A Fast Approach for Relative Orbital Determination in

Spacecraft Formations
Silvano Sgubini Giovanni B. Palmerini
Sapienza Università di Roma Sapienza Università di Roma
DIAEE Dipartimento di Ingegneria DIAEE Dipartimento di Ingegneria
Astronautica, Elettrica ed Energetica Astronautica, Elettrica ed Energetica
via Salaria 851/881 via Salaria 851/881
Roma, Italy Roma, Italy
[email protected] [email protected]

Abstract—The design and operation of spacecraft formation- 1. INTRODUCTION

flying missions require a suitable knowledge of the relative
dynamics of the platforms belonging to the formations. The The increase in autonomous orbital control capabilities of
perturbed, non-Keplerian, orbital environment can be easily spacecraft introduces the problem of suitable orbital
faced by numerical approaches, like Cowell integration. propagation to be carried out on board by means of
However, that solution is accurate but quite slow to obtain, and necessarily limited computation resources. This issue is
indeed more appealing in the design phase. Instead, in a even more relevant when dealing with formation flying, as
number of applications, above all from the operation point of the individual trajectory of every satellite, each of them
view, a faster propagation would have a strong interest. facing different environmental conditions, independently
This paper proposes a special writing of the equations of
concurs to the acquisition and keeping of the formation
motion which does provide a closed form solution for orbits
including the oblateness, i.e. a more realistic representation configuration, and should be known to assess its overall
than the Keplerian one for higher LEO and medium altitude behavior. Different solutions can be proposed, always
formations environment. The idea, which originated from providing an indication of future behavior without the need
previous literature, is to express the variables of interest of a continuous link with the ground segment and the
(radius, node, inclination, etc) as a series, which can be limited control center.
to the desired accuracy level in terms of eccentricity. The
adoption of symbolic mathematics currently allows for exact Interestingly, similar problems of fast orbital prediction are
computation of the parameters of interest at every time. More also found while dealing with other aerospace applications
important, the paper offers this formulation in terms of of current interest. As an example, faster time-to-first-fix
relative dynamics, i.e. in terms of differences in the orbital
techniques for global navigations satellite systems receivers
parameters of the platforms. Such a writing, which is the
novelty of the paper, is actually what is required in the could be based on a reduced, but carefully chosen set of
spacecraft formation case. ephemerides so that a resource-limited receiver should be
The strong advantage of the proposed formulation is the quick able to roughly but quickly locate the source-satellite and
availability of the solution, orders of magnitude faster than therefore provide a rough estimate of its own position. Fast
numerical integrators of similar accuracy. In such a way the orbital propagation is also a mandatory issue dealing with
future behavior of the platforms belonging to the formation debris-alert systems, where accurate knowledge of the
can be easily computed, and these data are available for on-line debris position in a close future is important in order to
control strategy evaluation as well as for payload operation evaluate the collision risks and possibly implement evasive
schedule. The paper details this special writing of the equation
manoeuvres.
of motion, provides the solutions for eccentricities up to 0.01
while validating the solutions with respect to standard,
accurate numerical propagators.
Accurate, onboard-based evaluation of a formation
evolution seems to be however the most compelling case for
TABLE OF CONTENTS fast orbital determination. Formations, including a number
of platforms, should require extensive (and expensive)
1. INTRODUCTION .................................................1 ground support with antenna network tracked throughout the
2. THE EQUATIONS OF MOTION FOR A SINGLE mission. However, their continuous targeting of the desired
configuration is usually mandatory to correctly exploit
PLATFORM .......................................................2 payload tasks. To this aim, a control should be permanently
3. NUMERICAL VALIDATION ................................4 implemented, with its performance strictly depending on the
4. THE EQUATIONS FOR A SPACECRAFT capabilities to forecast the system evolution taking into
FORMATION .....................................................5 account all orbital perturbations involved. In fact, the
5. FINAL REMARKS ...............................................5 simpler model, including only the differences in the
gravitational attraction acting on the platforms (Euler-Hill or
REFERENCES.........................................................6 Clohessy-Wiltshire equations, [1]) is certainly not enough
APPENDIX..............................................................6 precise for current applications. The typical approach is
BIOGRAPHIES........................................................6 therefore to introduce simplified dynamical models for the
978-1-4577-0557-1/12/$26.00 ©2012 IEEE
1

