v

I
j{K,

FLIGHT SIMULATION MOTION-BASE DRIVE ALGORITHMS:

PART 1 - DEVELOPING AND TESTING THE EQUATIONS

by

'\,

L. D. Reid and M. A. Nahon

Submitted March 1985

,.

UTIAS Report No. 296

December 1985 eN ISSN 0082-5255

Bibliotheek TU Delft / LR

1111111111111
C 2119439
 Acknowledgement
The work presented in thi s report was performed during the fi rst year
of a three year contract hel d wi th the Canadian Transportati on Development
Centre (Contract No. OS083-00101, Project No. 6069). Technical assistance
with certain aspects of the algorithm development was kindly provided by CAE
Electronics Ltd. The authors would also like to thank Mr. W. O. Graf of
UTIAS for his valuable inputs concerning the practical side of software
implementation.

ii
 Abstract

This report covers the first phase of a project aimed at developing and
test i ng fl ight si mul ator moti on-base al gorithms suited to commerc ia 1 jet
transports. Full six degrees-of-freedom motion of a synergistic motion
-base is being studied. Three forms for these algorithms are being
considered:

(1) classical linear washout,

(2) optimal control ,
(3) coordinated adaptive washout •

It is felt that the latter two techniques may provide some advantages over
the classical, which is currently employed in most commercial flight
simul ators. The aims of thi s proj ect are to:

(1) develop the necessary equations,

(2) implement the necessary software in real-time
(3) evaluate the performance of the software with the help of
test pilots in a complete flight simulation.

The present report describes the development of the equations and the
software used for testing them. At this stage the evaluation of the
performance of the algorithms is based on predicted pilot motion sensations
and motion-base travel. The motion sensations are formulated from linear
model s of the human vesti bul ar system. At present the actual moti on-base
hardware is not employed.

The results of many test runs of the al gorithms are presented to

demonstrate their proper operation and their effectiveness. Several
modifications to existing formulations are suggested.

iii
 TABLE OF CONTENTS
Acknowledgement ii
Abstract iii

Notation vii
1.0 INTRODUCTION ..... ·. ..... 1.1
2.0 BACKGROUND INFORMATION • 2.1
2.1 Reference Frames ·................... 2.1

· . . . . . . . . . . . . . . . ·. .. .. .. .. ..
2.1.1 Frame FD · . . . . . . 2.1
2.1. 2 Frame FC 2.1
2.1. 3 Frame FS ·........ ·.....
2.1.4 Frame FA · . . . . . . . . . . . . . . . . . . . . .
2.1
2.1
2.1. 5 Frame Fps ·....... 2.1
2.1.6 Frame Fpa · ....... 2.1
2.1. 7 Frame Fa · . . . ........... 2.2
2. 1. 8 Frame FI
2.1. 9 Reference Frame
· . . . ................
Locations .............. 2.2
2.2
2.2 Specific Force ••••• ........... 2.2
2.3 Inertial Acceleration •••• · . ....... 2.2
2.4 Euler Angles • • • • • • • • •
2.5 High-Pass Filters . ·. .. .. .. .. .. .. .. 2.3
2.4
3.0 VESTIBULAR MODELS .. .. .... ... .. 3.1
3.1 Rotational Motion Sensation . . .. ....... 3.1
3.2 Specific Force Sensation ••• · . . . . . . . . . . . . . . 3.3
3.3 Threshold Values • • • • ............. 3.7
4.0 FAST INTEGRATION METHODS ••• ·..... 4.1
4.1 Introduction ••• . . . . . . . ·. 4.1
4.2 Results - Case 1: Fifth-Order Optimal Washout Filter . . . 4.3
4.2.1 Running Time - Case 1 •• • • • • • • • • • •• 4.3
4.2.2 Sinusoidal Input Performance - Case 1 • • • • •• 4.3
4.2.3 Step Input Performance - Case 1 • • •• 4.4
4.2.4 Ramp Input Performance - Case 1 ••• • • • • •• 4.4
4.2.5 Filtered Random Noise Input Performance - Case 1 • •• 4.4
4.3 Results - Case 2: Second-Order Classical Washout Filter •• 4.4
4.3.1 Running Time - Case 2 • • • • • • • • • • • • • • • • 4.5
4.3.2 Sinusoidal Input Performance - Case 2 • • • • • • • • 4.5
4.3.3 Step Input Performance - Case 2 • • • • • • • •• 4.6
4.3.4 Ramp Input Performance - Case 2 • • • • • • • • • • • 4.6
4.3.5 Filtered Random Noise Input Performance - Case 2 • •• 4.6

