GUIDANCE, NAVIGATION

AND CONTROL

Approved, JLd. �
Datd
G. M. LEVINE, DIREC OR, GUIDANCE ANAL
APOLLO GUIDANCE AND NAVIGATION PROGRAM
S
� 7/

Approved:
R. H. BATTIN,
J!. . l � \"(, ;:;J5;;,
Date: .._. ·r

DIRECTOR, MISSION DEV ELOPMENT

1'1 0, 1

APOLLO GUI NCE AND NAVIGATION PROGRAM

..u..i[£-'-'<4PJ...L..(,.<j;;__,...:::..;r+=::J..loo�b-- Date: /�t..IH1l(

ATION PROGRAM

Approved:
R. R. RAGAN,
)ft�ji( PUTY DIR
Jl�
CTOR
Date:/8 8'-- 71

CHARLES STARK DRAPER LABORATORY

R-695

APOLLO LUNAR-DESCENT GUIDANCE

by

Allan R. Klumpp

JUNE 1971

CHARLES STARK DRAPER

CAMBRIDGE, MASSACHUSETTS, 02139LABORATORY
 AC KNOW L E DG ME N TS

This report w as prep ared under DSR P roj ec t 5 5 - 2 38 9 0 , sponsored by the

Manned Spacecr aft C enter of the N ation al Aeronautics and Spac e Administr ation
th rough C ontrac t N AS 9 - 4 0 6 5 .

The P 66 vertic al channel was developed b y C r aig W . Schulenberg. The

an alytic al design and gain setting of the P 6 6 horizontal ch annels w as done by N icholas
J. Pippenger using concep ts suggested by Jerrold H. Suddath. The concep t of
analytically extr apolating to yield the p r edictive guidance equation for P 63 and P 6 4
w as con c eived by W illiam S . W idnall. The existence of an an alytic al solution for
the guidance frame orientation to yield zero c rossrange target j erk w as recognized
by Thom as E. Moore. The thrust direction filter configur ation for eliminating
thrust-pointing errors due to attitude bias within the digital autopilot deadband was
conceived by W illiam S . \Vidn all and Donald W . Keene.

The public ation of this report does not constitute app r ov al by the N ational
Aer,onautcs and Space Administr ation of the findings or the conc lusions contained
therein. It is published only for the exchange and stimulation of ideas.

(D C opyright by the M assachusett s Inst it ut e of Technology

Published by the C ha rles Stark Draper Labo ratory of the
Massac husetts Institute of Tec hnology.
Printed in C amb rid ge, Massachusetts, U. S. A. , 1971

ii
 R-695

APOLLO LUNAR -DE SC ENT GUIDANC E

ABS T R AC T

This report records the technology associated with Apollo lunar- descen t
guidance. I t c ontains an introduction plus five m aj or sec tions:

1. Braking-ph ase and app roach - ph ase guidan c e. Braking- phase guidance begins
in lun ar orbit pr ior to engine ignition and transfers the lun ar module ( LM ) to a
terminus typically 7 8 0 0 - m slant- r ange before the landing site. Approach-phase
guidance begins at braking-ph ase terminus and transfers the LM to a terminus
typ ic ally 30 -m above the landing site. The b r ak ing-ph ase transfer is near - optim al.
when�as th e app ro ach - ph ase tr ansfer sac rifices propellant-utili zation efficien cy to
p rovide landing-site visibility and landing- site redesign ation c ap ability.

2. Term inal- descent- phase guidanc e. Initi ated autom atic ally at approach - ph ase
term inus, or m anually any time during the app r o ach ph ase. termin al- descent-phase
guidan c e automatic ally nulls hori zontal velo c ity and controls altitude r ate to a
reference value. The referen c e v alue is inc remented or decremented by ast ron aut
m anipulation of a rate-of- descent control switch.

3. Powered - flight Attitude- m aneuver Routine. The routine connects all powered­
flight guidance p rogram s- descent and ascent- to the digital autopilot. A dep artu re
from traditional app r o aches, the routine transfers the L M from any cur rent attitude
to any comm anded attitude while avoiding gimbal lock by inherent char acteristics
of the m aneuver algo rithm. No gimbal- lo ck - avoidance str ategy is required.

4. Throttle R outine. The routine connects sever al guid ance progr ams to the
descent p ropulsion system ( DPS ) . DPS h ardw are limitations r equi re operation either
at a m aximum - thrust point or within a sep ar ate p ermitted- thrust region. The routine
alters thrust commands from the guidan c e progr am s when nec essary to meet DPS
c onstr aints and issues cor rected th rust inc rement commands.

iii
5. Braking-phase and A pproach- phase Targeting P rogram . This groun d - b ase d
p ro g r am is use d at the Drap e r Laboratory and at the N ASA M anned Spacecraft C enter .
T h e p r og ram supplies descent targets which are lo aded into guidance computer
m emory shortly before laun ch.

by
Allan R . Klumpp
June 1 9 7 1

iv
 TABLE OF CONTENTS

Section

INT RODUCTION 1

Descent Phases 1
Navigation, Guidance, and Control Configuration 4

Navigation . . . . . . . . . . . . . . • . . . . . . . 4
Guidance • . . . . . . . . . . . • . . . . . . • . . 4
Powered - flight Attitude-maneuver Routine 6
Throttle Routine . . . . . . . . . . . . . . . . . . 6
Braking-phase and App roach-phase Targeting P rogram 7
Digital Autopilot . . . • . . . . . • • • . . • . . • • . • . 7

DE FINITIONS OF LUNAR DESCENT COORDINATE FRAMES,

ATTITUDE ANGLES , AND GIMBAL ANGLES . . . . . . . • . 9

BRAKING- PHASE AND APPROACH-PHASE GUIDANCE 13

Guidance Equation Derivatio n • . • . . . 13

Braking-Phase Targeting Objectives 18
Approach -Phase Targeting Objectives 19
Landing-Site Redesignation P rocedure 19
P 6 3, P64 Guidance Algorithm . . . . . . 21
P 6 3 Ignition Algorithm . . . . . . . . . • . • . . . . . . . • . • • . . . . . . . 27

T E RMINAL-DESCENT-PHASE GUIDANCE 31

P 6 6 Horizontal Guidance Algorithm . 31

P 6 6 Vertical ( ROD) Guidance Algorithm 32

POW ERED - FLIGHT ATTITUDE - MANEUVE R ROUTINE 37

THROTTLE ROUTINE . . • . • . • . . . . . . . . . . . . • . . . . • . • . . . . . . . 45

v
 TABLE O F CONT ENT S (Cont . )

Section Page

B RAKING-PHASE AND APPROACH-PHASE T ARGETING PROGRAM 51

Constraints . • . . . • . • . . . . • • • . . , . • . . . . . . . . • . . • . . • . . 53
Approach-phase Targeting . . . . • . • . • • . . . . . . . . . . . • . . . . . 54
Braking-phas e Targeting . . . . . . . . . . . . . . . . . • . . . • . . . . . . 56

R E F E RENCES . . . • . • . . . . . . • . . . . • . • • . • . • • . . • . • . • . . . . . . 61

vi
NO MENC L AT U R E

Symbols a r e norm ally defined whe r e first introduced. The refore it is necessary
to define here only those symbols use d in more than one section of this report.

Most symbols are self- defining by being constructed of standard i dentifiers

as follows:

1. Type of variable. Position and its deriv atives veloc ity, accele r ation,
je rk, and snap are denoted R , V, A, J , S . Thrust is denoted F, thrust­
accele r ation AF , unit vecto rs 1J.N, c lock-times by lo wer c ase t, and
target- refe renced times (times with respect to the target point of a
particular mission ph ase ) by uppe r c ase T.
2. Mission ph ase. The braking ph ase ( P 6 3 ) is denoted B R , the app r o ach
phase ( P 6 4 ) AP.
3. Applic able point in phase. Inception is denoted I, terminus F , and tar get
point T .
4. C oordinate fram e of referen c e . P latform coor din ates are denoted P ,
guidance G , an d L M body B .
5. Achieved ( as opposed t o nomin al) v alues ar e denoted A .

Thus b y construction, R B R FG A i s the p osition vector e xpressed i n guidance

coo rdinates achieved at the b r aking ph ase terminus. Without a ph ase i dentifier,
B_TG to �TG represent the b r aking or appro ach ph ase targets R BR T G t o §.BRTG o r
B_AP TG t o .§.AP TG, whichever phase is current. Vector elements a r e denoted by
subsc ript, e . g. , V
· Vector m agnitudes are implied whenever a symbol reserved
z
for a vector lacks the unde rscore. Row vectors of 3 x 3 m atrices are denoted

�X'�Y'�Z '
and m atrix e lem ents are identifie d by row and column, e . g. , C y z is the
Z component of the row vector �y·

Symbol Definition

_AFC P Thrust- ac c e le r ation command

AF P Thrust- accele r ation me asurement ( computed by dividing
the inerti ally sensed velocity inc rement by the guidance
sam p le inte rval)
AGS Abort Guidance System
ATT P owered- flight Attitude - maneuver R outine
C BP T r ansform ation to body from p latfor m coordinates
CGP T r ansform ation to guid ance from p latform coordin ates

vii
D.\P Digital autopilot

DEC 1\ Descent engine c ontrol assembly. A digital to analog

interface between the LGC and the DPS thr ottle
DPS Descent p ropulsion system

oF.\ Tht·ust cor rection in c r ement which must be added to the

thrust measurement aver aged over the sample interval
to obtain the samp le-instant thrust.

G:\1 Lunar g r avity ( ac celeration of p ositive sign val i d at the

lun a r s u r face)

GP Cur rent g r avity valid at the cu r rent position .B_P

I:\ I u Inertial measurement unit consisting of a three - gi m b alled
stable member and three acceler o meters

u: :\DTII\IE A time interval ( typically 2.2 seconds) added to the

target-referenced time T in P63 and P64 to account fo r
the effective t r ansp o r t lag due to c omputation and system
response times

LGC Ll\1 guidance computer

LII\IIT c\ function of two arguments yielding the fi rst argument

limited in magnitude to the value of the second argu ment

Ll\1 Lun ar m odule

LP L anding site

LPD L a n d in g p o int designator c onsisting of two ret i c les, one

on the inside w in d o w p anel, and one on the outside

I\ L\.'\. I:\IU :\1 A function of n arguments yielding the algeb r ai c ally

highest argument

:\1I;'>JI!\1UM A function of n a r guments yiel ding the algeb r ai c ally lowest

a r gument

N A.. S,\ National Aeronautics and Sp ace Administ r ation

Pitch "
L:\1 attitude angles. See Definitions o f Lunar Descent
Yaw "
Coordinate F r ames, Attitu de Angles, and Gimbal An gles
Boll

P63 B r aking ph ase o r b r aking-phase g u i dance p rogr am

P64 Ap p r oach phase o r app r o ach-ph ase guidan c e p r o g r am

P66 Term i n al - descent phase or te r m i n al - descen t - p h ase

gui d an c e p r o g r am

R .�PTG, RBRTG, ll_TG,

Y:M>TG, YBRTG, YTG,
T ar get position, veloc ity, acceleration, jerk, an d snap of
.:::ir\PTG, ABRTG, ATG,
the app r o ach phase, b r aking phase, or c u r rent phase
JAPTG, JBRTG, JTG,
�A.PTG, �BRTG, .§.TG
.B.G, .B_P Cu r rent position

ROD R ate of descent

t c lo c k - t i m e

T T arget- referenced time ( t i m e w i t h respect t o the t a r get

p o int of the c u r rent m ission phase )

viii
THROT Throttle Routine
!JN F C P Unit thrust command
!J.N W CP Unit window comm and
.YG. Y.P C u r rent veloc ity

ix
IN TRODUC TION

This report d e s c r ibes how th e Apollo lunar-desc ent guidance works, why it
w as designe d th is w ay, and, in seve r al c ases, how it might h ave been desi gned
d iffe rently. The c onc epts desc ribed c an be applied to l anding on any planetary b ody,
w ith or without atmosphere, should m an r e s olve to c ontinue this adve nture. The
s olutions p resented offe r ample opportunity for checking the theory. Such c hecks
h ave been m ade, and all algorithm s ar e known to work as c onceive d,

Lunar -desc ent guid ance begins with the lunar module ( LM ) at about 15 - km
altitude in a s lightly elliptic al c oasting lunar orbit, and ends with the L M on the
lunar surface. The guidance is pe rformed by the onboard LM guidance c ompute r
( LGC ), which t akes input d ata and c om m ands directly from the LM c rew and via
the uplink from N AS A's R e al Time C omputation C ente r in Houston, Texas, The
c r ew c onsists of a c ommander and a L M pilot. ( See Figure 1. ) Standing on the
left, th e c ommander m onitors and c ontrols the descent using visu al cues and v ar ious
h and c ontrolle rs and s witch e s . St anding on the right, the LM p ilot monitors the
c ompute r display, vocally relays pe rtinent d ata to the c om m ander, and e nters any
nece s s ary d ata into the computer via the keyboard.

The primary guidance m ode for the lunar de s c ent i s automatic; the LGC c ontrols
both attitude and th rust. The c ommander c an, tempor arily or permanently, select
nonautom atic guidance m odes if he wishes to c ontrol, m anually, attitude or thrust
or both . The nonautom atic m ode s , not d e s c r ibed fur th e r in this report, provide
attitude and th rust refe renc e s for the c om m ande r to follow if he chooses to fly the
L:\1 manually along the autom atic guidanc e p r ofile.

D e s c e nt P h ases

The lunar d e s c ent i s a nom inally-planar trajectory c onsisting of three phases

illustrated in F igure 2 and d e s c ribed as follow s :

1. T h e b r aking ph ase ( P r ogr am 6 3 , o r P 6 3 ) is initiated b y keyboard entry aboat

1 0 minutes befor e nominal ignition tim e . P 6 3 fir st c omputes the p re c ise time an d
attitude for ignition. N ext, at typic ally 49 2 -km s l ant-r ange from t h e l anding site,
P 63 ignites the DP S . F i nally, P 6 3 t r ansfe r s the LM to the terminal state required
as initial c onditions for the succeeding app r oach phase. The tr ansfe r takes typically
5 1 4 second s and is ne ar- optim al.

1
 LANDING POINT
DESIGNATOR
RETICLES
SUPERIMPOSED

LM
PILOT

"
/;'/k.__
LM
COMMANDER

Figure 1 C o mponents of the Luna r - d e s c e nt Guidance System

 IGNITION
1 5 k m. ALTITUDE

11 min 31. 3 sec.

492k m. SLANT RANGE

16.2" CENTRA L ANGLE TO TOUCHDOWN

t
22 MINUTES
START P63

30m ALTITUDE
START P66
U LLAGE
11 m GR OUND RANGE
45" CENTRAL ANGE 15km ALTIT UDE
31 sec TO T O UCHDOWN
TO TOUCHDOWN
t ....;:-4
--..._
11 m;n. 38.8 sec.
50 5 k m SL ANT ANGLE
START P64 '-... , PpRo401
2231 m AlTITUDE ' , .... Pf.t45f p64
16.6" CENT RAL AN GlE
TO TOUCHDOWN

w ..--.:;�--. t--�......;-;".:-.-.-
---�.:.$ .--�-:-.-
. -·

...
·.•., �-:: �---�·;;,:-:--;-� .

AL L NUMBERS ARE TYPIC AL AND DO NOT REPRESENT ""

'
ANY SPECIFIC APOLLO MISSION PPROACH-PHASt · -
B RAKING-PHASE )
29m ALTITUDE
T ARGET POI NT
TARGET POINT
-541 m ALTITUDE 4.8m GROUND R ANGE
4416 m GROUND R ANGE TO LANDING SITE
TO L ANDING SITE

Figure 2 Luna r - descent -guidance Phases and Targets

2. Approac h - phase ( P 6 4 ) guidanc e begins w ith initial conditions c ons isting of,
typically a) 2. 2- km altitude and 7. 5-km ground r ange and b) - 44 - m / s ec vertical
velocity and 129 - m / s ec fo rward velo c ity. In typically 1 46 s econd s , P 6 4 t r an sfer s
the Ll\I to a point almost directly above the landing s ite. P 6 4 provides continuous
visibility of the lunar surface and, spec ific ally, of the landing s ite until around 5
s e c onds befo re terminus . Du ring P 6 4 the commander c an direct the LGC to land
at any visually chosen point on the lunar s urfac e b y a landing- s ite r edes ignation
p r o c edure which c an be continued until initiation of the terminal- des cent phase,

3. Th e term inal- des c ent phase ( P 6 6 ) b egins autom atic ally at typic ally 3 0- m
altitude and 11- m ground r ange from the landing s ite, o r it m ay b e initiated b y the
c om m ander any time during P 6 4 . The P 6 6 guid an c e algorithm c ontrols velo c ity
o nly; there is no position control. P 6 6 nulls the forward and lateral velo c ity
c omponents to p roduce a vertical app ro ach to the lunar surface, an obj ective which
c annot be achieved from visual cues when the s urface is obscured by a sheet of
radially moving dust. P 6 6 controls altitude r ate to a r eference v alue that is
inc remented or decrem ented by 0. 3 - m / sec each time the commander r ai s e s or lowe r s
a three - p o s ition r ate - of - d e s c ent ( RO D ) c ontrol s witch lo c ated ne ar his left h and.

I'\ avig aticn, Guid anc e , and C ontrol C onfigur ation

Th e N avigation, Guidanc e, and C ontrol configur ation illu strated in F igure 3

applie s to all LM powered- flight guidance maneuver s . This repo rt des c ribe s only
the s olid portions of F igure 3 . All routines are p r o c e s s ed once eve ry two s econds,
except the ve rtical ch annel of the P66 guidance algo rithm is p rocessed once per
s e c ond, and the digital autopilot is proc e s s ed 10 times per s e cond.