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.
 most relevant effects, like the drag [2] and the oblateness, inertial frame related to γ and N directions (i.e. Aries or
with a linearization with respect to the current configuration. vernal equinox and the North Pole).
This model, included in a suitable controller, allows
propagation of the state and identification of a short-term
control strategy which can be even optimal in the frame of
the limiting hypothesis. The role of the time window
duration is of course crucial in order to identify the more
efficient sequence of commands, but a numerical
propagation of suitable dynamical models is prohibitive in
terms of computation time and resources, while simpler
models intrinsically lose their meaning as they are extended
from the original linearization point and relevant time.

Better results can be expected by having a propagation

providing higher confidence for a longer time, and
computing on this basis a strategy which operates in an
extended time-window. The crucial point is the quick but Figure 1 – The sequence of positions of a spacecraft in a
accurate propagation of the state, which could be obtained perturbed (non-Keplerian) orbit.
by means of a closed form solution. To this aim, also
building on a previous paper by the authors [3], this work
investigates the possibilities of a propagation technique
based on a series expansion in terms of the eccentricity. This
technique, based on a original concept by Broglio and
reported in [4], has been applied in [3] to a single platform,
with the aim to help spacecraft operations when the control
center should quickly compute the orbit based on a limited
number of observations.

In the present paper the theory is recalled and reported in a

more extended way (Section 2), validated numerically
(Section 3) and then extended to the case of a formation by
means of the linearization about the leader spacecraft
(Section 4). It has to be remarked that with respect to Figure 2 – The unit vectors (which orientation
previous work [2], calling for a semi-analytical approach, changes in time) defining the reference frame
the current presentation involves only analytical steps. Even intrinsic to the trajectory.
if these steps are far from trivial and quite elaborate, they do
not represent a problem as they can be computed by means
of symbolic mathematical software (the symbolic
capabilities of MATLAB from Mathworks have been adopted
to this aim) work.

2. THE EQUATIONS OF MOTION FOR

A SINGLE PLATFORM

The real, non-Keplerian trajectory of a spacecraft can be

r
represented as the evolution of a position vector (radius r ),
centered at the Earth, during a sequence of time steps. This
evolution identifies a sequence of angular steps (φ) among
r
the time-varying r (Figure 1). In contrast to the Keplerian
case, such a trajectory does not lie on a plane, and the
sequence of the unit vectors l̂ , m̂ , n̂ , which define a
coordinate frame linked to the specific trajectory (Figure 2),
can be written in terms of the angular step φ. Introducing
(Figure 3) the set of Euler angles typical of the orbital Figure 3 – The (classical) set of Euler angles
mechanics, namely Ω (right ascension of the ascending defining the (instantaneous) orbit with respect to the
node), i (inclination) and ξ (argument of latitude, i.e. (assumed to be) inertially fixed directions of the vernal
argument of perigee plus the anomaly), this frame, intrinsic equinox and of the Earth rotation axis.
to the motion, can be also expressed with respect to an

