SponDored by -

American Astronautical Society (AAS)

American Institute of Aeronautics and Astronautics (AIAA)
+
..

NO.72- 931

I N T E G R A L S O F T H E MOTION F O R O P T I M A L
'TRAJECTORIES IN A T M O S P H E R I C F L I G H T

Astrodynamics conference
P A L 0 ALTO, CALIFORNIA / SEPTEMBER 11-12, 1972

First publication rights reserved by American Institute of Aeronautics and Astronautics.

1290 Avenue of the Americas, New York, N. Y. 10019. Abstracts may be published without
permission if credit is given to author and to AIAA. (Price: AIAA Member $1 50. Nonmember$2.00)
Note: This paper available at AIAA New York office for six months;
thereafter, photoprint copies are available at photocopy prices from
Technical Information Service, 750 3rd Ave., New York, N. Y. 10017
 A bs t ra i:t f o r c c i: orthogonal to i t . W e s h a l l use t h r ~u s u a l
assumption
T h o problem of optimizing the flight t r a j r c -
t o r y of a rocket vchicle moving in a r e s i s t i n g nic-
d i u m . in R g c n c r a l gravitational force f i e l d , is
c o n s i d e r e d . Optimal laws f o r t h e modulation o f
the aerodynamic and propulsivc f o r c c s a r c formu-
lated in t c r m s of the p r i m e r v e c t o r , the vector a d - w h z r r p is the a t m o s p h u r i c m a s s d e n s i t y , function
joint to the velocity. ltrlations f o r flight along a n of r , S i s a r e f c r e n c e a r e a . In hypersonic flight,
i n t e r m e d i a t c - t h r u s t a r c , and integrals of thii mo- thc: lift coefficient C L , and thc d r a g coefficicnt C,)
tion for s e v e r a l cases of p r a c t i c a l i n t e r e s t a r e a r c a s s u m e d indi!pcndent of the Mat:h n u m h e r and
derived thc Reynolds n u m b e r . F o r s i m p l i c i t y , we s h a l l
assumc! a parabolic d r a g p o l a r , defined by
1. Introduction
'The problem of determining optimal t r a j e c
C,, = C I), + kC2L (3)
t o r i e s f o r a rocket powcrcrl lifting vehicle flying
whc,.c cDo and a r c constants, It i s convrnient
inside the a t m o s p h e r e of a planet h a s received
to ripfine a lift control p r a m e t r r A s u c h that
ronsidcrablc attrintion in recent y e a r s . G e n e r a l
control laws f o r t h e lift and t h r bank angle, 2nd
f o r tlir thrusting p r o g r a m havc been obtained in X=jkC
I~, (4 1
t e r m s of t h e p r i m e r v e c t o r , the adjoint vect.or a s -
sociated to the velocity vectorf'). In this papcr w c Thcn, when h = 1 , th? flight i s a t maximum l i f t - t o -
modify the r e s u l t s in Ref. 1 to apply to the case drag ratio.
when the t h r u s t direction i s constrained i.o align
'The motion of the vchiclc, flying in a g c n r r a l
with the vclocity v e c t o r . This i s t h e r a s e when
gravitational f o r c e field, and subjcct to a e r o d y n a m -
the engine is mounted fixed with r e s p e c t to thc vc-
ir f o r c e and thrusting f o r c e , i s govcrncd !hy the
h i c k and t h e flight i s effectuated in the dcnsc lay-
eqaations
er o f thc a t m o s p h e r e and a t s m a l l angle of attack
to validate the assumption('). F r o m the g e n e r a l dy
- <
t h e o r y , solutions for optimal maneuvers f o r flight
in a uniform gravitational f i e l d will be obtained by -
dt
d V - T ?
-
A ++
canonical t r a n s f o r m a t i o n s .
dt v +m + g(r.t) (5)
11, The Optimal Controls
-~
d m - . -T
Consider the motion o f a powered, lifting vc- dt c
hicle in a g e n e r a l gravitationa~tf o r c e field. At the
w h r r r c i s the constant cxhaust velocity of the gas
t i m e t , the state of thc vehicle i s defined by
.r ( t ) position vector
ejected f r o m the engine and; i s the accolcration
of thc gravitational Ticld. The optimal t r a n s f e r
V ( t ) = velocity vector
problem is defincd a s follows.
- - . - -
,