iv
 4.4 Conclusions and Recommendations ·.......... 4.6

4.4.1 Case 1 - Fi fth-Order Optimal Washout Fil ter • 4.6

4.4.2 Case 2 - Second-Order Classical Washout Filter 4.7
4.4.3 General Concl usions ••• 4.7

5.0 CLASSICAL WASHOUT FILTER ALGORITHM • 5.1

5.1 Transl ati onal Moti on Channel ·............ 5.2

5.2 Rotational MJtion Channel ·....... 5.3

5.2.1 Simulation of Aircraft Rotational Velocity • 5.3

5.2.2 Tilt-Coordination •••••••• ·.. 5.5

5.3 Selecting alocation for FS • • • • • • • 5.8

5.4 Inputs for the Vestibular Response Models 5.9
5.5 Filter Equations • • • • • • • • • • ·....·.. 5.10

5.5.1 High-Pass Filters 5.10

5.5.2 Low-Pass Filters ·.. 5.12

5.6 Test Runs 5.13

5.6.1 Algorithm Configuration • • • • • • • • 5.13

5.6.2 Input Test Signals • • • • • • • • • • • •• • • 5.14
5.6.3 The Effects of Filtering Frame Selection ••••• 5.15
5.6.4 The Effects of the Location of FS ••••••••• 5.16
5.6.5 Results • • • • • • • • • • • 5.17

6.0 OPTIMAL CONTROLLER ALGORITHM • ·..... 6.1

6.1 System Oynamics • • • • • • • • • • · . . . . . . . . . . . 6.1

6.2 Optimal Controller Equations • • • • · . . . . . . . . 6.2
6.3 The Uncoupled System Equations ·. ..... 6.5

6.3.1 Vestibular Models •••••• · . . . . . . 6.5

6.3.2 Collected System Equations • ·... 6.11

6.5 Optimal Controller· Desi

· · gn· · ·· ·· ··. · · · · · · · · ·• ·· · · ·
6.4 Filter Equations 6.13
6.15

6.5.1 Expressi ons for F

··• ··
6.5.2 Vestibular MJdel Parameters ·····
· · · · · · · · · •
6.15
6.15
6.5.3 Random Input Parameters •
6.5.4 Se 1ec t i ng the Order of the Fi · · · ·· ·· ·· · · · · · · ·
· lters 6.16
6.16
6.5.5 The Generation of W(s)
··• ··· • • · 6.18

6.6 Test Runs. . . .

········· ·········· 6.20

6.6.1 Algorithm Configuration •

6.6.2 Inputs for the Vestibul ar Response · · · · Model
· · ·s · ·· ·· ·• ·· 6.20
6.21

6.6.4 Results . ··
· · · · · · ·· ·· ·· ·· ·· · · ·· ·· ·• ·· · · ·
6.6.3 Input Test Signal s 6.21
6.21

v
7.0 COORDINATED ADAPTIVE WASHOUT ALGORITHM ••
······· 7.1
7.1 System Equations •• ··· 7.1

. · · · · ·· ·· · ·
7.1.1 PitchjSurge ••••• 7.2
7• 1. 2 Ro 11 Sway • • • • • •
j 7.7
7.1.3 Yaw • • • • • • • • • ····• ···· 7.10
7.1.4 Heave • • • • • • • • • • • • •
7.1.5 Including Vestibular Terms in J .
• · · · · ······ 7.12
7.14
7.1.6 Including IpPAI(PPA - PPS)2 in J y · · · · ··· 7.18
7.2 Test Runs . . . . . . . . . . . . . 7.19

7.2.1 Algorithm Configuration · · · · · · ·· · · 7.19

7.2.2 Inputs for the Vest i bul ar Response
7.2.3 Input Test Signals . . . .
~1odels.

· · ·· ·
·
7.19
7.20
7 • 2. 4 Re sult s . . . . . . . . . . . . ·····
···· ····
7.20
7.2.5 Comparisons with Reference 7.3 . 7.22

8.0 SUMMARY . . . . . . . . . . . . . . ·... 8.1

APPENDIX A: EQUATIONS PROGRAMMED FOR FAST INTEGRATION METHODS A.1

APPENDIX B: SCALING AND LIMITING B.1

APPENDIX C: FILTERING FRAME EFFECTS C.1

APPENDIX D: ACTUATOR EXTENSIONS 0.1

APPENDIX E: TILT-COORDINATION EXAMPLE E.1

APPENDIX F: PROGRAM DESCRI PT ION FOR WASHOlJT FILTER EVALLIATION F.1

APPENDIX G: SINE AND COSINE APPROXIMATIONS G.1

APPENDIX H: THE OPTIMAL SIMULATOR PROBLEM H.1

APPENDIX I: THE UNCOUPLED SYSTEM EQUATIONS 1.1

APPENDIX J: PROGRAM DESCRIPTION FOR THE DESIGN OF THE OPTIMAL WASHOUT
FILTERS J.1

vi
 Notation

a acceleration of a point with respect to FI

+
intermediate acceleration variables in the classical
washout algorithm
ae intermediate acceleration variable in the coordinated
adaptive washout algorithm

system matri ces

A· S the coordinates of the upper bearing block of the i-th

-1
actuator in FS components
otolith vestibular model parameters

control matrices

B· I the coordinates of the lower bearing block of the i-th

-1
actuator in FI components
white noise gain matrix
otolith vestibular model parameters

