The dynamics of satellites formation is of great interest for the space mission. This work discusses a more efficient model of the relative motion dynamics of satellites formation. The model is based on employing the concepts restricted three-body problem (R3BP) and for more accuracy, it considers the effects of both oblateness and radiation pressure on deputy relative motion w.r.t the chief satellite. A model of deputy relative motion w.r.t the chief satellite is derived in the local-vertical local-horizontal system and simplified assuming the concept of the circular restricted three-body problem (CR3BP). The deputy equations of motion were rewritten in the form of recurrence relations and solved numerically using the Lie series approach. Assuming that the formation is revolving around the Moon in the Earth-Moon system, the effects of both oblateness and radiation pressure on the deputy satellite orbit were assessed through a particular example of satellites formation. A comparison between the perturbed and unperturbed R3BP shows a significant difference in the deputy relative position that has to be considered for the formation dynamics.
Parallel to the start of the space programs, the study of spacecraft relative dynamics became one of the most important aspects in designing and analyzing space missions. Several authors were interested in this study to analyze rendezvous and docking of two spacecrafts in addition to maintenance of spacecraft formation. The most famous models of relative motion, Clohessy-Wiltshire and Tschauner-Hempel, have been used to analyze relative guidance, navigation and control systems. However, these two models assumed that the spacecraft relative distance is significantly small compared with its position w.r.t the centre of mass of the primary. Moreover, they assume pure Kepler motion (e.g. two-body problem without including any perturbations) [
In 2019, Giovanni Franzini and Mario Innocenti studied the relative motion dynamics using the classical restricted three-body problem (unperturbed problem). They found that a three-body scenario is more suitable than a two-body scenario to describe the dynamics of satellite relative motion and more efficient for studying the relative guidance and navigation system [
The current study aims to get a more accurate formulation of the relative motion employing the perturbed restricted three-body problem. The dynamical model assumes that the primaries are radiating and oblate spheroids considering only zonal harmonics. Based on constructing recurrence formulas and applying the Lie series approach, the model is solved numerically. Finally, the numerical application is performed assuming a circular three-dimensional problem.
The restricted three-body problem is one of the most famous dynamical modellings of celestial systems. Extensive studies were performed using a variety of methods assuming that the primaries are spheres and only their gravitational attraction is considered [