m(t) 2 instantaneous m a s s .
T h e flight is controlled-hy a thrusting f o r c e T
- At the initial time, t 0 , r = To, V = Vo, m =
+
~

mo. The v e c t o r s ro. Vo and the sra+lar4mo~rrprc-

s c r i b e d . At thc final t i m e , t = t f , r = rf. V V f ~

and the aerodynamic f o r c e A (Fig. 1). It is a s - and m = m f . T h e problem is to k n d the t i m e h i s -

s u m e d that T i s aligned with the vcloeity, and its t o r y of B(t) and T ( t ) s u c h that s o m e s c a l a r function
magnitude is hounded by of thc final s t a t e i s a minimum.
o ZT z T
max (1) --
Using the maximum principle, we introduce t h e
adjoint clements q , p and pm to f o r m the Hamilton-
F u r t h e r m o r e , we a s s u m e that thc vehicle h a s a ian
plane of s y m m e t r y , both the t h r u s t and the a e r o -
dynamic f o r c e a r e applied a t the c e n t e r of m a s s
a n d , in coordinated flight, the aerodynamic f o r c e
H =
+
q.V-1. p . - v- -
- -.(2 +) + - - t g T (p.V) - m p m ]
mi c
(6)

and thc vclocity a r e contained in that p k n e of s y m - - +

m e t r y . It i s c u s t o m a r y to decompose A into a whcrc q , p and pm a r e defined by the adjoint cqua-
d r a g f o r c e D , always opposite to V and a lift tions

W. h P r e p a r e d under A i r F o r c e Grant N o . AFOSI2-71-2129.

* a P r o f e s s o r , Department of A e r o s p a c e Engineering, T h r University of Michigan, Memher AIAA
The following general r e s u l t s f o r optimal t r a j e c -
t o r i e s have been obtained in lief. I .
v2 = V c o s y s i n + (14)
I . I f the direction of the t h r u s t can be takcn a r b i -
v, = V s i n y
t r a r i l y , whenever the engine is operating, the
t h r u s t m u s t be directed along the vector:, called It is s e e n that y is the flight path angle, and + the
the p r i m e r vector. T h i s is the cxtension of thc heading. We consider an exponential a t m o s p h e r e
c l a s s i c a l r e s u l t obtained by Lawdcn for t r a n s f e r of the f o r m
in a vacuum. P = Po exp ( k P X 3 ) (151
whcre i s the constant height s c a l c . For the t r a n s
2. T h e thrusting p r o g r a m is governcd by thc
formation (14) to be canonical, while conserving
switching function, which, f o r the c a s e considered
the s a m e Hamiltonian
in t h i s paper when the t h r u s t i s always aligncd
with the velocity, is rlcfined by Pi dv, + Pz d V 2 + ~3dv3 pvdV 4 PydY + P,)d+

T h i s gives the l i n e a r transformation f o r thc adjoint

components of the p r i m e r vector
If K > 0 , w c s e l e c t T = Tmax (boosting phase)
+ +

1[] ]
K < 0 , wc s e l c c t T = 0 (roasting phasc)
cos v cos cos y s i n
K = 0 , f o r a finitr time i n t c r v a l , -sinycos + - s h y sin+ cosy p2 = py/V (16)
w e s e l c c t T = variable [sustaining phasc)
-cosysin c!- cosycosy 0
3 . The optimal lift modulation i s such that CI, = sin P!",
C o r , f o r variable lift p r o g r a m lteversing the m a t r i x cquation
1.Nlax'
B c11
tan e ~~ - '19)
aCL
where e is thc angle bctween 9 and -6
4,_The Eptimal bank angle is such that the vc,ctor>
V , p and A a r e coplanar, that is
In p a r t i c u l a r , w e can verify the following rclations
(I;XT;, .A' I: 0 (10)
5. F o r optimal flight constantly a t maximum lift-
t o - d r a g r a t i o , % is orthogonal to A", that is

6.
z.;i=o
Along a v a r i a b l e - t h r u s t a r c , w e have
(11) and
->
p.
-
v = PlV, +PlV, + mv3 ~~ VP" (1 9)
For the t r a n s f o r m a t i o n of the aerodynamic f o r c e s ,
we define the bank angle $ a s the angle between the
By taking the derivative of this equation, m d u s i n g plane of s y m m e t r y of the a i r p l a n e 5nd thc vertic.ai
( 5 ) , (6) and (7) tocether with (12) and r e a r r a n g i n e , plane passing through the velocity V ( F i g . 2 ) . Thnn,
f r o m the g c o m e t r y of the figure
( A , c o s + i A,sin+) c o s y ~ A
l ,siny = -D
s i)n y - A 3 c o s y ~- I , c o s +
(AIcos~riALsin+ (LO)
The v a r i a b l e t h r u s t magnitude control will be ob-
tained upon taking the derivative of t h i s equation. A, s i n + - A,cos = -Id s i n $