system state observation matrices

d otolith displacement
dTH otolith threshold value
o output of otolith threshold

Q., 0 system control observation matrices

e motion sensation error

E system kinematics matrix
f specific force = a+ - 9+
+

fp specific force sensed by pilot

+

f1 intermediate specific force variable in the classical

washout algorithm
fEl, fE2 estimated specific force
+ +
see Equations 7.16 and 7.41

vi i
F optimal control feedback matrix

FX reference frame X

9 the acceleration due to gravity

~

G weighting matrix

Gox ' Goy ' hOZ otolith vestibular model gains

r,S p' hSq ' (\r semicircular canal vestibular model gains

GXj' hyj , Gz ' h<\l steepest descent gains

h time step in the numeri cal integration algorithm

hny see Equation 7.45

H white noise gain matrix
I the identity matrix
J, .J x ' J y ' IJZ' J<\I system cost function or functional
kxi ' kyi ' kzi ' k<\l coordinated adaptive washout algorithm gains
K otolith vestibular model gain
K system kinematics matrix
limiting and sealing matrix
~.
~1
displacement of the i-th hydraulic actuator

,Uil see Equation 5.39

lXY rotation matrix going from fral'1e Y into fral'1e X

components
n white nai se
p roll rate component of w

Pxi ' Pyi ' Pz' P<\I adapt i ve parameters

PxiO' PyiO' PzO' P<\IO

initial values of adaptive parameters
p solution to the Riccati equation or the adaptive p
parameters (according to the context)

viii
q pitch rate component of w
weighting matrix
r yaw rate component of w
weighting matrices
I

~l' B.2' B.12' B.l intermediate system matrices

location of Fp with respect to FA
a
location of Fp with respect to Fe
a
location of Fp with respect to FS
s
see Equation 2.14

s Laplace variable

II location of FS with respect to F I

t t i me (s)

TL' Ta· TS semicircular canal vestibular model parameters

Ix see Equation 2.12

Tip' Tiq • Tir semicircular canal vestibular model parameters

u x-component of velocity i
u system control vector

v y-component of velocity V
velocity with respect to FI

w z-component of velocity i
~(s ) optimal control open-loop transfer function matrix

weighting parameters in the cost function

system state vector
input and output of a differential equation

xd kinematic state variables

output of vestibular model

ix
 simulator kinematic variables included in the cost
functional J
Euler angles. ! = [~ e ~]T
~c, ~, ~SH, ~SL intermediate system angular variables. ~ = [~x ex, ~X]T
• •
an estimate of 15
y weighting parameter in the cost function
ö cupula deflection in the semicircular canal model
öTH semicircular canal threshold value
output of semicircular canal threshold
6t integration time step
second-order system damping ratio
e Euler pitch angle component of ~

•
tilt-coordination pitch rate limit
p, Pi weighting parameters in the cost function
otolith vestibular model parameters
semicircular canal vestibular model parameters
Euler bank angle component of ~

•
~LIM tilt-coordination roll rate limit
~ Euler yaw angle component of ~

w angular velocity. ~ = [p q r]T

+
A
W angular velocity sensed by pilot
+

an estimate of w
+

intermediate system angular velocities.

wX = [pX, qX, rX]T
first-order high-pass filter break frequency
first-order low-pass filter break frequency
second-order system undamped natural frequency

x
Superscri pts
)X, ()y, ()z x, y or Z cO/llponents
)s rel ated to the sifTIul ator
)a related to the aircraft
)n related to the white noise input
Subscripts
( )e equilibrium value
(Euler Angles)X the Eul er angles rel ating fraMe FX to frame FI

)X cOJTlponents in fraf11e FX (note that subscript is delp.tec1 in

the case of FB)
a property of frame F~ expressed in Fy components (X=P
stands for Pa or Ps) '(note that thp. first subscript is
deleten in the case of FX = FB)
Other SYf'lbols
"x value of x sensed by the vestibular system

x dxjdt
N N successive differentiations (integrations) of x with
x
respect to time if N is positive (negative)

x Lapl ace transfonn of x(t)

I.~I magnitude of ~

x a vector (with components given by a column matrix x)

+

X a matrix
xT transpose of X
X-I inverse of X
DIAG(X I X2 X3) a diagonal matrix
E{x} expected value of x(t)
LH1( x) a limited and scaled version of x(t)
SGN(x) the polarity associated with x
[x]h the output of a high-pass filter with input x

xi
 the output of a low-pass filter with input x
perturbation in x about x = xe
a x b cross-product
~ ~

Units
Standard internati onal units are employed in thi s report. Angul ar measure
is normally given in radians except in the plots of angular displacement and
velocity where degrees are used.

xi i
1.0 INTRODUCTION

Flight simulators are playing an expanding role in the training of

ai rl ine crews. The industry and the regul atory bodies feel that thi s has
resulted in a cost-effective means of improving the aviation safety record.
Advances in technology are resulting in a steady improvement in the fidelity
and hopefully the effectiveness of flight simulators. The motion generation
system is one of the major subsystems employed in modern cOOlmercial
simul ators. Recent hardware developments have improved the fi del ity of
these moti on systems si gnifi cantl y. However the cOOlputer software that
commands the motion system needs further refinement. The lack of realistic
motion cues during a number of flight conditions is still a constant problem
experienced by the ai rl ines. In some instances they resort to effecti vely
turning off the moti on system in order to avoid improper cues. The present
project is directed towards investigating the possibility of applying
improved software al gorithms to commerc ial fl ight simul ators in order to
solve these problems. The new techniques are based on optimal control and
adaptive control methods.