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.
 As a result, the equations of motion of a spacecraft along the and it is possible to write down for the radius:
perturbed orbit read as
p [5]
3 = A( θ ) cos θ + B( θ ) sin θ
1 1 r 1 1 r
( )' '+ = - 2 [ al - ( )' am ]
r r k r r
with p semilatus rectum and A and B auxiliary functions,
3 introduced by means of the typical variation of the constants
r parameter technique. A and B are expressed by means of
k' = am [1]
k2 closed form integrals for any number of the terms accepted
and evaluated in their power series expansion for θ,
r2 notwithstanding the fact that the convergence of these series
t' = is quite fast. Also, after introducing some initial condition
k
[3], the radius can be expressed as
r3
i' = an cos ξ po [6]
k2 r= 2
3 Req
1 + e o cos θ + J F( θ )
r 3 sin ξ [2] 2 2 p o2
Ω' = an 2
k sin i
where F is a periodic function in the angular variable θ,
r 3 sin ξ which can be completely evaluated till the desired order of
ξ ' = 1 - an 2 expansion in θ
k tan i
where, in addition to the symbols already defined, k is the 3 2 eo2 eo2 eo2
angular momentum, am, an, al the components of the forces F( θ ) = (1- sin io )[(1+ ) - (1+ ) cos θ - cos 2θ] +
2 2 2 6
along the three axes, and the apex (') indicates the derivative 2
with respect to the angular step φ. These equations of 4eo eo 1 27 eo
+ sin2 io [(1+ + ) cos2ξo - (5+ - e2 ) cos2ξo cosθ -
motion can be integrated in an approximate, but accurate at 3 8 6 4 o
till the required level, form if the forcing term can be
1 5e 2e2 1 3e2
expressed as the leading gravitational term plus perturbing - ( + o + o ) sin 2ξo sinθ - (1+ o ) cos( 2ξo + 2θ ) -
term [3]. In the following, interest is focused on the higher 3 8 3 6 2
belt of LEO (above approximately 700 km, so to neglect the 5 e o2
atmospheric drag) and MEO orbit. Therefore, the main - e o cos (2ξ o + 3θ ) - cos (2ξ o + 4θ )] = [7]
24 24
perturbation is given by the harmonics of the Earth
gravitational potential, led by the oblateness term. 4
Considering the potential up to the second order, the = Fo + ∑ Fk cos kθ + Fk' sin kθ
components of the force along the three axes read as k =1

2
µ 3 Req
al = - - µ J [1 - 3 sin 2 i sin 2 ξ ]
r2 2 2 r4 The effect of the orbital perturbations, i.e. the difference
with respect to the Keplerian case (which relevant quantities
2
3 Req [3] are indicated with the index “o”) is given by
a m = - µJ 2 4 sin 2 i sin 2ξ
2 r
2
po po 3 Req
3
2
Req - = 1 + eo cos θ + J 2 2 F ( θ ) - 1 - eo cos φ [8]
an = - µJ 2 4 sin 2i sin ξ r ro 2 po
2 r
which is approximately equal to
µ
to simplify as a l = - , am=0, an=0 in the Keplerian case.
r2 2
po po 3 Req  3  [9]
Substituting φ with a new angular variable θ [difference in - = J 2 2 [ F (θ ) + eoθ 1 − sin 2 io  sin θ ]
the order of (J2)2], we obtain
r ro 2 po  2 

2
The orientation of the position vector with respect to the
R  3 [4] inertial reference frame is defined by the three Euler angles
θ = φ 1 − 3J 2  eq  (1 − sin 2 i)
 p  2 Ω, i and ξ :

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.
 2 3
3 Req T0
Ω = Ωo - J 2 2 cos io [θ + B( θ )] t= (θ − θ 0 ) + ∑ Tk sin kθ + Tk′ ( cos kθ − 1) 
2π
2 po k =1
4
r
= C0 + ∑ [ Ck cos kθ + Sk sin kθ ]
2 a0
3 Req [10]
k =1
[14]
i = io + J 2 2 sin 2io A( θ ) 3
4 po i = I 0 + ∑ [ I k cos kθ + I k′ sin kθ ]
k =1
3
ξ = X 0 + X 0′θ + ∑ [ X k cos kθ + X k′ sin kθ ]
2
3 Req 5
ξ = ξo + θ + J2 2 [cos 2 io B( θ ) + θ (2 - sin 2 io )] k =1
2 po 2 3
Ω = ϒ 0 + ϒ 0′θ + ∑ [ϒ k cos kθ +ϒ k′ sin kθ ]
where k =1
where all coefficients are known functions of i0, e0, a0, ξ0,
A(θ ) = cos 2ξ 0 − cos(2ω0 + 2θ ) + e0 [ cos 2ξ 0 + ω0, J2 and the Earth radius.
1 The variable θ can be expressed as function of time
− cos(ξ 0 + θ + ω0 ) ] + [cos 2ξ0 − cos(3θ + 3ω0 − ξ0 )] = (actually an implicit function of θ itself):
3
3 ∞ 
 