Navigation

1\' avigation
1
( s ee K riegsman ) p rovides an estimate of the current state vector
based on data from an inertial measurement unit (ll\IU ) and a l anding r adar. IMU
data are used throughout all thrusting m aneuvers, but, to avoid accumulatiun of
inertial errors, I M U d ata are not used dur ing coasting flight except for a m inimum
perio d immediately preceding and following e ach thrusting maneuver, The l anding
r adar p rovides altitude data below typically 1 0-km altitude, and velocity d ata below
6 1 0 - m /s e c .

Guidance

G uidance transfe rs the LM from the current state to the terminu s of the current
phas e . In addition to the current state e stim ate from Navigation, Guidance is based

4
 �
BRAKINCH'HAS£ AND
APPROACH-PHASE r------,
I
TARGETING PROGRAM
THRUST INCREMENT

� r4
DESCENT
I GROUND BASED I COMMAND
THRUST VECTOR
THROm£ ROUTINE

I
PROPULSION
SYSTEM

J L
�-P64
UIDAI'I:E TARGETS __ ____ _j

,------l r--- - ---,

I
THRUST -ACCELERATION

:
GUIDAJC:£
COMMAND

!STATE VECTOR UPDATE

NAVIGATION

I
I
:
SPACECRAFT
P63, P64 GUIDA!'�:£ ALGORITHM

I UNIT THRUST COMMAND

CURRENT v
DYNAMICS

L ______ _j L ___ _j
P66 GUIDANCE AlGORITHM

UNIT WI NOOW COMMAND

ROUTINE STATE
GRAVITY __

-:!'>I
r
--
D�;��l
--
,---if
�
be
POWERED -FliGHT

I TORQU[ VECTOR
AUTOPILOT

ROUTINE
AT!' I TUDE-MANEUVER "'
GIMIAL INCREMENT ANO CONTROl

L �� � .:__ - _j
COMMANDS C R
PHASE PlANE DATA

CJl NAVIGATION
MEASUREMENTS

,-------i
lb===================== I INERTIAl SENSORS �I :::!:========:::::= = ==== =======::1
::: ::::=�
� I
LANDING RADAR
I
L _______ _j
LINEAR ANO ROTATIONAl MOTION

Figure 3 Navigation, Guidance, and C o ntrol C onfiguration

on p r ecomputed targets from th e groun d -based targeting p r ogram . The outputs
from the Guid ance algorithm are a unit thrust c ommand and a unit window c om m and
is sued to the Powered- fli ght Attitude- maneuver R outine, and a thrust- ac celeration
c om m and is sued to the Throttle R outine. Through the s e routin e s , Guidance c ontrol s
th e thrust vector magnitude and direction with respect to ine rtial spac e .

Powered-flight Attitude - maneuver Routine

The Powe red- flight Attitude - m aneuver R outine ( AT T ) c onnects all guidan c e
p r ogram s , de s c ent and ascent, to the digital autopilot ( D AP). ATT inputs a r e two
c om m and vector s ; a unit th rust c om m and and a unit window c ommand. AT T e stimates
a unit thrust vector from ac celerometer m e asurements, an d i s sues incremental
c om m ands to the DAP. The s e commands c ause the DAP to dr ive the e stimated
unit th rust vector into c oinc idence w ith the unit thrust command and the symmetry
plane of the LM into c oinc idence with the unit window c om mand .

During P64, as long a s the landing s ite would b e vi sible, the unit w indow
c om m and is sued to ATT by Guidance i s the lin e - of- sight vector to the current landing
s ite . By rotating the Ll\1 symmetry plane into coinc idence with the line - of- sight
ve ctor, A.TT sup e r impos e s the landing-point designator reticles of F igu re 1 on th e
current landing s ite.

Th rattle Routine

The Throttle R outine ( T HRO T ) connects several powered- flight guidance

p r ogram s to the D PS . The DPS is used by all d e s c ent guidan c e p rogram s , one of
the two abort p rogram s, and on e guidance program whose purpose is to p r ovide a
velocity -vector inc rement c omputed by the R e al Time C omputation C ente r in Hou ston
and tran sm itted to the LM.

The DPS must be op erated e ither at the m aximum - thrust point ( about 92'7o of
th e rated th rust of 46 7 06 newtons ) or within a p erm itted-thrust region ( 1 1 to 6 5o/o
of rated th rust ) . The intervening region ( 6 5 to 93o/o) is forbidden because in this
region oxidi zer flow an d fuel flow m ake independent tran s itions from c avitating to
n oncavitating regime s . The independent tran s ition s c ause gros s deviations from
th e required mixture ratio and produc e exc e s s ive erosion of the DPS noz zle.

U s ing a c omputed mas s estimate, a thrust- ac celeration me asurement, an d

th e th rust- ac cele ration command from the guidan c e equation s , THROT computes
the cur rent and c ommanded th rusts and i s sues th rust inc rement commands to the

6
DPS. These c omm ands e ithe r 1 ) drive the com puted th rust into c oincidence w ith
the c om m ande d thrust whenever the c omman d e d thrust lies within the permitted­
thrust region , 2) p roduce m aximum thrust whenever the commanded thrust lie s above
the p e rmitte d-thrust region, or 3 ) p roduce minimum thrust whenever the com m anded
th rust lie s be low the perm itted - thrust region.

Braking- ph ase and Approach-phase T ar geting P r ogr am

The targeting p r og r am provide s t argets for P 6 3 and P 6 4 . The targets define

b r aking- and app r oach- phase reference t r aj e ctories as independent vector
p olynom ials centered at individu al target p oints as illu strated in Figure 2 . Although
only P 6 3 and P6 4 are t argeted, the targets are des igned to achieve all the guidan c e
objectives of P 6 3 , P 6 4, and P 6 6 .

Digital Autopilot

2
The DAP ( s ee Widn all ) c ontrol s the attitude of the LM during pow e re d flight
by m e ans of c ontr ol effectors c onsisting of a r e action contr ol system and a trim
gimbal system. As the n ame implie s , the trim gimbal s ystem is a slow s ystem
u s e d p rim ar ily for trimm ing the DPS thrust vector through the LM c enter of m as s ,

7
DE F INITIONS O F LUN AR DESC ENT C OOR DINATE F R AMES,
A T T I T U DE AN G LES, AND G IMBAL A1�G LES

Three coordin ate fram es are required for lunar des c ent guid an c e bec au s e 1)
all inertial m easurements are with resp ec t to the s table p latform of the I MU , 2 )
P 6 3 and P64 guidance is with respect to a l anding site which rotates with, and c an
b e redesignated along, the lunar surface, and 3 ) thrust- vector determin ation an d
landing- site redesignations are with res p ec t to L M body axes . These coordin ate
s y s tem s are illus trated in F igure 4 and defined as follo w s :

1. P latfo rm coordinates . V ariables in p latform coordin ates are tagged P . The

o rigin i s at the center of the moon, the XP- axis pierces the nominal (unred esign ated )
landing s ite at the nom inal landing time, the Z P - axis is p arallel to the orbital p l ane
*
of the C omm and Module and points forward, and the YP- axis completes the
right- hand triad. Platform coordinate s are nonrotating.

2. Guidance coordin ates . Variables in guidance coordin ates are tagged G. The
o r igin coincides continuously with the current landing site ( the frame rotates with
the m oon ); the XG -axi s i s vertic al; the ZG - axis lies in, or near, the plane containing
th e I....M and the landing s ite and points forward; and the YG - axi s completes the
r ight- hand triad. Thu s , the origin and orientation of the guidance frame are altered
each time the landing s ite is redes ignated, Guidance- frame unit vectors exp res s ed
in p latform coordinate s are the row vectors �G P , C G Py • C G P of the m atrix
X Z
CGP.

3. Body coordin ates . V ariables in body coordinates are tagged B. Thes e are
the generally accepted LM coordinates . The X B- axis is in the direction of the
nominal thru s t vector, the ZB - axis is directed forward, and the Y B- axis completes
the right- h and triad. Body- frame unit vectors expres s ed in platform coordinates
a r e the r ow vecto r s C BP ' �BP y • C BPz of the m atrix C BP .
X

F rom the s e definitions i t i s noted that if the L M lands at the nominal s ite at
the nominal tim e in a nomin al erect attitu de, the three frames w ill be p arallel at
the ins tant of touchdown.

The following conventions are defined for orthogonal m atrices:

*
The LGC transfers state vectors in pl atform coordin ates to an abort guidance sys tem
( AGS). The AGS requires the s tate to be expres s ed in a frame who s e Z axis p arallels
the orbital plane of the command module.

9
 XB
XP XG
ZB
LUNAR ROTATION TO OCCUR
DUR ING TIME REMAINING TO
NOMI NAL LANDING TIME.

NOMINAL
LANDIN G
SI T E AT
NOMINAL
LANDING
TIME

Exage rated Scale

'------+- zP.

XP, YP, ZP-Piatf o rm ( Inertial Measu rement Unit) Coordinates

XG, YG, ZG-Guidance Coo rdinates

X B, Y B, Z B -Body ( LM Spacec raft) Coo rdinates

Figure 4 Lunar-des cent Coordinate Frames

10
 1) A m atrix element i s denoted b y two sub s cripts which indic ate the row
and column res pe ctively of the element. Thus C BP denote s the
XY
Y- component of the row vector �BP .
X
2) A m atrix transpose ( inve r s e ) i s denoted b y interchanging t ags.
3) F rom the definitions , it follows that m atrix p ro ducts are obtaine d b y
deleting internal tags .

By c onventions 2 and 3 , a vector V is transformed to body from gui d ance c oo rdin ates
by

-1 V
VB C BP C G P G = C BP C PG V G = C BG V G .

LM attitude angle s are a s et of E uler angle s defined as c lockwise rotation s

about the XB- axis ( yaw ), the displac ed YB- axi s (pitch), and the displaced ZB- axi s
( r oll).

LM gimb al angles are a s et of E ule r angle s defined as c lockwise rotation s

about the YB- axis ( inne r ) , the d i s plac ed ZB-axi s ( m iddle) , and the displac ed XB- axis
( oute r).

11
B R AK I NG -PH A S E AN D APPRO AC H - P H ASE G U IDANC E

The guidance p rograms for P6 3 and P6 4 are almost identi c al . The two p h ases
use the s am e guidan c e algorithm, the s ame Throttle R outine, and the s ame Powered­
flight Attitude- m an euver R outine. The differences are 1) the guidan c e equ ation s elects
different s ets of tar get s , 2 ) the erection of the guidance coordin ate f r ame is slightly
different, and 3 ) landing s ite r edes ignation c ap ability is available only in P6 4 .

Gu idan c e E quation Derivation

To guide a spacec r aft from any initial or current state to a s p ecified target
state c an be viewed either as an explic it guidance p roblem or as an implicit guid an c e
p r oblem. Explicitly, we c an r epetitively determine, as the m i s s ion progres s es , a
vector p olynomial function of time that inter s ects the cur rent and target states .
G uidance then commands the cor responding p rofile of ac celeration vs time.
Implicitly, we c an define, in advance of the mis s ion, a refer en c e traj ectory as a
vector p olynomial function of time that evolves backward from the target state but
c annot be expected to inter sect a disper s ed initial ( o r disper s ed current ) state.
Onboard guidance then commands an ac celeration vector p rofile c omposed of th r ee
term s, n amely the ac c eleration along the reference t r aj ectory m inus two feedb ack
term s p rop ortional to the deviation s in veloc ity and position of the actual t r aj ectory
with respect to the referen c e traj ectory. In either the explicit o r the implicit c as e,
rep etitively solving the guidan c e p roblem p roduces c onvergence upon the spec ified
tar get state even though the target point m ay be redes ignated in flight and th e
c ommanded ac c eleration is not p reci sely achieved bec au se of cont rol erro r s .

The implicit guidance equation derived here i s c ategoric ally superio r t o the
explicit guidance equation bec au s e the explicit equation is merely a spec ial c as e,
as will be shown. Besides bein g intellectu ally more s atisfying, the implicit equ ation
h a s demon strated in s imulations significantly fas ter r eduction of deviations from
the reference tr ajec tory. Deviation s come from navigation erro r s an d from disp lac ing
the reference traj ectory to inter s ect a redes ignated l anding s ite. R ap id reduction
of deviations restores a nom in al app roach to the r edesignated landing s ite.
U nfortun ately, the implicit equation had not b een developed when the program for
Apollo 1 1 was c oded, and the advantages were ins uffic ient to recode the guidance
p rogram for later mis sion s .

C her r /3) derived the exp licit guidance equ ation .

K lumpp •
(4 5)
simplified it
(6)
for LGC coding and gener ali zed it to nth o rder. Moore et a1 derived an impli c it
equation which did not generali ze the explicit equ ation. The gener al impli cit guidance
equation is now derived and spec i alized to the explicit c ase.

13
 It is convenient to think of the reference traj ectory as evolving b ackwards in
tim e from the target point, with the time variable T reaching zero at the target
point and negative prior to th at point. Thus target- referenced time ( T ) is to b e
dis tin guished from clock- time ( t ) . Bec au s e guidan c e gains w ould become unbounded,
th e target point is never reached . In s tead, a guided phase is terminated at a negative
time T and the succeeding phase is started . Both the terminu s and the target point
lie on the referen c e traj ectory, but the target point lies beyond the portion that is
(? )
actually flown, s im ilar to a suggestion of McSwain and Moore.

In term s of a vector polynomial func tion of target- referen c ed time, we wish

to define a referen c e traj ectory that satisfies a two - point bound ary value p roblem
with a total of five degrees of freedom for each of the three components . This
number of degrees of freedom is required in order to cons train terminal thru st in
PG3 and to shape the traj ectory des ign in P 6 4, as is discu s s ed in connection with
the targeting program.

A quartic polynom ial is the m in imum order with which five cons traints on
th e reference traj ectory c an be satisfied. W ith the reference traj ectory evolving
b ackw ards in tim e from the target point, it c an be defined as

(1)

where .B_RG is the position vector on the referen ce traj ec tory in guidance coordinates
at the negative tim e T, and R TG, .Y_ TG , ATG , .i_ TG , and � TG are the target pos ition,
velocity, acc eleration, j erk, and snap .

T h e acceleration to be commanded a t any poin t in s p ac e consists o f three

term s : the acceleration of the reference traj ectory at the p articular time T, minus
tw o feedback term s p roportional to velocity and position deviations from the referen c e
traj ectory. T aking derivatives of Eq. ( 1) a s the veloc ity an d acceleration o n the
referen c e traj ectory yields the three- term guid anc e equ ation

+
2
AC G ATG + .i_ TG T .§. TG T / 2

2 3
- ( .Y_G - V TG - ATG T -.l_ TG T / 2 - .§. TG T / 6 ) K V/ T (2)

where AC G is the comm anded acceleration, V G and R G are the current velocity
and pos ition, and K V and K are the nondimensional feedback gains .
R

14
 C o mbining like terms in E q. ( 2 ) yields

AC G = - RG K / T 2 - Y.G K / T
R v

+
2
B_TG K I T
R

+ Y:TG ( K V + K )/ T
R
(3)
+ ATG ( 1 + Kv + K / 2)
R

+ 1.T G ( 1 + K VI 2 + K R I6 ) T

+
2
.§.TG ( 1 / 2 + K vl 6 + K I2 4 ) T .
R

E quation ( 3 ) is the implicit guidan c e e quation. Although the refe r ence tr aj ec tory
is quartic, the traj e ctory gene rated by th e implicit guidance equation is obviously
not. The implicit equ ation c an, howeve r , be spe c i alized to the explicit equ ation -
which does gener ate a quartic - b y a spe c ific choice of the feedback gains K v an d
K F i rst we note that E q. ( 2 ) m ay be identified with the lin e ar second-order
W
differ ential equation

+
• • •

2 !; w X
2
X + w = 0
n n

by the associations (noting T is negative )

(4)

where w i s the undam ped natu r al frequen c y an d !: i s the d amping r atio . Of c ourse
n
the system is time varying. Howeve r, this assoc i ation does affor d some intuition
on gain setting. Solving Eqs. (4) yields

(5)

where P is the undamped period 2 rrlw and ,

n

(6)

Equ ation ( 5 ) p rovides a m eans of c ontrolling the system response time in terms of
th e nondimensional r atio P / T, and E q. ( 6) p rovides a m e ans of setting the d amping
ratio .

15
 An interesting set of values to choose for response and damping is

P/ T = - rr/f3', .t = {3' / 2 .

This choice yie lds

W h en th ese v alues are substituted into E q. (3 ), the result is the explic it guid ance
equation derived in refe ren c e s 3 to 5,

+ + +
2
AC G = 1 2 ( R TG - B_G /
) T 6 ( Y:TG Y.G /
) T ATG . ( 7)

The discussion of implicit vs explicit guidance is c oncluded by introdu c in g

the concept o f a sp ace c ontaining all p e rmissible combinations o f guidance p ar ameters.
I mplicit guidan c e - p arameter sp ace is one qu adr ant of the t, PIT plane or, equivalently,
one quadrant of the KR' Kv plane. E xplicit guid an c e - parameter sp ac e is a single
p o int in eith e r plan e .

E quation ( 7 ) presents the expli cit guidance equation assuming negligible

transport t ime d elay. The explicit equation progr amed in Apollo is cor rected to
c ommand an acceleration app ropriate fo r the time at which the ac c el e r ation is
p redicted to be achieved. Let this predicted target-refe renced time be

T T + LE ADT I M E
p

w h e r e LE ADT IME i s the transport delay due to computation and c o mmand execution.
An explicit guidance equation will now be derived that fits a qu artic polynomial
through the target position, veloc ity, and ac cele r ation, and through the current
position and ve loc ity. The ac celeration of the qu artic at the p redicted time T
is
P
the ac celeration to command at the current time T in order to realize the qu artic
p r ofile. It w ill be shown that the resulting guid ance e qu ation reduces to E q. ( 7) for
T = T , and the refo re E q. ( 7 ) gen e r ates a qu artic profile when the tran sport delay
p
is ze ro.