Because a ground-based fl i ght simul ato r mot i on system cannot

dupl icate the motions of an actual aircraft it beCOOles necessary to
determine the best way to util i ze its somewhat 1 imited capabil ities. Thi s
in turn requi res a determination of how motion inputs are detected and
interpreted by human pi 1 ots. In the context of the present report thi s
determination takes the form of mathematical models predicting human
specific force and angular velocity sensation in response to translational
and rotati onal moti on. The mot; on-base dri ve software al gorithms are
intended to maximize the motion cueing effect while restricting the physical
motion to be within the displacement, velocity and acceleration capacity of
the motion system hardware.

The hardware employed in the present study is a synergistic six

degrees-of-freedom motion-base incorporating hydrostatic bearings, delivered
by CAE El ectronics Ltd. in 1983. The cab i s a DC-8 unit provided by Ai r
Canada (see Figure 1.1). The system is run by a Perkin-Elmer 3250 digital
computer.

A sampling of the recent literature on motion-base drive algorithms is

contained in References 1.1 to 1.12. Typically the simulated aircraft's
motion output is modified by software in order to stay within the rnotion
constraints of the simulator by:

(1) scaling the motion commands,

(2) limiting the motion commands,

(3) high-pass filtering the motion commands.

Sustained translational acceleration is simulated by tilting the simulator

cab. Additional positive software 1 imiting of the motion-base actuator
signal s is provided to take care of any excessive motion signal s which are
not blocked by items (1) to (3) above.

1.1
_
 During the first year of this project three software algorithms have
been stud ied :

(1) classical algorithm of the type currently ef!1ployed in airline flight

simulators (References 1.7, 1.10 and 1.12),

(2) optimal algorithm of the type recently proposed in research journals

(Reference 1.2),

(3) adaptive algorithn of the type employed in some NASA research simulators
(Reference 1.9).

For each of the three algorithms the following steps were carried out:

(1) Literature survey

(2) Study of underlying theory

(3) Formul ati on of equati ons for a si x degrees-of-freedom system

(4) Writing computer programs

(5) Debugging computer programs

(6) Correcting numeri cal problems

(7) Preliminary parameter selection

(8) Run test cases

(9) Modify equations and parameters as required

(10) Determine the influence of parameter variations

(11) Invest i gate novel formul ati ons

(12) Carry out production test runs.

During thi s phase of the study . the test runs have been restricted to
cOOlputer simul ati ons.

1.2
 References

1.1 Cardullo, F. M. and Kosut, R. L., "Old Problem/New Solutions: Motion

Cueing Algorithms Revisited". AIAA Paper 83-1082, ,June 1983.

1.2 Sivan, R., Ish-Shalom, J. and HJang, J-K., "An ~timal Contro 1
Approach to the Design of Moving Flight Simulators". IEEE
Transacti ons on Systems, Man, an<;l Cyberneti cs, Vol. SMC-12, f'b. 6,
Nov./Dec. 1Q82, pp. 818-827.

1.3 Jex, H. R., Magdaleno, R. E., Jewell, W. F., Junker, A. and Mcmllan,
G., 11 Effects on Target Tracking of Motion Simul ator Drive-Logic
Fil ters". AF M1RL-TR-80-134, August 1981.

1.4 Hofmann, L. G. and Riedel, S. A., "Manned Engineering Flight

Simulation Validation, Part I: Simulation Requirements and Simulator
Motion System Performance". AFFDL- TR-78-192 Part 1, ~vember 197R.

1.5 Bitner, M. E., 11 Investigation of Motion Base Drive Techniques".

NADC-77306-20, Harch 1978.

1.6 Sinacori, J. B., "The Determination of SOme Requirements for a

Hel i copter Fl i ght Research Simul ati on Fac il ity" • NASA CR-152066,
Se ptember 1977.

1. 7 Baarspul , M., "The Generation of ~tion Cues on a Six-Degrees-of-

Freedom Motion System". Delft University Report LR-248, June 1977.

1.8 Sinacori, J. B., "A Brief Survey of r1:>tion Simul ators' Drive Logic
With 8nphasis on the Roll Axis". sn Horking Paper ~. 1094-2, May
1977 •

1.9 Parrish, R. V. and Martin, D. J. Jr., "Comparison of a Linear and a

Nonl inear ~Jashout for Motion Simul ators· Util izing O:>jective and
Subjective Data From CTOL Transport Landing Approaches". NASA TN
0-8157, June 1976.