ϕ k = − G m k r k [ 1 − ∑ n = 2 ∞ j n k ( R k r k ) n P n k ( sin δ ) ] (1)
G is the universal gravitational constant, R k is the mean radius of each body, δ is the latitude of the infinitesimal body, r k is the separation between every two bodies and j n k is the dimensionless coefficient which represents the non-spherical components of the potential. The Legendre polynomials “ P n k ( sin δ ) ” of degree n is given by:
P n k ( sin δ ) = 1 2 n n ! d n d δ n ( sin 2 δ − 1 ) n
The gravitational force “ F ¯ g r ” exerted by the body is:
F ¯ g r = − ∇ ϕ k (2)
where ∇ denotes the vector differential operator. Apart from the gravitational potential, if the body is radiating, then it exerts a radiation force “ F ¯ r a d ” in the opposite direction of its gravitational force. Consequently, the total force “ F ¯ k ” is:
F ¯ k = F ¯ g r − F ¯ r a d = q k F ¯ g r (3)
where q k = ( 1 − F r a d / F g r ) is the radiation factor ∈ ( 0 , 1 ) . Substitute (1) and (2) into (3):
F ¯ k = − q k G m k ∑ n = 2 ∞ [ 1 r k 3 − A n ( k ) ( 1 r k ) n + 3 ] r ¯ k (4)
where
A n ( k ) = ( n + 1 ) j n ( k ) R k n P n ( k ) ( sin δ )
Consider a system of three bodies that have masses m 1 , m 2 and m 3 such that the masses of primaries m 1 > m 2 and m 3 is the mass of the chief satellite in a formation. Let ( I : i ^ , j ^ , k ^ ) is a sidereal (inertial) coordinate system with the origin lies in the centre of mass of the two primaries, the motion of each body is given by:
[ r ¯ ¨ k ] I = − G ∑ i = 1 3 ∑ n = 2 ∞ q i m i [ 1 ‖ r ¯ i k ‖ 3 − A n ( i ) ( 1 ‖ r ¯ i k ‖ ) n + 3 ] r ¯ i k
where r ¯ i k = r ¯ k − r ¯ i , i = 1 , 2 , 3 and k = 1 , 2 , 3 and i ≠ k . For the restricted case of three bodies, the primaries are not affected by the gravitational influence of the satellite. Assuming that the formation is revolving around the smaller primary, the equation of motion of the chief satellite is (see
[ r ¯ ¨ 23 ] I = [ r ¯ ¨ 3 ] I − [ r ¯ ¨ 2 ] I = − q 1 G m 1 ∑ n = 2 ∞ [ 1 ‖ r ¯ 13 ‖ 3 − A n ( 1 ) ( 1 ‖ r ¯ 13 ‖ ) n + 3 ] r ¯ 13 − q 2 G m 2 ∑ n = 2 ∞ [ 1 ‖ r ¯ 23 ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ 23 ‖ ) n + 3 ] r ¯ 23 − q 1 G m 1 ∑ n = 2 ∞ [ 1 ‖ r ¯ 12 ‖ 3 − A n ( 1 ) ( 1 ‖ r ¯ 12 ‖ ) n + 3 ] r ¯ 12
As clear in
[ r ¯ ¨ c ] I = − q 1 G m 1 ( ( r ¯ 12 + r ¯ c ) ‖ r ¯ 12 + r ¯ c ‖ 3 + r ¯ 12 ‖ r ¯ 12 ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ ( r ¯ 12 + r ¯ c ) ‖ r ¯ 12 + r ¯ c ‖ n + 3 + r ¯ 12 ‖ r ¯ 12 ‖ n + 3 ] ) − q 2 G m 2 ∑ n = 2 ∞ [ 1 ‖ r ¯ c ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ c ‖ ) n + 3 ] r ¯ c (5)
Equation (5) can be normalized by assuming that the gravitational constant G = 1 , the distance between the two primaries is unity and the sum of their masses is unity. Let the normalized mass of the small primary parameter be μ = m 2 m 1 + m 2 , then that of the big primary is “ 1 − μ ”. Consequently, the chief equation of motion will be:
