JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN

ELECTRONICS AND COMMUNICATION ENGINEERING

IMPLEMENTATION OF SLIDING MODE ALGORITHM FOR CONTROLLING
SATELLITES POSITION IN GEOSTATIONARY ORBIT
SHETH DHRUMIL H.
M.E. (Elect!"#c$ % C!&&'"#c(t#!") St'*e"t, De+t. O, Elect!"#c$
%C!&&'"#c(t#!"-C.U.S.(. C!lle/e O, E"/#"ee#"/ A"* Tec."!l!/0-W(*.1("
C#t0-S'e"*("(/(-G'2((t-I"*#(.
[email protected]
ABSTRACT In this paper, in control theory, sliding mode control, or SMC, is a form of variable structure
control (VSC). It is a nonlinear control method that alters the dynamics of a nonlinear system by application
of a high-freuency s!itching control. "he state-feedbac# control la! is not a continuous function of time. It
s!itches from one continuous structure to another based on the current position in the state space. $ence,
sliding mode control is a variable structure control method. "he multiple control structures are designed so
that tra%ectories al!ays move to!ard a s!itching condition, and so the ultimate tra%ectory !ill not e&ist
entirely !ithin one control structure. Instead, the ultimate tra%ectory !ill slide along the boundaries of the
control structures. "he motion of the system as it slides along these boundaries is called a sliding mode and
the geometrical locus consisting of the boundaries is called the sliding (hyper) surface. 'sing this la! !e can
control the Satellite(s position in )eostationary *rbit.
e! "ord#$ Slidi%& Mode Co%'rol( )a#' Trac*i%&
I.INTRODUCTION
In recent years, the sliding mode control
methodology has been widely used for robust control
of nonlinear systems (Slotine and Li, 1991). Sliding
mode control, based on the theory of variable
structure systems, has attracted a lot of research on
control systems for the last two decades. A
comrehensive survey on variable structure control
was given in !ung et al. (199"). #he salient
advantage of sliding mode control is robustness
against structured and unstructured uncertainties. In
ath trac$ing systems, however, the system
invariance roerties are observed only during the
sliding hase. In the reaching hase, trac$ing may
be hindered by disturbances or arameter variations.
#he straightforward way to reduce trac$ing error and
reaching time is to increase the control discontinuity
gain.
#ra%ectories from this reduced&order sliding mode
have desirable roerties (e.g., the system naturally
slides along it until it comes to rest at a desired
e'uilibrium). #he main strength of sliding mode
control is its robustness. (ecause the control can be
as simle as a switching between two states (e.g.,
)on)*)off) or )forward)*)reverse)), it need not be
recise and will not be sensitive to arameter
variations that enter into the control channel.
Additionally, because the control law is not a
continuous function, the sliding mode can be
reached in finite time (i.e., better than asymtotic
behavior).

+nder certain common conditions, otimality
re'uires the use of bang,bang control- hence, sliding
mode control describes the otimal controller for a
broad set of dynamic systems
.Sliding mode control must be alied with more
care than other forms of nonlinear control that have
more moderate control action. In articular, because
actuators have delays and other imerfections, the
hard sliding&mode&control action can lead to chatter,
energy loss, lant damage, and e.citation of
unmodeled dynamics. /ontinuous control design
methods are not as suscetible to these roblems and
can be made to mimic sliding&mode controllers.
II. SLIDING MODE CONTROL
+. T.e!0 !, Sl#*#"/ M!*e c!"t!l
/onsider a nonlinear dynamical system described by
0

ISSN3 4567 8 9665: NO; <4 TO OCT << 1 ;OLUME 8 4<, ISSUE = 4> P(/e 6?
JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN
ELECTRONICS AND COMMUNICATION ENGINEERING
d,
dt
2 f(3,t)4((3,t)u(t)
5here,
3(t)2
1
6
"
1
( )
( )
(")
.....
( )
( )
n
n
& t
& t
&
& t
& t

and +(t)2
1
6
"
1
( )
( )
(")
.....
( )
( )
m
m
u t
u t
u
u t
u t

A common tas$ is to design a state&feedbac$ control
law u(3(t))(i.e., a maing from current state 3(t) at
time t to the inut u ) to stabili7e the dynamical
system in 8'uation around the origin 32
[ ]
9, 9, 9, 9...9
"
. #hat is, under the control law,
whenever the system is started away from the origin,
it will return to it. :or e.amle, the comonent .1 of
the state vector may reresent the difference some
outut is away from a $nown signal (e.g., a desirable
sinusoidal signal)- if the control can ensure that .1
'uic$ly returns to .1 2 9, then the outut will trac$
the desired sinusoid. In sliding&mode control#his
reduced&order subsace is referred to as a sliding
(hyer) surface. #he sliding&mode control scheme
involves
(1) Selection of a hyer surface or a manifold (i.e.,
the sliding surface) such that the system tra%ectory
e.hibits desirable behavior when confined to this
manifold.
(6):inding feedbac$ gains so that the system
tra%ectory intersects and stays on the manifold.
#he sliding&mode designer ic$s a switching
function that reresents a $ind of )distance) that
the states 3 are away from a sliding surface.
A state 3 that is outside of this sliding
surface has (.) ;9.
A state that is on this sliding surface has