T h e s e r e s u l t s are valid f o r a general gravita- T h e r c f o r e , we have f o r the C a r t e s i a n components

tional force liuld. In p a r t i c u l a r , when 2 is time of the aerodynamic f o r c e
invariant, the Hamiltonian given by (6) is constant.
A, 7 - D cos $ c o s y - I.(cos + c o s + s i n y 4~ s i n + s i n + )
111. Application to the C a s e of a
A, - D s i n $ c o s y - y c o s + s i n + s i n y - s i n $ cos b) (21)
Uniform Gravitational Field
Consider thc case w h e r e 2 i s a constant vec- A, = -1Isiny +~ L C O S+ c o s y .
tor.
-r
In Cartesian rectangular coordinates, let
=
(x, ,x~ ,x, ) with x3 along g
- T h e aerodynamic optimal controls can be obtained
directly f r o m the general t h e o r y , Using formula
 ( 9 ) , with relation ( 3 ) . we havc f o r thc optimal
v a r i a b l e lift control
tan c 2kCL (22)

For t h e bank nnglc, hy rxpanding the d c t e r m i n a n t

( I O ) , using ( 1 4 ) and ( 1 7 ) f o r the components of 9
L.

1
and p, we havr!
P
p [A, s i n + - A,cos $ 1
Y
-
cos Y
A
The Hamiltonian (6) becomes, with ncw variahlct;,
. [ ( A , cos + + A , s i n $ ) s i n y - A , c o s y ] 0 in a uniform gravitational field

Hy relations (20). we have for thc optimal variable €1 = ( q l c o s + + q z s i n + ) V c o s yt q 3 V s i n y

hank anglo

(29)

>.
p.v
-
W e can rpwritu relation ( 2 2 ) by observing that

~ pvc.osr ~ vp
Along a n optimal t r a j c c t o r y , with variahlc lift and
bank controls given by (23) and (27) we have
Thcrc-forc
13 ~ ( q , c o s $ . I . q 2 s i n + ) Vc o s y I g V sin)

On th? o t h e r hand, f r o m ( l a ) , and using the opt,

m a l l a w (23)

0 L' We notice t h a t , if t h e flight is a t m a x i m u m lift-to-

PL d r a g r a t i o kCf,- Cu, = 0 , and the constant Hamil-
(25) tonian i s f r e e of aerodynamic components. T h i s
Y r e s u l t h a s been shown to he valid f o r n g e n e r a l
By eliminating p' between (24) and ( 2 5 ) , we hnvc gravitational f o r c c field in Ref. 1 The switching
function is s e e n to hc

'The optimal lift c o n t r o l , c x p r e s s c d in t c r m s o f the T h e adjoint equations, with optimal variable l i f t a n d

adjoint v a r i a b l e s , i s thcn hank controls, a r e

LkC
I,
=
vi'VPYc o s + (27)

Wc sei! that the acroclynainic controls a r e govrrn-

ed by the adjoint components p and p
v'Py 41

1V. T h e l n t e p a l s of the Motion

A numbcr o f the integrals of the motion,
mostly associatcd with a flight a t maximum l i f t - t i t
d r a g r a t i o , has been displayed in R e f . 1 . Here w e
shall derive them d i r e c t l y , and f u r t h e r m o r e oh-
tain s o m e additional integrals. T h e equations o f
lhc motion, written in components f o r m , and using
the velocity transformation ( 1 4 j , a r e

-
dr,
2 v c o s y cos*
dt
-
dxz v cosy sin 4
it t

Y
W c f i r s t have the integrals
ql = a, ~ constant
(33) P
q, ~ aL ~ constant ~~ -9, v s i n y +p'pV s i n y $cosy)
1
( 4 ~ )
If I h r longitudinal rangc and the l a t e r a l rangc a r c
fi.c<r, t h c s c constants a r c zcco. It i s secn that <b With this intcgral we can delete the diffcrcntjal v
i s a l s o constant if thc flight i s a t maximum l i f l - t m equation f o r p y . Using (42) w e writc the cquation
d r a g r a t i o . Next, using (33) we havc for q3