= A0 + ∑ [ Ak cos kθ + Ak′ sin kθ ] θ − θ0 =
2π
t + ∑ Θ n sin n
2π
t + Θ n′  cos n
2π
t − 1 
k =1 T0 n =1  T0  T0 
T0
[15]
4π 3
2π
[11] Θn = −
T02 ∫0 ∑
k =1
Tk sin kθ + Tk′ ( cos kθ − 1)  sin n
T0
tdt

e0 4π
T0 3
2π
B (θ ) = e0 sin(θ + ω0 − ξ 0 ) +[sin 2ξ0 − sin(ξ0 + θ + ω0 )] Θ n′ = − ∫ ∑ T sin kθ + Tk′ ( cos kθ − 1)  cos n tdt
2 T02 0 k =1
k
T0
1 e
+ [sin 2ξ 0 − sin(2ω0 + 2θ )] + 0 [sin 2ξ 0 + where the integrals, previously [2] intended for a numerical
2 6 integration, can be actually solved analytically, with the
3
help of a symbolic mathematical software.
− sin(3θ + 3ω0 − ξ 0 )] = B0 + ∑ [ Bk cos kθ + Bk′ sin kθ ]
k =1
By changing the integration variable from t to θ it is
The time can be obtained through integrating the equation possible to express the argument of the integral by means of
a series in terms of the eccentricity which coefficients are
p3  3  Req 2  3 trigonometric functions of increasing multiples of θ (like
dt 1 2 
= 0 1 + J 2   1 − sin i0  + [12]
dθ µ (1 + e0 cosθ )2  2  p0   2 
sinθ, sin 2θ, sin 3θ…and cos θ, cos2θ,…). The integration
 k k
of resulting series like (e cos kθ , e sinkθ ) can be stopped
F (θ ) 
2 2 at the desired level of accuracy in the eccentricity, level
3  Req   Req 
+ J2   sin 2
i0 A(θ ) −3J 2   dictated by the accuracy already selected for the
4  p0   p0  1 + e0 cosθ  n+1 2
gravitational potential so that e ≈ (J2) and the n+1 term
could be therefore neglected. Due to the intrinsic
where all resulting terms are like orthogonality of the trigonometric functions selected for the
cos kθ sin kθ expansions, the integrals, which should be performed
n
or n
with k = 0,1,2,3,4 n = 2,3 between 0 and 2π, are not difficult as almost all terms
(1+e0 cosθ ) (1+e0 cosθ ) disappear. The same solving procedure can be applied to all
variables of interest, like the quantities r, i, Ω and ξ which
and are analytically integrable. can be expressed as trigonometric series expansions in the
time t. Therefore, the position of the orbiting body is
In the hypothesis that the eccentricity satisfies the condition completely defined by means of functions which can be
e03 ≈ J22 it is possible to obtain time and geometric
evaluated at every desired time and, in the inertial reference
quantities according to
frame linked to the vernal axis will be given by
4
2π 2 kπ 2 kπ
θ −θo = t + ∑ C k cos t + S k sin t [13] x (t ) = r (cos Ω cos ξ − sin Ω sin ξ cos i )
To k =1 To To
y (t ) = r (sin Ω cos ξ − cos Ω sin ξ cos i ) [16]
where To is To= f (J2,initial conditions) z (t ) = r (sin Ω sin ξ + cos Ω sin ξ cos i )

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.
 0.02
3. NUMERICAL VALIDATION

O rb ita l rad ii differen c e (K m )

Previous relations have been validated by comparison to a 0.01
traditional, numerical integration performed with a Runge-
Kutta algorithm (namely the ode45 MATLAB solver), and 0
checked for robustness with respect to the variations in the
integration step. The test case consists in a 7000 km radius, -0.01
20° inclination orbit. The eccentricity has been varied in an
adequately wide range, even if data available in literature -0.02
seem to consider 0.0002 an upper limit in eccentricity
variation for tight formation flying. Notice how the -0.03
0 0.5 1 1.5 2 2.5 3
differences between the proposed analytical formulation and number of orbits
the numerical integration obtained from this comparison are
extremely small all along the simulation, which spans three Figure 6- Difference between the radii computed with
orbits, and with a cost in computation time which is about the proposed model and the (“true”) result of the
500 times less for the proposed closed form solution with numerical integration.
respect to the numerical integration (see Figures 4-6). In
such a way, the time window to compute the control action
could be easily extended to several orbits, therefore leading Once implemented in software, this approach offers a way
to more efficient control strategies. to compute, in an approximate but closed form, the position
-3
of the spacecraft orbiting around the Earth and subject to the
x 10
1 oblateness effect. Moreover, due to the closed form, it easily
O rbital right as c ens ions differenc e (°)

