaternion Based Fault-Tolerant Control For Spacecraft Formation Flyin
Dual-Quaternion Based Fault-Tolerant Control For Spacecraft Formation Flyin
ISA Transactions
journal homepage: www.elsevier.com/locate/isatrans
Research Article
Department of Control Science and Engineering, Harbin Institute of Technology, Harbin 150001, China
School of Automation Science and Electrical Engineering, Beihang University, XueYuan Road No. 37, HaiDian District, Beijing 100191, China
art ic l e i nf o
a b s t r a c t
Article history:
Received 1 December 2014
Received in revised form
28 February 2015
Accepted 14 December 2015
Available online 14 January 2016
Study results of developing control system for spacecraft formation proximity operations between a
target and a chaser are presented. In particular, a coupled model using dual quaternion is employed to
describe the proximity problem of spacecraft formation, and a nonlinear adaptive fault-tolerant feedback
control law is developed to enable the chaser spacecraft to track the position and attitude of the target
even though its actuator occurs fault. Multiple-task capability of the proposed control system is further
demonstrated in the presence of disturbances and parametric uncertainties as well. In addition, the
practical nite-time stability feature of the closed-loop system is guaranteed theoretically under the
designed control law. Numerical simulation of the proposed method is presented to demonstrate the
advantages with respect to interference suppression, fast tracking, fault tolerant and practical nite-time
stability.
& 2015 ISA. Published by Elsevier Ltd. All rights reserved.
Keywords:
Spacecraft formation
Dual quaternion
Fault tolerant control
Adaptive control
Finite time
1. Introduction
Spacecraft proximity operation like rendezvous and docking is
an important part of multi-spacecraft formation ying tasks. The
technology of spacecraft proximity has a pretty good application
prospect, and the advanced space transportation plan (ASTP) of
USA has identied this technology as a primary technology, and
many studies are focused on this. A leader-follower strategy was
applied to spacecraft formation in Ref. [1] even with actuator
saturation constraint. Ref. [2] introduced architecture for the
multi-agent coordination system including the leaderfollower
strategy, agent behavioral and virtual-structure approaches. Jiang
et al. [3] studied the two-point boundary value problem of a
leader-follower spacecraft formation ying in unperturbed elliptical reference orbits, and showed this problem can be solved just
like the classical Lambert's problem. Breger et al. [4] presented a
schedule for online generation of safe, fuel-optimized rendezvous
trajectories. Minimum-fuel, impulsive, time-xed solutions for the
problem of orbital rendezvous are further presented with path
constraints in Ref. [5].
Modeling and control are both core technologies of spacecraft
formation ying tasks. For the traditional formation control methods, the orbit control system and the attitude control
system are often considered separately, but in practice, due to the
n
Corresponding author.
E-mail address: huql_buaa@buaa.edu.cn (Q. Hu).
http://dx.doi.org/10.1016/j.isatra.2015.12.008
0019-0578/& 2015 ISA. Published by Elsevier Ltd. All rights reserved.
strong coupling between these two sub-systems, they are generally required to be regarded as a whole six-degrees-of-freedom
(6-DOF) system, and this has better accuracy than traditional
separated system. Many works on the modeling and control problems of the 6-DOF motion of spacecraft has been paid more
attentions recently [612]. In Ref. [6], a 6-DOF synchronized control of spacecraft formation ying in the presence of input constraint and parameter uncertainties is presented. Zhang et al. [7]
presented an integrated relative position and attitude control
strategy for a pursuer spacecraft ying to a space target in proximity operation missions. Gaulcher [8] took advantage of small
angle approximation and a series of assumptions, established a
linear coupling dynamic model for spacecraft. Sinclair et al. [9]
used Cayley to transform a coupling model in the same mathematical framework with Ref. [8]. Comparing with other coupling
modeling methods mentioned above, dual-number description is
considered as one of the most widely used one, due to its clear
physical meaning and high calculation efciency. In addition, the
emergence and development of dual quaternion makes the dualnumber method is very suitable to describe the coupling motion of
spacecraft. Brodsky et al. [13,14] established a general coupling
dynamic model for the motion of rigid body by using dual number.
Wang et al. [15] proposed a 6-DOF relative-motion model by using
dual quaternion for spacecraft formation ying, and then presented several nite-time controllers. Under this, a nonlinear
adaptive coupling tracking controller based on dual quaternion
was presented in Ref. [16], which also considered the uncertainties
of mass and inertia.