C onstr aining the actual traj e ctory to be a quartic fun ction of time allows the
cur rent position and velo c ity to be expr essed as

16
 l
[:: [:
T T2/2 T3/6 T4/24] RTG
1 T T2/2 T3/6 Y:TG (8)
ATG
.J.TGA
§.TGA
where !I TG A and .§ TG A are the j erk and snap which would be achieved at the target
point, and are not targets loaded into LGC m emory. Solving Eq. (8 ) for the j erk

l
and snap yields

[
[�TGA -24/T3 -18/T2 -6/T 24/T3 -6/T2] RTG
.STGA 72/T4 48/T3 12/T2 -72/T4 24/T3 Y:TG
(9)
ATG
RG
VG

The acceleration to be commanded at the current time T and realized at the predic ted
tim e T pi s

A CG = A TG + !L TG A T
P
+ §.. T G A T � / 2. (1 0)

Subs tituting E q. (9) into E q . ( 1 0 ) and s implifying yields the A pollo lun ar- des c ent

l
gui d ance equation

ACG [3 (T;) - 2 C:) "

+ [ {;y - C:)]· VG/T [ {:y - 6 C:) + + 1] ATG.

N ote that when time T is large in m agnitude comp ared to the transport delay,
T
p / T approaches unity. all bracketed coefficients in E q. (11) appro ach unity. and
Eq. (11)
approaches Eq. ( 7 ) identic ally. The net effect of E q . (11) not ach i eved by
Eq. (7) is a gain reduction as the target point is approached. Bec au s e Eqs. (7) and
(11) are identical for T = T, Eq. (7) generates a quartic profile when the tran s port
P
delay i s zero.

17
 In the derivation of the guidance E q. ( 7) or ( 1 1 ) , nothing c on strained the time
T. At any point in a guided phase, T c ould be s et to any arbitrary negative v alue,
and Eq. ( 7) or (1 1 ) would satisfy the boundary-value problem from that point forward.
Lan ding- site redesignation, which c an arbitrarily stretch or shrink the traj ectory,
would produce an unnec essarily s evere guidance response if T were not c or­
respondingly adj u sted. Becau s e T is arbitrary, it can be computed to s atisfy an
additional boundary c on straint. In Apollo, this additional c on s traint is imp o s ed on
the downrange (Z) c omponent of j erk . Thus the Z - c o mp onent of the j erk polynomial
of E q. ( 9 ) is solved for T by u s ing a target Z- component of j erk JTG 2 . Separating
th is scalar cubic p olynomial from E q. ( 9 ) yields

JTG
3
+
2 + + 6 VG z ) T + 24 ( R TG z - R G z )
z T 6 ATG z T ( 1 8 VTG z = 0. (1 2)

One root of thi s cubic i s th e required time T.

An altern ate c riterion for c omputing T reduces the propellant-consumption

p enalty of downrange landing- s ite redesignation s . Although exten s ively tested, the
altern ate was developed too late for inco rporation in the LGC p rogram. The altern ate
criterion sets the downrange pos ition error to zero. Thu s T is one root of the
quartic
2 +
T /24 + ATG z T /2
4 3
STG J TG Z T /6 + + VTG z T R TG
Z
- RG
z = 0.
Z

Braking- P h a s e T argeting Obj ectives

The near-optimal tran sfer provided by P 6 3 targeting mus t s ati sfy a throttling
c on s traint th at the DPS be operated within the permitted - th rust region for, nomin ally,
the final two minutes of the phase. This throttling duration ab s orbs dispersions in
D P S p erformance and errors in lunar terrain modeling. With a total p ropellant
(8)
c ons umption of over 6 6 0 0 - kg, Y ang
shows that the Apollo guidan c e and targeting
are w ithin 16 - kg of an optimal traj ectory s ati s fying the s ame throttling constraint.

To provide thru s t within the 1 1 to 6 5 o/o region for the last two minutes of P 6 3 ,
th e targets are chosen to p roduc e a constant thrust level o f about 57o/o o f rated
thrust at P 6 3 term inu s . The targeting p rogram accomp li shes thi s by con s training
the magnitude of the terminal thrust- acc eleration vector to be F I M, and con straining
the Z- component of terminal j erk to be ( es s entially ) K F i\1/M2, where F is the
required terminal thrust, l'vl is the es timated terminal mas s , M is the estimated
termin al mas s flow rate (negative), and K is a j erk coefficient to account for the
verti c al component of thru s t. The targeting program achieves the two minute duration
of constant thru s t by adj u s ting the initial range.

P rop erly targeted, the guidance algorithm commands during mo st of P 6 3 a

th ru s t- acc eleration in exc es s of what can be achieved . The throttle routine multiplies
thi s thru s t- ac c eleration command by the estimated mas s to yield the guidance thrus t

18
c om m and (GTC ), and p rovide s m aximum thrust until the G TC falls below 5 7%.
Figu re 5 illustrates the profiles of GTC and actual thrust for a p roperly targeted
P 6 3 phase and shows the effects of adj ustments of the ignition time and of the DPS
th rust level.

The targeting program computes the P 6 3 targets by p r oj ecting computed

termin al conditions forward typic ally 60 seconds . Although the targets are p roj e cted,
they are computed to p roduce the required terminal condition s on a nomin al
traj e ctory. T r aj ectory disp e r s ions c annot be elimin ated prior to the target p oint,
but they c an be reduced suffic iently by terminus to achieve the targeting obj ective s .
Both the ignition time and the p rojected targets are c omputed iter atively us ing a
d e s c ent s imulation in the ite ration loop.

App roach- Phase T argeting Obj ective s

The P 6 4 targets are computed to provide lunar - surface vi sibility until about
5 s e c onds before terminus, and c ontinuous throttling. In addition, the P 6 4 t argets
are computed to p roduce at te rminus a m atched s et of values for the Z-components
of ac cele r ation, veloc ity, and position such that, in the nominal c as e , the P 6 6 algo rithm
w ill p ro duce no initial p itch trans ient and will s imultaneou s ly null the Z- compon ents
of veloc ity and position. Unlike P 6 3 , the P 6 4 reference traj e ctory c an be determin e d
i n c lo s e d form from specified traj e ctory constraints. Thus t h e proj ected targets
are computed w ithout nume ric al ite r ation.

Landing-Site R edes ignation P roc edure

To steer th e LlVI via the autom atic P 6 4 guidance to a visually s elected landing
s ite, the com m ander u s e s an ite r ative proce dure akin to steering an automobile.
The proce dure c on s ists of 1 ) identifying the current landing s ite where the LGC
w ould take him in the abs ence of intervention and 2 ) steering the c urren� s ite into
c oinc idenc e w ith his visually s elected s ite by c omm anding incremental l anding- s ite
displac ements ( rede s ignation s ) . Be c ause the P 64 targets are define d in the guid an c e
c oordinate fram e , which i s repetivively e rected through the landing s ite, t h e P 6 4
target p oint is disp lac e d ac cordingly.

To identify the currently s elected l anding s it e to the astronauts, the LGC 1 )

o r ients the LlVI about the thrust axis to sup e r impo s e the landing point de sign ator
( L P D ) reticles ( s ee F igu re 1 ) on the current s ite and 2 ) display s a number which is
r e ad by the LM p ilot and voc ally relayed to the commander. By sighting through
th e indicated point of the LP D reticles, the c om m ander identifie s the current s ite.
H e registe rs his eye by superimp o s ing the two LPD reticles, one of which i s p ainted
on the inside window p anel, and one on the outs ide window p an e l. The s ep ar ation
between reticle s is 2 . 5 em .

19
 120 Guidance Thrust Command ( GTC) Profiles
--- NOMINAl
-=--=-
��----- ABOVE NOMINAl OP'; TH�U'lT
-�- EARliER THAN NOMINAl DPS IGNITION

100
� �, ::=::.:: :.::-
===-===A:::: ct� ua Thrust Pr ofi , e �
�-
:::.::; ::...:;:::-= --- .- --- -----
I-
V)
'�" , :
::::>
e:::
80 '�\ ii
1L_
:r:
I- \\. Braking phase terminus
0
approach phase in eption
I-
I...LJ
\
<(
60
e:::
""
0

V)
a_
0
GTC and actual thrust coincide
LL. 40 26 seconds to trim thrust through center of mass.
0
I-
z
u
I...LJ

e::: 20
a_
I...LJ

t)" Time of first Guidance pass

( GU IDTIME)
0 /.
50 100 150 200 250 300 350 400 450 500 550
TIME FROM IGNITION- Seconds

Figure 5 Actual Thrust Profile s and Guidance Thrust C ommand P rofile s During P 6 3
 By m an ipulating his c ontroller left, righ t, forward, or aft, the commander
d i r e c ts the LGC to displace the landing site ( and the P64 targets) along the lunar
°
surface by a c o rrespondingly dire c te d fixed angular inc rement ( 1 ) with respect to
the cur rent line of s ight.

The LGC redirects the th rust to guide to the rede signated (now current) s ite,
and reorients about the thrust axis to m aintain superp o s ition of the reticles on the
current site . The c om m ander c an c ontinue thi s rede signation p r o c e s s - s te e r ing
th e current l anding site into c oinc idence with his chosen s i te - until 1 0 s econds
b efore reaching the P64 target p oint, at which time P 6 6 is initiated.

F igure 6 illustrate s the P 6 3 , P 6 4 Guidance Algorithm. As shown in F igure 3 ,

the algo rithm receives guidance targets , the current s tate vector, and the current
g r avity vector as inputs and i s sues a th rust- ac celeration com m and, a unit thrust
c omm and, and a unit window command as outputs.

Be c ause the landing site move s due to lunar rotation and l anding- site
redes ignation, the L M is guided with respect to the guidan c e coordinate frame, which
i s e r e c ted th rough the landing site e ach p a s s . Guidance targets are fixed in thi s
flo ating frame. O ther inputs and all outputs are exp r e s s e d in platform coordinate s .

The landing site ve ctor 1-_ P i s up date d for lun ar rotation (Eq. ( 6 . 1 )) using an
approxim ate algorithm that avoids c omputation of tr igonometric func tion s, yet
p r e serves the m agnitude of the lunar r adiu s . The algo rithm accounts for the lunar
rotation rate W :YIOON P and the elap s e d c lock - tim e s ince the p r e ce ding upd ate (t -
tO LD ).

For the landing-s ite redesign ation algorithm ( Eq s . ( 6 . 2 ) - ( 6 . 7 ) ) , whenever

th e c omm ander m anipulate s the c ontroller ( F igure 1 ) in the autom atic mode, the
LGC is interrupte d and the azim uth command c ount ( N C AZ) or the elevation comm an d
c ount (NC EL) i s inc remented or decremented ac cording to the dire c tion o f
m anipulation. The rede sign ation algorithm fetche s and r e sets to zero the NC AZ
and N C EL ac cumulators and rotate s LOS P ( the unit line - of- sight vec tor to the cur rent
°
landing s ite ) by 1 per c ount (Eq. ( 6 . 3 )). If NC A Z and NCE L are both zero, the
redes ign ation algo rithm has no effec t. G iven that attitude c on trol m aintain s
c oincidence of the ZB,X B p lane and 1-_0S P , the rotations of LOS P are about two
axe s normal to 1-_0S P . E levation redesign ation s rotate LOS P about the Y B- axi s ,
and azimuth redesignations rotate LOS P about an axis normal to LOS P i n the ZB,X B
p lane . The landing- site redes ignation geometry shown in F igure 7 depends upon
the defined LM platform orientation, n amely that the X P - axis i s n e ar verti c al through
th e landing site. The con s tr aint th at LOS P be at least as negative as - 0 . 0 2 ( E q .
X

21
( 6 . 5 ) ) p revents redes ign ating the landing s ite beyond the horizon. E qu ation ( 6. 7 )
c o m pute s th e di splac e d point near the surface shown in F igu re 7 and plac e s the
r e d e s ignated s ite directly b ene ath this point.

The displaye d LPD angle ( OL P D, Eq. (6. 8 ) ) is the angle between LOS P and
the ZB- axi s .

The co mputation o f the s tate vector in guidance coordinates ( Eq. ( 6,9 ) ) places
th e origin of the guidance frame at the landing site and yields the veloc ity of the
LM re lative to the lunar surfac e .

Target-refe renced time T i s c ompute d u sing N ewton's m e thod s tarting with a

good estim ate ( E q s . ( 6 . 1 2 )- ( 6 . 1 3 )). N o te that the denomin ato r of Eq. ( 6. 1 2 ) i s the
de rivative of the num e r ator .

T h e guidance equation ( Eq . ( 6 . 1 5 )) is identi c al t o E q. ( 1 1 ) . The thrust­

acceleration c om m and ( Eq s . ( 6 . 1 6 ) - ( 6 . 1 7 ) ) is merely the total acceleration command
m inus current gravity G P , and the unit thrust command ( Eq . ( 6.1 8 ) ) i s the direction
of the thrust- ac cele r ation comm and. The s o - c alled r adial - acceleration guidance
c o rrection des cr ibed in reference 9 is rende red unne c e s s ary by current targeting
te chniques and is om itted from thi s report, although presen t in the LGC p rogram .

T h e c omputation o f the unit window command presented h e re i s a s implific ation

of the LGC c oding which produ c e s the s ame result. The obj ect is to keep the landing
s i te in th e c enter of vis ion ( superimp o s e the L P D reticles on the current site ) whenever
the geom etry permits and, otherwise, to command a forw ard - fac ing attitude. F igure
8 show s why the landing site c annot always be kept in the c ente r of vision. F igure
9 shows the geometry p e rtinent to c omputation of the unit window command QN WC P .
C om m anding the line - of- s ight vector ( !JN WC P = LOS P ) alines the retic les with the
landing s ite; c omm anding the forward vector ( .!J.N W C P = FO R P ) produces a forward
fac ing attitude. If the fi r s t alternative is chosen ( _!2N W C P = LOS P ) the LM will
rotate about the X B-axis to aline the Y B- axis with the ve c tor LOSP x
C BP . Thus
X
the direction of LOS P X C BP indic ates whether a normal forward- facing attitu de
X
o r an abnor m al attitude would result from the c ommand _!2N WC P = LOS P . In addition,
the m agnitude of LO S P x C BP
measures the degree of indete rmin acy in the
X
c om m and QNWC P LO S P . The proj e c tion ( P ROJ, Eq. ( 6 . 20 ) ) of LOS P x c;2BP
=
X
on the Y G - axis detects both the m agnitude and the direction of LOS P x C BP .
X
Thus P RO J i s used as the c rite rion for mixing LOS P and £.0R P into UN W C P . If
the d e s c ent traj ectory i s planar, the m ixing ( Eq. ( 6 . 2 1 )) yields .!J.N W C P LOSP for
=

FORP for OLP D � 7 5 , and U N W C P a mixture line ar w ith

° °
nL P D $ 6 5 , .!J.N W C P =

c o s rJL P D for 6 5 < ALP D < 7 5 . R egardle s s of whether the tr ajectory i s p l an ar

° °

o r nonplan ar, it is neve r p o s s ible to com m and a side - facing or a r e ar - facing attitude.

22
 ------ ---- --- -�
--- - -
+
COMPUTI T SUCH AS TO SATISFY TIRMINAL JERK CONSTRAINT
JTG T3 • 6 ATG Tz +I 18 VTG • 6 VG IT+ 24 I RTG - RG l
INPUT VARIABLES: t.T. --�-
Z Z --
� �--�· ---
Z Z
�-
z z
- --- ---- - 1!21
3 JTG T2 + 12 ATG T +18 VTG • 6 VG
NCAZ, NCEL
Z Z Z Z
REDESIGNATION COMMAND COUNTS
CURRENT STATE �.� T • T +t.T 1131
CLOCK-TIME TAG OF CURRENT STATE t
!!.!'