allows for writing the differences between the position of

0 two (or among more than two) spacecraft flying in
-1 formation.
-2

-3
4. THE EQUATIONS OF MOTION FOR A SPACECRAFT
FORMATION
-4
The relations for the radius, the node, the inclination and the
-5 ξ parameter should be now expressed for two close
-6
spacecraft to compute their differences. These differences
0 0.5 1 1.5 2 2.5 3 can be linearized, assuming that the two spacecraft are close
number of orbits and aim to maintain in time the same or a similar
configuration, so that their orbital period (and therefore their
Figure 4- Difference between the right ascension of the semiaxis) should be the same. Therefore, the differences
ascending node (RAAN) computed with the proposed will be expressed as functions of the variation in the
model and the (“true”) result of the numerical remaining five parameters, namely inclination, eccentricity,
integration. node, argument of perigee and the anomaly reckoned form
-3 the ascending node. The relations for the parameters needed
x 10
1.5 to define the relative position of a deputy spacecraft with
respect to a real or virtual chief spacecraft read as:
O rb ital in c lin atio ns diffe ren c e (°)

1
∂r ∂r ∂r ∂r ∂r
∆r = ∆e o + ∆i o + ∆ω o + ∆Ω o + ∆ξ o
0.5
∂eo ∂i o ∂ω o ∂Ω o ∂ξ o

∂i ∂i ∂i ∂i ∂i
0 ∆i = ∆eo + ∆io + ∆ω o + ∆Ω o + ∆ξ o
∂e o ∂i o ∂ω o ∂Ω o ∂ξ o
-0.5 ∂Ω ∂Ω ∂Ω ∂Ω ∂Ω
∆Ω = ∆eo + ∆i o + ∆ω o + ∆Ω o + ∆ξ o
∂e o ∂i o ∂ω o ∂Ω o ∂ξ o
-1
0 0.5 1 1.5 2 2.5 3
∂ξ ∂ξ ∂ξ ∂ξ ∂ξ
number of orbits ∆ξ = ∆e o + ∆io + ∆ω o + ∆Ω o + ∆ξ o
∂e o ∂i o ∂ω o ∂Ω o ∂ξ o
Figure 5- Difference between the inclination computed
with the proposed model and the (“true”) result of the [17]
numerical integration.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.
 where the derivatives can be left as a task of the symbolic APPENDIX
software. The computation of these derivatives, which are
The detailed relations for the coefficients reported in
obtained as a series expansion, can be limited to a defined
Equations (14), obtained by means of MATLAB symbolic
accuracy in order that, for each parameter variation ∆(.), the
toolbox, are reported in the following, holding the
following condition holds
assumption e03 ≈ J22, for the radius:
e k ∆ (⋅) ≈ J 2 [18] r 4
= C0 + ∑ [Ck cos kθ + S k sin kθ ]
a0 k =1
Once the parameters have been obtained, it is possible to 2
1 J R 
evaluate the position of the deputy spacecraft in an intrinsic {
C0 = 1 − e02 + 2  ⊕  12  −2 + sin 2 i0 ( 3 − 2 cos 2ξ0 )  +
2 16  a0 
reference frame centered in the chief spacecraft. In such a
frame, the position can be usefully expressed as range,
azimuth α and elevation ε, which are defined as
+4e0  −6 + sin 2 i0 ( 9 − 13cos 2ξ 0 )  + 3e02  −24 + sin 2 i0 ( 36 − 29 cos 2ξ0 )  }
2
∆r  R  3 1  1  71 
C1 = −e0 + J 2  ⊕   − sin 2 i0 ( 9 - 5* cos 2ξ 0 ) + e0 3 − sin 2 i0  9 - * cos 2ξ 0   +
ε = arcsin z  a0   2 4  2  8 
∆r 1 2 
+ e0 86 − sin i0 (129 - 125* cos 2ξ 0 )  
2
[19] 16 
∆r y 1  R  1
2
1
tan α = C2 = e02 + J 2  ⊕   cos 2ξ 0 sin 2 i0 − e0 6 − sin 2 i0 ( 9 - 5* cos 2ξ 0 )  +
∆ rz 2  a0   4 4
  13  
+e02  −2 + sin 2 i0  3 - * cos 2ξ 0   
  4  
and actually computed by means of 2
1  R⊕ 
C3 = J2 
16  a0 
{
 e0 cos 2ξ 0 sin i0 + 3e0 6 − sin i0 ( 9 - 5* cos 2ξ 0 ) 
2 2
}
r(∆i sinξ − ∆Ωsini cosξ ) 2
sinε = 1  R⊕  2 2
C4 = − J2   e0 sin i0 cos 2ξ 0
r 2 (∆i sinξ − ∆Ωsini cosξ )2 + ∆r 2 + r 2 (∆ξ + cosi ∆Ω)2 16  a0 