[ r ¯ ¨ c ] I = − q 1 ( 1 − μ ) ( ( r ¯ 12 + r ¯ c ) ‖ r ¯ 12 + r ¯ c ‖ 3 + r ¯ 12 ‖ r ¯ 12 ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ ( r ¯ 12 + r ¯ c ) ‖ r ¯ 12 + r ¯ c ‖ n + 3 + r ¯ 12 ‖ r ¯ 12 ‖ n + 3 ] ) − q 2 μ ∑ n = 2 ∞ [ 1 ‖ r ¯ c ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ c ‖ ) n + 3 ] r ¯ c (6)
Similarly, the motion of a deputy satellite of the formation will be:
[ r ¯ ¨ d ] I = − q 1 ( 1 − μ ) ( ( r ¯ 12 + r ¯ d ) ‖ r ¯ 12 + r ¯ d ‖ 3 + r ¯ 12 ‖ r ¯ 12 ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ ( r ¯ 12 + r ¯ d ) ‖ r ¯ 12 + r ¯ d ‖ n + 3 + r ¯ 12 ‖ r ¯ 12 ‖ n + 3 ] ) − q 2 μ ∑ n = 2 ∞ [ 1 ‖ r ¯ d ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ d ‖ ) n + 3 ] r ¯ d (7)
Let ( S : i ^ 2 , j ^ 2 , k ^ 2 ) is a synodic coordinate system with an origin that lies in the centre of mass of the smaller primary which is defined as follows:
i ^ 2 = − r ¯ 12 r 12 , j ^ 2 = k ^ 2 × i ^ 2 , k ^ 2 = − h ¯ 2 h 2
where h ¯ 2 = r ¯ 12 × r ¯ ˙ 12 is the angular momentum of the smaller primary w.r.t. the bigprimary. Let this frame rotates with angular velocity ω ¯ 2 = ω 2 k ^ 2 w.r.t the inertial frame ( I : X , Y , Z ) . Then, the equation of motion of the chief w.r.t. smaller primary is:
[ r ¯ ¨ c ] S = [ r ¯ ¨ c ] I − 2 ω ¯ 2 × [ r ¯ ˙ c ] S − [ ω ¯ ˙ 2 ] S × r ¯ c − ω ¯ 2 × ( ω ¯ 2 × r ¯ c ) = − q 1 ( 1 − μ ) ( ( r ¯ 12 + r ¯ c ) ‖ r ¯ 12 + r ¯ c ‖ 3 + r ¯ 12 ‖ r ¯ 12 ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ ( r ¯ 12 + r ¯ c ) ‖ r ¯ 12 + r ¯ c ‖ n + 3 + r ¯ 12 ‖ r ¯ 12 ‖ n + 3 ] ) − q 2 μ ∑ n = 2 ∞ [ 1 ‖ r ¯ c ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ c ‖ ) n + 3 ] r ¯ c − 2 ω ¯ 2 × [ r ¯ ˙ c ] S − [ ω ¯ ˙ 2 ] S × r ¯ c − ω ¯ 2 × ( ω ¯ 2 × r ¯ c ) (8)
Similarly, the motion of the deputy satellite will be:
[ r ¯ ¨ d ] S = − q 1 ( 1 − μ ) ( ( r ¯ 12 + r ¯ d ) ‖ r ¯ 12 + r ¯ d ‖ 3 + r ¯ 12 ‖ r ¯ 12 ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ ( r ¯ 12 + r ¯ d ) ‖ r ¯ 12 + r ¯ d ‖ n + 3 + r ¯ 12 ‖ r ¯ 12 ‖ n + 3 ] ) − q 2 μ ∑ n = 2 ∞ [ 1 ‖ r ¯ d ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ d ‖ ) n + 3 ] r ¯ d − 2 ω ¯ 2 × [ r ¯ ˙ d ] S − [ ω ¯ ˙ 2 ] S × r ¯ d − ω ¯ 2 × ( ω ¯ 2 × r ¯ d ) (9)
where
[ r ¯ ˙ c ] S = [ r ¯ ˙ c ] I − ω ¯ 2 × r ¯ c
To describe the relative motion of the deputy w.r.t. the chief, let ( L : i ^ 3 , j ^ 3 , k ^ 3 ) is a Local Vertical Local Horizontal frame “LVLH” with an origin that lies in the chief centre of mass and is defined as (As is clear in
i ^ 3 = j ^ 3 × k ^ 3 , j ^ 3 = − h ¯ c h c , k ^ 3 = − r ¯ c r c (10)
where h ¯ c = r ¯ c × [ r ¯ ˙ c ] S is the chief angular momentum w.r.t. the second primary. The position of the deputy w.r.t. the smaller primary is given by (see
r ¯ d = r ¯ c + ρ ¯ (11)
Let ( L : i ^ 3 , j ^ 3 , k ^ 3 ) rotates with angular velocity ω ¯ 3 w.r.t. the inertial frame ( I : X , Y , Z ) , then
[ r ¯ ˙ d ] I = [ r ¯ ˙ c ] I + [ ρ ¯ ˙ ] I = [ r ¯ ˙ d ] I + [ ρ ¯ ˙ ] L + ω ¯ 3 × ρ ¯
[ r ¯ ¨ d ] I = [ r ¯ ¨ c ] I + [ ρ ¯ ¨ ] L + 2 ω ¯ 3 × [ ρ ¯ ˙ ] L + [ ω ¯ ˙ 3 ] I × ρ ¯ + ω ¯ 3 × ( ω ¯ 3 × ρ ¯ ) (12)
Introducing (6) and (7) into (12), then