��'
CURRENT GRAVITY
BEGIN I

� UPDATE LANDING SITI VECTOR FOR LUNAR ROTATION

lP. LP UNIT [.b.P tIt- tOLD I �MOONP X .b.P J Ill
I
I
GUIDANCE EQUATION
Tp • T +LEADTIME
2 z
1141

�CG ·(3(f-)-z {;)]

T T T T
121J!,TG J!GIIT2 •[4(�f) 3(nl6YTGIT

<$,?�'
LANDING-SITI REDESIGNATION ALGORITHM
J.OSP ·UNIT I J.P- J!PI
E

64

121
•(z( n n]
T ?
-(
T
6YGIT •[6 ( {)
T 2

COMPUTE THRUST-ACCELERATION COMMAND AND UNIT THRUST COMMAMJ

· 6
T
(-f-) +1] �TG 1151

J.OSP • J.OSP + 0. 01745 NCAZ !;,SP y

.lfCP • CPG ACG-liP 1161
131 AFCP • I�Fcpr
+0. 01745 NCEL I �BPy X lOSP I !J.tf'CP ·UNIT I �FCP I
1171
1181
l'.:l NCAZ • 0, NCEL • 0 141
w
LOSf'x • MINIMUM I LOSf'x . • 0. a? I 151 •
)
COMPUTE UNIT WINDOW COMMAND
!,OSP ·UNIT llOSP I 161
£0RP • UNIT I �BP X !;_GPy I IJq)
LP - RP
!lJ' +,lOSP
X

(
X X
!,P LP UNIT 171 PROJ ·l,OSP X £8Px • £GPy
•
1201
LOSP
X l
yNWCP ·UNIT MAXIMUM I PROJ ·COS 65� 0 I J.OSP 1211

• +MAXIMUM I COS 75 °- PROJ. 0 I £0RP]

L LANDING POINT DESIGNATOR ANGLE DISPLAY -'
9 LPD ·- ARCSIN ll,OSP • £8P I
I 181 1
�
r
X

�
K ·I L P63 P64 J K- 0
; I I
COMPUTE CURRENT STATI VECTOR IN GUIDANCE COOROINATIS I
.!!,G • CGP I.!!!'· .IJ' I, ':J.G • CGP l,l!P- �MOONP X J!.P I 1q11 ERECT GUIDANCE COORDINATE FRAME FOR SUCCEEDING PASS

+ �GPx ·UNIT I J.P I IZ21

UPDATE TARGET REFlRENCED T I ME T & COMPUTE CONVERGENCE CRITERION FOR T �GPy ·UNIT X {J!P -lP- K I yp · �MOONP X J!P I T /4
[1r 1231 B
T • T + It- tOLD I, tOLD • t 1101 �GPz ·kGPX X kGPy 1241
t.TCRIT • T/128 1111 THRUST-ACCELERATION COMMAND AfCP
OLJTPUTSo UNIT THRUST COMMAND J.!NFCP
UNIT WINDOW COMMAND J.!NWCP

END

Figure 6 P63, P64 Guidance Algorithm

 XP

pp
X X
R p + LOSP
LP. -
DISPLACED POINT _

X
NEAR THE SURFACE-- - LOSP.

PLANE

NEW LANDING SITE LP

O LD LANDING
SITE LP

CENTER
OF MOON

Figure 7 Landing- s ite Rede signation Geometry

24
 XG
THRU ST VECTOR

THRUST V EC TOR

LM ATTITUDE
B AC KWARDS

'

�
..
.
N
.

·�
· . · /. ·
Ul

.
:�. _ ,:, N·.z.
;�.
41':'
.
I
- .
-- - ·- -

:-
_ :...: --- • '
= C � RRE NT S I T E
--

.. -
--.

�.-::� "'-"-;'i , ·- _/ · -;c f'

' -.1: :.
.

-�:.� -- ..: �,· , '.:

.
- ,,.,.;

·1� 7
.

·:,r -:
::. : :: . ·..:
� · :
::--: �.:� ·:- . . .

,. �·�·: ···;y
I

;.'"
.

.·. � : .f-.:.--�--
. .

� --
-, ..

.;_��} /
.

/. - - ,. .--e-·:z--·--·�"t.--:-- , :.·
-
'\ . .

- -
-

.
. .. .... - --· ·- ·
· • •

..
-

-· � -.:. ""' _ --- ---

�>:r. - · :..
--
-
·
.. �· -
,. �=
·,
- " :·
.. ... .
.- .. :f- : .2
.

a"' � i ..
.

,� :;: )- .
.
. .

('\ w � � ,
.

<:. ·;�:;=:-� .,., - '·_;;; ,

- -

-:: · : - -
"
-
. .

...�-
, , , -
-

�
o·• - --"'-' - - : :_ ,
____ __

,. � "-:J:?. ' ',...;,. ·;- - .

-

?
, -
.

-
-<-
• . . ' -· ·

c
'
.
,•
.

"vf>.' ""
.
__

--
_,;,., , ": ·"<-""
. ..

-�
-
· < '

-
.�/;;-_;-:/"· .._
-
:-
·
.
�

•V"- -
.
'

�· y-- '- .:·.:- ! �

t� :<
. •- - ' p
.. -
·

-.-" '�... 2- �
... - ... ·- '

�
- · •• •

Figure g W hy K e P p i n g t h e Landing Sitf• i n t lw C c ntl· r o f V i s i o n C a nnot be t h e Sole

C r ite r i on fo r C ont roll i n g Attitude . \bout the T h r u st <'\ x i s
 XG

HT VECTOR
(
LI NE- OF-SIG APPROACH
PHA SE
TE RMI N US
X r AXI :l � BFx
/FO
!:_O S P = U N I T
{ A f Pf< OX THf\
=
( !:_P - � P )
UST D / � )

----/----f=Jt::�::��:::j=;���
' RwARD VEC TOk
0 FO R P = U NIT (5;8P lC
£G Py }
��
X

RANGE

CU RRENT
S I T E LP
�
� YG fG Py
=

CR O SS RANGE

Fi gu re 9
Pe rti ne nt to
Ge om etry
C o m pu tin g
the Un it Win
do w C o m m and

26
 E r ection of the guid ance c oo rdinate fr ame ( E q s . ( 6 . 2 2 ) - ( 6 . 2 4 ) ) is illustrated
in F igure 1 0 . W ith K = 1 in P 6 3 , the gui d ance coordinate fram e orientation about
the verti c al X G - axis is such that the Y G - component of j e rk would reach z e ro at the
target p o int if the traj e ctory were flown there ( s ee reference 5 ) . With K = 0 in
P 6 4 , the ZG - axis i s in th e verti c al plane c ontaining the line -of- s ight vecto r. For
c ro s s r ange landing- s ite rede signation s , s etting K = 0 in P 6 4 w as found to c onsume
le s s DPS propellant than s etting K = 1 to null the c r o s s r ange j e rk at the tar get
p oint.

P 6 3 Ignition Algorithm

T r aj e ctory dispe rsions p rece ding P 6 3 require an ac c u r ate ignition time an d

attitude to be com puted to 1 ) avoid exce s s ive variation s of the time dur ation of
th rottle control in P 6 3 and 2 ) to avoid commanding an exc e s s ive attitude trans ient
the fi rst time the P 6 3 , P 6 4 Guidance Algorithm is proce s s e d. The P 6 3 ignition
p r ocedure consists of:

1) C omputing onboard the p re c i s e ignition time and attitude about 1 0 minutes

in advan c e of ignition
2) O rienting the LM to the ignition attitude
3) Initi ating re action control system ullage 7 . 5 se c onds p r ior to ignition
4) Igniting the DPS at m inimum thrust and holding con stant th rust and
c on stant attitude for 2 6 s ec onds, the m aximum time required for the
D AP to orient the DP S trim gimbal system to point the th rust vector
through the LM c ente r of m as s
5) Conne cting the guidance algor ithm , which immedi ately c omm an d s
m aximum thrust and b egins commanding an attitude p rofile ac cor ding
to the current state vector and the P 6 3 targets .

To dete rmine the required ignition attitude, the ignition algorithm ( F igure
1 1 ) c alls the guid ance algorithm as a subroutine. The ignition algorithm supplies
inputs c on s i sting of an ac curate extrapolation of the state vector an d the corre sponding
gravity vector (both valid at G U IDTIME, the e stim ated c lo ck-time of the fi r s t P 6 3
guidance pas s ) . In p reparation, Eqs. ( 1 1 . 1 ) - ( 1 1 . 5 ) initi ali ze gui d ance algorithm
input s . On the first ite r ation, the state vector ext r apolation rep resented by Eq.
( 1 1 . 6 ) i s p e rfo rmed by an o rbital integration rouhne and, on sub s equent ite r ation s ,
h y a K e p l e r routine . Equation ( 1 1 . 9 ) corr ects the extr ap olated velocity vector by
the veloc ity inc rement imp arted during the 2 6 s econds of minimum th rust p r e c e ding
th i s p oint . ( The errors due to not correcting the extrapolated position vector and
not c o r recting for ullage are negligible . ) The gui d an c e algorithm p rodu c e s a unit

27
 x a === £
Z G ::= C
�
G P.
- z . ------ G U i
do n c e
ap - --- I YG ==: c
Coa �at·
n ate F
x cp ra m e
Ve rt ica XG YG
l - Y zc
Th ro u g
h
Cu r re n
t
L an d i n
g
S ite

Cu rre
n t Lan
d i ng
S ite .L p

_
Trajec
to ry to
!Redes Cu rre n
igna te t
d 1 La n
d in g S
ite

- K I J(p
- WM
OONp
X BP
LM Ve 1 T/4
loc ity
Re lat iv
L u na r e to
K � {1 0 P64
P63
S u rtac
e

T • Tar
g et - Re
fere nc
u rre n t ed Tim
L M Po e
Fi gu
s it io n
re 10
BP
 INPUT S
C U R iif NT STAT£ ON COASTING TRAJECTORY RP_ yP
CLOCK - T I M£ TAG Cl CU RRENT STAT!
t
E STIMATtD Cl OC K - T I M£ FOR f I RST P6J GU I D ANCE PASS GU I DTIME
THRUSJ.ACCEI F A A T I ON Cl Z o S F C C1 M I NIMUM THRUST AFTRIM

BEG I�

SAVE COPY (I WARfNT STArt J!P

l YP ANO ITS C l OCK-TIME TAG I

t
I N I T I Al \ l£ FOR P6l. P6il GU I D ANCE AlGORITHM
T · · oM � 'fC 1\1
tOtO • G U I D T I ME IZI
LP · L P I G U I D T I ME
111

I
'
I

0I
I 0 0:
CGP • 0 I
} I
14 1
0 0
U Nf C P • I 0. D. -I I Ill

t
SET LOOP COU NTE R

�: : �: \G�� �DT;;::/1 y• · _y:P

EXTRAPCUTf COASTING STATE AND GRAVITY T O GIJ I DT I ME
.utD ,.,E 1 t • 161

"N
GUIDTIMI

-
171
· l 181
�-

t
CORRECT EXTRAP(lATtD V I L OC I T Y FOR 16 SECONDS Cl M I N I MUM THRUST

I VP • VP • Af R I M UNIT I Z6 SEC _191

J
I CAL P6l P6il GU I D ANCE ALGORITHM O E C REM£NT LOOP COUNT!R I
I N • N - I I lUI
I

NO

<?
ADJUST GUIDTIME FOR TRAJECTORY D I S PE R S I ON S
s

� GU I D T I 'IE • ' KX I RG - R R R I G
- X X 1
, K y Rr.
/ 1111

' I RG - RAR IC I � K I 1/G VRR I G t l

7 l V
r � vc 7 + K x \J f, x l

GUIDT IMF GIJ I D T I MF • �Gli i D T I M F f l;'l

�
YES

PR[PARf fOR I G N I T ION

I I I G N I T ION I · GU I O T I MI 16 SFC l}3f
I I ULLAGE 1 • \ I I G N i fiON I - 7. 1 SEC 1 14 1

�
Rf S TOR!. CURRENT RP. VP ANO ITS ClOCK - T I ME TAG I

�
OUTPUT� :
END

t 1 1 c "'J 1 1 1 or� f
1 l lJI I I\r,f I

T H R l r ) l O I RI Cl tON R f () I I I R I D Al t r; N i l l r H � i)Nf r _ f1

Pfi 3 I gnition Algorithm

29
th rust comm and !l_f',; FC P , which is the direction to point the XB- axi s . Bec ause the
d i r e ction of the velo city correction is unknown on the fir st iter ation, the above
p r o c e dure is iterated th ric e .

An outer ignition - algorithm loop accounts for dispersions with re spect t o the
nom inal t r aj e ctory. E quations ( 1 1 . 1 1 ) - ( 1 1 . 1 2 ) adjust G U IDTIME to correct the

Z
HG component of pos ition at G U IDTIME as 1 ) a line ar function of the dispersion
in o rbital speed VG and of the dispe r s ion in the R G X component of position ( e s s entially
altitud e ) and 2) as a quadratic function of the out-of-plane position R G y · R B R IG
X
and R B R IG are nomin al initial altitude and r ange components of p o s ition in guidan c e
Z
c oordin ate s ; V B R IG is th e nom inal initial speed; and K , K , an d K a r e correction
x y v
c oefficient s . The nominal initial altitude , range , and speed a re computed by
the targeting p rogram. The c o r rection coeffi cient s are c omput e d u s ing a m anual
p rocedure based on descent simulations.

\V hen c onve rged, this proc e s s yield s a precise time and attitude for igniting
the D P S . T r aj ectory dispe r s ions r e s ult in typ ical variation s of 2 seconds in the
time duration of throttle control and typ i c al attitude trans ients of 2 millir adians
c o m m anded by the guidance algorithm on the first P 6 3 p as s .

30
T E R l\I IN AL- DESC E N T - P H ASE G U IDAN C E

Hori zontal and ve rtical velocity are controlled in P 6 6 by completely independent

algorithm s . P 6 6 p rovide s a non autom atic attitude -hold m ode in which the commander
c an c ontrol the L M att itude to translate or not, as he wishes, hori zontally over the
luna r surface . P 6 6 i s sue s no unit w indow command; yaw is controlled manually . A
( 1 0)
d e s c ription of P 6 6 including the nonautomatic mo de s is p rovided by Eyle s .

P 6 6 Hori zontal Guidance Algor ithm

The P 6 6 hori zontal guidance algorithm ( F igure 1 2 ), p roc e s s e d once every two
s e c onds, nulls the hori zontal components of velocity relative to the lunar surfac e
by directing the th rust vector a s m all angle aw ay from verti c al in oppo s ition to
hori zontal veloc ity. The hori zontal algorithm neither m e asures nor commands
th rust- accele ration m agnitude; the algorithm is derived on the as sumption that the
ve rtical comp onent of thru st- accele r ation equals lun ar g r avity.

Just as veloc ity feedb ack damp s a p o s ition control loop, ac celeration feedbac k
dam p s a velocity control loop . Because of the s ampled - d ata char acter o f the system,
a good m easure of current ac celeration is the ac cele r ation commanded the p re c e ding
p as s . The P 6 6 hori zontal algorithm feeds back the veloc ity error ( current velo city

Z minus lunar surface veloc ity V MOON P y , V MOON P z > an d, to p rovide the
VP 'VP
Y
r equired dam p ing, feeds back a fr action of the thrust- acceleration command from
the preceding pass ( Eq s . ( 1 2 . 2 ) - ( 1 2 . 3 ) ). On the first P 6 6 p a s s , the thrust­
acceleration fed back i s th at com m anded the final P 64 p a s s .

°
The direction o f the thrust - acceleration comm an d i s limited t o 2 0 from
verti c al ( E q s . ( 1 2 . 4 ) and ( 1 2 . 5 ) ) to m aintain a ne arly e rect L M attitude. fhe LIMIT
fun ction of two arguments limits the m agnitude of the first argument to the v alue o f
the s e c ond argument.

The unit th rust com m and ( E q . ( 1 2 . 6 ) ) i s the direction of the limited th rust­
accele ration com m and.

The as sumption in gene r ating hori zontal comm ands that the verti c al c omponent
of thrust- accele r ation equals lun ar gravity ( E q . ( 1 2 . 1 )) is re ali zed only if the L M
i s not ac celerating verti c ally. The purp o s e o f ignoring ve rtic al acceler ation i s to
elim inate coupling from ROD inputs to LM attitude . The effect of verti c al acceler ation,
which occurs wheneve r the commander m anipulates the R O D switch, is to modulate
the gains of the hori zontal chann e l s . T h i s g ain modulation i s negligible b e c au s e

31
only limited changes in the d e s c ent rate will eve r be commanded; the vertical
accele ration c an be s ignific antly non ze r o only for short p e riods of time.

P66 V e rtical ( R O D ) Guidance Algorithm

The ROD guid ance algorithm, proce s s e d onc e p e r s e c ond, c ontrols altitude
r ate to the refer en c e value by throttling the D P S. The RO D algorithm has no control
ove r the L :\1 attitude; the thrust- ac celeration command it i s su e s accounts for any
non- vertical ori entation of the thrust vector.

The obj ect of the ROD guid ance is to re spond r apidly without ove r shoot to
R O D in c r e m ent com m an d s . The algorithm p rovides a time constant o f 1 . 5 seconds,
even though the s ample interval i s 1 . 0 s e c ond, by c ap itali zing on the s ampled- d at a
characte r o f the system. U s ing a computed e stim ate o f the total acceleration at
the R O D s ample instant, the R O D algorithm extrapolates s ampl e - instant me asured
veloc ity by the effective transport lag of 0 , 3 5 second and thus comm ands an
acc eleration app ropriate for the velocity error at the time the ac celeration comman d
will be realized. A s ampled data analysis ( referen c e 1 1 ) shows that the compen s ation
fo r effective tran s port lag is highly effective in stabilizing the verti c al channel.
The signific ant s ystem dynamics reduce to a s ingle zero and two poles in the Z
plane . The zero i s

ZZ = - L AG / ( s ample interval - L AG ) = - 0 . 3 5 / ( 1 - 0 . 3 5 ) - 0 , 53 8 .

One pole i s at the or igin, and the s ec ond pole is

Zp = (time c onstant - s ample inte rval ) / time c on stant = ( 1 . 5 - 1 ) / 1 . 5 = 1 / 3 .

T h e poles are the s ame as f o r an ideal system containing neith e r a t r ansport l ag

n o r· an extr apolation.

The HOD algorithm has b een s implified for this report as follo w s :

1. I n the LG C c oding, the ROD algorithm begin s e ach p a s s by reading the

acc elerometers and recording the time at which they are r e ad. This time is c alled
the H O D s ample in stant. ROD sample instants oc cur irregularly, but the interval
b etween them, c alled the ROD s ample interval, ave r ages 1 s ec ond. The accelerometer
r e adings are used to c ompute a) the th ree - component current velocity vector valid
at the R O D sample instant, based on updating the velocity vector supplied by the
state vector update routine (SVUR, F igure 3 ), and b) a thrust- ac cele r ation

32
m e a s u r e m ent w h i c h is the ave r age ove r the ROD s ample inte r v al . To c ompute the

veloc ity vecto r it s upp l i e s , the SVUR al s o reads the ac c e l e r o m e te r s . e ac h p as s . at

the S V U R s am p l e instant o c c u r ring at r e gular 2 - s e c ond i nt e r v al s . The i r regular

H O D s a m p l e i n s t ants are e s s enti ally asyn c h ronou s with the r e gular SVUR s am p l e

instan t s . C on s equently t h e int e r ac t i o n s b e t w e e n t h e R O D algo r i th m an d the SVUR

in up d ating the S V U H - s up p l i e d velo c i ty v e c t o r are ext r e m ely int r i c ate . In th i s r ep o rt.

the d ata p r o c e s s e d by the R O D algo r i thm are s h own as input s . H o w the s e inputs

a r e o btained i s d e s c r i b e d in r e fe r e n c e 9.

