https://ntrs.nasa.gov/search.jsp?

R=19690027552 2018-08-11T17:32:30+00:00Z

NASA TECHNICAL NOTE

00

m
Lm
I
n
z

LOAN COPY: RETURN TO

AFWL I W a - 2 )
)(\RTLAND AFB, N MkX

A UNIFIED FORM

OF LAMBERT’S THEOREM

by E. R. Lancaster n~2dR. C. Blamhard

Goddaid Space Flight Center
Greenbelt, M d ,

N A T I O N A L AERONAUTICS AND SPACE A D M I N I S T R A T I O N WASHINGTON, D. C. SEPTEMBER 1 9 6 9

/
i
 TECH LIBRARY KAFB, NM

111l1l1ll11lll
llI1
1l1l1ll1lI1l111110132294 l

._
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.
NASA T N D-5368 ~

4. T i t l e and Subtitle 5. Report Date

A Unified F o r m of Lambert's Theorem September 196 9
6. Performing Organization Code
-
7. Author(s) 8. Performing Organization Report
E. R. Lancaster and R. C. Blanchard
9. Performing Orgonizotion Name and Address 10. Work U n i t No.

Goddard Space Flight Center 11. Contract or Grant No.

Greenbelt, Maryland 20771
.~
13. Type of Report and Peiiod Cove
2. Sponsoring Agency Name and Address
Technical Note
National Aeronautics and Space Administration __
Washington, D. C. 20546 14. Sponsoring Agency Code

5. Supplementary Notes

6. Abstract
A unified form of Lambert's theorem is presented which is valid for elliptic,
hyperbolic, and parabolic orbits. The key idea involves the selection of an inde­
pendent variable x and a parameter q such that the normalized time of flight T is :
single-valued function of x for each value of q . The parameter q depends only upc
known quantities. For l e s s than one revolution, T is a monotonic function of x for
each q , making possible the construction of a simple algorithm for finding x, givei
T and q . Detailed sketches are given for T( x. q ) and formulas developed for the
velocity vectors at the initial and final times. Also included is a careful derivatic
of the classical form of Lambert's equations, including the multirevolution cases.

7. Key Words Suggested by Author 18. Distribution Statement

Lambert's theorem
orbit determination Unclassified- Unlimited
rendezvous
targeting
9. Security Clossif. (of t h i s report) Classi(. (of t h i s page) 21. No. of Pages 22. Price

Unclassified Unclassified 14 $3.00

*For sale by the Clearinghouse for Federal Scientific and Technical Information
Springfield, Virginia 22151
I

CONTENTS

INTRODUCTION ......................... 1

THE PROBLEM ......................... 1

THE CLASSICAL FORM O F LAMBERT'S

EQUATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

A UNIFIED FORM O F LAMBERT'S EQUATIONS . . . 7

AUXILIARY FORMULAS . . . . . . . . . . . . . . . . . . . . 9

References ............................. 13

iii
 by

E. R. Lancaster and R. C. Blanchard

Goddard Space Flight Center

INTRODUCTION

Lambert's problem, as it a r i s e s in most applications, is concerned with the determination of

an orbit from two position vectors and the time of flight. It has important applications in the a r e a s
of rendezvous, targeting, and preliminary orbit determination. In this paper a unified form of
Lambert's theorem will be presented which is valid for elliptic, hyperbolic, and parabolic orbits.

The key idea involves the selection of an independent variable x and a parameter q such that
the normalized time of flight T is a single-valued function of x for each value of 9. The parameter
q depends only upon known quantities. The problem then is to find x for given values of q and the
time of flight. For l e s s than one revolution, T is a monotonic function of x for each value of q . Thus
it is an easy task to design an algorithm for finding X . For multirevolution cases, T( X) has a single
minimum for each q .

This idea was presented in a previous paper (Reference l), where a unified formula was given
f o r the computation of T from x and q . Detailed sketches were given for T ( x , q ) , and a simple
formula developed for the time derivative of the magnitude of the radius vector at the initial time
in t e r m s of x and given quantities.