[ ρ ¯ ¨ ] L = − q 1 ( 1 − μ ) [ 1 ‖ r ¯ 12 + r ¯ d ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ 1 ‖ r ¯ 12 + r ¯ d ‖ n + 3 ] ] ( r ¯ 12 + r ¯ d ) − q 2 μ ∑ n = 2 ∞ [ 1 ‖ r ¯ d ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ d ‖ ) n + 3 ] r ¯ d + q 1 ( 1 − μ ) [ 1 ‖ r ¯ 12 + r ¯ c ‖ 3 − ∑ n = 2 ∞ A n ( 1 ) [ 1 ‖ r ¯ 12 + r ¯ c ‖ n + 3 ] ] ( r ¯ 12 + r ¯ c ) + q 2 μ ∑ n = 2 ∞ [ 1 ‖ r ¯ c ‖ 3 − A n ( 2 ) ( 1 ‖ r ¯ c ‖ ) n + 3 ] r ¯ c − W ( x , y , z ) (13)
where
W ( x , y , z ) = 2 ω ¯ 3 × [ ρ ¯ ˙ ] L + [ ω ¯ ˙ 3 ] I × ρ ¯ + ω ¯ 3 × ( ω ¯ 3 × ρ ¯ ) = ( ω 3 x ω 3 y y + ω 3 x ω 3 z z − ( ω 3 y ) 2 x − ( ω 3 z ) 2 x + 2 z ˙ ω 3 y − 2 y ˙ ω 3 z + z ω ˙ 3 y − y ω ˙ 3 z ) i ^ 2 + ( ω 3 x ω 3 y x + ω 3 y ω 3 z z − ( ω 3 x ) 2 y − ( ω 3 z ) 2 y − 2 z ˙ ω 3 x + 2 x ˙ ω 3 z − z ω ˙ 3 x + x ω ˙ 3 z ) j ^ 2 + ( ω 3 x ω 3 z x + ω 3 y ω 3 z y − ( ω 3 x ) 2 z − ( ω 3 y ) 2 z + 2 y ˙ ω 3 x − 2 x ˙ ω 3 y + y ω ˙ 3 x − x ω ˙ 3 y ) k ^ 2
The angular velocity of “ L ” w.r.t “I” is given by:
ω ¯ 3 = ω ¯ L / S + ω ¯ 2
[ ω ¯ ˙ 3 ] L = [ ω ¯ ˙ L / S ] L + [ ω ¯ ˙ 2 ] L = [ ω ¯ ˙ L / S ] L + [ ω ¯ ˙ 2 ] S − ω ¯ 3 × ω ¯ 2 (14)
where ω ¯ L / S and ω ¯ 2 are angular velocities of L w.r.t. S and S w.r.t. I respectively. To determine ω ¯ L / S and ω ¯ ˙ L / S , a simple scheme based on the time derivatives of LVLH w.r.t. the synodic frame “ S ” [
[ i ˙ ^ 3 ] S = ω ¯ L / S × i ^ 3 , [ j ˙ ^ 3 ] S = ω ¯ L / S × j ^ 3 , [ k ˙ ^ 3 ] S = ω ¯ L / S × k ^ 3 (15)
Multiply (15) by the relative unit vector as follows:
i ^ 3 × [ i ˙ ^ 3 ] S = i ^ 3 × ( ω ¯ L / S × i ^ 3 ) = ω ¯ L / S − ( ω ¯ L / S ⋅ i ^ 3 ) i ^ 3
j ^ 3 × [ j ˙ ^ 3 ] S = j ^ 3 × ( ω ¯ L / S × j ^ 3 ) = ω ¯ L / S − ( ω ¯ L / S ⋅ j ^ 3 ) j ^ 3
k ^ 3 × [ k ˙ ^ 3 ] S = k ^ 3 × ( ω ¯ L / S × k ^ 3 ) = ω ¯ L / S − ( ω ¯ L / S ⋅ k ^ 3 ) k ^ 3
By summing up the previous equations we get
i ^ 3 × [ i ˙ ^ 3 ] S + j ^ 3 × [ j ˙ ^ 3 ] S + k ^ 3 × [ k ˙ ^ 3 ] S = 2 ω ¯ L / S (16)
Considering (10), the time derivative of the unit vectors of the LVLH frame is:
[ k ˙ ^ 3 ] S = − 1 r c ( [ r ¯ ˙ c ] S ⋅ i ^ 3 ) i ^ 3
[ j ˙ ^ 3 ] S = − r ¯ c h c ( [ r ¯ ¨ c ] S ⋅ j ^ 3 ) i ^ 3
[ i ˙ ^ 3 ] S = r ¯ c h c ( [ r ¯ ¨ c ] S ⋅ j ^ 3 ) j ^ 3 + 1 r c ( [ r ¯ ˙ c ] S ⋅ i ^ 3 ) k ^ 3 (17)
Substitute from (17), into (16), then:
ω ¯ L / S = [ 0 − h c r c 2 − r c h c 2 ( [ r ¯ ¨ c ] S ⋅ h ¯ c ) ] and ω ¯ ˙ L / S = [ 0 − ( h ˙ c r c 2 − 2 r ˙ c h c r c 3 ) − ( r ˙ c h c 2 − 2 h ˙ c r c h c 3 ) h ¯ c ⋅ [ r ¯ ¨ c ] S − r c h c 2 h ¯ c ⋅ [ r ¯ ⃛ c ] S ] (18)
where [ r ¯ ¨ c ] S and the jerk [ r ¯ ⃛ c ] S can be obtained by Equation (8) and its differentiation. Equation (13) along with (14) and (18) represent the equation of relative motion of the deputy w.r.t. the chief in the frame of the perturbed restricted three-body problem. It noted that the equation of relative motion is a nonlinear 2nd order differential equation with time-varying parameters which can be simplified assuming the circular case of the restricted three-body problem.
