The Six Constants of Orbital Motion: From Initial Conditions to Orbital Elements

1 Where GPSORB02 fits in the derivation chain

GPSORB01 derived the ideal Earth-satellite relative equation

|------μ---| |¨r = − -3r | -------r---|

(1)

from Newton’s second law and universal gravitation. Written in Cartesian coordinates,

ẍ

= −

, ÿ

= −

, z

= −

(2)

Kaplan’s discussion then leads toward six orbital elements. The natural question is:

Why does Eq. (1), which looks like one vector equation, require six independent constants to determine one satellite trajectory?

The short answer is that Eq. (1) contains three coupled second-order scalar equations. Each second-order degree of freedom contributes two initial data: position and velocity. The purpose of this article is to make that counting rigorous and then connect it to orbital geometry.

2 Begin in one dimension: why second order means two constants

Consider the simplest second-order differential equation,

¨x = 0.

(3)

Integrating once gives

˙x = C , 1

(4)

and integrating again gives

------------------ | | -x(t) =-C1t-+-C2.-

(5)

Two arbitrary constants appear because the equation is second order. If the position and velocity are specified at t = t₀,

x(t0) = x0, x˙(t0) = v0,

(6)

then

C1 = v0, C2 = x0 − v0t0.

(7)

Thus the two constants may be represented either as abstract integration constants (C₁,C₂) or as physically meaningful initial data (x₀,v₀).

A harmonic oscillator makes the same point in a less trivial way:

2 ¨x + ω nx = 0.

(8)

Its general solution is

|----------------------------| |x(t) = A cosωnt + B sinωnt. | -----------------------------

(9)

Again there are two constants. One may instead use amplitude and phase,

x(t) = C cos(ωnt − ϕ),

(10)

which shows an important principle for orbital mechanics:

The number of independent constants is fixed by the differential equation, but the coordinates used to represent those constants are not unique.

The pair (A,B) and the pair (C,ϕ) describe the same two-dimensional family of solutions.

3 Three dimensions: the two-body problem has a six-dimensional state

Introduce the velocity

v = r˙

(11)

and define the state vector

|----[-]---------------------------| x = r = [x y z v v v ]T . | v x y z | ------------------------------------

(12)

Equation (1) is then equivalent to the first-order system

|-----------[--------]--| | v | x˙= F(x) = − μr ∕r3 .| -------------------------

(13)

In components,

	= v_x,		= v_y,		= v_z,	(14)
	= −,		= −,		= −.	(15)

A unique solution is selected by specifying all six state components at one epoch t₀:

|------------[---]--| | r0 | x0 = x (t0) = v . | ----------------0----

(16)

Thus the initial condition consists of

|------------------------| |x0, y0, z0, vx0, vy0, vz0. -------------------------

(17)

These are six independent scalar quantities.

For r≠0, the right-hand side of Eq. (13) is smooth, so the usual existence-and-uniqueness theorem for ordinary differential equations guarantees a locally unique trajectory through a specified state. Symbolically one may write the resulting flow as

|------------------| |x(t) = Φ(t,t ;x ).| -------------0--0---

(18)

The solution family therefore has six parameters because x₀ belongs to a six-dimensional state space.

Figure 1. The two-body equation is a six-state first-order system. Six scalar initial values select one trajectory from the six-parameter family of possible solutions.

4 What “six constants of integration” really means

For a general coupled nonlinear system such as Eq. (13), one should not imagine literally integrating the x, y, and z equations twice independently. The components are coupled through

r = (x2 + y2 + z2)1∕2.

(19)

The phrase “six constants of integration” is better understood structurally:

1.: the system has six first-order state equations;
2.: a general local solution therefore depends on six independent parameters;
3.: specifying six independent initial values fixes those parameters;
4.: any other nonsingular set of six coordinates may be used instead.

We may denote a general solution abstractly by

x(t;C1,C2, C3,C4, C5,C6 ).

(20)

The constants C_i need not look like Cartesian coordinates. They may be transformed into quantities with much more physical meaning.

This is precisely what orbital elements do.

5 Constants of integration are not the same as constants of motion

Two phrases sound similar but should be distinguished.