The present paper expands upon the previous one by:

1. giving complete derivations which were only sketched before;

2. giving a careful derivation of the classical form of Lambert's equations, including the
multirevolution cases;
3. deriving a number of useful auxiliary formulas, e.g., for the semilatus rectum and for the
velocity vectors at the initial and final times.

THE PROBLEM

Suppose a particle in a gravitational inverse-square central force field has distances r l and
r2 ,from the center of attraction at times t and t 2 . Let c be the distance and e the central angle
between the positions of the particle a t the two times, where 0 5 e 5 2 ~ .

, ,,.,,..,.. ,,, ,.. ....- ...-..., . ..__..-.

- L

Lambert’s problem is that of finding the semimajor axis or some related quantity for the orbit
of the particle, given t, , r l , t 2 , r2, and 8 . When Lambert’s problem has been solved, other quanti­
ties associated with the orbit a r e easily found, as will be later discussed. Using the law of cosines,
we can express c in t e r m s of r l , r 2 , and 8:

We define

G = universal gravitational constant,

M = mass of attracting body,

P = GM,

a = semimajor axis of transfer orbit,

e = eccentricity of transfer orbit.

We will follow the common sign convention for a, i.e., a > 0 for elliptic orbits and a< 0 for
hyperbolic orbits.

Definitions of other symbols will be given as they a r e introduced.

THE CLASSICAL FORM OF LAMBERT’S EQUATIONS

The path of a particle in an inverse-square central force field is an ellipse, parabola, or

hyperbola. With origin at the center of attraction, we have, for elliptic motion,

ri = a ( l - e c 0 s 4 ~ ),

n(t2-tp) = 6, - esin&2 , (5)

where 9, and b2 a r e the eccentric anomalies a t times tl and t , , tp is the time at pericenter, and

n = (,u/a3)1’2 .

2
 If 2, is a unit vector pointing towards periapsis and 2, is a unit vector in the plane of motion
90" ahead of Tl in the dil'ection of motion, then for the position vectors and 2; at times t l and t ,

= a(cosc#J1- e);, + a ( 1 -e'"" (sin+,)Z, ,

r,
+ = a(cos+,-e)Zl + a(l-e')V'(sin+,)Z, .

Substituting these equations in

c2 = r t -t r: - 27, * 7, ,

where the dot indicates scalar product, we have

c2 = a2 (cos +2 - cos + 1 ) 2 -t a2 (1 - e') ( s i n + ' - s i n 4,)'

= 4a2 [I- e2 c o s 2
1
2 (b1 + +, )] s i n 2 21 (+2 - 9,) .

Adding Equations 2 and 3,

1 1
rl + r, = 2a[l - e cos 3 (+1 + + 2 ) c o s2 (+2 -+1)] .

Subtracting Equation 4 from Equation 5,

Equations 6, 7, and 8 determine the three unknowns a , 92- and e COS (1/2) (+1 + 4,). Let

where m is the number of complete circuits made by the particle between times t l and t , .

Equations 6 , 7, and 8 become

 The two inequalities for Equations 9 and 10 a r e
P geometrically equivalent to the shaded region of
Figure 1, from which it is evident that 0 a < 2 n
and -n(P <n. We can also obtain 0 5 a < 271by adding
the inequalities for Equations 9 and 10; and, i f we
add P - a to each part of the inequality for Equation 9
and divide the result by 2, we obtain - ( a - P ) / 2 5 p < T
+ ( a - P ) / 2 , or -n(p<n.

With appropriate trigonometric identities, Equa­

tions 11and 12 become

cosp - cosu = c/a ,

c o s b + cosa 2 - ( r l +r2)/a .

(0, -2lT Solving these two equations,

I cosu = 1 - s/a = 1 + 2E ,

Figure 1. cosp = 1 + 2KE ,

where we have defined

E = - s/2a ,

K 1 - c/s.

Since c o s a = 1- 2 s i n ’ ( a / 2 ) , Equation 14 can be changed to

E = - sin’(a/2) , 0 5 a < 2n.

..
Similarly, Equation 15 becomes

KE = - s i n 2 (P/z) , - 7 7 5 p < v ,

4
 Introducing Equation 1, we have

K = ( r l r2/2s2) (1 +cos S )

= (r, r2/s*)cos2 ( ~ / 2 ) .