Assuming that the two primaries revolve in a circular orbit around their common centre of mass, then the following simplifications will be considered [
r ¯ 12 = − i ^ 2 , [ r ˙ 12 ] S = 0 , ω ¯ 2 = k ^ 2 , ω ¯ ˙ 2 = 0 , ω ¯ ¨ 2 = 0
Consequently, the angular velocity and acceleration of the LVLH frame w.r.t. the inertial frame is simplified as follows:
ω ¯ 3 = ω ¯ L / S + k ^ 2
[ ω ¯ ˙ 3 ] L = [ ω ¯ ˙ L / S ] L − ω ¯ L / S × k ^ 2
For more simplifications, assume that both primaries are radiating and only the second zonal harmonic is considered. Then
A 2 ( k ) = 3 j 2 ( k ) R k 2 P 2 ( k ) ( sin δ )
Let
r c = ‖ r ¯ c ‖
r ¯ c = − r c k ^ 2
ρ ¯ = x i ^ 2 + y j ^ 2 + z k ^ 2
r ¯ d = r ¯ c + ρ ¯ = x i ^ 2 + y j ^ 2 + ( z − r c ) k ^ 2
G 1 = ‖ r ¯ d ‖ − 3 = ‖ r ¯ c + ρ ¯ ‖ − 1 = ( x 2 + y 2 + ( z − r c ) 2 ) − 1 2
G 2 = ‖ r ¯ 12 + r ¯ c + ρ ¯ ‖ − 1 = ( ( x − 1 ) 2 + y 2 + ( z − r c ) 2 ) − 1 2
G 3 = ‖ r ¯ 12 + r ¯ c ‖ − 1 = ( 1 + r c 2 ) − 1 2
Under these assumptions, the equation of relative motion (13) will be reduced to:
[ ρ ¯ ¨ ] L = [ − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] ( x − 1 ) − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] x − ( 1 − μ ) q 1 [ G 3 3 − A 2 ( 1 ) G 3 5 ] − ω 3 x ω 3 y y − ω 3 x ω 3 z z + ( ω 3 y ) 2 x + ( ω 3 z ) 2 x − 2 z ˙ ω 3 y + 2 y ˙ ω 3 z − z ω ˙ 3 y + y ω ˙ 3 z ] i ^ 2 + [ − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] y − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] y − ω 3 x ω 3 y x − ω 3 y ω 3 z z + ( ω 3 x ) 2 y + ( ω 3 z ) 2 y + 2 z ˙ ω 3 x − 2 x ˙ ω 3 z + z ω ˙ 3 x − x ω ˙ 3 z ] j ^ 2
+ [ − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] ( z − r c ) − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] ( z − r c ) − ( 1 − μ ) q 1 [ G 3 3 − A 2 ( 1 ) G 3 5 ] r c + μ q 2 [ 1 r c 2 − A 2 ( 2 ) ( 1 r c ) 4 ] − ω 3 x ω 3 z x − ω 3 y ω 3 z y + ( ω 3 x ) 2 z + ( ω 3 y ) 2 z − 2 y ˙ ω 3 x + 2 x ˙ ω 3 y − y ω ˙ 3 x + x ω ˙ 3 y ] k ^ 2 (19)
where ω 3 x , ω 3 y , ω 3 z and ω ˙ 3 x , ω ˙ 3 y , ω ˙ 3 z are the components of the angular velocity ω ¯ 3 and ω ¯ ˙ 3 .