A constant of integration is an arbitrary parameter required to identify one member of a family of differential-equation solutions. Initial position and initial velocity are examples.

A constant of motion, also called a first integral or invariant, is a function of the state that remains constant along a trajectory. If

I = I (x, t),

(21)

then along the motion

dI ∂I --- = ---+ ∇xI ⋅x˙. dt ∂t

(22)

|--------| |dI | |---= 0,| -dt-------

(23)

then I is constant along that solution.

The distinction matters because one can have several components of conserved vectors without all of them being independent. Conversely, not every constant needed to specify the time history must appear as a familiar conserved scalar.

6 The special constants of the Kepler problem

The inverse-square central-force problem has unusually rich structure. Later articles derive the conservation laws in detail; here we introduce them to understand the counting.

6.1 Specific angular momentum

Define

|----------| h--=-r ×-v.-

(24)

For the central two-body force, h is constant. Its direction is normal to the orbital plane, so two directional degrees of freedom orient that plane, while its magnitude contains information about the orbit’s transverse motion.

Although h has three Cartesian components,

[ ] h = hx hy hz T ,

(25)

it is one conserved vector, not three unrelated conservation laws.

6.2 Specific mechanical energy

Define

|------------| | v2 μ | |ℰ = ---− --.| ------2----r--

(26)

For the ideal two-body problem, ℰ is constant. For an ellipse it later gives the semimajor axis through

μ ℰ = − ---. 2a

(27)

6.3 Eccentricity vector

Define

|----------------| | v × h r | |e = ------ − --.| --------μ-----r--

(28)

For the inverse-square two-body problem, e is constant. Its magnitude

e = ∥e∥

(29)

is the orbital eccentricity, and its direction points toward periapsis.

At first glance we now seem to have

3 (h ) + 3 (e) + 1 (ℰ ) = 7

(30)

constants. But the original differential equation only requires six parameters, and even those six include orbital phase. Something is clearly being overcounted.

7 Why the conserved quantities are constrained

The conserved quantities in the Kepler problem are not all independent.

7.1 First constraint: h is perpendicular to e

Take the dot product of Eqs. (24) and (28):

h ⋅ e

h ⋅ (v × h) − h ⋅ r-
r

(31)

The first term vanishes because v × h is perpendicular to h. The second vanishes because

h = r × v

(32)

is perpendicular to r. Hence

|----------| -h-⋅ e-=-0.|

(33)

Thus the six components of h and e satisfy at least one algebraic constraint.

7.2 Second relation: energy is determined by h and e

Square the eccentricity-vector definition:

| |2 2 ||v-×-h- r|| e = | μ − r| .

(34)

Expanding,

|v × h |2 2 r e2 = ----2--- + 1 − --(v × h) ⋅-. μ μ r

(35)

Because v ⋅ h = 0,

|v × h|2 = v2h2.

(36)

Next use the vector triple-product identity

v × (r × v ) = rv2 − v(r ⋅ v).

(37)

Dotting with r gives

(v × h) ⋅ r	= r²v² − (r ⋅ v)²,	(38)
	= \|r × v\|²,	(39)
	= h².	(40)

Therefore

2 (v × h ) ⋅ r-= h-. r r

(41)

Substituting Eqs. (36) and (41) into Eq. (35),

e²	= + 1 −,	(42)
	= 1 + .	(43)

Using Eq. (26),

v2 − 2μ-= 2 ℰ, r

(44)

|--------------| | 2ℰ h2 | e2 = 1 + -----.| -----------μ2---

(45)

Equivalently,

|----------------| | μ2(e2 − 1 )| |ℰ = -----2----.| ---------2h------

(46)

Thus energy is not independent once the magnitudes of h and e are known.

Figure 2. The conserved angular-momentum vector, eccentricity vector, and energy contain redundant information. Their algebraic relations leave five independent constants describing the fixed Keplerian conic. A sixth phase constant is still required for the complete time-dependent solution.

8 Five constants determine the orbit curve, but not the satellite’s location

This is the central conceptual point of GPSORB02.

The fixed Keplerian conic in three-dimensional space requires five independent geometric parameters:

1.: one parameter for size;
2.: one for shape;
3.: two to orient the orbital plane;
4.: one to orient periapsis within that plane.