(3)29
#he sliding&mode&control law switches from one
state to another based on the sign of this distance. So
the sliding&mode control acts li$e a stiff ressure
always ushing in the direction of the sliding mode
where (3)29.<esirable 3(t) tra%ectories will
aroach the sliding surface, and because the control
law is not continuous (i.e., it switches from one state
to another as tra%ectories move across this surface),
the surface is reached in finite time. =nce a
tra%ectory reaches the surface, it will slide along it
and may, for e.amle, move toward the 329 origin.
So the switching function is li$e a toograhic ma
with a contour of constant height along which
tra%ectories are forced to move.
#o force the system states to satisfy (3)29 , one
must>
(1)8nsure that the system is caable of reaching
(3)29 from any initial condition
(6)!aving reached (3)29 , the control action is
caable of maintaining the system at (3)29.
/ondition for e.istence of sliding mode
/onsider a lyaunov function candidate
?( (3))2
1
6
@
"

(3)@ (3)2
1
6
@
6
6
( ) &
5here . is a 8uclidean norm (i.e.,
6
( ) & is the
distance away from the main fold where (3) 29
Sufficient condition for the e.istence of sliding mode
that
dV
dt
A9 5here,
dV
dt
2 @
dV d
d dt

Boughly sea$ing (i.e., for the scalar control case

when m 2 1), to achieve
"

d
dt

A9, the feedbac$

control law u(3) is ic$ed so that C and
d
dt

have
oosite signs. #hat is,
(1) u(3) ma$es
d
dt

negative when (3) is

ositive.
(6) u(3)ma$es
d
dt

ositive when is (3) is

negative. u(3) direct imact on
d
dt

.
B. T.e Sl#*#"/ M!*e C!"t!l (l/!#t.&
In sliding mode control, the systemDs reresentative
oint is constrained to move along a surface (hyer
lane or line) located in the state sace.
#he first roerty is the fact that alication of SE/
does not re'uire an accurate model of the lant.
Secondly, SE/ is robust in the sense that it is
insensitive to arameter variations and bounded
disturbances. #hirdly, SE/ is characterised by
accurate and fast resonses. Lastly, the algorithm is
simle. A tyical hase&lane resonse of a second&
order system is shown in :ig which illustrates the
following shortcomings of SE/ schemes.
:irst of all, there is a )reaching) hase in which the
systemFs reresentative oint (BG) tra%ectory starting
from a given initial state &(9) away from the sliding
line C 2 H&- 4 &. moves towards the sliding line.
#hus the BG in this hase is sensitive to lant
arameter variations and disturbances.
ISSN3 4567 8 9665: NO; <4 TO OCT << 1 ;OLUME 8 4<, ISSUE = 4> P(/e 67
JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN
ELECTRONICS AND COMMUNICATION ENGINEERING
C. Sl#*#"/ M!*e C!"t!l L(1 E@'(t#!"
/onsider a second order lant described in the
controllable canonical form
d,
dt
2A.4(u
y 2 /.
5here,
A 2
(2 and / 2 I1 9J
5here,
. is the state vector,
u is the control signal,
y is the outut signal,
And a1 and a6 are constants.
Let the switching surface is defined as
5here constant is strictly ositive
/onsider the swithing control low
@
d
dt

2(&H
6
4a1 & Ha6 4K9 ) y9
5here,
Satisfies the condition
@
d
dt

9
on the sliding line the system has first order
dynamics
314 H3629
Lote that the sliding mode can be obtained with only
the rior $nowledge of the bound
M

9 and that a1
and a6 could even be time varying. #hus for the lant
in e'uation the sliding mode control law is given by
e'uations .
Lote that the sliding mode can be obtained with only
the rior $nowledge of the bound KN 9 and that a1
and a6 could even be time varying. #hus for the lant
in e'uation the sliding mode control law is given by
e'uations .
III. A MATHEMATICAL MODEL FOR
GEOSTATIONARY ORBIT
A mathematical model for geostationary orbit can be
defined by the Lagrangian function. 5hich is
alied to the sliding mode control algorithm to
stabili7ed the satellite. +r, +

, +

are the
thrusters are the satellite which is shown in fig..
#he Lagrangian function is defined as L2O&G and
the dynamic behavior of the system is secified by
LagrangeDs e'uations>
#he otential energy as>
G2
#m
r

#o derive the e'uations of motion of the system, we

e.ress the $inetic energy as>
O2
"
u
6
6
mv
2
6
m
I(
dr
dt
)
6
4(r
d
dt

)
6
4(r
d
dt

cos
)
6
J
where O is a $nown hysical constant (P Q191P
R.m6 * $g).
d
dt
(
d/
dr
)&
d/
dr
2+r
d
dt
(
d/
d
)&
d/
d
2(r cos )+

d
dt
(
d/
d
)&
d/
d
2r+

<efine state inut and outut vectors, S(t), ?(t) and

y(t), resectively, as>
ISSN3 4567 8 9665: NO; <4 TO OCT << 1 ;OLUME 8 4<, ISSUE = 4> P(/e 69
JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN
ELECTRONICS AND COMMUNICATION ENGINEERING
S(7,v)2
( )
( )
( )
( )
dr
t
dt
r t
t
d
dt
t
d
dt