Power series approaches are widely used to solve different celestial mechanics problems. Many authors depend on that algorithm to find an approximate solution for their problems [
Step I: Construction of the Lie operator
Let x = g 1 x ˙ = g ˙ 1 = g 2 y = g 3 y ˙ = g ˙ 3 = g 4 z = g 5 z ˙ = g ˙ 5 = g 6
Then
g ˙ 1 = g 2
g ˙ 2 = − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] ( g 1 − 1 ) − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] g 1 − ( 1 − μ ) q 1 [ G 3 3 − A 2 ( 1 ) G 3 5 ] − ω 3 g 1 ω 3 g 3 g 3 − ω 3 g 1 ω 3 g 5 g 5 + ( ω 3 g 3 ) 2 g 1 + ( ω 3 g 5 ) 2 g 1 − 2 g 6 ω 3 g 3 + 2 g 4 ω 3 g 5 − g 5 ω ˙ 3 g 3 + g 3 ω ˙ 3 g 5
g ˙ 3 = g 4
g ˙ 4 = − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] g 3 − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] g 3 − ω 3 g 1 ω 3 g 3 g 1 − ω 3 g 3 ω 3 g 5 g 5 + ( ω 3 g 1 ) 2 g 3 + ( ω 3 g 5 ) 2 g 3 + 2 g 6 ω 3 g 1 − 2 g 2 ω 3 g 5 + g 5 ω ˙ 3 g 1 − g 1 ω ˙ 3 g 5
g ˙ 5 = g 6
g ˙ 6 = − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] ( g 5 − r c ) − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] ( g 5 − r c ) − ( 1 − μ ) q 1 [ G 3 3 − A 2 ( 1 ) G 3 5 ] r c + μ q 2 [ 1 r c 2 − A 2 ( 2 ) ( 1 r c ) 4 ] − ω 3 g 1 ω 3 g 5 g 1 − ω 3 g 3 ω 3 g 5 g 3 + ( ω 3 g 1 ) 2 g 5 + ( ω 3 g 3 ) 2 g 5 − 2 g 4 ω 3 g 1 + 2 g 2 ω 3 g 3 − g 3 ω ˙ 3 g 1 + g 1 ω ˙ 3 g 3
In general, the components of the equation of relative motion can be rewritten in a matrix form as:
g ˙ = P + W
where g ˙ , P and W are 6 × 1 matrices defined as follows:
g ˙ = ( g ˙ 1 g ˙ 2 g ˙ 3 g ˙ 4 g ˙ 5 g ˙ 6 ) T
and
W = ( W 1 W 2 W 3 W 4 W 5 W 6 ) T
With W 1 = W 3 = W 5 = 0 . Based on (18), ω 3 g 1 = 0 , then
W 2 = − ( 1 − μ ) q 1 G 3 3 + ( ω 3 g 3 ) 2 g 1 + ( ω 3 g 5 ) 2 g 1 − 2 g 6 ω 3 g 3 + 2 g 4 ω 3 g 5 − g 5 ω ˙ 3 g 3 + g 3 ω ˙ 3 g 5
W 4 = − ω 3 g 3 ω 3 g 5 g 5 + ( ω 3 g 5 ) 2 g 3 − 2 g 2 ω 3 g 5 + g 5 ω ˙ 3 g 1 − g 1 ω ˙ 3 g 5
W 6 = − ω 3 g 3 ω 3 g 5 g 3 + ( ω 3 g 3 ) 2 g 5 + 2 g 2 ω 3 g 3 − g 3 ω ˙ 3 g 1 + g 1 ω ˙ 3 g 3 − ( 1 − μ ) q 1 G 3 3 + μ q 2 [ 1 r c 2 − A 2 ( 2 ) ( 1 r c ) 4 ]
However, P is defined as:
P = U V
where is 6 × 6 matrix given by
U = ( g 2 0 0 0 0 0 0 ( g 1 − 1 ) 0 g 1 0 1 0 0 g 4 0 0 0 0 g 3 0 g 3 0 0 0 0 0 0 g 6 0 0 ( g 5 − r c ) 0 ( g 5 − r c ) 0 r c )
V = ( 1 − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ] 1 − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ] 1 − ( 1 − μ ) q 1 [ G 3 3 − A 2 ( 1 ) G 3 5 ] ) T
The Lie-Operator is defined as [
D = d d t = ∑ i = 1 6 ∂ ∂ g i d g i d t + ∂ ∂ t
For explicit time variables, then
D = ∑ i = 1 6 g ˙ i ∂ ∂ g i