88
where a; a 3 A R are called the real part and the dual part of a^
respectively, and is the dual unit with rules as
0; but a 0
2
Multiplication:
3
a^ b^ ab ab a 3 b
a^ a a 3 a a 3
where A R is a scalar.
Dual vector is a class of dual number, and its real and dual parts
are both vectors, it can be described as
a^ a a 3
6
3
a^b^ ab a b
Norm
^ j j j aj j j j a 3 j j
j j aj
Corresponding product
^ a^ a 3 a 3
10
^ a^ Aa A 3 a 3
A
11
13
where
q1 3 q2 1 2 1 U 2 ; 1 2 2 1 1 2
14
Cross product
3
q^ 1 q^ 2 q1 q2 q1 q23 q13 q2
15
Conjugate
h
i
q^ ^ ; ^
16
89
21
_^ f
^ f
^ f q^ 3
^ f
^ f
^ ff M
^ ff M
^ ll 3 q^ f l M
^ ff l
M
fl
fl
f
^ l 3 q^ f l F^ g F^ d F^ u
q^ f l 3
22
In virtue of Eq. (17), the logarithmic operation of dual quaternion are introduced in Ref. [15], which can be described as
follows
ln q^ ln q q 1 3 q 3
f
f
lower with respect to the leader, and also, here F^ g f g fg is the
18
19
zlb
24
xlb
23
i 1;2;3
ylb
zfb
yI
xfb
yfb
oI
xI
f
and velocity error respectively. Design a control input F^ u , such that
for all physically realizable initial conditions, the errors of states of
the closed-loop system are nite-time bounded, this can be
expressed as follows
t ZT r ;
j j e^ q tj j r and j j e^ tj j r where Tr is a nite
time, and are small and bounded positive constants.
90
t 4 T Vx r
26
1
where 0 o r 1 is a scalar, and T o 1 is a nite time satises
1p
Vx0
, where Vx0 is the initial value of Vx.
T r
1 p
^
Lemma 2. [15]: For a dual quaternion qt
and the corresponding
^ then the following equation can be
^ t, dene ^ 2 lnq,
motor
established.
1d
^4
o ^ ; ^ 4 o ^ ;
2 dt
27
28
i3 1 1 3 j 0i j 3 i3 , and also
^ 0 0 diag01
3
0
3
01 ;
j e^ i j 1
02
3
02 ;
1
03
^ diagc^ i
M
f0
^ f 0
^ ff
^ M
e^ ; e^
f
f
^ diagc^ i e^ i 1 e^ ; e^
^ 3 M
f
^ f
^ ff l q^ f l 3
^ ll 3 q^ f l q^ f l 3
^ ll 3 q^ f l
M
f
^
^ ^
f
^
^ f
^ f l;d F d f M f f
F^ g0 M
f
32
^ e^ ; e^ e 1e
!
eT e 1=2
1
1=2
T
1 0:5e e
e T
cot
e
2
e e
e e ev e ep
with j e^ i j 1 j e i j 1 j e3 i j 1 .
Then the main result of this paper is summarized as:
Theorem 1. Given the closed-loop dynamics in Eqs. (21) and (23),
then the designed control law Eq. (29) and adaptive law Eq. (30) can
achieve the control objective proposed in Section 3.3.