For an ellipse these are naturally represented by

|--------------------| |a, e, i, Ω, ω. | ---------------------

(47)

But these five parameters specify only the curve. They do not tell us where the satellite is on the curve at a particular time.

Imagine two satellites occupying different points on exactly the same ideal Keplerian ellipse with the same direction of motion. They can have the same a,e,i, Ω, and ω. They also share the same h, e, and specific orbital energy. Yet their position vectors at the same clock time are different.

A sixth parameter is therefore required to specify orbital phase.

Figure 3. Five constants define the oriented Keplerian conic. One additional phase constant specifies where the satellite is on that conic at a chosen epoch.

9 The sixth constant: phase or epoch information

Several equivalent quantities can supply the missing phase information. Common choices include

ν = true anomaly at t , 0 0

(48)

E0 = eccentric anomaly at t0,

(49)

M0 = mean anomaly at t0,

(50)

or the time of periapsis passage

τ = epoch at which the body passes periapsis.

(51)

These are not four additional degrees of freedom. For a known Keplerian orbit they are alternative representations of the same one-dimensional phase information.

For an ellipse, later articles derive

M = E − e sin E

(52)

and

M (t) = M0 + n(t − t0),

(53)

where

∘ -μ- n = --. a3

(54)

Thus M₀ is especially convenient because its ideal two-body propagation is linear in time.

Another common representation is

M (t) = n (t − τ),

(55)

so that

|----------------| |M0 = n(t0 − τ).| -----------------

(56)

The constant M₀ and the constant τ therefore encode the same phase degree of freedom once n and t₀ are fixed.

10 The six classical orbital elements

For a nondegenerate elliptical orbit, one standard six-element set is

|----------------| (a,e,i,Ω, ω,M0 ).| ------------------

(57)

Their roles can be organized as follows:

Element	Role	Physical meaning
a	size	Semimajor axis; sets orbital energy and characteristic period.
e	shape	Eccentricity; distinguishes circular and elliptical shape within the bound case.
i	plane orientation	Inclination of the orbital plane relative to the reference plane.
Ω	plane orientation	longitude or right ascension of the ascending-node direction, depending on the chosen reference frame.
ω	in-plane orientation	Argument of periapsis; rotates the ellipse within its orbital plane.
M₀	phase	Mean anomaly at the chosen reference epoch t₀; locates the satellite in time along the orbit.

The decomposition is therefore

|------------------------------------------------| |2 (size/shape ) + 3 (orientation ) + 1 (phase ) = 6. -------------------------------------------------

(58)

This is the geometric meaning behind the six constants of the two-body solution.

11 Cartesian state and orbital elements are two coordinate systems for one state

At an epoch t₀, the Cartesian description is

[ ] x0 = r0 ∈ ℝ6. v0

(59)

The classical-element description is

[ ] α0 = a e i Ω ω M0 T .

(60)

Away from classical-element singularities, there is a reversible transformation

|-----------| x0-←-→--α0.-|

(61)

The state has not acquired or lost information. We have merely changed coordinates in the six-dimensional space of initial conditions.

Figure 4. Cartesian state components and classical orbital elements are alternative six-parameter descriptions of the same physical state, provided the orbital-element representation is nonsingular.

This viewpoint is valuable because it removes some of the mystery from orbital elements. They are not six new physical laws. They are a coordinate transformation chosen because the resulting numbers expose orbit geometry and make Keplerian propagation convenient.

12 How the state begins to reveal the elements

The complete state-to-element derivation is reserved for GPSORB05 and GPSORB07, but the structure can already be previewed.

Given r and v, compute

h = r × v.

(62)

The direction of h establishes the orbital plane and later yields i and Ω.

Compute

v2 μ ℰ = ---− --. 2 r

(63)

For an ellipse,

|----------| | -μ- | |a = − 2ℰ .| -----------

(64)

Compute the eccentricity vector

v × h r e = ------ − --. μ r

(65)

Then

|--------| |e = ∥e∥ | ---------

(66)

and the direction of e identifies periapsis, allowing ω to be determined after the node direction is known.

Finally, the angle or anomaly locating r relative to periapsis supplies the phase. Thus the six Cartesian numbers are systematically reorganized into the six orbital-element numbers.