2. Although the ve rt i c al o r i entation of the X P - ax i s is c ap itali z e d upon by s eve r al

LG C routin e s , including the P66 h o r i zontal algo r ithm an d the l anding- s it e

r e d e s ignation algo r it h m . t h e LG C R O D algo r i th m l ab o r io u s ly m an ipul ate s c om p l e t e ­

v e c t o t · s t ate d at a to m aintain validity for any p l atfo r m align m ent. Presented here

i s t h e s c alar equivalent val i d f o r the lun ar - l an d ing platfo rm ali gn m ent.

F i gure 13 shows the HOD algo r ithm . The input s are all v al i d at the ROD

s am p l e ins tant. E qu ation ( 1 3 . 1 ) c o m p u t e s the s amp l e - i n s t an t total ve rti c al ac ­

c e l e r ation by adding, to the t h r u s t - ac c e l e r ation m e a s u r ement ( ave r aged ove r the

ROD s am p le interval ) , c u r rent g r avity and a c o r r e ct ion fo r the throttle c h an ge

c oncluding the p r e c e ding H O D p as s . The th r u s t c o r r e c t i on i n c r e m ent oF A is supplied

by t h e Th rottle R o utin e . l.c:quation ( 1 3 . 2 ) ext r ap o l at e s t h e s am p l e - in s t an t m e as u r e d

v e lo c ity . The c o mm an d e d ve rti c al veloc ity ( r e f e r e n c e alt itude r at e ) i s initi ali z e d

a s the ve rt i c al veloc ity exi sting a t t h e time P66 i s initiated. and i s in c r e m en t e d o r

d e c r e m e nted b y E q . ( 1 3 . 3 ) e ach R O D p a s s ac c o r ding t o the R O D c o m m an d s i s s u e d

by t h e c om m an d e r s i n c e t h e p r e c e d ing R O D p as s . E q u ation ( 1 3 . 5 ) f i r s t c ompute s

th e total v e r t i c al a c c e l e r ation requ i r e d as the n e g ative of the ext r ap o lat e d velo c i ty

e r r o r divided by the R O D time c o n s tant ( 1 . 5 s e c on d s ) . The equation then obt ain s

the r equi r e d v e r t i c al t h r u s t - ac c e le r ation by s ub t r acting c u r r ent g r avity. F in ally.

dividing by C B F'xx • w h i c h is the c o s in e of the angle between the X B - ax i s and the

v e r t i c al X P - axi s , Eq. ( 1 3 . 5 ) yield s the t h r u s t - ac c e l e r ation c o m m an d A F C P . To

av o i d an e m p i r i c ally d i s c ov e r e d i n s t ab ility wh i ch o c c u r s when the th rottle routine

o r the DPS c annot c o mp ly with the th r u s t - ac c e le r ation c o m m an d f r om P66, Eqs.

(13. 6) an d ( 1 3 . 7 ) r e s t r i c t AF C P t o p r o d u c e th r u s t within the p e r m it t e d - th r u s t r e gion .

33
 I NPUT,
CU RRENT VHOC I T Y V P

BE G I N

j
C CWI PUT£ UNL I M ill O THRUST-ACCELERA T I ON COMMAND
----- - - -

AFCP GM
Ill
•
X

AFCP · - I V P - VMOONP I I 5 SEC - 0. 4 AFCP (21

I V P - VMOONP I I 5 SEC - 0. 4 131
y y Y Y
AFCP · - AFCP
Z Z Z z

L I M I T COMMANDED THRUST D I RE C T I ON T0 2D° FRCWI VERTICAL

AFCP · L I M I T I AFCP , AFCP tan 20 ° I 141

y Y X

AFCP
z
· L I M I T I AFCP
z
, AFC P tan 2D0
x
I 151

I S SUE U N I T THRUST COMMAND

VNFCP · U N I T I AFC P I 161

OUTPUT:
U N I T THRUST COMMAND L!NFCP

£NO

Figure 1 2 P 6 6 H o riz ontal Guidance Algorithm

34
 I NPUTS :

COUNT Of ROD I NP I I T <;

S MIPLE l �i S T A NT M E A S U R E D VELOC I TY

C U R R E NT GRAV I TY

T Y R l! S T � >IC C : : E R AT I n c ; \lE A S U R f l.lf'H

I Averaqed over t h e ROD sample l nl n 'l a l l
T H R U S T CORRECT I ON I N C REMENT
I f rom the Throttle Rouline I

C U R RF NT M A S S E ST I MATT M

BE 1. 1 �

-�- � -�-�� ___ ____ ...______________,

C OMPUll EX I ST I N G V E R T I CAL A C C E L E RAT I ON AT ROD S AM P L E I N STANT

l
E X T R A POL ATE S AM PLE I N'iTANT MI A S I I R F D VELOC I T Y BY f F F FC T I VE T R A N S PORT l A

U P D ATI COMMANDED VE RT I C Al VFl OC I T Y I N C O RPORAT I NG ROD I N P U T S

VCP - VC P • NROD D. l mlsec ill

X X

NROD - D 141

COMPUll THRtJ S T � A ( Cf lE RAT I ON COM M AND F O R T HR OIT lf ROI I T I NI

- I VP - v r rx 1 / 1. 5 SEC - GP
X X
AF CP •
�J I
C BP
XX

I
R E S T R I C T T H R U S T ACCELFRATION COMMAND TO PRODUCE
T H R U ST W I T H I N Pf R M I TIED � T H R l ! S T R E G I O N

q I/ 3 "', 46706 NFWTONI M I 161

M I N I MU M I AF C P . (:{) ··�
A F C P · MA X I M U M I A F C P .
AfCP 467DIJ NEWTON/ M I 171

O U T PU T :

T HR I J '> I AC C f l FRAT I O N C O MMAND A f C P

E NO

Figu re 1 3 P fi 6 V e rt i c al ( ROD) Guidance Algo rithm

35
l,O W E H E D - F LIG H T ATTITUDE - M AN E U VE H ROU TIN E

A link in the attitude control ch ain of command, the Powered-flight Attitud e ­

m aneuver r outine ( AT T ) connects t h e various powe r e d - flight guidance p r og r am s t o
th e D AP . T h e functions o f A T T are :

1. F o r the sm all attitude changes norm ally r equired e ach guid ance cycle,
ATT c omm ands a maneuve r of con stant r ate such as to achi eve the
r equired attitude 2 s e c onds late r .
2. F o r gross attitude m aneuve r s which m ay b e required at phasic inte rfac e s
o r upon abort, ATT com m ands a r ate- limited m aneuver which m ay extend
ove r s eve r al guidan c e cycle s .
3. F o r all attitude m aneuve r s AT T avoids th e gimbal - lock region ( middle
gimbal angle > 70° m agnitude ) . ATT i s sues a gimb al - lock alarm c o de
if and only if the comm anded attitude c o mputed from guid ance inputs
lies within the gimbal - lock re gion. AT T com m an d s a m aneuver which
c i rcumvents the gimb al - lo c k region and i s sue s no gimbal - lo c k alarm
c ode when the most direct p ath to the comm anded attitude p a s s e s through
the gimbal- lock region.

Switching from a d e s c ent program to an abort p ro g r am m ay p ro duce up to

°
1 8 0 change in comm anded thrust direction. A b r e ak with tr aditional app roaches,
ATT m akes gimbal lock during any m aneuve r inhe rently impos sible by 1 ) c omputing
c o m m anded gimbal angle s , 2 ) lim iting the m agnitude of the middle comm an de d
gimb al angle , and 3 ) is suing t o the D A P a s e ries o f inc r emental attitude - m aneuve r
*
commands that monotonic ally drive the gimbal angles from their cur r ent values
to their c omm anded value s . P rovided the attitude i s not currently in gimb al lock,
and given that the m iddle c ommanded gimb al angle i s m agnitude limited at the
gimbal- lock boundary, it is inherently impo s s ible to m aneuver through gim b al lock;
the m iddle gimbal angle is c onfined to the r ange between its cur r ent and c omm anded
v alue s . Other attitude- m aneuver s chemes with appended gimbal- lock avoidance
r equire more computation to p roduce s imilar m aneuve r s .

F igure 1 4 p r e sents an overview o f the L M powered- flight attitude c ontrol

p r o c e s s , including some inform ation on the p ro c e du r e s on the D AP side of the

*
Exc ept the outer gimb al angle p rofile m ay not be m onotonic in the geometric ally
c omplex case of a large m aneuve r about multiple axes at substantial middle gimb al
angle and with m agnitude limiting of the X - axis attitude angle change on at l e ast
one p a s s through ATT.

37
 I
I
I
m ��: l
R

RATES

PERM !TIED
lAG ANGLE I

1
OMPUTt COMMA�£
G l 'oi B AL ANGLE S,
LIMlT MAGNITUDE CX
U N l T TI< RIJ S T

I
COMMAND
COMM ANDED MIDDU
IMBAl ANGU TO 700
tiNlT W l NOOW
I
I
COMMA�D
COMMANDE D

I
DAP SPACECRAFT

I
G I MBAL

�:� I
ATTITUDE
ANGUS REFERENCE COMMANDS
G L

I L
INCREMENTS!
w
co
I I
I I
MEASURED
G I MBAL

I I
ANGLES

I REF E RE NCE G I MBAL ANGLES I

THRUST- I ! I�
� 19
ACCElERA TION
III'A SUREIII'NT
I
�� 1 5�
J>1 1 :0,:,
� E� ���
�ISs �§ �I:�
���= � �
�!!I I � \l!�
�
�lg
�li� ���
G
l� ���

Figure 1 4 LM P owe red - flight Attitude C ontrol

inte rfac e . Two c omputational c oordinate frame s are introduced. F rom gui d an c e
and n avigation inputs , A T T c ompute s a c om m anded- body frame ( tag C B ) t o repres ent
the c om m anded attitude inherent in the input vecto r s . F rom ATT inputs, the DAP
com putes refe rence gimbal angles to compare with m e asured gimbal angle s for
c om puting the attitude e r r o r s . The reference gimbal angles define a refe rence- body
fram e (tag R B). ATT c omputes that the attitude errors are zero when the s e two
c om putational coordin ate fram e s coincide. Of cou r s e , the re m ay be DAP c ontrol
e r r o r s undete cted by ATT, but any thrust pointing error is detected in the ste ady
s tate by a thru st- direction filte r, and c o r r ected.

The guidance and navigation inputs to ATT, shown in F igure 1 4, consist of a

unit th rust comm and, a unit window command, and a thru st - ac c ele r ation measur ement.
,\TT proce s s e s the thrust- ac celeration m e asurement in a thrust- direction filter to
dete rmine an e stimated unit th rust vector with respect to the referenc e - body fr ame.
C o rrecting for the offset of th e e stimated unit th rust vector w ith respect to the
X R B- axi s , ATT u s e s the unit thrust command and the unit window comm and to e r ect
the c om m anded -bo dy frame. F rom the commanded- body fram e m atrix, ATT extracts
commanded gimbal angle s which it compares with the refe ren c e gimb al angle s to
generate inputs to the D A P . T en tim e s per se cond, the D1\P update s the refe rence
attitude and gene rate s the corre sponding control command s . The dyn am ic r e sp on s e
i s suffi ci ently fast and tight that the refe renc e attitude i s a good me asu re o f
instantaneous space c r aft attitude.

A feature of thi s c onfiguration is that, although ATT runs at a s ample r ate of

2 s e c onds, clo s e to the fue l- slosh resonant frequency at cert ain points in the mis s ian,
it avoids exc iting fuel slosh by avoiding all c oupling with the actual spacecraft attitude
except th rough the s lo w thrust- dire ction filter.

F igure 1 5 details the Powe r e d - flight Attitude - m aneuve r routine. The th rust­
d i rection filter compute s the thrust - ac celeration m e asurement in referenc e - body
coordinates by constructing th e required t r an sfor m ation from the referen c e gimb al
angles ( Eq . ( 1 5 . 1 ) ) . The change in th rust direc tion i s limited on e ach cycle to 7 - mr
( E q s . ( 1 5 . 3 ) and ( 1 5 . 4 ) ), the m aximum travel of the trim gimbal in 2 secon d s . The
total excurs ion of the e stimated unit thrust vector i s limited to 1 2 9 - mr ( E q s . ( 1 5 . 5 )
and ( 1 5 . 6 )), the mechanical exc u r s ion limit o f the trim gimb al plus mechanic al
deflection and thrust offset with respect to the nozzle. The X - component of the
e stimated unit thrust vector is not needed and not computed.

If e ithe r 1 ) guidance p rovid e s a unit w indow c ommand too c lo s ely alined with
the unit thrust command to ade qu ately dete rmine the att itude o rientation about the

39
 I 'Jf't· l '.
: 1 "' 11 TH�t·)J CtlM��"-jO U�CP

--
!'NIT W INlHl'll (l�Y.ANO I UM'fCP

C I Mfi:A.l .l.N�;[ f ')

TIIRI • 'll�A.U l l t R II, TiON W A <., U Rf W NT MP
!?Hi R! "''rl �GA y
VM U[
RGA x llGA z
. , l i M I T MIDDLE COMMA NOLO G IMBAL AM;lE SAVING O R I G I NAl
_g_ R8
1\�'A II Afll! A.l'\ \ t f ll ii, I ION,Ri r i � ! �'i
AllflY 1 nnR�l t NAH '.
FOR POSSIBL� AlARM

Q • CGAz ()91

CGA • l iMIT I CGA

z ro• 1 1101
1
B EG I N

L THRliS 140 I R�Cl iON f 1. Tf R

�fR8 · C I R t; A ); R f. A. y RGAf
1
l�fP 'II � t
I
I SSut G I M BAL-l OC K
ALARM COOl

I
· II� II I �fR8
I
�·�Af R � I 111

[ D l l lJ�f R B y u ,.,- R B y I 0 001]

COMPUTE Ulll IMITEO �EFE�EM:E G I MBAL ANGlE CHANGES US I NG
t. l l 'f R B y · l l �tl m· MODULAR SUBT�ACTION

6 U"' R s7 [ 0 � liNAf RB l Uti-"RBz I 0 001]

I I Y. l l 141
• !

11"11- R By · 1 1'-IIT [1 Ur,ifRBy t..U �RBy I. 0 12']

L 6URGA • CGA
M
RGA 1211

t1NF RB 1 · : I M I T [t L'Nfli'Sz AU�R: 8z I , 0 12'9 ]

• II'
r---- ·-
l
If COMMAND lNG Y Z
I >41�
� 161 RES£1 W I NDOW HAG [)(CESS lYE OR G I MBAL ANGLE CHANCES

If I �URGAy FLAG . 0 1221

l >4��
-- -·-

• 1f I
l 1R; r�- Ac-f
- -
---
�-- --
AURGA FLAG . 0 i2}t
t l r'ol tt W'f t r-.oow (OMMANo tr w i T H I N ts0 or uNn THRUST coMMAtfD
.z

o\N[) � : \ I f VV I�Oll'W H A,f;

' • +
! If I :J\'Wt
� � ,!!,t
<'
1
r - UNH P f. <OS 1 , OR If P�NWC P • £8Pz FlAG o(l Ill
.. �05 , I "I c.l • .!l"MC P • - 'BPx
0 COMPUTE RHER£NCE G IMBAl ANGLE CHANGES l i M I T I NG ATIITUO£ ArK;tl RAH5 TO

l �
l f � �"-Wi'f' · NI · p �
_-
:
_ _
'
:·-- -
191
°
10 1 5lC i l0°1 l SEC I

- -- -
- - - I -RAMI
- -�l•OY
L --·
l'lRGA z • l iMIT (ll,URGA z 10° 1 1241

f Rf 1- ����MA.�Of 0
\JBD;( · \.!_M I- r
- - -
r !- -- - - l'lRGA y · L I MIT 1/j.URGA COS RGA
°
1 I COS RCA l ll�\

'lff
20

�'j
J
Y

[
Z
1101

�CRP,/ llr-.11 l � f\IW( p x !.;. C B PX

I
I Ill\
l'lATT X · L IMO ll.uRGAX
;.; ' t.. URGA y S I N RGA
z
I,

If HAG • 0 . 6 ATT X • 0 Q1\

�,- R P ,
l I
'::: < 'RPX )( ��l BP � I Ill fj,RGAX • l'l ATT X · l'lRGA
y
S I N �GA
l
i281
L...__________
..._
. J. - -- --
-------
-
+
�-���;ti;-��MM-ANOI [) BODY F R A .W: f RfCTION FOR THRUST OffSfT II SUE OAP COMMANDS

I I Ct:RP� i
I 6 ftGA • l:.ftGA I 10 1291
�C B P U N II

j
•
X
U Nf R B C.CBP U>HB z I,C BP I 1131
y Y

II �� B P 'I" · �i'!H') X �\B PX: z

;
11•1 S I N RGA

. [�
" RB COS RGA z COS RGA X liN GA 6B_GA I 2 SlC llOI
· �rRP)( )( �f.BPy I III
l
i,._C R P
/ COS RGA z S IN R:GA COS RGA

I
�
X )(

l
°
W�,,( T Cll�MAND£D
4>RB · L IMIT ( o.� R B X J �o� R B I I 2 a RB 10 1
t X X

CC B Pn
G I MBAL "NGLES X

1 (.!> ( · ARCH� I(, I I flO) $RB � I

Y
CCBP y · l i M I T I �J RB OI R B Jl o RB 10°
z yi y y 011
Cf B P)( l !1 1
I
°
J (,tl \ · ,!I, P [ T P ) f, l CC B PX
X
I UTI �•ez · LIMIT I • RBz I • R Bz o RB
1
. 10 1

y � )
I
j i ,A , A�\TRIC ! I r C BP
)(
C C BP UBI OUTPUTS :
)(Y
� RGA
N
REFfRENC[ G IMBAL ANGI( INfRI MFNTS
T
� RB
Rf CE All ITUOf RA I S iie

I
REn
PfRM InfO I AG ANGI • :,

END

Figure 1 5 P ow e re d - flight Attitu de - maneuve r Routine

40
XCB -ax i s , or 2 ) the guidance p rogram i s P66 (which p rovid e s no unit w indow com ­
mand ) , then ATT p rovid e s a unit window command suitable for e re ction of the
c ommand e d - body frame and resets a flag t o indicate that no attitude rotation i s
allowed about the XC B - ax i s . A T T first p rovid e s the c u r rent Z B -axis ( E q. ( 1 5 . 8 ) ) .
But th i s choice may also b e nea rly c olline a r w ith the unit thru st c ommand, so a
s e c ond poss ibility, the c u r rent negative X B - axis , is also o ffe re d (Eq. ( 1 5. 9)) .
Because the Z B - and X B -axes ca nnot both parallel the unit t h ru st command , no
fu rthe r checks need be made .