Proof. Consider the follower Lyapunov function candidate
1 ^ i
1
^
V s^ M
o 1^ ^ ^ ; ^ 4
fs
2
2k1
33
1
_
o 1^ ^ ^ ; ^ 4
k1
34
After substituting Eqs. (29) and (30) into Eq. (34), one can
obtain
2
3
^ d ^ sgns^ I^ E
^ ^ d ^ sgns^
0
d
d
6
7
7
V_ 6
4s^
5
d ^ sgns^ I^ E
d ^
^
^
^
^
^
sgn
s
E
s
d
d
1
_
o 1^ ^ ^ ; ^ 4
k1
2
3
^ d ^ sgns^ I^ E
^ ^ I^ E
^ d ^ sgns^
0
d
d
6
7
6
7
6
7
6 ^ d ^ sigs^ ^ d ^ sigs^ ^ E^ s^
7
r 6s
7
d
d
6
7
6
7
4
5
^
d ^
d
d sigs^ ^ d sigs^
0i
3
03
31
1
_
o 1^ ^ ^ ; ^ 4
k1
r
3
X
i j i3 j j si j i3 j i j j si3 j
i1
X
i1
i Ei si 2
1
i3 Ei3 j si3 j 2 o1^ ^ ^ ; ^ ^ 4
k1
35
According to the properties of the sum of square, for any constant number 4 12, one can further obtain,
^ ; ^ ^ 4 k11 o 1^ ^ ^ ^ ; ^ ^ ^ ^ 4
1
o 1^ ^
k1
2 1
o 1^ ^ ^ ^ ; ^ ^ ^ 4
2k1
2k1
o 1^ ^ ^ ; ^ ^ 4
r
36
2 1
o 1^ ^ ^ ^ ; ^ ^ ^ 4
o 1^ ^ ^ ; ^ ^ 4
2k1
2k1
1
r 1 V 21
3
X
2k1 i 1
3
2 1 X
i 1 2i i3 1 3 i3 2
2k1 i 1
i 1 2i i3 1 3 i3 2
2 1
V_ r 1 V 1
1 2i i3 1 3 i3 2
2k1 i 1 i
1
2
37
r V 2
1
3
3
2 1 X
X
i 1 2i i3 1 3 i3 2
1 2i
2k1 i 1
2k1 i 1 i
32
38
i3 1 3 i
n
o
1 min 1 12 ; 2 1 min 3 1 3 12 ,
with min 1 ; 22k
i
i
2k1
1
i 1, 2, 3.
For further analysis, the following two cases are considered:
12
Case (1). If
i 1 2i i3 1 3 i3 2 Z1, i 1, 2, 3, it
follows that
12 2 1
2 1
i 1 2i i3 1 3 i3 2
i 1 2i i3 1 3 i3 2 r 0
2k1
2k1
39
32
12
o 1, i 1, 2, 3, then based
Case (2). If i 1 i i3 1 3 i
on the conception of power function, one can obtain
12 2 1
2 1
0r
i 1 2i i3 1 3 i3 2
i 1 2i
2k1
2k1
2 1
32
40
i3 1 3 i r
8k1
2
41
3
3 P 1 2 3 1 3 3 2 .
where r 68k
i
i
i
i
2k
1
1
i1
44
45
V_ 2 o e^ ; e_^ 4
3
12
2 1 X
i 1 2i i3 1 3 i3 2
2k1 i 1
46
3
3
2 1 X
X
i 1 2i i3 1 3 i3 2
1 2i
2k1 i 1
2k1 i 1 i
32
12
3
12
2 1 X
i 1 2i i3 1 3 i3 2
2k1 i 1
i3 1 3 i
x 4T 1 j j sj j ob; j j s 3 j j o b ; b
42
1
with 1 min mini i3 ; mini3 i , i 1, 2, 3;
Then making the appropriate substitutions from Eq. (37), we
obtain
3
X
91
i3 j e3 i j 1
o e^ ; e^ 4
X
i1
r 2V 22
i j e i j 1
47
2 min min i ; min i3 ; i 1; 2; 3:
According to nite-time control theory, it can be guaranteed
that in nite time T 2 which satises T 1 r T 2 o1, the following
inequality can be established.
8
< j e i j r b=ci 1
48
1
: j epi j r b=c 3
i
(
49
Remark 1. By employing terminal sliding mode and some properties of dual quaternion, the tracking errors can be guaranteed to
converge to a small region of the origin in nite time under the
presence of actuators faults, extern disturbances and parameter
uncertainties, and this small region can be as small as desired by
choosing the appropriate control parameters.
Remark 2. In the controller, the adaptive law is used to estimate
and compensate the inuences caused by the uncertain fault, mass
and inertia as well. In addition, the linear term is introduced in the
adaptive law to limit controller gain and smooth the adaptive
process.
Remark 3. In comparison with other fault tolerant controllers
[1721], the controller proposed in this paper takes into account
the coupling of orbit motion and attitude motion, and regards the
faults of orbital actuator and attitude actuator as a whole problem
and then stabilizes the states of system in nite time. When
comparing with other coupling controller based on dual quaternion [15,16], the controller presented in this paper can guarantee the coupling motion of spacecraft formation to be stabilized
92
100
200
300
400
500
600
700
-0.5
1
100
200
300
400
500
600
700
100
200
300
400
500
600
700
100
200
300
400
500
600
700
0
-1
0.5
0
-0.5
Time (s)
85
Value
Eccentricity
Inclination
Longitude ascending node
Semi-major axis
Argument of perigee
Initial true anomaly
0.02
40
50
7000 km
45
10
3.986004 1014 (m3/s 2)
Table 2
Control parameters of DFF in simulation.