Substituting Equation 16 in Equation 17,

s i n (p/2) = q s i n (a/2) , -7rsp<7r,

where

9 = * 7% = [ ( r , r*)1/2/5] cos (8/2)

Note that the sign of q is taken care of by the angle 0:

We can introduce E into Equation 13, since

“(t, - t,) = ( w / z I ~ ) ”(~t , - t , ) = (-E)3’2T ~

where

T = ( ~ w / s ) ” (~t , - t l ) / s .

This can also be written as

T = (-E)-3’2 [2m7r+a-p-(sina-sin/J)] .

Substituting Equation 16 into Equation 22,

TsinJ(a/2) = 2m + a - p - sina + sinp

 Equations 18 and 23 with 0 5 a < 27r a r e Lambert's equations for elliptic motion. Given T and
9, they arp to be salved for a and P, after which i t is a simple matter to find all other quantities
a$soc/ated with the orbjt.

It is cpstomary in the literature (e.g., Reference 2) to consider T as a function of E (or a) and

break the ellintic case of Lambert's theorem into four cases, depending upon the sign of q and
whether a/q is taken in the first o r second quadrant in Equation 16. The choice of E as the indp­
pendent variable makes T a double-valued function. This problem can be avoided by choosing a as
the independent variable. However, an even better choice willbe discussed in the next section.

By a derivation very similar to that for the elliptic case, one finds for the hyperbolic case

E = sinh2(y/2) , (25)

When m = 0, Equations 22 and 24 break down for E = 0 and suffer from a critical loss of sig­
pificant dig-@ in the neighborhood of E = 0. To remedy this, Equation 22 is written in the form

T = a(-E) - qKa(-KE) , (27)

where

Replacing arcsin u112 and (1 - u ) l 1 2 by s e r i e s (Reference 3) with 0 5 a < 7r,

A similar proaedure produces the same s e r i e s for the hyperbolic case. For the parabolic case
we have E = 0, in which case the s e r i e s gives

Thus with m = 0 we have a s e r i e s which is valid for elliptic, hyperbolic, and parabolic transfer
provided ]El < 1 (with 0 5 a < 7r for the elliptic case),

6
A UNIFIED FORM OF LAMBERT’S EQUATIONS

As mentioned in the previous section, T is a single-valued function of a. However, a better-

behaved function is obtained if we choose as the independent variable

x = cos ( d 2 ) , - 1 5 x 5 1 ,

We then have, for elliptic, hyperbolic, and parabolic transfer,

i f for parabolic transfer, we let x = 1.

For the elliptic case, let

y = sin (a/2) = (-~)‘/2 ,

7. = cos ( P / 2 ) = (1+KE)”2 ,

h = 3 (sina- sinp) = y(x-qz) ,

A = arctan(f/g) , O 5 A l n .

It then follows from Equation 21 for the elliptic case that

T = 2 ( m t A - h)/y3

For the hyperbolic case, let

f = sinh2(y-S) = y(z-qx),

1
g = cosh 3 ( y - 8 ) = xz - qE .

Note that O’Y -6 < m, since 0’ f < a. Let

1
h = 2 (sinhy-sinh6) = y(x-qz) .

It follows that

1 ( y - 8) = a r c t a n h ( f/g) ,

Thus, for the hyperbolic case,

T = 2[h - i n ( f + g)]/y3

It is now apparent that, given q and x, the following steps produce T for all cases:

1. K = q2

2. E = x z - 1

3. If x is near 1, compute T from Equation 27

4. Y = lE11/2

5. z = (1 ~ K E ) ” ~

6. f = ~(z-qx)

7. g = xz - qE

 9. T = 2 ( x - qz -d/y)/E.

The following formula for the derivative holds for all cases except for x = 0 with K = 1, and
for x = 1:

dT/dx = ( 4 - 4qKx/z - 3xT)/E .

If x is near 1, differentiate Equation 27 to obtain

dT/dx = 2x[qK2 c' (-KE) -0' (-E)] ,

~ ' ( u )= du/du = znanun-'

"= 1

The derivative in the case of x = 0 with k = 1will be discussed in the next section.