The m atrix C C B P , whose row vecto r s are the comm anded- body fr am e unit
ve cto r s exp r e s s ed in platform coor dinate s , is computed to s atisfy the unit thrust
c o m m and, the unit window comm and, and the thrust offset (th e angu l ar displac e m ent
between the e stim ated unit thrust vector and the X R B - axis ) . C C B P is c o mputed in
two step s as illustr ated in F igur e 1 6 . The first step ( E q s . ( 1 5 . 1 0 ) - ( 1 5 . 1 2 ) ) u s e s
t h e unit th ru st c om m and and t h e unit window command b u t fails t o account f o r thrust
offset. The s e c on d step ( E q s . ( 1 5 . 1 3 )-( 1 5 . 1 5 ) ) corrects for thrust- o ffset components
U N F RB
y and UNF R B
Z
. Since the s e c o r re ctions are s m all, no unit need be t aken
in J:: q . ( 1 5 . 1 4 ). A s m all window pointing e r ro r , shown in F igur e 1 6 , i s introduced
by the th rust- offset c o r r e ction. Define d a s the angle between the ZC B,XC B plane
and the unit window command, the w indow pointing error i s the product of the sine
of the LPD angle and the thrust- offset angle about the Z C B - ax i s . Although the
°
t r im gimbal h as a m aximum displac ement of 6 , the m aximum thrust offset during
d e s c ent i s about 1 ° , which yields a m aximum window pointing e r ror of 0 ° at 0 °
L P D angle and 0 . 9 ° at 6 5
° L P D angle, the lowe r e dge of the LM window.

B e c au s e the m at r ix e e B P i s the t r an s formation from p l atfo r m to commanded­

body c oordinat e s , it c an be exp r e s s e d in terms o f the I M U gimbal angle s which
would place th e body axe s in the com m anded dire ction s . The r efore, c o m m anded
gimbal angle s c an be e xt r acted from the commande d - body m atrix. E xpre s s in g e e BP
as the p roduct of the three m atrices that c o r r e spond to rotations about the thr e e

[
gimbal axes yields

e e BP =
+C Z

-ex
+SX
ey
sz
sz
ey
ey
+

+
sx
ex
SY

SY
: +sz
I

I
: -e z
I

: +e x e z : +e x sz
I I

: - sx c z : - sx sz
I
SY

SY

SY
+ sx
+ ex ey
CY l (13)

where S and e indi c ate s in e and c o s in e , and X , Y , and Z ind i c ate the commanded X ,
Y , and Z gimb al angl e s . F rom E q . ( 1 3 ), i t i s apparent th at the comman d e d gimb al
an gle s are extracted from the elements of e e BP by E q s . ( 1 5 . 1 6 ) - ( 1 5 , 1 8 ), with
AR e TR IG defined a s follow s .

41
 -c s P
c 1
X
=
1
UN FCP

- U N FR B y ---
( U N I T T H R UST
COMMAND) J - U N FR B z

UNWC P
( U N I T W I NDOW
COMMAND )

- �N F R B y

YC B
�C B P
y

- U N F R Bz

ZCB
fC B Pz

Figure 1 6 Ge ometry of E re ction of C ommand e d - body Frame

Viewed on a LM - c e nt e r e d Unit Sphe re

42
 The AR C T H IG function of two arguments yields the angle whose tangent i s
t h e r atio of th e f i r s t and s e c ond argument s . ARC TR IG ext r acts t h e angle anywhe r e
in the c i rcle b y u s ing the r atio of t h e s m alle r - m agnitude argument t o the large r ­
m agnitude argum ent a s the tangent of the angle o r its c omplement, and b y u sing the
s i gns of the argum ents to dete rmin e the qu adr ant of the angle . Equations ( 1 5 , 1 6 )
an d ( 1 5 . 1 7) yield the outer and inn e r c o m m anded gimbal angle s anywh e r e in the
c i r c l e . B e cau s e the s e c ond argument i s always pos itive, imp lying a p o s itive c o s ine,
l·: q . 0 5 . 1 8 ) yields the m i ddle commanded gimbal angle between ±9 0 ° .

To p re c lude c o m m anding gimbal lock, E q . ( 1 5 , 2 0 ) limits the m iddle c om m anded

g i m b al angle to 70 ° m agn itude. B e c au s e the unlim ite d v alue lay between ±9 0 ° and
the oute r and inn e r c o m m anded gimbal angle s were c omputed consi stent w ith the
mi ddle c om m anded gimbal angle range, no quadrant s w itchin g of the oute r o r inn e r
c omm ands is r equi r e d by g im b al- lock limiting. If limiting c h anges t h e m i d dle
c o m m and, the gui d ance is com m an ding gim bal loc k, and the gimbal- lock alarm code
i s i s s ue d .

U nlimited reference gimb al angle c h anges are t h e changes w h i c h would b e

req 11 i r e d t o b ring the D ,; p 's r e f e r e n c e gimbal angle s into coincidence w ith th e
c o m m anded gimbal angle s . The se are computed by subtr acting, modularly, the
c u n ent reference gimbal angles from the comm anded gimb al angles ( E q. ( 1 5 . 2 1 ) ) .

The m odular subtractions yield the s m alle r angular diffe ren c e s , i . e . ,

If a Y o r Z gimbal angle change g r e ater th an 4 5

°
i s requi red, t h e flag i s
r e s e t i ndicating n o attitude rotation i s allowed about the XC B - axi s . T h i s i s n e c e s s ary
to p r event fals e starts about the XC B - axi s as de rived in the appendix of refe rence
1 2.

Equations ( 1 5. 2 4 ) - ( 1 5. 2 8 ) yield the reference gimbal angle c h anges oy lim iting

° °
th e m agnitud e of the attitude c h anges to 20 in 2 s e conds (1 0 / s e c ) about e ach of
th r e e o rthogonal axe s ; one axi s i s coinc ident with the XC B- axi s and the oth e r axe s
l i e i n t h e YC B , ZC B plan e . Th i s p e r m its an angular - r ate vector of length 1 0 {3
deg ; s e c . N ote that i f the flag i s r e s et, the attitude rotation about the XC B - ax i s i s
m ade z e r o , re sulting i n an outer gim b al angle change t o offset the inne r gimbal
an gle c h ange ( E q . ( 1 5 . 2 8 ) ) .

The DAP c omm an d s c on s ist of th e refe rence gimbal angle inc rements to be
app l i ed by the D .\P e ach 1 I 10 se cond, the referen c e attitude r at e s , and the p e rmitted

43
l a g angle s . T h e refe r e n c e g i m b al angle incr em ents are the r eferen c e gimbal angle
c h an g e s m u ltip l i e d by the r at i o of the I L\ P and ATT s am p l e inte rval s ( Eq . ( 1 5 . 2 9 ) ) .

The r·efe r e n c e attitude r at e s a r e c omputed b y the nonorthogonal transfo r m ation o f

t h e r · cfe rence g i m b al angle c h an g e s shown in E q . ( 1 5 . 3 0 ) , The p erm itted l ag angle s ,
which a c c ount fo r t h e angle s b y which t h e attitude will l a g behind a r amp angular
c o m m an d due to the finite ac c e le r ation s available, are computed u s in g the available
a c c e l e r · at i o n gR B , and then individu ally m agnitude limited ( E q . ( 1 5 , 3 1 ) ) . The DAP
avoi ci s attitude - r ate ove r s hoot by permitting lagging attitud e e r r o r s equal to the
p e r m i tt e d lag angle s .

44
T H HO T T LE H O U T IN E

The throttle routine connects the currently ope r ating guidance algorithm to
th e D P S , as illustrated in F igure 1 7 . F o r e as e of und e r s t anding, all thrust levels
are represented as p e r centages of the DPS r ated thrust of 4 6 706 newton s . T HROT
g enerates th rust inc rement com m ands to drive the input thru s t - ac celer ation
m e asurement into c oinc idence with the input th rust - acceleration c omm and wheneve r
the resulting thrust would lie within the illu strated p e rmitted - th rust region. W hen
the r e sulting th rust would lie below or above the p e rmitte d-thrust r egion, THROT
c au s e s m inimum o r m aximum thrust. The hyst e r e s i s - like region from 5 7 to 6 5 o/o
th rust a\'oids frequent alternation between the m aximum - thrust point an d the
p e r m itted-thrust region when the thrust command dwells at the boundary between
th e permitted-thrust and forb idden -thrust r e gion s .

:\ digital- to- an alog interface between the LGC and the D P S is p rovided by the
d e s c ent engine c ontrol assemb ly ( D E C A). E ach guidance cycle (once p e r two seconds,
except once per s e c ond for P 6 6 ) T HRO T gen e r ates the th rust inc r ement c omm and
� F C % which i s c onve rted to a pul s e t r ain and is sued to the D E C A. E ach pul s e
c au s e s about 1 2 . 5 newtons thrust change, an d the pulse r ate i s 3 20 0 I s e c ond. F ollowing
i s suance of a thrust inc rement c om m and, th e thrust therefore changes at the r ate
of 4 0 , 0 0 0 newton s I s ec ond ( 8 5% of r ated thrust per second ) until the thrust inc rement
is achieved. \V ith a guidance cycle as short as one s e cond and an engine r e spon s e
time which m ay be a substantial fr action of o n e s e cond, i t i s nec e s s ary f o r P 6 6 and
T H R O T to ac c ount for this transport delay.

c\s illustrated in the rightmost box of F igu re 1 7, in the region from 1 1 to 9 3 %

the DPS th rust i s a ne arly line ar function of the pulse c ount ac cumulated by the
D E C :\. :\i ot shown in F i gu re 1 7 i s the m anual throttle command, which is summed
w ith th e DEC A output command by the DPS and p rovides the m inimum 1 1 o/o thrust
w hen the D I,:C A c om m and i s zero. The DPS c ontains a m e c h ani c al stop at typ ic ally
9 3 o/o rated th rust. This thrust level m inimizes p ropellant c on sumption on a nominal
descent, con side ring the lo s s of specific impul s e at higher thrust. To en su r e that
th e DPS is driven to the m e c h ani c al stop, the DEC A s atur ates at a sub stantially
high e r thrust level ( about 9 9 o/o ) and the throttle routine drive s the DEC A into
s atu ration whenever m aximum thrust is required.

� onlinear iti e s in r e sponse and unc e rtainties in DEC A and DPS scal e factor s
a r e ove rcome b y the th rust in c rement command c on cept. N omin ally, THROT p rovide s
dead-beat respon s e to step input s , but w ith down stream nonlineariti e s and s c al e - factor
e r rors T H R O T d rive s the thru s t - ac celeration error to zero in the ste ady state .

45
 1HRUST-CONTRO. LOGIC TO DECA AND D P S RfSPONSl
COMPUH ?J;?OOUCE T Hl REQL I RED

�.��:- THRUST 1�
Af(P
THJlU� T O�HtALL SYSTE\1 RlSPO�;Sf
THRUST ACCEURATIQN
COMMA'NO
CCY>IMANO
:;; 'l'l
C<Y>I""''«l

'\
C ATIVl

. __
� 9),...,..._ -/, ,
�
-

MA S ) f S T I M A TI_
� M C•
� c c u ..u c � • c , '" " ' , "
OPS \'-' _,/ 1
I FCC\ fLC'J �6fC'I ll<J � fCC� � qq,. 0
M
�
'-----_l
T H RU S T Mt:_(HANJC OEI IVE RI D
"'
INCRfMENT STOP THRUST
>

I
C.-AND
0
COMf'UH
Af P lHRUST ' m
THRIJSHCCUERATI0N ,'v\£ASURfW"NT
COMMAt«l I C U M U l A T I VE

I
THRUST
THRUST COMMA"..O
MEASUREMENT

*""
Ol I
1 ,.
I I�
�I ���
� �
;:: I
I g :S
�Z I § !Oi l � �
I O< ;:: I � �
;I@ � I � a.:
� �
31� �I��
�I �I ��
I oo

Figure 1 7 DPS Thrust C o ntrol

 F i gure 1 8 illu strates the Throttle Routine computation s . The thrust comm an d
and th rust m e asurement are c omputed usin g the input m as s e stimate ( E q s .
( 1 8 . 1 )- ( 1 8 . 2 ) ). T h e input thrust- accele ration m e as urement i s the aver age o v e r the
p re c e ding s ample interval, du ring which a th rust inc rement com m and was i s sued
p roducing an instantaneous th rust p rofile as illustrated in F igur e 1 9 . Therefore,
to obtain the current s ample- instant thrust, Eq . ( 1 8. 3 ) corrects the th rust
m e asurement by addin g the th rust c orre ction inc rement c omputed the previo u s cycle.

The th rust- c ontrol logic for p roviding the r equired ove r all system re spon s e
illustrated in F i gure 1 7 is t o pick one o f four po s s ible thru sting policies acc or ding
to the regions of the p re c e ding and p r e s ent th rust commands ( FC O L D% and F C %).
and to r e s et the th rust command if nec e s s ary to s atisfy DPS constr aint s . Equ ation
( 1 8 . 4 ) or ( 1 8 . 8 ) re sets the thrust command to the thrust actually anti c ip ated . A
th rust c ommand augment ( F C AUG%) i s c o mputed that e ithe r drive s the DEC A into
s atu r ation if the policy is to initiate or retain m aximum thrust ( Eq. ( 1 8 . 5 ) or ( 1 8 . 9 )) ,
o r c o r re cts for the region between the D P S mechan i c al stop an d the DEC A s aturation
valu e if th e policy i s to initi ate th ru sting w ithin the permitted - thrust region ( Eq.
( 1 8 . 7 )) . No thrust command augment is required when the policy i s to continue
th rusting w ithin the perm itted-thrust region. No equivalent thrust- control logic i s
needed a t the m inimum - thrust point b e c ause minimum thrust would occur only i f
t h e c om m ander c ould i s sue five or more downward ROD c omm ands within a s ingle
P 6 6 guidance s ample interval, p ractic ally imp o s s ible.

The th rust increment command ( Eq. ( 1 8 . 1 2 )) i s comp o s e d of the actual thrust

in c r em ent �F Ao/o, plus the thrust command augment FC AUG% to drive the DEC A in
or out of s aturation, when requi red.

P re paratory to com puting the thrust corre ction increment for the s u c c e e ding
p as s , Eq. ( 1 8 . 1 3 ) compute s the total effective t r an s port lag. The terms in the effective
tr an sport lag are 1 ) the c omputation duration t - tSI, 2 ) the e stim ated DPS time
c on st ant of 0 . 0 8 s e cond, and 3) the effective DEC A d elay equ al to hal1 the time
r equi red to output the thrust inc r e m ent command pulse tr ain at 8 5 % thrust change
per s e c ond. As long as the actual thrust inc rement AF A ( F igure 1 9 ) i s c ontaine d
enti rely w i thin the s ample interval At, i t is c l e ar that. a s L AG app r o aches ze ro,
the th rust measurement ( obtained by diffe rencing accelerometer r e adings at the
s ample instants ) app roaches the s ample - instant thrust F . Similarly, as L AG
approaches the s ample interval At, the thrust m e asurement must be augmented by
an amount app roachin g the actual thrust inc r ement AF A to obtain the s ample - in stant
th rust F . F rom this heuristic argument, i t i s app arent that the thrust c o r r ection
inc r ement whi ch must be added to the thrust m e asurement to yield the s ample - in s t ant

47
 I NPUTS
THRUST ACCH I R A T i fl N COM\IAND Af C P
THRlJST ACI H I RAT I O N Mf A SU RI '.II NT Af P
C l OC K - T I MI
SAMPll I NSTANT Cl CURRENT
GU I D ANCf C YC l E lSI
C U R RI N! M A S S f S T I MAIT M

BI G I N

C OM P U IT TH RU S T COMMAND A N D THRUST Ml A S U REMENT

46-i�-Niwror.iS
IOO Af C P M
I] I
FC'I.