Parameters Value
c^ i
^ i
i
k1
^ i
qf l 0
f l 0
pf l 0
vf l 0
2/3
0:01 0:03
3 3
2
0.5
0:001 0:001
0:6245; 0:5; 0:5196; 0:3
0; 0; 0; 0rad=s
20; 50; 10m
0; 0; 0m=s
ADT
100
200
300
400
500
600
700
100
200
300
400
500
600
700
100
200
300
400
500
600
700
100
200
300
400
500
600
700
0
-20
0
o2
20
87
ADF
-20
o1
DFF
20
-10
-20
20
3o
ADT
ADF
DFF
1
0.9
0.8
0.7
0
-20
Time (s)
Fig. 3. Time response of qf3l .
DFF
p1 (N)
20
ADF
ADT
20
x 10
93
-3
-20
-40
15
0
100
200
300
400
500
600
0.4
s1
s2
0.2
s3
700
-0.2
s o1
-0.4
s o2
-40
-60
10
-20
100
200
300
400
500
600
700
si
si
p2 (N)
-0.8
p3 (N)
20
-1
-1.2
-20
-40
s o3
-0.6
-1.4
0
100
200
300
400
500
600
700
Time (s)
Fig. 4. Time response of
-5
200
400
-1.6
600
200
400
600
Time(Sec)
Time(Sec)
pff l .
0.05
1.5
DFF
ADF
ADT
(rad)
initial point
-0.05
0.5
f
u1
-0.1
f
u2
-0.15
uf (Nm)
0
-0.5
1
desired point
f
u3
-0.2
-0.25
-0.3
0.5
(rad)
0
-0.5
-0.5
0.5
1.5
-0.35
-0.4
(rad)
-0.45
100
200
300
400
500
600
700
Time(Sec)
Fig. 8. Time response of control torque.
10
DFF
ADF
ADT
z axes (m)
0
-10
desired point
initial point
-20
-30
-40
-30
-20
-10
x axes (m)
0
10
-50
-40
-30
-20
-10
y axes (m)
the action of DFF, and the steady-state error is less than 2 U104 m.
In general, by using the controller proposed in this paper, the
attitude and position of the follower all meet the requirements.
For purpose of comparison, Figs. 24 also show the simulation
results using ADF and ADT. It can be seen that under the inuence
of actuator faults, DFF has a much better performance with respect
to rapidity and stability compared with ADF and ADT.
6. Conclusion
An adaptive fault-tolerant control law based on dualquaternion description model is proposed for spacecraft formation proximity operations in this paper. The novelty of the proposed control strategy also lies in its robustness against inevitable
external disturbances and its independence of system parameters/
dynamics, and actuator fault tolerance. Furthermore, the closed-
94
fu1
3.5
fu2
fu3
2.5
1.5
fu (N)
1
0.5
0
-0.5
-1
100
200
300
400
500
600
700
Time (s)
Fig. 9. Time response of control force.
Acknowledgments
This work was supported partially by National Natural Science
Foundation of China (Project no. 61522301, 61273175, 61174200),
Program for New Century Excellent Talents in University (NCET11-0801), Heilongjiang Province Science Foundation for Youths
(QC2012C024), and Research Fund for Doctoral Program of Higher
Education of China (20132302110028). The authors greatly
appreciate the above nancial support. The authors would also like
to thank the reviewers for their valuable comments and constructive suggestions that helped to improve the paper
signicantly.
References
[1] Wang PKC, Hadaegh FY, Lau K. Synchronized formation rotation and attitude
control of multiple free-ying spacecraft. J Guid Control Dyn 1999;22(1):28
35.
[2] Beard RW, Lawton J, Hadaegh FY. A coordination architecture for spacecraft
formation control. IEEE Trans Control Syst Technol 2001;9(6):77790.
[3] Fanghua J, Junfeng L, Hexi B, Yunfeng G. Two-point boundary value problem
solutions to spacecraft formation ying. J Guid Control Dyn 2009;32(6):
18271837.
[4] Breger L, How JP. Safe trajectories for autonomous rendezvous of spacecraft. J
Guid Control Dyn 2008;31(5):147889.
[5] Der-Ren T, Coverstone-Carroll V, Prussing JE. Optimal impulsive time-xed
orbital rendezvous and interception with path constraints. J Guid Control Dyn
1995;18(1):5460.
[6] Lv Y, Hu Q, Ma G, Zhou J. 6 DOF synchronized control spacecr form y input
constraint parameter uncertainties. ISA Trans 2011;50(4):57380.