13 Why the six elements are constant only in the ideal two-body problem

For Eq. (1), the classical Keplerian elements may be treated as constants except for the anomaly that advances with time. More precisely, one may choose a constant phase-at-epoch such as M₀ and use time to propagate the instantaneous anomaly.

For a real satellite, however,

¨r = − -μr + a , r3 pert

(67)

where a_pert may contain nonspherical gravity, third-body gravity, solar radiation pressure, maneuvers, and other effects. The orbital elements then become time-varying quantities,

a(t), e(t), i(t), Ω (t), ω(t), M (t).

(68)

One may interpret them as osculating elements: at each instant they describe the Keplerian conic tangent to the real trajectory in state space.

This distinction is essential for GPS. The broadcast ephemeris uses parameters that are Keplerian in appearance but are augmented by secular rates and periodic corrections so that a compact user algorithm can reproduce the satellite’s fitted Earth-fixed trajectory over the intended interval.

14 The bridge to the GPS broadcast ephemeris

The six classical quantities explain the skeleton of the legacy GPS broadcast parameterization. In the idealized limit, the following correspondences are suggestive:

√ -- √ -- A ← → a, e ← → e, i ← → i, 0 Ω0 ← → node-orientation parameter, ω ← → ω, M0 ← → M0.

(69)

The word “correspondence” is deliberate. The broadcast quantities should not be interpreted as six untouched constants of an exact inertial Kepler ellipse. In the legacy LNAV user algorithm, for example, Ω₀ is defined relative to the GPS weekly epoch and is combined with Earth rotation and the transmitted node rate. The broadcast values are fitting parameters for a specified user model.

The additional legacy terms include

Δn, ˙Ω, IDOT,

(70)

and the harmonic correction coefficients

C , C , C , C , C , C . uc us rc rs ic is

(71)

These do not mean the underlying differential equation suddenly needs more than six initial conditions. Instead, they belong to a compact parameterization of a perturbed and fitted trajectory model. The fundamental dynamical state is still six-dimensional; the broadcast message uses additional coefficients because it approximates time-varying departures from one fixed Keplerian conic.

This distinction resolves an apparent paradox:

Six numbers are enough to specify an ideal two-body state, while more than six transmitted coefficients may be useful to approximate a real GPS orbit over time.

15 A useful dimensional argument: orbit geometry versus orbital phase

There is another way to see the 5 + 1 structure without using any formulas.

An ellipse in its own plane requires two numbers to specify size and shape. Placing that plane in three-dimensional space requires two orientation angles. Rotating the ellipse within its plane requires one additional angle. This totals five numbers for the geometric curve.

A moving satellite is not merely a curve. One must also specify a location along that curve at a reference time. That is the sixth number.

This geometric counting exactly mirrors the differential-equation counting:

|--------------------------------------------| | ◟◝5◜◞ + ◟◝1◜◞ = ◟◝6◜◞ .| |oriented conic phase at epoch state degrees of freedom| ----------------------------------------------

(72)

The agreement is not accidental. Both descriptions encode the same physical state.

16 Special cases and singular classical elements

The classical six-element set is intuitive but not globally well behaved.

16.1 Circular orbit

e = 0,

(73)

there is no unique periapsis direction. Consequently, ω is undefined. One instead uses a combined angle such as argument of latitude or true longitude, depending on the geometry.

16.2 Equatorial orbit

i = 0,

(74)

the orbital plane coincides with the reference equatorial plane. The line of nodes is not uniquely defined, so Ω is undefined.

16.3 Circular equatorial orbit

If both

e = 0, i = 0,

(75)

several classical angles lose separate meaning, although the Cartesian state itself remains perfectly well defined.

This is an important mathematical point: the physical state space remains six-dimensional, but a particular coordinate chart can become singular. Alternative element sets such as equinoctial elements are designed to avoid some of these singularities.

GPS satellites normally have small but nonzero eccentricity and substantial inclination, so the classical geometry remains very useful for understanding the broadcast equations. Still, one should not confuse a coordinate singularity with a physical singularity.

17 A second caution: not every six-number list is independent

A set containing six numerical entries is not automatically a valid six-coordinate representation. Independence matters.