4blliil NfWTONS
100 Af P M
-
��

COMPUH SAMPlf I NS TMIT T HR \ I S T BY A D D INC TH> ACTUAl THRUST

THF PREC f D INC CYCI I

I N I T I AH
MAX I MU M
THRUST

{ R! SH THRUST COMMAND &

S £ 1 TH RU S T COMMAND AUGMENT
}

COMPUTE ACTUAl T H R U S T I �C RIMFNT AND S A V l

THRUST COMMAND FOR S U C CH D lNG P A S S
I'. I A'' I C' ' F � I 101
r ca o ' r c,, 1111

l]ll

1 14 1

THRUST C O R R f C T I ON I N C R!Mf �T f O R P 66 V f RT I C A! � OD} G U I D A NC l ALGORI THM

bfA btA• 46 706 NI WTONS I I OO

OUTPUTS,
THRUST I NC RFMENT COMMAND M C '\
THRUST C O R R fC T I O N I NC R f MoNT bfA

I N il

Figure 1 8 Throttle Routine

48
 �--- Guidance Sampl e Interval �t

Computation Del ay
Effective
P l us D P S Time
C onstant
r-; DECA Del ay
.1�J .
•
1 nstantaneous Thrust P rof i l e

_ Sampl e - 1 nstant
Thrust F

..,.
co i
14---- LA£ ------.1
Actual Thrust Inc rement 6FA

LAt
Th rust C o rrection Increment o FA = NA
6t
P revious
C ur rent
Sampl e
Sampl e
Instan t
Instan t

Figure 1 9 Thrust Dynamic s Within a Single Guidance Sample Interval

th rust is proportional to L AG as c omputed by E q . ( 1 8 . 1 4 ). A r igorous derivation of
th is re sult is p r e s ented in Appendix A of reference 1 1 . The s ole purpo s e of E q.
( 1 8 . 1 5 ) is to inte rfac e the P 6 6 V e rtic al ( RO D ) Guidance Algorith m .

\V ith the thrust command FC % either within t h e permitted- thrust region o r

r e s et to the value which will actually b e achieved, A F A o/o is an ac c u r ate p rediction
o f the actual thrust increm ent, and 8 F Ao/o o r c5F A is an ac c u r ate thrust c o r r e ction
inc rement . a F A o/o or aFA is s lightly in e rr o r w hen initiating thrusting w ithin the
permitted -thrust region. The slight e rro r is due to ne glecting FCAUG% in the
c o m putation of LAG.

50
B R AK IN G - P H ASE AN D A P P RO AC H - P H ASE T AR G E TING P ROG R AM

The targeting program gene r ates m i s s ion - dependent d at a for the P 6 3 Ignition
Algorithm an d for the P 6 3 , P 6 4 Guidan c e Algo rithm . All d at a are exp r e s s e d in
gui d an c e coor dinate s . The ignition alg orithm requ i r e s nominal initial altitude, r ange,
and speed d at a that dete r m ine ignition time and indi rectly determine the throttle­
c ontrol du ration. The guidance algorithm requ i r e s t ar gets for P 6 3 that p rovide an
efficient t r ansfer and targets for P 6 4 th at p r ovide a t r aj ectory m eeting s ev e r al
c on s t r aints on geometry, vis ibility, and thrust. De s c r ibed in detail in referenc e 5,
the P 6 4 c onstraints p r ovide a fast s h allow app roach phase m o r e akin to an ai rplane
app r o ac h th an a h elicopter app roach, although the termin al - des cent phase is
e s s entially vertical i n helicopter fashion. T h e landing s ite m u s t b e appr o ached
°
along a ne arly s t r aight- line p ath depre s s e d typic ally 1 6 from hori zontal, terminating
typic ally at 30 m altitude 1 1 m ground r ange. The l anding s ite must be vis ible
c ontinuously until a few s e c on d s before app r o ac h - ph ase terminus , and the DPS thrust
must begin at around 5 7% and must lie c ontinuou s ly in the 1 1 to 65% region.

G eometry, visibility, and thrust during app r o ac h c annot b e specified explicitly.

V i s i b ility depends upon the p o s ition and attitude p rofile s , and the s e p rofiles (with
the th rust and m as s p rofile s ) are con s t r ained to s atisfy the l aw s of phy s i c s . The
guidan c e algo rithm will p rovide the t r an sfer from any arbitr ary initial s t ate ( w ithin
b ound s ) without regard to any vis ibility o r thrust constraint s . The task of the t ar geting
p r og r am is to set up the P 6 4 initial s t ate and guidan c e t argets such th at suit able
v i s ibility an d thrust p rofiles are r e al i ze d implicitly.

During the final p ortion of P 6 3 , and th roughout P 6 4 , the guidance algo r ithm
will gene rate a t r aj e ctory who s e p o s ition vector is a quartic polynomial function of
time, as shown in F igure 20. T argeting consists of 1 ) defining e ach of the two
p olynomials and 2) ext ractin g the guidan c e t argets as the p o s ition vector and its
d e r ivative s at a t arget point, lying on the polynomial, sub s tantially beyond p h a s e
t e r m inu s .

The P 6 4 targeting c on c ept i s to c on st ruct the app r o ac h - ph ase quartic b y

i m p o s in g nec e s s ary and suffi cient constr aints . W ith quartic degree, five independent
c on st r aint s m ay be im posed in e ac h of three axe s . The nomin al t r aj ectory i s
arbitr arily m ade planar, requi r in g the Y - components i n guidan c e coordinates o f
p o s ition and all its derivative s t o be z e r o and leaving two axes t o specify . Bec au s e
t h e initial state c an be c ontrolled by t h e p r e c e ding b r aking- phase guidan c e , all five
c o n s t r aints in e ac h of the two rem aining axes m ay be specified arbitr arily. Sinc e
these ten const r aints - c alled a P 6 4 c on s t r aint s e t - completely deter m in e the

51
 T o - 1 8 0 SEC T o -6 0 S E C T o - 1 56 S E C
T H ROTTLE BRAKING- APPROAC H -
CONTROL PHASE PHASE
RECOVERY T E R M I NUS INCEPTION

'\.. /

AR l iC
N O N OU
IGNI T I O N

. -· -

(J1
N

T o 0 SEC
BRAKING ­
PHASE
TARGET
POIN T

Figure 2 0 Composition of the Luna r - d e scent T raj e ctory

P 6 4 t r aj ectory, the geometry and vis ibility p rofiles c an be dete rmined in c l o s e d
fo r m , and the th rust p r ofile c an be dete rmined from a p r i o r knowledge of m as s .
Thus P 6 4 targeting consists of gen e r ating c l o s e d - form s o lutions for a number of
P 6 4 constraint s et s and pickin g one which provide s adequate vis ibility and th rust.
S p e c ification of P 6 4 con straint sets i s reduced to a two - dimen s ion al s e arch, as
w ill be d e s c ribed in the following s e ction .

The P 6 3 targeting p roc e s s is not so c le an . B e c au s e the engine must be run

at fixed m aximum th rust fo r most of the phase, the guidan c e commands are not
s ati sfied, and the refo r e , as shown in F igure 2 0 , the m aximum- th rust po rtion of
P 6 3 i s not quartic . W hen th rottle c ontrol i s recovered, gene r ation of a qu artic is
b e gun. But the th rottle recovery point i s not c lo s e to any target point. Therefo re
the state vector at this point c annot b e controlled, and we h ave no c lo s e d - fo r m
solution f o r it . S i n c e t h e initial p o s ition and velo c ity on the braking-phase quarti c
must be free, there remain only th ree constraints , in e ach of two axes , which c an
be imposed arbitr ar ily. The guidance algorithm p ermits a fourth constraint i n one
axis by s olving for the current target- refe ren c e d time such as to s atisfy a c on st r aint
on the ZG - component of j e rk . Thus a P 6 3 c on straint set comp o s e d of seven constraints
is spec ified arbitrarily, and the rem aining three condition s required to define th e
b r aking- phase quartic are det e r m ined ite r atively by simulation. Three or four
ite rations are gen e r ally requi r e d bec au s e there i s b ilate r al inter action b etween the
targets and the simulation.

C onstraints

The P64 constraint s et i s c onstructed as follow s :

1. F ou r constr aints at a spe c ified target- refe renced term in al time T AP F : Two
te r m in al verti c al const r aint s , spec ified by the m i s s ion comm ander, are the termin al
altitu de ( R AP F G = 3 0 - m typ ic ally ) and altitude r ate ( V AP FG =- 1 -m / sec
X X
typ i c ally ) . T w o termin al hori zontal con s t r aints , imp o s e d b y the choice d effective
P66 hori zontal time constant T, are that the terminal pos ition, veloc ity, and
2
ac celeration shall be re lated by R AP FG A A P F G r , V AP F G
= - AAPFG =r.
z z z z
T h e s e P 6 6 compatibility constraints c au s e the p itch comman d s at P 6 4 terminus
and P 6 6 inception to be identical ( avoiding a p itch tran sient at the p h as i c interfac e )
and c au s e the P 6 6 algorithm t o null the hori zon t al p o s ition e r r o r a s well a s the
hori zontal velocity e r ror, without p o s ition feedback. B e c au s e the P 6 6 hori zontal
algorithm fee d s b ack the prior accele r ation c o m m and, and b e c au s e of the tran sp ort
de lay, an effective r of 8 s e conds h as been found s atisfactory r ather th an the 5
s e c onds u sed by the P 6 6 algo rith m .

53
2. Four c on st raints at an unspec ified t arget - referenced m idpoint time T AP M :
T h e m i dpoint constraints are specified by the commander according t o h i s sense o f
s afety a n d compatability with a pos s ible m anual t r an sition to P 6 6 . Typic ally, h e
°
s lope, completely
m ay spec ify - 5 - m / s e c altitude r ate at 1 50 - m altitude, a n d a 1 6
determ ining the m idpoint state R AP MG ,..Y_APMG given R AP MG = V APMG = 0.
y y

3. T w o con st raints a t a n unspecified t ar get- referenced initial time T AP I : The

°
initial pos ition is arbitrarily spec ified to lie on the 1 6 p ath and to p rovide an app r o ac h
p h a s e o f typic ally 7 . 5 -km length, dete rmin ing the initial p o s ition vector R AP IG given
R AP IG = 0.
y

This c ompletes the P 6 4 constraint set except for spec ifying the time s T AP M
and T AP I at which the midpoint and initial constr aints apply. These times are
determined by running the App roach -phase Targeting R outine over the two­
dimens ional s weep of value s of T AP M and T AP I . F rom the c as e s run, one i s p icked
that exh ibits suitable attitude and th rust behavior ( b as e d on an a - p r iori P 64 initi al
m as s e stimate ) . If sub s equent s imulation p rove s the m a s s estimate exc e s s ively in
e r ror, the initial thrust w il l be un s atisfactory, and an alt e rn ate c as e must be p i cked.

The seven P 6 3 constr aints are specified as follow s : Four constr aints are
spec ified by c omp atibility of the terminal state on the b r ak ing-phase quartic with
the initial state on the app roach- phase quartic . Two constraints are impos e d on
termin al acceleration by requirin g the termin al thrust to be 57% and by specifying
the term inal pitch angle . The final constraint is imp o s e d on the hori zontal componen t
of termin al j e rk b y requiring zero rate of c h ange of thrust at terminu s , The term in al
°
pitch angle, typic ally around 60 , is chosen by t r i al and e r r o r to m inimize p ropellant
c on s umption.

App roach - phase T argeting

F igure 2 1 illustrate s the App roach - phase T argeting Routine. N o r mally, this
routine is first run separately in search of t argets for the approach phase, and
then run j o intly with th e B r aking- phase T argeting R outine ( F igure 2 2 ) to deter m in e
t a rgets fo r t h e entire lunar d e s c ent.

In the XG - axi s ( altitude), the te rminal ac celeration, j e rk, and snap are computed
by Eq. ( 2 1 . 2 ), which is obtained immediately from

54
 RAPMGX 1 TMF TMF2 / 2 TMF3/6 TMF4:3 / 2 4 RAPFGX
VAPMGX 0 TMF TMF2 / 2 TMF /6 VAPFGX
1

RAPIGX 1 T TIF2 / 2 TIF3 /6 TIF4/24 AAPFGX (14)

IF

JAPFGX
SAPFGX
where T l\TF and TIF are the m i dpo int and initial te rm inu s - referenced time s c omputed
by E q s . (2 1. 1).
I n the YG - axi s , pos ition and all its deriv ative s are z e ro t o p rodu c e a planar
t i· aj ectory.

In the ZG - ax i s ,
Eq. ( 2 1 . 4 ) i s obtained by substituting the
2
comp atibility P66
const raints H AP FG A. A P F G r , V A P F G
= - A APFG r into th e ZG-axi s ve r s ion
=
Z z Z z
of Eq. (14)
and inv e rting. E quation s ( 2 1 . 5 ) and ( 2 1 . 6 ) c om p lete the definition of the
app r oac h-phase quartic . It r e m ains to c ompute the app r o ac h - p h as e t argets as the
p o s ition vector and its d e r ivativ e s at th e target po int on the quartic .

F o r a quartic polynom i al, a 5 x ID ) c an b e

d efined by
5 state t r an s ition m atrix (T 1, T
0

R1 1 (T 1 -T 0 )
2
(T 1 -T 0 ) / 2 (T 1 -T 0 ) 3 / 6 (T 1 - T 0 ) 4 / 2 4 Ro
Y.1 0 1 ( T 1 -T 0) (T 1 -T o > 2 / 2 (T 1 -T 0 ) 3 / 6 .Yo
X1 = Al 0 0 (T 1 -T 0 ) (T 1 -T 0 ) 2 / 2 Ao = m (T l , T O )X O,
0 0 0 1 (T 1 -T 0 )
I. 1 I.o
0 0 0 0 1
�1 �

whe r e H . to S . are row vecto r s . _m T ) c an b e de rived u s ing linear syst e m-s

theory,
-1 -1
(T
1, 0

r�
R 0 1 0 0 0 R
v 0 0 1 0 0 y
0:
X A
.

0 0 0 1 0 A X
j
� li 0
0
0
0
0
0
0
0
1
0
J
.s._

55
w ith the s olution

where I is the 5 x 5 identity m atrix. The exponential s e r i e s is zero after the fifth
i
term b e c ause o 0 for i > 5 . All the p roperti e s of s t at e t r an s ition m atr i c e s c an
=

be applied t o s c alar and vector polyn o m i al s .

Equations ( 2 1 . 7 ) yie ld the comp lete target and initial states by u s in g state
t r an s ition m at r i c e s and the definition of target- referenced t arget time as ze r o .

B r aking- phase T argetin g

To target P 6 3 , w e m u s t c o mpletely dete rmine the b r aking- p h as e quartic shown

in F igu re 2 0 . Seven of the ten n e c e s s ary c onditions are dete rmined in c l o s e d form,
although th ree are based on a P 6 3 termin al m as s e stimate which must be upd ated
by simulation . The rem aining three c on d itions n ec e s s ary to define the quartic are
determined ite r atively by s imulation. The terminal p itch angle IJ P BR F is a fixed
input to the B r ak ing - phase T argeting R outine.

*
F i gure 22 illu s t r ates the routine . Four conditions are s p e c ified by s etting
the P 6 3 termin al pos ition and veloc ity equ al to the P 6 3 initial state (Eqs. ( 2 2 . 1 )) .

A unit vector in the term inal- th rust direction is c omputed from the terminal pitch
angle O P B R F ( Eq . ( 2 2 . 2 ) ), and the terminal ac c e l e r ation i s c alculated by Eq. ( 2 2 . 3 )
u s ing th e term inal thrust F B R F , the P 6 3 term in al m as s e stim ate M B R F , and allowing
fo r lunar gravity GM. The XG - component of terminal j erk must be dete rmin e d by
s imulation an d is therefo r e set to zero for the first ite r ation (Eq. ( 2 2 . 5 ) ) . The
Z G - c omponent of termin al j e rk is c omputed by E q . ( 2 2 . 5 ) to p ro duce z e r o r ate of
c h ange of th rust at terminu s , accounting for the e stim ated terminal m a s s flow r ate
c omputed by Eq. ( 2 2 . 4 ) ; the j e rk c oeffic ient KJ, typ i c ally 1 . 2 , accounts for the
XG - component of thrust. The te rmin al snap must b e dete rmined by s imulation and
is the refore set to zero ( E q. ( 2 2 . 6 ) ) for the first ite r ation. This c ompletes the
fi r s t- ite ration definition of the b r aking- phase quart i c .

*
N ot s hown i s the c ap ab ility of the targeting program to set the P 6 3 terminal s t ate
to a b ackwards extr apolation of the P 6 4 initial s t ate to allow for a short tran s ition
du ring which the ac c e l e r ation is as sumed to c h ange line arly with time. Thi s c ap ab i lity
is n ot alw ay s used, and to show it would unn e c e s s arily c o mp l i c ate the p re s entation
of F igu re 2 2 .

56
 B r aking-phase targets are c omputed by E q . ( 2 2 . 7 ), u s in g the state t r an s ition
m at r ix and the definition of target - referenced target time a s zero. U s ing the
c om puted t arget s , a s imulation i s run to produce c o r r e cted d ata.

The nom inal initial r ange used by the ignition algorithm is c o r r e cted by E q .
( 2 2 . 8 ) to c o r r e ct t h e error i n t h e target- refe ren c e d t i m e o f throttle control recovery.

The s imulation p roduces a b r ak in g - ph as e quartic s atisfying the t arget v alues

of pos ition, veloc ity, acceleration, and Z G - component of j erk. The rem aining
conditions nec e s s ary to define the quartic c an be obtained from the cur rent s t ate
on the l ast pass of the b r akin g - p h a s e s imulation. The equation for the curren t
state,

[::l [: T RBRTG

1 T Y:BRTG

ABRTG

.J.B RTG

§B RT G

is re adily s olved to yield the achieved t arget j e rk and s n ap accor ding to Eq. ( 2 2 . 9 ).
Solving for the ZG - c omponent of achieved target j erk provides a check on the
com put ation of T by the guidan c e algorithm; agreement between achieved and input
values is typic ally to s even plac e s .