For j = 1 , assuming that P i j = P i , W i j = W i and g ˙ i j = g ˙ i then g ˙ i = ( P i + W i ) and
D = ∑ i = 1 6 ( P i + W i ) ∂ ∂ g i = ( P 1 + W 1 ) ∂ ∂ g 1 + ( P 2 + W 2 ) ∂ ∂ g 2 + ( P 3 + W 3 ) ∂ ∂ g 3 + ( P 4 + W 4 ) ∂ ∂ g 4 + ( P 5 + W 5 ) ∂ ∂ g 5 + ( P 6 + W 6 ) ∂ ∂ g 6 (20)
Step II: Construction of the recurrence relations for each variable
Applying the Lie operator Equation (20) on g i , we obtain that
D n g i = D n − 1 ( P i + W i )
1) The recurrence formulas for D P i
P i = ∑ j = 1 6 U i j V j i
D P i = ∑ j = 1 6 [ U i j D V j i + V j i D U i j ]
Generally,
D n P i = ∑ m = 0 n ∑ j = 1 6 [ D n − m V j i ] [ D m U i j ] (21)
a) The recurrence formulas for D V j i
V 21 = − ( 1 − μ ) q 1 [ G 2 3 − A 2 ( 1 ) G 2 5 ]
D V 21 = T 1
T 1 = ( 1 − μ ) q 1 [ 3 G 2 5 − 5 A 2 ( 1 ) G 2 7 ] [ ( g 1 − 1 ) D g 1 + g 3 D g 3 + ( g 5 − r c ) D g 5 ] = ( 1 − μ ) q 1 [ 3 G 2 5 − 5 A 2 ( 1 ) G 2 7 ] ( g 1 g 2 + g 3 g 4 + ( g 5 − r c ) g 6 )
V 41 = − μ q 2 [ G 1 3 − A 2 ( 2 ) G 1 5 ]
D V 41 = T 2
T 2 = μ q 2 [ 3 G 1 5 − 5 A 2 ( 2 ) G 1 7 ] ( g 1 D g 1 + g 3 D g 3 + ( g 5 − r c ) D g 5 ) = μ q 2 [ 3 G 1 5 − 5 A 2 ( 2 ) G 1 7 ] ( g 1 g 2 + g 3 g 4 + ( g 5 − r c ) g 6 )
The higher powers of D V 21 and D V 41 are computed as
D n V 21 = D n − 1 T 1 and D n V 41 = D n − 1 T 2 (22)
The recurrence relation of the rest of the elements of V is zero.
b) The recurrence formulas for D U i j
By definition of g i , its noted that D n g i = D n − 1 g i + 1
U 11 = g 2 then D n U 11 = D n g 2
U 22 = g 1 − 1 and U 24 = g 1 then D n U 22 = D n U 24 = D n g 1 = D n − 1 g 2
U 33 = g 4 then D n U 33 = D n g 4
U 42 = U 44 = g 3
then
D n U 42 = D n g 3 = D n − 1 g 4 (23)
U 55 = g 6
Then
D n U 55 = D n g 6 (24)
U 62 = U 64 = g 5 − r c
Then
D n U 62 = D n U 64 = D n g 5 = D n − 1 g 6 (25)
The recurrence relation of the rest of the elements of U is zero.
2) The recurrence formulas for D W i j
D n W 1 = D n W 2 = D n W 3 = 0 (26)
D W 2 = β 2
β 2 = [ ( ω 3 g 3 ) 2 + ( ω 3 g 5 ) 2 ] D g 1 − 2 ω 3 g 3 D g 6 + 2 ω 3 g 5 D g 4 − ω ˙ 3 g 3 D g 5 + ω ˙ 3 g 5 D g 3 = [ ( ω 3 g 3 ) 2 + ( ω 3 g 5 ) 2 ] g 2 − 2 ω 3 g 3 D g 6 + 2 ω 3 g 5 D g 4 − ω ˙ 3 g 3 g 6 + ω ˙ 3 g 5 g 4
D W 4 = β 4
β 4 = − ω 3 g 3 ω 3 g 5 D g 5 + ( ω 3 g 5 ) 2 D g 3 − 2 ω 3 g 5 D g 2 − ω ˙ 3 g 5 D g 1 = − ω 3 g 3 ω 3 g 5 g 6 + ( ω 3 g 5 ) 2 g 4 − 2 ω 3 g 5 D g 2 − ω ˙ 3 g 5 g 2
D W 6 = β 6
β 6 = − ω 3 g 3 ω 3 g 5 D g 3 + ( ω 3 g 3 ) 2 D g 5 + 2 ω 3 g 3 D g 2 + ω ˙ 3 g 3 D g 1
The higher powers of D W 2 , D W 4 and D W 6 are computed as
D n W 2 = D n − 1 β 2 , D n W 4 = D n − 1 β 4 and D n W 6 = D n − 1 β 6 (27)
The recurrence relation of the rest of the elements of W is zero.