2
1  R⊕ 
 ( 8 + 19e0 + 38e0 ) sin i0 sin 2ξ0
2 2
r S1 = J2 
tan α = (∆ξ + cos i ∆Ω ) [20] 16  a0 
∆r 2
1 R 
S2 = − J 2  ⊕  ( 2 + 4e0 + 13e02 ) sin 2 i0 sin 2ξ0
8  a0 
This set of relations finally allow for computing the relative
2
position of the deputy at every time with the required 1  R⊕ 
 e0 ( −1 + 6e0 ) sin i0 sin 2ξ 0
2
S3 = J2 
accuracy, and can be therefore used for a closed form 16  a0 
representation of the formation trajectory. 2
1  R⊕  2 2
S4 = J2   e0 sin i0 sin 2ξ 0
16  a0 

5. FINAL REMARKS
The implementation of space missions based on the
formation flying concept requires the determination of the
relative trajectories of the spacecraft. Such a knowledge is
usually obtained by means of a propagation software based
on a numerical integration. This process requires remarkable
computation time and resources to be available onboard.
Therefore, an analytical, closed form solution can be really
important, as it involves limited spacecraft capabilities. The
paper details how such a closed form solution can be
obtained and validates this solution by comparison with a
standard Runge-Kutta propagator. The position of the
deputy satellite with respect to the chief is expressed as a
function of an expansion of the eccentricity, which can be
carried on until the desired level of accuracy. A symbolic
mathematical software has been used to evaluate – once and
for all - the elevation, the azimuth and the distance of the
satellite beginning to the formation in an intrinsic reference
frame centered on the – real or virtual - chief.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.
 For the inclination [2] T. Carter, M. Humi, “The Clohessy-Wiltshire Equations
3 can be Modified to Include Quadratic Drag”, Paper AAS
i = I 0 + ∑  I k cos kθ + I k′ sin kθ  03-240, 13th AAS/AIAA Space Flight Mechanics
k =1
2
Meeting, Ponce (Puerto Rico), Febr. 2003.
J R 
I 0 = i0 − 2  ⊕  sin 2i0 cos 2ξ 0 ( 3 + 4e0 + 6e02 )
8  a0  [3] S. Sgubini, G.B. Palmerini, “Ground-Based Orbit
2 2
3J  R  3J  R  Determination for Spacecraft Formations”, paper
I1 = 2  ⊕  e0 sin 2i0 cos(ω0 + ξ 0 ) I1′ = − 2  ⊕  e0 sin 2i0 sin(ω0 + ξ 0 )
8  a0  8  a0  IEEEAC1480, Proceedings of the IEEE Aerospace
3J 2  R⊕ 
2
3 J 2  R⊕ 
2 Conference, Big Sky (MT, USA) (2010).
 sin 2i0 cos 2ω0 (1 + 2e0 ) I 2′ = −  sin 2i0 sin 2ω0 (1 + 2e0 )
2 2
I2 =  
8  a0  8  a0 
2 2 [4] S.Sgubini, Appunti del corso di Meccanica del Volo
J 2  R⊕  J 2  R⊕ 
I3 =   e0 sin 2i0 cos( −3ω0 + ξ 0 ) I 3′ =   e0 sin 2i0 sin(−3ω0 + ξ 0 ) Spaziale, Scuola di Ingegneria Aerospaziale, Università
8  a0  8  a0 
di Roma La Sapienza (in Italian).