I n p rep aration for c o r r e cting e stimates at the terminus, the complete s t ate
at t e r m inus is computed by Eq. ( 2 2 . 1 0 ). E quation ( 2 2 . 1 0 ) yields a term in al state at
the specified termin al time T B R F p re c i s ely, whe r e as the state R G , Y:G appli e s at
the time T which m ay differ from T BR F by up to the 2 - s e cond g r anularity.

E quation ( 2 2 . 1 1 ) corrects the P 6 3 t e rminal mas s estimate u s in g the rocket

equation. Equation s ( 2 2 . 1 2 ) - ( 2 2 . 1 4 ) c o r r ect the termin al ac celeration, j erk, and
s n ap u s ing the corrected P 6 3 termin al m a s s e stimate, the achieved X G - component
of terminal j e rk, and the achieved X G - and Z G - c omponents of t e rminal s n ap .

F inally, state convergen c e test quantiti e s are computed b y E q s . ( 2 2 . 1 5 )

( 2 2 . 1 7 ) . Since only three conditions ( J B R F G A:x_ , S B R FG AX ' and S B R FG A z ) defining
th e b r aking - ph as e quartic are sought ite r atively, only three convergen c e c rite r i a

57
 I NPUTS:
T E R M I NAL TARGH-REFI:RENCEO T I ME TAPF
TER M I NAL ALTI TUDE RAPFGx
TERM I NAL ALT ITUOE RATE VAPFG X
EFFECTIVE P66 HOR I ZONTAL T I ME CONSTANT T
M I D PO I NT TARGET -REFERENCED T I ME TAP M
M I D PO I NT STATE RAPMG, \IAPMG
I N IT I AL TARGE'HIEFERENCED T I ME TAP I
I N IT I AL POS I T ION BAPIG

BEG I N

f-----
COMPUTE TERM I NUS-REFERENCED T I ME'>

"IMF • TAPM - TAPF, T I F · TAP I - TAPF

(\)

O] ;:::�:
'�
TERM I NAL X COMPONENTS

1 I
3
TMF 16 - TMF
-1
2

IZI
TMF 12 o

I
3
TIF 16 -TIF RAPMG X
VAPMGX
RAPIG
X

TERMINAL Y COMPON NTS

RAPFG y • 0, VAPFGy • 0, AAPFGy • 0, J A PFGy • 0, S A PFGy • 0

["'"1 [ 'L,
.-----

[ l
T
TERM I f.IAI 7 �" '"""c"'
'" ' "" l/ 2 TMF 3 1 6 • w '' ""ffi z
-T + TMF TMF 2 1 2 TMF 31 6 141

l
JAPFGz � VAPMGz
3
SAPFG z t 2- t TIF + T I F 2 1 2 TIF 16 T I F 4 1 24 R A P I Gz

R A PFG l • AAPFGz t2 151

['""]
VAPFG z •-.\APFG z T 161

COMPUTE COMPLHE TARGET AND I N I T IAL STATES

l'""l r·l
r'"'"
VAPTG
�APTG ·¢ 10, TAPFI l VAPFG
�APFG
VAP I G
�APIG · ¢> 1TA P I , O I
VAPTG
fAPTG 171

S APTG S
JAP IG
- -APFG .
JAPTG

"[
JA PTG JAPFG
�APIG -APTG

OUTPUT S :
COMPLHE TARGET AND I N ITIAL STATES

['""] "j
V A PTG

JAPTG
AAPTG

SAPTG
- ,
A
'A P I G
VAPIG

JAP I G

-
SAPIG

f NO

Figure 2 1 Appro a c h - phase T argeting Routine

58
 I N PUT S
Pb4 IN l�l Sl�TE ft�PIG. ��PIG
TERM I l T�RGET-REHREN:ED T I ME TBRf
TfRMI I PilCH A�GlE CotiSTRAINT 6 PBRF
TERM I .L lHRU S.l C ON STR� I NT FBRF
P63 TE I I N�l MASS ESTIMATE MBRF CDMPUH COMPtm SI Alf Al llRMINUS ACHIEVED B Y 5 1 MlllATED TRAJECTORY

r'·l ['" ' J

i
DPS EX �UST Vl'LOCITY ESTIMATE fOR P6l TERMINUS VEXBRF
J£RK Cl JF ICIENT T O �CCOUNT fOR THE Vl'RTIC�l COMPONENT Of THRUST
�B RfG A ·+ (18RF,OJ XBRTG
KJ
MJMINA. INITI�l AniTUDE, RANGE, ANO SPEED RBRIG , R B R I Gz VBRIG VBRFGA VBRlG
!lOI
X
FO THE IGNITION ALGOR ITHM

REOU I I D TARGET REHRFN:ED T I ME Of THROTILE RECOVl'RY

I B R T <. Ax JBRTGAy JBRTGz
COEFFI ENT fOR CORRECTING MJMINAl iNITIAl RANGE R B RIGz KlG JBRIGA
ITHROT BRfC.A �BRTC.A

ff
CONVl' :ENGE CRITERION fOR TAR<.ET ·REFERENCED T I M£ OF lHROTILE R EC OVER Y f TIHROT
CONVl' :EN:E C R I TERION FOR ITRMINAl STAT£ ERRORS f t
CORRICT Al l ESTIMATES Al TfRMINUS

MBRF • M rxPI t lyBRFG I 1 �G p/Vf x BRF l

SEG I N

�B RI G · l B R l MBRf I �M' B R fG f GM , Q. O l
1111

lBRfG · I J BRF GA x , 0, KJ A B RF G z MBRFI MBRF I

1111

2B R Fr; - j
<Ill
SBRfGA X . 0. S B R F G Az l 114)

COMPUTE FIRST ITERATION Of P6l COMPLETE TERMINAl STATE t

COMPUTE I l ATE CONVERC.fNCE T E S T QUANT IT IES

f
U l " I ' AOKIGA z . A B R f G 7 I I I �BRfG\1
RBRFG · RAP I G. YBRFG • YAPIG Ill
111 bl

· I I ABRfGA x I I I A BR f G I I
YNFBRFG • I COS 8PBRF , O, - S I N B P BRF I
ABRFG · I FBRFI MBRF I UM'BRFG · I GM. O. 0 I Ill
02 1161

I YB Rf G I I
�BRf • - FBRFI VEXBRF I(I
ABRfC.
X
lBRFG • I 0. 0, - KJ �BRFGz MBRFI MBRF I 151 OJ • )1 VB R GA X VBRIG I I !171

[]
X

[
�BRFG · I D,O.O I 161

t
COMPUTE COMPl£1£ TARGET STATE YES
01 · E
�BRTG RBRFG

ABRTG
VBRTG VBRFG
· 4> 1 0. TBRf I ABRFG Ill NO

SIRTG im G
JIRTG JBRFG
c.n YES
01 , r
L
CD
+
S I MULATE DESCENT US I NG TARGETS Of CURRENT ITERATION
NO
S I MULATION OUTPUT DATA USED FOR CORRECTING IGNIT ION ALGORITHM PARAMEHRS!
ACTUAl TARGET ·REFEREM:ED liME Of THROTILE CONTROL RECOVl'RY ITHROTA YES
OJ · f
ACTUAL STATE ON FIRST GUIDAM:E PASS BBRIGA. YBRIGA

S IMULAT ION OUTPUT DATA USED FOR CORRECTING GUIDAM:E �lGORITHM TARGETING;
NO

i nH ROTA TIHROT I , E TIHROl

EXACT TARGEHIEHREN:ED l i ME Of fiNAl P6l PASS T
STATE AT TARGET REffREN:EO TIME T II_G . yG YES
MASS AT T�RGET-REHR£NC[O Tlr.t: T M

• TNo
KlG TIHROTA · TIHROT I
RESET NOMINAl I NI TI AL RANGE TO CORRECT THE Tfft!OTILE CONTROL DURATION CORRECT FOR IGNITION AlGOR ITHM MJMINAl INITIAL �LTITUOE. RANGE
AND SPEED
Ill
""" ''z " """ z
'"
RBRIG · R BRIGA . R8R1G z · RB R I GA , VBRIG · I YB R I G A I 1181
X X Z

t OUTP UTS :
RBRTG

COMPl£TE TA RGn STATE.

�8RTG
[!�"W''
VBRlG
COMPUTE JERK AND SNAP AT TARGET POINT ACHIEVl'D 8Y S llllJLATED TRAJECTORY

� 8RTG

•l�l
w •'
JBRTG

nt T 4 411 Tl 111 r ·nt T4

"'' "
J
}.IRTGA 141 T �IRTG 191
�BRTG
MJMINAl INITIAl AlT ITUDE AND RANG RBRIG . RBRlG
X
COMPONENTS Of THE STATE VECTOR
Z
!!.G
yc I AT HIE f I RS! P6J GUIDAM:E PASS
MJMINAl INITIAL SPUD VBRIG
END
I ---

Figure 2 2 Braking - phase T argeting Routine

a r e needed. The three c r iteria chosen are important for guidan c e per forman c e
a n d are related nons ingularly t o three conditions s ought. If any one of t h e state
c onvergence tests fails, or if the throttle control r ecovery time conve r gence t e s t
fails, the b r aking- phase t argets a r e c o r r e cted an d anothe r s im ulation i s run; otherwise
the targeting i s concluded by corre cting the ignition algorithm inputs p e r E q . ( 2 2 . 1 8 ).

60
 REFERENCES

1. K ri e g s m an , B . A. ," R adar- Updated Inertial N avigation of a C ontinuou s ly ­

Powered S p a c e Vehicle ' ' , I E E E Ae rospac e S y s t e m s C onfe rence, S e attle,

W ashington, July 1 1 - 1 5, 1 9 6 6 .

2. W i dn all, W . S . , " Lunar Module Digital Autopilot " , Journ al o f Space c r aft an d
Rocket s , Vol. 8 , N o . 1 , J anuary 1 9 7 1 .

3. C h e r ry , G . W . , 1 1 E G uidance - A G en e r al E xpli c it, Optimizing Guidan c e Law

for Rocket- P ropelled Spac e c r aft " , M I T Instrumentation Lab o r atory R eport
R - 4 5 6 , Augus t 1 9 6 4 .

4. Klumpp, A. R . , 1 1 A M anually Retargeted Automatic De s c ent and L anding System

for LE M 1 1 , M I T Instrumentation L ab o r atory R eport R - 5 3 9 , M ar c h 1 9 6 6 ,

5. K lumpp, A. R . , ' ' A M anu ally R etargeted Automatic L anding System for the Lun ar
M o dule ( L M ) 11 , Journal of Spacec r aft and R ocket s, F eb ruary 1 9 6 8 ,

6. ?vl oore, T. E . , G.G. M cS w ain, and J . D . Montgomery, '' Guidan c e Laws for
C ontrolling Off- N om inal LM Powered De s c ent T r aj ectorie s B ac k to the
I\ omin al 1 1 , Inte rnal N ote MSC - E G - 69 - 9 , P roj ect Apollo, N AS A, M anned
Sp ac e c raft C enter, Houston, Texas, F ebru ary 2 8 , 1 9 6 9 ,

7. l\l c S w ain, G . G . and T . E . Moore, 1 1 F al s e High G ate T argeting fo r L M Powe red

Des cent" , Internal N ote M SC - EG - 6 8 - 0 7 , N AS A, M anned Space c r aft C enter,
Houston, Texas, M ay 2 7 , 1 9 6 8 .

8. Y ang, T . L . , 11 A T ar geting Scheme for F u e l Optim al R o c ket T r aj e ctories -

W ith App lic ations to the LM De s c ent B r aking P h a s e " , Technic al M e mo r an du m
T M - 7 1 - 20 1 4- 1 , Bellcomm J anuary 2 2 , 1 9 7 1 .

9. MIT C h arle s Stark D r ap e r Laboratory R ep o rt R - 5 6 7, " Guidance System

Operation s P l an for M ann e d L M E arth Orbital and Lun ar M i s sions U s in g
'.
P rogram Luminary 1 D . S e ct ion 5, Guidance Equation s , R ev . 9 , De c e m b e r
1 9 70 .

61
1 0. Eyle s , D. E . , 11 Apollo LM Guidan c e and P ilot- A s s i stan c e During the F in al Stage
of Lunar De s cent - Soft w are C onsideration s " , F ourth IF AC Symp o s ium on
Automatic C ontrol in Space, Dubrovnik, Yugos l avia, September 6 - 1 0 , 1 9 7 1 .

11. K lumpp, A. R . and G . R . K alan, " E limin ation of N oi s e and E nhanc ement of
Stab ility and Dyn amic R e sponse of the Apollo L M R at e - of - De s c ent P ro g r am " ,
MIT C h arle s Stark Dr ap e r L ab o r atory R eport E - 2 54 3 , October 1 9 7 0 .

1 2. K lumpp, A. R . , " F IN DC DU W - Guidan c e Autopilot Interfac e Routine", M IT

Instrumentation Labor atory L U M IN ARY Memo N o . 2 7, R ev . 1 , September 2 6 ,
1 968.

1 3. Bennett, F. V., " Lunar De s cent and As cent T r aj e ctorie s " , AIAA E i ghth
Aerospace Sciences M eeting, N ew York, J anuary 1 9 - 2 1 , 1 9 70 .

1 4. Bennett, F . V . , " M i s s ion P l anning for Apollo Lunar Module D e s c ent and Ascent",
to be publi shed as a N AS A Techni c al Note.

62
 R-695

DIST RIBU T ION LIST

Internal:

P. Adle r G . K alan
S . Albe rt D . Keene
R . Battin J. Kernan
L . Berman A. Klumpp ( 1 5 0 )
H. Blair -Smith R . Lar son
F . Brun swick J . L aning
S . Copps B . M c C oy
L. Drane R . Metzinger
T . E delbaum D. Millard
A. E ngel P. Mimno
D. E yle s D. Moore
D. Farrar J. N evins
P. F e lle man N. Pippenger
S. F e mino J. R e e d
T. Fit zgibbon P . Rye
D. F r a s e r J . Scanlon
R . Gilbert MIT / KSC R. S chlundt
K . Glick c. S chulenberg
E . C. Hall N . Sears
M. Hamilton R . Stengel
P. He ine mann P. V ol ante
D. Hoag R . W e athe rbee
A. Hopkins P. Wei s s man
I. Johnson CSDL / T D C ( 1 0 )
M. Johnston Apollo Library ( 2 )
Group 2 3A ( 1 8 )

D-1
Ext e rnal:

N A SA / RA SPO (1 )

D E LC O ( 3)

Kolls man (2)

Raytheon (2)

M SC : ( 3 7 & 1 R)

National Ae ronauti c s and Sp ace A dmini stration

M anne d Spacec raft C ente r
Houston, T exas 7 7 0 5 8
ATTN : A pollo Document C ontrol Group (BM 8 6 ) ( 1 8 & 1 R)
M . Holley (EG 1 4 ) (2)
T . Gib son ( FS 5 ) (1)
B . Kirkland ( FM 2 ) (1)
B. Taylor ( FM 2 ) (1)
J . Alphin ( FM 2 ) (1)
W . Bolt ( FM 2 ) (1)
D . J e z e w s ki ( FM 8 ) (1)
F . Be nnett ( FM 2 ) (1)
J . Garman ( F S 5 ) (1)
D . Scott (C B ) (1)
A . W o rden (C B ) (1)
P . C onrad (C B ) (1)
J . Young (C B ) (1)
K . Mattingly (C B ) (1)
F . Raise (C B ) (1)
E . Mitchell (C B ) (1)
R . Gordon (C B ) (1)
V . Brand (C B ) (1)
D. Cheatham (EG) (1)
R . Chilton (EG) (1)

KSC : ( 1 & 1 R)

National Ae ronautics and Space Administration

J . F. Kennedy Space C ente r
J . F. Kennedy Space C ente r, Florida 3 2 8 9 9
ATTN : Te chnical Document C ontrol Office
R . Pearson (C FK) (1)

LRC : (2)

National Ae ronautic s and Space A dministration

Langley Re search C ente r
Hampton, Virginia
ATTN : M r . A. T . Matt son

GA : (3& 1R)

Grumman A e rospace C o rporation

Data Ope ration s and Se rvice s , Plant 2 5
Bethpage , Long I sland , New York
ATTN : Mr. E . Stern

D-2
NAR: (4& 1 R)

N o rth Ame rican Rockwell, Inc .

Space Division
1 2 2 1 4 Lakewood Boulevard
Downey, C alifornia 9 0 2 4 1
ATTN : T . R . W atson; C SM Data Management
D / 0 9 6 - 40 2 AE 9 9

N A R RASPO : (1)

NASA Re sident Apollo Spacec raft P rogram Offi c e

N o rth A me rican Rockwell, Inc .
Space Divi sion
1 2 2 1 4 Lakewood Boulevard
Downey, C alifornia 9 0 2 4 1

GE: ( 1)

General Electric Company

Apollo Syste m s
P . 0. Box 2 5 00
Daytona Beach, Florida 3 2 0 1 5
ATTN : E . P . Padgett, J r . / Unit 5 0 9

H DQ : (1)

National A e ronaut i c s and Space A dministration

Washington DC 2 0 5 4 6
ATTN : G . C he rry

J PL: ( 2)

J et P ropulsion Laboratory
4 8 0 0 Oak Grove Drive
P a s ad e na, C alifornia 9 1 1 0 3
ATTN : R . Mo rris
W . B re cke nridge

D-3