AUXILIARY FORMULAS

In this section we will show how to obtain a number of useful quantities associated with the
two-body orbit, assuming Lambert's problem has been solved for x.

In the derivation of Lambert's equation for the elliptic case, J. and .A+ a r e defined in such a way
that

4, - .p, a - ,A t 2m = 2(X + m ) .

From Equation 22, the eccentric anomaly difference can also be written in the form

d', - = (-E)3'2T + sina - sinp

= y 3 T + 2y(x-qz)

From Equation 28 we have

 4,.- - sin (4,- = y3T - 4 y 3 q ( z - 9x1 , (2 9)

1 - cos(q52-q51) = 2yZ(z-qx)Z . (3 0)

We now obtain a formula for the time derivative i at time t 1. Kepler’s equation in the elliptic
case can be written in the form (Reference 4)

Substituting l / a = 2 y 2 / s , t, - t, = s 3 / ’ T/(~P)”* , and making use of Equations 29 and 3 0 gives

( 2 / p ~ ) ~ ” ( z - q x )r l ;, = 2q - ( 2 r l / s ) ( x z - q E ) .

Multiplying through by z + qx, we have, since ( z - qx) ( z + qx) = 1 - K = c/s and ( z + qx) ( x z - qE)
x+qz,

( ~ / , L L S ) ~ / ~ =C ~2 q, s; (~z + q x ) - 2rl(x+qz)

= 2 q z ( s - r l ) + 2x(Ks-rl) ;

Ks - rl = (l-c/s)s - rl s - c - rl = r, - s

Thus we have finally that

At time t , we find

By a similar procedure we can show that Equations 3 1 and 3 2 hold also for the hyperbolic and
parabolic cases.

Having X, we can find the semimajor axis a, o r its reciprocal,

10

 We know that

where r is the magnitude of the position vector, and 4 is the eccentric anomaly, at time t ,

Thus for the eccentricity we have

For the semilatus rectum P we have

and for the value of r at the point of closest approach to the center of attraction we have

For the speed we have

“ 2 = & -a, .
For the component of velocity perpendicular to a radius vector,

= (pp)”‘/r

The above formulas hold for any type of two-body motion. Either of the subscripts 1 o r 2 can
be placed on r , i , and V.

If i, < 0 and i, > 0, or if i, and i, have the same sign with B > n , periapsis passage will occur
,
between times t and t 2 , in which case it may be of interest to compute r p .

,
If the plane of motion is known, the velocity 5 at either time t or time t can be written in
any convenient coordinate system, since the components i and v Q a r e known. The plane of motion
can be found from the position vectors ,
and 2’ at times t and t ’, provided they are not parallel.
Since this is a common case, we will express 2 , and i;, (the velocities at t l and t,) in t e r m s of y 1
and 2‘.

11

 For the velocity G1 we have

where Grl is along T1 and Gel is in the plane of motion perpendicular to 1

; and in the direction of
motion, Le., in the direction of increasing true anomaly. We have

where c1 and c, a r e to be determined.

-r l . v-O 1 -
­ o = c 1 r: + c, +
rl . *
r,

- .- -
r~ - r , v O 1s i n @ =
+
c1 r l e
+
r2 + c2 rg

r l c 1 + ( r , cos B)C, = o ,

( rl c o s @ )c1 + r 2 c, = ve s i n 0 .

Solving for c1 and c2, we have

In a similar way, we find

Figure 2 shows T as a function of X . Note

the discontinuity in the slope for x = 0 with
K = 1. ForK = 1, z = 1x1. Thus w e a r e l e d
to consider four cases: q = 51 with x 2 0,
q = r t l with x 5 0. Examination of the formulas
for dT/dx in these cases reveals that i f q = 1
we have a left-hand derivative of -8 and a right-
hand derivative of 0 at x = 0. If q = -1 we
have a left-hand derivative of 0 and a right-
Figure 2. hand derivative of -8 at x = 0.