, ?(t)2
( )
( )
( )
r
u t
u t
u t

,y(t)2
( )
( )
( )
r t
t
t

And
Low consider the e'ution
d,
dt
2A.4(u And T2/3
5here,
A2U
df
d0
And (2U
df
dv
#hus the lineari7ed and normali7ed e'uations of
motion in the geostationary orbit are given by
d,
dt
2
And
5here the states vector V reresents the
erturbations about the nominal orbit and u are the
forces re'uired to correct the satelliteDs osition. As
can be seen, the system is highly non&linear and
ultivariable (i.e. it has many inuts and many
oututs

I?. SIMULATION RESULTS AND CONCLUSION
#he ma%or aim of the study was to design an otimal
sliding mode controller for geostationary
communications satellites that have become vital
tools of modern global communications. #he study
begun with a survey of the technical literature on the
ma%or arts of communications satellites, the forces
that affect satellite orbit and attitude in sace as well
as the control strategies that have been roosed to
maintain correct satellite osition and orientation.
started with a brief descrition of the functions of the
following subsystems of communications satellite>
attitude and orbit control- telemetry, trac$ing and
command- ower suly- and communications
electronics and antennas. It was ointed out that
geostationary satellites allow use of small and fi.ed
earth antennas in global communications networ$s.
#he geostationary orbit is circular, aro.imately
"W,XYZ $m above 8arth, and coincides with the
e'uatorial lane. #he ma%or factors that cause
satellite to change osition and attitude in sace
include> the ellitical shae of 8arth around the
e'uatorial lane causes satellites to e.erience
acceleration towards latitudes XW
o
8 and 19W
o
5-
variations in the gravitational forces of the moon and
sun cause satellites to drift from orbit- solar radiation
ressure on the solar anels cause the satellite orbits
to be more ellitical than circular. #here is
increasing congestion of the geostationary arc as
more and more countries launch satellites for global
and domestic communications. :or these reasons
there is a growing need for effective and efficient
satellite control algorithms. An overview of satellite
attitude and orbit control methods available in the
technical literature was resented . #he essential
features of the robust and fast sliding mode control
method were resented with an overview of
techni'ues that have been devised to overcome its
ma%or shortcoming of signal chattering in the sliding
mode. #he Lagrangian method was utili7ed to
derive the orbital dynamic model of a geostationary
satellite. #he obtained si.th&order statesace model
comrised highly non&linear and couled
differential e'uations. #he system of e'uations was
normali7ed such that the nominal mass of the
satellite is unity and its nominal orbital radius is
unity. #he e'uations were then lineari7ed, using the
#aylor series method, about a nominal orbit. <etails
of the design of sliding mode controllers for such
systems were resented. sliding mode controllers
were also resented based on the theories of linear
'uadratic regulators and sliding mode control.
Simulation results revealed that the sliding mode
control algorithm emloying outut feedbac$ alone
could not handle the roblem of satellite orbit
control.. Secifically, the study has achieved the
following>
[ An u&to&date survey of the technical literature on
satellite attitude and orbit control was comiled.
[ A state feedbac$ decouling control law was
designed for the system. sliding mode controllers
were designed for the system.
ISSN3 4567 8 9665: NO; <4 TO OCT << 1 ;OLUME 8 4<, ISSUE = 4> P(/e 66
JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN
ELECTRONICS AND COMMUNICATION ENGINEERING
[ A structured modular Eatlab rogram was coded
for simulation of a satellite system with the designed
controllers. It emloys a fourthorder Bunge&Outta
numerical integration algorithm with fi.ed ste si7e.
;. SIMULATION RESULTS3
.
;I REFERENCES
I1J <unlo, \. and Smith, <. ]. et al>
#elecommunications 8ngineering, #hird 8dition,
Stanley #hornes, /heltenham, +O, 199P.
I6J Eiya, O. (8d.) et al> Satellite /ommunications
#echnology, 6
nd
8d., O<< 8ngineering and
/onsulting, #o$yo, 19ZW.
I"J. Gratt, #. and (ostian, /.5. et al> Satellite
/ommunications, 5iley ^ Sons, Lew Tor$, 19ZY.
IPJ =mi7egba, 8. 8. et al> _:u77y Attitude /ontrol
of =rbiting Satellite`, Gh< #hesis, Abuba$ar #afawa
(alewa +niversity, (auchi, =ctober 699".
IWJ . ]lobal Gositioning System =verview,
htt>**www.colorado.edu*geograhy*gcraft*notes*gs.
html.
IYa/ommunications Satellites,htt>** ctd.grc.nasa.
gov * rleonard*regslii.html.
ISSN3 4567 8 9665: NO; <4 TO OCT << 1 ;OLUME 8 4<, ISSUE = 4> P(/e 6A