IIcncc
The equation can be integratcd if s k i p trajectoi'y
'1, x2 q , x l
~

*
= p ~1

F o r f r u c longitudinal and l a t e r a l r a n g e s
constant (34) assumption i s u s e d . We havc

p ~ a b = constant (3 5)
'I, Upon integrating
If the final h r a d i n g i s not p r c s c r i h e d , t h i s con-
9, = const. X exp(-px,) = c o n s t . X p (43)
stant i s z e r o , and f r o m ( 2 3 ) , : 0. +
T h e flight
i s effectuated in a v e r t i c a l planc containing the By the change of variable f r o m xg to p, using thr:
initial vrlncity. exponential law (15), a canonical t r a n s f o r m a t i o n
requires
Now r o n s i d r r
Qdx, = q dp = - p p q dx,
P P
Hence
Qi = - P P q P
F i r s t , along a coasting a r c , T = 0 , o r along a s u s - and t h e r e f o r e
taining a r r , K 0 , and i f thc flight i s a t maximum
~

lift- to - d r a g r a t i o 9 = constant (44)

P
mpm = constant (37) T h i s shows the advantage of using the a t m o s p h e r i c
m a s s density a s the altitude v a r i a b l e .
F o r flight a t v a r i a b l e lift coefficient, and along a
coasting a r c or a sustaining a r c , by eliminating T h e E g g e r s assumption, in the f r e e range and
thc t i m e between (36) and the equation for q, , we f r e e t i m e case, gives explicit l a w s for the moduia-
tion of the lift and the bank.
U'
have
~1% pd(mpm)
~
F i r s t , along a coasting a r c , we w r i t e (43)
By intcgrating, wc have
9, = b 3 ~ (45)
% = PmPm + a3 (381
where b, is a constant. The Hamiltonian relation
where a, i s a new constant of integration. (42) gives
Consider the d e r i v a t i v c of Vp,, 2b3 m s i n y
"v = S(CD, - k C L )

Since in this c a s e Vp = a 4 , we have

V
kCt
-~
- 1 -
2b, m siny
Sa, C
Along a coasting a r c , T = 0 , and using E g g c r s Do Do
a s s u m p t i o n by neglecting the g r a v i t y c o m p a r e d to That is, by o b s e r v i n g that m is constant along a
a e r o d y n a m i c f o r c e r t a n assumption generally valid coasting a r c
for a s k i p o r pull-up m a n e u v e r , we h a v e , upon A' = 1 + a siny (46)
integration
where a is a constant, We have the c l a s s i c a l re-
Vpv = - H t + a4 (40) s u l t , f i r s t obtained by C ~ n t e n s o u ( ~ ) a nl adt e r ex-
where a 4 is a constant of integration. In particu- tended to three-dimensional c a s e by Griffin and
l a r , when t h e final t i m e is f r e e , I1 = 0 , and V i r ~ h ( and
~ ) Speyer and Wornble('). W e now r e w r i t e
the l a w f o r the bank angle
V pV = a 4 (4 1)
tan+ = p* = a6
We w r i t e the Hamiltonian f o r the c a s e of f r e e py c o s y 2ka C cos+cosy
4 L
longitudinal and lateral r a n g e , and free t i m e , or
along a coasting a r c o r a sustaining a r c . Relation b
(30) b e c o m e s sin+ = ~
(47) Lj
A cosy