For example, suppose one tried to use the six components

(hx, hy,hz,ex,ey,ez).

(76)

These six quantities satisfy

h ⋅ e = 0,

(77)

so they cannot vary independently. They describe only a five-dimensional family of Keplerian orbit geometries. A separate phase quantity must still be supplied.

Similarly, adding energy to the list does not create a new independent degree of freedom because of Eq. (45).

This is why careful orbit mechanics counts independent constants rather than merely counting symbols.

18 Worked conceptual example: same orbit, different state

Consider two satellites A and B on the same ideal Keplerian ellipse. Suppose they have the same

a,e,i,Ω,ω

(78)

and move in the same direction. Then they share the same orbital plane, semimajor axis, eccentricity, periapsis direction, specific energy, angular momentum vector, and eccentricity vector.

Let their mean anomalies at epoch differ:

M0,A ⁄= M0,B.

(79)

Then

rA(t0) ⁄= rB(t0), vA (t0) ⁄= vB (t0)

(80)

in general. The five geometry constants are identical, but the six-dimensional states are different.

This example shows precisely why the invariants that determine the conic are not enough to specify a time-tagged satellite state.

19 What will be derived next

GPSORB02 has established the dimensional and mathematical structure without yet relying on the detailed derivation of each conservation law. The next articles can now proceed in a clean sequence.

GPSORB03 will start from Eq. (1) and derive

d ---(r × v) = 0, dt

(81)

so that

h = r × v

(82)

is constant. From this, the planarity of two-body motion and Kepler’s area law will follow.

GPSORB04 will derive conservation of

2 ℰ = v--− μ- 2 r

(83)

and the eccentricity vector, leading to the conic equation

r = ----p----- 1 + ecosν

(84)

and the relations

h2 μ p = --, p = a(1 − e2), ℰ = − --. μ 2a

(85)

Those derivations will turn the abstract “six constants” discussed here into concrete orbital geometry.

20 Summary

The ideal satellite equation

μ ¨r = − -3r r

(86)

is three coupled second-order equations, equivalently six first-order equations. Therefore one unique trajectory requires six independent scalar initial values,

[ ] r0 x0 = v0 .

(87)

These six numbers may be replaced by any nonsingular six-parameter coordinate system describing the same state.

The Kepler problem supplies conserved angular momentum, eccentricity, and energy, but those quantities are not all independent:

2 h ⋅ e = 0, e2 = 1 + 2ℰ-h-. μ2

(88)

Five independent constants specify the fixed conic’s size, shape, and orientation. A sixth phase constant specifies where the satellite is on the conic at an epoch. This produces the classical structure

|----------------| -(a,e,i,Ω,ω,-M0-)-

(89)

and explains why these six quantities are a natural bridge between Newton’s equation of motion and the GPS broadcast ephemeris.

The key conceptual chain is

|----------------------------------------------------------------------------| |3 second-order equations ⇐ ⇒ 6-state first-order system | | ⇐ ⇒ 6 independent initial data | | | | ⇐ ⇒ 5 orbit-geometry constants + 1 phase constant | -----------------------------------------------------------------------------

(90)

References

[1] E. D. Kaplan and C. J. Hegarty, editors, Understanding GPS: Principles and Applications, 2nd ed., Artech House, 2006.

[2] R. R. Bate, D. D. Mueller, and J. E. White, Fundamentals of Astrodynamics, Dover Publications, 1971.

[3] D. A. Vallado, Fundamentals of Astrodynamics and Applications, 4th ed., Microcosm Press, 2013.

[4] R. H. Battin, An Introduction to the Mathematics and Methods of Astrodynamics, revised ed., AIAA Education Series, 1999.

[5] H. D. Curtis, Orbital Mechanics for Engineering Students, 4th ed., Elsevier, 2020.

[6] O. Montenbruck and E. Gill, Satellite Orbits: Models, Methods, and Applications, Springer, 2000.

[7] Global Positioning Systems Directorate, IS-GPS-200N: NAVSTAR GPS Space Segment/Navigation User Interfaces, 1 August 2022.

"The Six Constants of Orbital Motion: From Initial Conditions to Orbital Elements" is owned by bloftin.
(view preamble)