12

 With further analysis we find that the c a s e s (q = 1, x 5 0) and (q = -1, x 2 0) represent recti­
linear orbits.
For m = 0, T is a monotone function of x, making possible a simple numerical procedure for
solving Lambert's problem. Figure 2 is for the elliptic case, Figure 3 for the hyperbolic case,
the parabolic case occurring at x = 1 in both figures. Figure 4 shows a small region of Figure 2

e.
",I
9

0 I

1 1.2 '
i 1.4 1.6 1.8 2
X

Figure 3.

where d 2 T/dx2 is negative. If the Newton-

Raphson method is being used to find x , a
switch should be made in this region to the s e ­
cant (regula falsi) method.

No solutions of Lambert's problem exist +

in the shaded regions of Figures 2 and 3.
x = 1 (m > 0) and x = -1 a r e vertical asymp­
-
totes. T 0 as X - - J J .

Goddard Space Flight Center -0.20 -0.15 -0.10 -0.05 0 0.05 0.10 0.15 0.20
National Aeronautics and Space Administration X
Greenbelt, Maryland, May 28, 1969
188-48-01-05-5 1 Figure 4.

REFERENCES

1. Lancaster, E. R., Blanchard, R. C., and Devaney, R. A., "A Note on Lambert's Theorem,"
7
Joumal of Spacecvaft and Rockets 3 (9):1436-1438, September 1966.
2. Breakwell, John V., Gillespie, Rollin W., and Ross, Stanley, "Researches in Interplanetary
Transfer,'' ARS J. 31:201-208, February 1961.
3. Gedeon, G. S., "Lambertian Mechanics,'' Proceedings of the X I t h Intevnntio?zal Astvonautical
Congress (Springer-Verlag, Wien, and Academic P r e s s Inc., New York, 1963).
4. Battin, Richard H., "Astronautical Guidance," New York: McGraw-Hill Book Co., Inc., 1964,
page 46.

13
NASA-Langley, 1969 - 30
NATIONAL AND SPACE ADMINISTRATION
AERONAUTICS
D. C. 20546
WASHINGTON,
OFFICIAL BUSINESS FIRST CLASS MAIL

POSTAGE A N D FEES PAID

NATIONAL AERONAUTICS A N D
SPACE ADMINISTRATION

If Undelivtrdble (Section 156

Postal Manual) Do Not Return

“The aerotzrutticnl niad space nrtii’ities of ihe United Stntes shnll be

. coizditcted so ns to coiztribute . . . t o the expnnsion of hitiiiaia kia01~1­
edge of pheuotiieizn iia the ntiiiosphere nizd spnce. The Adiiiiiaistratioiz
shrill proi fide f o r the widest prncticnble ntzd nppropsinte disseiiiim!io?z
. of iiiforiiintioiz coiicerizing its rrctiimitirs nitd the resnlts thereof.”
-NATIONALAERONAUTICS
A N D SPACE ACT OF 1358

NASA SCIENTIFIC AND TECHNICAL PUBLICATIONS

TECHNICAL REPORTS: Scientific and TECHNICAL TRANSLATIONS: Information
technical information considered important, published in a foreign language considered
complete, and a lasting contribution to existing to merit NASA distribution in English.
knowledge.
SPECIAL PUBLICATIONS: Information
TECHNICAL NOTES: Inforrnation less broad derived from or of value to NASA activicies.
in scope but nevertheless of importance as a Publications include conference proceedings,
contribution to existing knowledge. monographs, data compilations, handbooks,
sourcebooks, and special bibliographies.
TECHNICAL MEMORANDUMS:
Information receiving limited distribution TECHNOLOGY UTILIZATION
because of preliminary data, security classifica­ PUBLICATIONS: Information on technology
tion, or other reasons. used by NASA that may be of particular
interest i n cornrnercial and other non-aerospace
CONTRACTOR REPORTS: Scientific and npplications. Publications include Tech Briefs,
technical information generated under a NASA .,-echnctlogy Utilization Reports and Notes,
contract or grant and considered an important and Technology Surveys.
contribution to existing knowledge.

Details on the availability of these publications may be obtained from:

SCIENTIFIC AND TECHNICAL INFORMATION DIVISION

NATI 0 NA L AERO N AUT1C S AND SPACE ADM I NISTRATI0 N
Washington, D.C. 20546