4
 wherc b is a constant. The constants a and I) in
the optimal lift and bank l a w s a r e detcrmined by
specifying t e r m i n a l conditions a t t h e ends of a
coasting a r c
Next, along a sustaining a r c , we have the r e -
-. lations (12) and ( 1 3 ) , which for a constant gravity
field. hccome
and
F o r a f r e e t i m e problem, H 0 , and neglecting
gravity, we havr the following s i m p l e law f o r opti
m a l v a r i a b l e lift control along a sustaining arc
(501
An i n t e r e s t i n g fact is that variable lift control for
sustaining flight is only optimal f o r high velocity,
V 7 c, and f o r vchiclc with high lift p e r f o r m a n c e ,
x> 1.
Finally f o r the boosting a r c , by eliminating
t h e t i m e hetwcen the equations f o r V and m , W E
have
dV.
dm
c
m
Tmax
[g +gsinJ
v with Tmax -* rn (impulsive approximation) the
equation can h e integrated to give
wherc 7% is a constant m a s s
V. Conclusions
t i e n e r a l control l a w s for the modulation of the
lift and t h c hank anglc, and f o r the t h r u s t i n g pro-
g r a m along the optimal t r a j c c t o r y of a rocket
powered, lifting vehiclc, flying inside the a t m o -
s p h e r e of a planct, with the t h r u s t aligned along
t h r velocity v c c t o r , have been obtained in t e r m s of
the p r i m e r vector, the adjoint vector a s s o c i a t e d
to t h e velocity vcc.tor. From t h e general t h c o s y ,
solutions f o r optimal maneuvers f o r flight i n a uni-
f o r m gravitational field a r e ohtained hy canonical
t r a n s f o r m a t i o n s . The problem is completely
solved f o r the f r e e range and f r c c t i m e c a s c , using
E g g e r s assumption f o r s k i p t r a j e c t o r y . It should
he noted that this l a s t stringent condition can be
r e m o v r d hy using Loh's second o r d e r theory('), as
applicd by Speyer and Womble for f r e e flight t r a -
jectory(6). 1,oh's sccond o r d e r theory which a l s o
includes the c u r v a t u r e of the flight path h a s a l s o
heen applied successfully to the c a s c of thrusting
flight by Griffin and Vinhi8).
Acknowledgments: T h e subject developed in Rcf. 1
h a s hcen proposed to the author by P r o f e s s o r P.
5
R r o u s s e of t h e University of Paris, and General P.
Contensou, Central Scientific Director a t the
ONERA, F r a n c e , as a minor t h e s i s f o r a "Doctor-
a t d'Etat 2s Sciences Math6matiques" submitted in
June 1972. An English v e r s i o n of t h e paper will
a p p e a r i n the J o u r n a l of Optimization Theory and
Applications.
References
1. Vinh, N. X. ,"Optimisation des T r a j e c t o i r e s e n
Vol Atmosphe?ique." D.Sc. t h e s i s , University
of Paris, June 1 9 1 2 .
2. Bryson, A. E. , and Lele, M . M . , "Minimum
F u e l L a t e r a l T u r n s a t Constant Altitude,"
AIAA J o u r n a l , V01.7, N o . 3 , M a r c h 1969, pp.
559- 560.
3. E g g c r s , A.J . , Allen, H . J., and Neicc,S. E. ~
"A Comparative Analysis of t h e P e r f o r m a n c e
of Long-Range Hypervelocity Vchicles. 'I
NACA r e p o r t No, 1382, 1958.
4. Contensou, P. ,"Contribution 1'Etude Schg-
matique d e s t r a j e c t o i r e s Semi- halistique 'a
Grandc Port&," p a p e r communicated to the
Association Technique M a r i t i m e et A6ronau-
__
tique, P a r i s , April 1965.
5. Griffin, J . , and Vinh, N . X . , "Threc ilimension-
a1 Optimal Maneuvers of Hypcrvelocity Ve-
hicles," paper No.71-920, AIAA Guid-
a n c c , Control and Flight Mechanics Confer-
e n c e , Hofstra University, Hempstead, New
Y o r k , August 16-18, 1971.
6. . " . J . ., and Womble ..M. .."Annroximate
SDever. ..
Optimal Re-Entry T r a j e c t o r i e s . ' ' AIAA paper
No. 71-919. AIAA Guidance. Control and
Flight Mechanics Conference, Ilofstra Iiniver-
s i t y , Henipstead, New York, August 16-18,
1971.
7. 1,oh. W . T . . " E n t r v Merhamcs and D v n a m i c s . "
i n R e - E n t r y and Planetary Entry Physics and
Technology, W . T . I.oh, ed. S p r i n g e r - V e r l a g
Inc., New York, 1968.
8. Griffin, J., and Vinh,N. X . ,"Optimal T h r e e
Ijimensional Maneuvering of a Rocket Powered
Hypervelocity Vehicle," IAF papcr, 2 3 r d
C o n g r e s s of the International Astronautical
Federation, Oct. 8-15, 1972. Vienna A u s t r i a .
 F i g u r e 1. State and Control Variables

Figure 2. Aerodynamic F o r c e s T r a n s f o r m a t i o n