Step III: Find the Lie-series solution
The solution is given by:
X ¯ ( g 1 , g 2 , g 3 , g 4 , g 5 , g 6 ) = [ { exp [ ( t − t o ) D ] } X ] X ¯ = X ¯ o = ∑ k = 0 ∞ [ D k X ¯ ] X ¯ o τ k k !
Then
g i = ∑ k = 0 ∞ τ k k ! [ D k g i ] g i = g o = [ g i ] g o + ∑ k = 1 ∞ τ k k ! [ D k − 1 ( P i + W i ) ] g i = g o = [ g i ] g i = g o + τ [ P i ] g i = g o + τ [ W i ] g i = g o + ∑ k = 2 ∞ τ k k ! [ D k − 1 W i ] g i = g o + ∑ k = 2 ∞ ∑ j = 1 6 ∑ m = 0 n τ k k ! [ D k − m − 1 G j i ] [ D m U i j ] g i = g o (28)
Consider a circular restricted three-body problem in the Earth-Moon system where the period of the Moon about the Earth is 27.23 days. The parameters of the primaries, the properties of the chief orbit and the initial condition of the deputy in the LVLH frame are tabulated in
The calculations have been made using 10−2 step size and five terms calculations over 24 hours. To assess the motion of the deputy employing the concepts of the restricted three-body problem, the solution algorithm is applied for both the unperturbed (classical) and the perturbed cases. A set of curves represent the relative motion of the deputy satellite w.r.t. the chief in both cases.
As is clear in Figures 3-5, in all panes of motion there is a significant difference in the components of the deputy position between the classical and perturbed cases. Consequently, the magnitude of its relative position vector is changed as is clear in
| The primary | Parameters | Values |
|---|---|---|
| Earth | j 2 ( 1 ) | 1.083 × 10−3 |
| R 1 | 6.357 × 103 km | |
| q 1 | 0.8 | |
| Moon | μ | 0.0121534 |
| j 2 ( 2 ) | 202.7 × 10−6 | |
| R 2 | 1.738 × 103 km | |
| q 2 | 0.45 |
| Satellite | Parameters | Values |
|---|---|---|
| Chief | Semi-major axis (km) | 8600 |
| Eccentricity | 0.00011476 | |
| Mean motion (rad) | 0.00266 | |
| Mean anomaly (degree) | 275.8850 | |
| Deputy | Relative position ( x o , y o , z o ) ( km ) | ( 0.710 , 210.135 , − 2.2151 ) |
| Relative velocity ( x ˙ o , y ˙ o , z ˙ o ) ( km / h ) | ( 2.264 , − 3.521 , 5.527 ) | |
| Latitude “δ” | 30˚ |
The motion of the deputy satellite w.r.t. the chief of the formation is modelled in frame of the perturbed restricted three-body problem using the LVLH frame. The model is a more accurate formulation compared with the previous work where it considers the effects of radiation of both primaries in addition to their oblateness second zonal harmonic. The system is simplified assuming the circular problem and solved numerically using the Lie series approach. The solution is tested using suitable initial conditions and applied for both the classical and perturbed restricted three-body problems. By comparing the two cases, the results show that there is a significant difference in the deputy relative distance. Consequently, the model will be suitable for a more accurate study of the different space mission issues (e.g. satellite rendezvous and control).
The author declares no conflicts of interest regarding the publication of this paper.
Khalifa, N.S. (2022) Dynamics of Satellite Formation Utilizing the Perturbed Restricted Three-Body Problem. Journal of Applied Mathematics and Physics, 10, 219-235. https://doi.org/10.4236/jamp.2022.101016