1.10 Parrish, R. V., Dieudonne, J. E. and ~-1artin, D. J. Jr., "~1otion

Software for a Synergi st i c Si x-Deg ree-of-Freedom ~t>ti on Base" • NASA
TN 0-7350, December 1973.

1.11 Friedland, B., Ling, C. K. and Hutton, ~1. F., "Quasi-Optimum Design of
a Six Degree of Freedom Moving Base Simulator Control System". NASA
CR-2312, October 1973.

1.12 Schmidt, S. F. and Conrad, B., "Motion Drive Signals for Piloted
Fl ight Simulators". NASA CR-1601, May 1970.

1.3
 FLIGHT RESEARCH SIMULATOR
UTIAS

F IGURE 1.1
2.0 BACKGROUND INFORMATION
This section contains material on the reference frames employed in
thi s report and summari zes the general mathemati cal fonnul ati ons on which
many of the subsequent developments are based. Most of these expressions
are derived in Reference 2.1.
2.1 Reference Frames
In describing the development of motion drive algorithms it is
convenient to employ a number of reference frames. These are shown in
Fi gure 2.1.

2.1.1 Frame FD
Reference frame FD is located with its orlgln at the centroid of the
simulator's upper frame bearing attachment points (see Figure D.1)~ The
x-axis points forward and the z-axis downward. The x-y plane is parallel to
the upper simulator frame which will be assumed to be parallel to the floor
of the cockpi t.
2.1. 2 Frame Fe
Reference frame Fe is an aircraft-fixed frame which has its ongln
at the same relative cockpit location as the simulator reference frame FO.
Frame Fe has the same orientation with respect to the cockpit as the
simulator frame FO.
2.1. 3 Frame FS
The simul ator reference frame FS has its orlgln at a reference point
selected to suit the motion drive algorithm under study. It is attached to
the simulator cab and is parallel to frame FO.
2. 1. 4 Frame FA
The aircraft reference frame FA has its orlgln at the same relative
cockpit location as the simulator reference frame FS. Frame FA is
parallel to frame Fe. Frame FA has the same orientation with respect to
the cockpit as the simulator frame FS.
2.1.5 Frame Fps
This is a frame attached to the simulator captain's head with its
ori gi n located at a poi nt mi dway between hi s 1eft and ri ght vest i bul ar
systems. The x-axis points forward and the z-axis downward along the spine.
The x-z plane contains the captain's vertical plane of symmetry. In the
present study it will be assumed that Fps is parallel to FO.
2.1.6 Frame FPa
This is a frame attached to the aircraft captain's head with its
origin located at a point midway between his left and right vestibular
systems. Both the aircraft captain and frame Fpa have the same location

2. 1
and orientation with respect to the cockpit as the simulator captain and
frame Fps. In the present study it will be assumed that FPa is
parallel to Fe.

2.1. 7 Frame FB

The aircraft reference frame FB has its orlgln located at the

aireraftls centre of gravity with the x-axis pointing forward and the z-axis
downward. The x-z plane contains the aireraftls vertical plane of symmetry.
The exact orientation of the x-axis wil 1 depend upon the form of the
aircraft equations of motion. This frame is fixed with respect to the
aircraft.

2. 1. 8 Frame FI

The inertial reference frame FI is Earth-fixed with its z-axis

aligned with the gravity vector~. lts origin and the orientation of
its x-axis are selected to suit the problem under study.

2.1.9 Reference Frame Locations

Figure 2.1 shows the relative locations of the reference frames
e mp 1 0 y e din th i s rep 0 rt. Th e vee tor s ~A a n d ~A de fin e
their locations within the aircraft while S and KS do the
same for the simul ator. From the above defi nit i ons gtven for +the vari ous
reference frames it follows that

&A = Bss (2.1)

2.2 Specific Force

Specific force f is defined to be the vector difference between

inertial acceleration and the acceleration due to gravity. That is

f = a - 9 (2.2 )
+ + +
A linear accelerometer responds to specific force as does the human sensory
system (see Section 3).

2.3 Inertial Acceleration

Consider a body having both translational and rotational motion with

respect to FI. The inertial translational acceleration of a point lil on
the body located at a distance ~ from the origin of any body-fixed
reference frame FB can be expressed in FB components by:
•
~iB = ~I+ ~Es} + .f$l!:iB (2.3)
where

2.2
 •
uB + qBwB - rBvB
•
vB + rBuB - PBwB (2.4 )
~B =
•
wB + PBvB - qBuB

0 -rB qB

0 (2.5)
.93 = rB -PB

-qB PB 0

2.4 Euler Angles

Fo 11 owi ng standard aeronaut i cal practi ce the orientat i on between two

reference frames FI and FB can be spec ifi ed by three Eul er Angl es

~ = [<p e <\!J T ( 2.6)

which define a sequence of rotations which will carry FI into FB' The
components of a general vector V expressed in the two frames are
related by +

(2.7)