[7] Feng Z, Guang-Ren D. Integrated relative position and attitude control of
spacecraft in proximity operation missions. Int J Autom Comput 2012;9
(4):34251.
[8] Gaulocher S. Modeling the coupled translational and rotational relative
dynamics for formation ying control. In: Proceedings of AIAA guidance,
navigation, and control conference; August 1518, 2005; Month of Conference.
p. 25949.
[9] Sinclair AJ, Hurtado JE, Junkins JL. Application of the Cayley form to general
spacecraft motion. J Guid Control Dyn 2006;29(2):36873.
[10] Pan H, Kapila V. Adaptive nonlinear control for spacecraft formation ying
with coupled translational and attitude dynamics. In: Proceedings of the 40th
IEEE conference on decision and control (CDC). December 47, 2001; Month of
Conference. p. 205762.
[11] Subbarao K, Welsh S. Nonlinear control of motion synchronization for satellite
proximity operations. J Guid Control Dyn 2008;31(5):128494.
[12] Kristiansen R, Nicklasson PJ, Gravdahl JT. Spacecraft coordination control in
6 DOF: integrator backstepping Vs passivity-based control. Automatica
2008;44(11):2896901.
[13] Brodsky V, Shoham M. Dual numbers representation of rigid body dynamics.
Mech Mach Theory 1999;34(5):693718.
[14] Brodsky V, Shoham M. The dual inertia operator and its application to robot
dynamics. J Mech Des 1994;116(4):108995.
[15] Wang J, Liang H, Sun Z, Zhang S, Liu M. Finite-time control for spacecraft
formation with dual-number-based description. J Guid Control Dyn 2012;35
(3):95062.
[16] Filipe N, Tsiotras P. Adaptive position and attitude-tracking controller for
satellite proximity operations using dual quaternions. J Guid Control Dyn 2014.
[17] Bustan D, Pariz N, Hosseini Sani SK. Robust fault-tolerant tracking control
design for spacecraft under control input saturation. ISA Trans 2014;53
(4):107380.
[18] Huo X, Hu Q, Xiao B. Finite-time fault tolerant attitude stabilization control for
rigid spacecraft. ISA Trans 2014;53(2):24150.
[19] Cai W, Liao XH, Song YD. Indirect robust adaptive fault-tolerant control for
attitude tracking of spacecraft. J Guid Control Dyn 2008;31(5):145663.
[20] Jin J, Ko S, Ryoo C-K. Fault tolerant control for satellites with four reaction
wheels. Control Eng Pract 2008;16(10):12508.
[21] Chen W, Saif M. Observer-based fault diagnosis of satellite systems subject to
time-varying thruster faults. J Dyn Syst Meas Control -Trans ASME 2007;129
(3):3526.
[22] Clifford WK. Preliminary sketch of bi-quaternions. Proc Lond Math Soc
1873;4:38195.
[23] Study E. Von den Bewegungen und Umlegungen. Math Ann 1891;39(4):441
556.
[24] Kim MJ, Kim MS, Shin SYA. Compact Differential formula for the rst derivative of a unit quaternion curve. J Vis Comput Animat 1996;7(1):4357.
[25] Filipe N, Tsiotras P. Rigid body motion tracking without linear and angular
velocity feedback using dual quaternions. In: Proceedings of the 12th European control conference, ECC. July 1719, 2013; Month of Conference. p. 329
34.
[26] Man Z, Paplinski AP, Wu HR, Robust Mimo A. Terminal sliding mode control
scheme for rigid robotic manipulators. IEEE Trans Autom Control 1994;39
(12):24649.
[27] Ding SH, Li SH. Stabilization of the attitude of a rigid spacecraft with external
disturbances using nite-time control techniques. Aerosp Sci Technol 2009;13
(45):25665.
[28] Zou AM, Kumar KD, Hou ZG, Liu X. Finite-time attitude tracking control for
spacecraft using terminal sliding mode and Chebyshev neural network. IEEE
Trans Syst Man Cybern Part B-Cybern 2011;41(4):95063.
[29] Du HB, Li SH, Qian CJ. Finite-time attitude tracking control of spacecraft with
application to attitude synchronization. IEEE Trans Autom Control 2011;56
(11):27117.
[30] Zhu Z, Xia YQ, Fu MY. Attitude stabilization of rigid spacecraft with nite-time
convergence. Int J Robust Nonlinear Control 2011;21(6):686702.