[5] M. Sabatini, G.B. Palmerini, “Deployment Strategies for a

For the anomaly reckoned from the node: Formation of Pico-Satellites”, paper IAC-09-D1.4.4, 60th
International Astronautical Congress, Daejeon, Korea
3 (2009).
ξ=X0+X0′θ+∑Xkcoskθ+Xk′sinkθ
 
k=1
2
3 R 
4  a0 
{
X0 =ω0 + J2 ⊕ (1+2e02)( 4−5sin2i0) (ω0 −ξ0) +
3 R 
2

X0′ =1+ J2 ⊕ (1+2e02)( 4−5sin2i0) BIOGRAPHIES

 4  4  a0 
+cos2i0sin2ξ01+ e0 +2e02 Silvano Sgubini is full Professor of Space Flight Mechanics
 3 
2 2 at the University of Rome, and previously at the University
3J R  3J R  of Pisa, in Italy. A Mechanical and Aerospace Engineer, he
X1 =− 2  ⊕ e0cos2i0[sin(ω0 +ξ0)−2sin(ω0 −ξ0)] X1′ =− 2  ⊕ e0cos2i0[cos(ω0 +ξ0)−2cos(ω0 −ξ0)]
4  a0  4  a0  has been especially interested in space structures and
3J R 
2
3J R 
2
thermal problems. His current research deals with trajectory
X2 =− 2  ⊕ (1+2e02) cos2i0sin2ω0 X2′ =− 2  ⊕ (1+2e02) cos2i0cos2ω0 design, with special reference to solar sails and to lunar
4  a0  4  a0 
2 2
observation mission, and includes plans for a lunar rover.
J R  J R 
X3 =− 2  ⊕ e0 cos2i0sin(3ω0 −ξ0) X3′ =− 2  ⊕ e0cos2i0cos(3ω0 −ξ0)
4  a0  4  a0  Giovanni Battista Palmerini received the Laurea Degree
(summa cum laude) in Aeronautical Engineering from the
University of Rome, Italy, in 1991, and the Ph.D. in
Aerospace Engineering in 1996. Formerly a visiting scholar
at Stanford and Professor at the University of Bologna, he is
For the right ascension of the ascending node: currently Professor of Aerospace Guidance and Navigation
Systems at the University of Rome. His research interests
3 are in the field of formation flying, robotics and space
Ω =ϒ0 +ϒ0′θ +∑ϒk coskθ +ϒk′ sinkθ systems.
 
k=1
2
3  R⊕ 
ϒ0 = Ω0 − J2 
4  a0 
{
 cosi0 2(1+ 2e0 ) (ω0 −ξ0 ) +
2

3  R⊕ 
2

 cosi0 (1+2e0 )
2
ϒ0′ =− J2 
 4  2  a0 
+sin2ξ0 1+ e0 + 2e02 
 3 
2 2
3J2  R⊕  3J2  R⊕ 
ϒ1 =   e0 cosi0 [sin(ω0 +ξ0) −2sin(ω0 −ξ0)] ϒ1′ =   e0 cosi0 [cos(ω0 +ξ0) −2cos(ω0 −ξ0)]
4  a0  4  a0 
2 2
3J2  R⊕  3J2  R⊕ 
  (1+ 2e0 ) cosi0 sin2ω0   (1+2e0 ) cosi0 cos2ω0
2 2
ϒ2 = ϒ2′ =
4  a0  4  a0 
2 2
J  R⊕  J  R⊕ 
ϒ3 = 2   e0 cosi0 sin(3ω0 −ξ0) ϒ3′ = 2   e0 cosi0 cos(3ω0 −ξ0)
4  a0  4  a0 

REFERENCES
[1] W.H. Clohessy, R.S. Wiltshire, “Terminal Guidance
System for Satellite Rendez-Vous”, The Journal of
Aerospace Sciences, Vo.27, No.9, pp.653-658, 1960.
7

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on May 08,2024 at 20:50:40 UTC from IEEE Xplore. Restrictions apply.