(2.8 )
~I = 1.IB Is
where -1
1.1 R = !:.s I (2.9)
T
= lsI
The rotation matrix lIB is given by

cose cos<\!
sin<p sine.cos<\! cos<p ~ine ços<\! ]
- cos<p sln<\! + sln<p sln<\!
sin<p sine sin<\! cos<p ~ine sin<\! (2.10)
lIB = cose sin<\! + cos <p cos<\! - Sln<p cos<\!
[
-s i n e sin<p cose cos<p cose

The angul ar vel oc i ty of FB wi th respect to FI can be rel ated to the

Euler angle rates by the fo11owing. Let ~ represent the components of
this angular velocity in frame FB, then

(2.11)

2.3
where
sin4» tane

~
1 cos'.tane

and
la= [
cos4»
sin4» sece
-sln 4»
cos4» sece J (2.12)

•
~= Eaj! (2.13)
where
0 -sine ]
cos4» sin4» cose (2.14)
-sin4» cos 4» cose

2.5 High-Pass Filters

High-pass linear filters are used in many motion-base drive algorithms
in order to bloek the low frequency aircraft motion signals which tend to
over-extend the motion-base actuators. These filters generally have
transfer funtions of the form

TF = snN{s) (2.15)
O(s)
where
N(s) = LK ak sk (2.16)
k=O

O(s) = LJ bJ-sj (2.17)

j=O
and
J ) K+ n (2.18)

ao, b o t- 0 (2.19)
In the case of translational motion, for commercial jet transports the most
demanding low frequency maneuver for the simulator is a constant linear
acceleration. Consider the case where this is represented for the aircraft
by
x = A (2.20)
A

where A is a constant. Assume that the simulator is asked to generate a

translational motion in the same direction as that of. the aircraft but based
on an acceleration command Xs formed by passing xA through the high-pass
filter represented by Equations 2.15 to 2.19. In LapTace transform notation

2.4
for zero initia1 conditions

.. s nN( s)
Xs = xA D(s)
(2.21)

anct taking the Lap1 ace transfonn of Equation 2.20

xA = Als (2.22)

al so
Xs = xS/s 2 ( ? • 23)

Thus the displacement of the simulator based on Equations 2.21 to 2.23 is

given by

( 2.24)

lIsing the fina1 va1ue theorem it follows that

1 i m xS( t) = 1 i m As n-2 N( s) ( 2.25)

t-+<x> s-+O O(s)

A proper1y designed high-pass (or washout) filter is expected to return the

f1ight simulator to its neut ral position (xS = 0) in the long term and
thus from Equation 2.25 it follows that in general it is necessary that
n ~ 3 ( 2.26)

A common fonn adopted for the trans1 at iona1 washout fil ters is

S3 (2. (7)
TF = - - - - - - - - - - - - -

In the case of rotationa1 motion, for commercial jet transports the

most demanding low frequency maneuver for the simu1 ator is a constant
angu1ar velocity. If the above ana1ysis is repeated for this case it is
found that it is necessary that
n ) 2 (2.28)

A common fonn adopted for the rotationa1 washout filters is

2.5
 S 2
TF = ---.-,;;---- (2.29)

.. •
Figures 2.2 to 2.4 give the simulated response xS, xS
and xS for XA = 1 m/s 2 in the case of a filter of the fonn given by
Equation 2.27.

References

2.1 Etkin, B., "Dynamics of Atmospheric Flight". Wiley, New Vork, 1972.

2.6
 Aircraft

Simulator

~J X1_~~//~T777
~/77777777/7 /
/777777777777777777777777777777/

ZI

F1GURE 2.1 AXIS SYSTEMS

 o
o
......

o
ro
o
Wb =O.10 RIS

o WN=O.25 RIS
(!)

o
ZETR=1.7S
(\J
Cf)
.......... 0
L::t'
o
z
o
........
I-
0
0: N
a:
WO
---l
W
U
u 0
0: o

o ---

o
N

o
I

o
::t'

~ I i i i i i i i i i
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00
TIME (SEC)

FIGURE 2.2 RESPONSE OF A THIRD-ORDER HIGH-PASS FILTER TO AN ACCELERATION STEP INPUT

 0
0
....... ,
Wb =O. 10 RIS

U") , ,WN=O.25 RIS

I"'-

0
.
\ZETR=1.75

0
U")
.
/ \
,........ 0
(J)
..........
L
U")
(\J
>-- •
1- 0
I---l

U
0
---1
WO
>~
0

U")
(\J

o
I

o
U")

910.00 I
10.00
I
20.00
I
30.00
I
40.00
I
50.00
I
60.00
I
70.00
I
80.00
I
90.00
T I ME (SEC)

FIGU~~ 2.3 RESPONSE OF A THIRD-ORDER HIGH-PASS FILTER TO AN ACCELERATION STEP INPUT

 0
0
(!)
.
Wb =O.10 RIS

Ol IWN=O.25 RIS
0
.
I..(')

ZETR=1.75

o
O ·

:::r

ZO
o~
~(Y)

I-
~

(J)
o 0
0.... o

ru

o
o
-
O~ _
~ I I I I I I ~ I I
0.00 lU.OO 20.00 30.00 ijO.OO 50.00 60.00 70.00 80.00 90.00
T I ME (SEC)

FIGURE 2.4 RESPONSE OF A THIRD-ORDER HIGH-PASS FILTER TO AN ACCELERATION STEP INPUT

3.0 VESTIBULAR MODELS

The need to predict the sensation of actual physical motion by the

human pilot arises during the evaluation of motion-base drive algorithms and
in the formulation of the optimal control algorithm. Although many sensors
throughout the body play a role in this sensing process it is apparent that
the vestibular system located in the head provides the dominant
signals (Reference 3.1). In order to obtain the desired time histories of
sensation in response to motion, c1ynamic models in the form of differential
equations suitable for implementation on a digital computer are desirable.
Such model s have been developed and are reported in the review paper of
Reference 3.2. It i sapparent that the model representi ng the response to
rotati onal moti on is much better documented than that for transl ati onal
motion. As reported in Reference 3.2 there is some uncertainty concerning
certain aspects of these models, notably:

(i) should motion thresholds be applied to velocity or

acce 1 era t i on?

(ii) do translational thresholds depend to some extent

on the head's orientation with respect to the
9 ravity vector?

(ii1) under what conditions is body tilt interpreted as

sustained translational motion?

(iv) how should the human awareness of linear and

angular displacement be model led?

The foll owi ng model s, al though they do not represent all aspects of
moti on sensati on, are considered to be reasonab ly good eng i neering
approximations to the primary sensations.

3.1 Rotational Motion Sensation

The primary sensing organs for rotational motion are the semicircular
canal s located in the head as shown in Fi gures 3.1 and 3.2. It has been
demonstrated that this system can be interpreted as an angular velocity
sensor and in the present model s three orthogonal sensi ng axes are used to
represent the sensation. For the frequency range of interest in the present
study the model of Young and Dnan (with the lead operator omitted) reported
in Reference 3.2 i s fel t to be adequate. Thi s model is represented in
Fi gure 3.3 and each of the three channel s has the same form although the
parameter val ues may di ffer. The block representi ng the dynami cs of the
semicircular canal itse1f has as its output the deflection ó of the cupula.
Thi s system is model 1 ed as an overdamped torsi ona1 pendul urn. As suggested
in References 3.2 and 3.3 a cupu1a def1ection threshold has been inc1uded in
the model. The fi rst-order hi gh-pass fil ter b10ek accounts for the observed
washout of human response to steady-state rota ti onal accel eration inputs.
The corresponding transfer function for the 1 inear portion of the model is
(for any one channe1)

3.1
 w= Tl Ta S2 (3.1 )
ij) (Tls + l)(T Ss + l)(T a s + 1)

where w is the angular velocity input about any one of the three axes. The
threshold is represented by
II = 0 for 1öl ( öTH (3.2)

II = 6 - SGN(ö)öTH for 16 1 > öTH (3.3)

The differential equation corresponding to Equati on 3.1 can be written as

(following Section 6.4)
• (3.4)
x=Ax+Bw

where (3.5)
w = xl
and

!J
-a2 1
-a 1 0 (3.6)
A= [
-a 0 0

(3.7)

(3.8 )

(3.9)

(3.10)

-1
b 2 = TS (3.11)

The frequency response function corresponding to Equation 3.1 is plotted in

Figure 3.4 for typical yaw channel parameter values of
TL = 10.2 s, TS = 0.1 s, Ta = 30.0 s
lt can be seen from the amplitude ratio plot that this system is a good
sensor of angul ar velocity -in the frequency band 0.2 rls to 10 rls. lts
response w tends to zero for both steady-state angular velocity and
steady-state angular acceleration inputs.

3.2
 A number of past experiments have been performed to determi ne the
val ues for the parameters of thi s model. In these experiments the test
subjects were generally rotated about the local vertical (i.e., the gravity
vector). The axis of rotation passed through the subject's head and they
were oriented 50 that the axis of rotation was also aligned with one of the
three axes of the head frame shown in Fi gure 3.1. Thus it is seen that
under these conditions the models contain effects from all of the body's
sensors and not the semicircular canals alone. This condition is compatible
with conditions which prevail in an aircraft or a simulator. In the present
study parameter values for subjective sensation summarized in Reference 3.2
were employed. These are listed in Table 3.1.
lf the semicircular canal model of Figure 3.3 is to be implemented on
a computer with the limiter in place then the differential equations
representing the system must be solved in two steps. The fi rst part rel ates
ö to wand has the form (following Section 6.4)
• (3.13)
x=Ax+Bw

where (3.14)

and

~J (3.15)

(3.16)

(3.17)

(3.18)

(3.19)

The second part relates hl to /1 (which comes from Equations 3.2 and 3.3) and
has the form
• (3.20)
x = - 1 (x + /1)
Ta

w=x+/1 (3.21)

3.2 Specific Force Sensation

The otolith is the other major sensor contained within the vestibular
system. It senses specific force f defined as the vector difference
-+

3.3
between translational inertial acceleration and the acceleration due to
gravity, that is

.. ..
..f = a - 9 (3.22)

The active portion of the otolith system in man is thought to be contained

within the utricle, the COOlmon base of the semicircular canals. The
response of the otol ith is combined with that from other body sensors to
generate the perceived specific force f with cOOlponents along the
three head axes of Fi gure 3.1. In the" present study the subj ecti ve
sensation model of ~iry and Young slJllmarized in References 3.2 and 3.4 is
employed. Although this model was developed based on horizontal motion of
seated and supine subjects (along the x and z-axes of Figure 3.1) it will be
applied to each of the three axes in the present study.

The model for subj ecti ve spec ifi c force sensati on is presented in
Fi gure 3.5 and each of the three channel s has the same form al though the
parameter values may differ. The blod representing the dynamics of the
otolith itself has the output d, the displacement of the otolith with
respect to the macula. A mechanical sensor threshold follows this sytem.
The fi nal step is neural processi ng represented by a fi rst-order 1ead. The
corresponding transfer function for the 1 inear portion of the model is (for
any one channel)
-=
f K('tas + 1)
(3.23)
f ( 'tL s + 1)( 'tSs + 1)

where f is the spec i fi c force along any one of the three axes. The
threshold is represented by

o= 0 for Id I ( dTH (3.24 )

o= d - SGN( d)dTH for Id I > dTH (3.25)

The different i al equati ons corresponding to Equati on 3.23 can be wri tten as
(foll owi ng Section 6.4)
•
x=Ax+Bf (3.26)

where
(3.27)

and
-a 1]
A= (3.28)
[
-.: 0

-
B = [b 1 boJT (3.29)

a o -- ('t L .'tS)-1 (3.30)

3.4
 -1 -1
a1 :: 'tL + 'tS (3.31)

(3.32)

(3.33)

The frequency response function corresponding to Equation 3.23 is plotted in

Figure 3.6 for typical parameter values of

'tl = 5.33 s, 'tS = 0.66 s, 'ta = 13.2 s, K = 0.4

It can be seen from the amplitude ratio plot that the system is a good
sensor of specific force in the frequency band 0.2 rls to 2 rls. It has a
steady response to constant val ues of f. Because thi s system wi 11 respond
to the gravity vector alone when the translational inertial acceleration of
the head is zero, it can be used to determine one's tilt orientation with
respect to the local vertical.

It appears that only a few studies have been performed in the past to
determine the form of Equation 3.23 and its parameter values (e.g.,
References 3.4 and 3.5). In the work of tv'eiry reported in Reference 3.4,
the otolith responses along the x and z axes of Figure 3.1 were studied by
placing subjects in the seated and supine position on a trolley that could
be translated in the horizontal plane. Based on this work tv'eiry and Young
(as summarized in Reference 3.2) produced the parameter val ues 1 isted in
Table 3.2. The same set of parameters are applied to all three axes. Again
the model represents an aggregate of the response of the compl ete set of
sensors available to pilot.

The val ues of the threshol d 1evel, dTH in Equati ons 3.24 and 3.25,
have been determined from the response latency times reported in Reference
3.4 and the model of Figure 3.5. These latency times are plotted in Figure
3.7 and they represent the time following the aplication of a translational
acceleration step at which the subjects reported the onset of motion
sensation. The same acceleration step was applied to the model (using the
model parameters of Table 3.2) and the level of d evaluated at

t = latency time - reaction time

(Note that the response of the subject to the sensation of motion wil 1
involve a certain internal reaction time as we"). It was found that for an
assumed reaction time of 0.25 s, the values of d (p1otted in Figure 3.8)
were fairly constant over each data set. These values were taken to be dTH
and are recorded in Table 3.2 as wel1. Note that it has been assumed that
the threshold along the y-axis is the same as that along the x-axis.

If the otolith model of Figure 3.5 is to be implemented on a computer

with the 1 imiter in pl ace then the differenti al equati ons representi ng the
system must be sol ved in two steps. The fi rst part rel ates d to f and has
the form (fol1owing Section6.4)

3.5
 •
x=Ax+Bf (3.34)

where
(3.35)

and
'A =[ -all] (3.36)
-aa 0

~ = [0 (3.37)

(3.38)

al = ~
-1
+ 'tS
-1
(3.39)

(3.40)
The second part relates f to 0 (where 0 comes from Equations 3.24 and 3.25)
and has the fonn (following Section 6.4)
-1
A
f =0
•
+ 'ta 0 (3.41)
•
In the computer algorithm 0 is approximated by OA where

OA = w2 ·s (3.42 )
o s + w2

Thus from Equations 3.41 and 3.42

- -1 -1
f = (w2 + 'ta )s + t&'ta (3.43 )
o (s + w2)

and the corresponding differential equation is (following Section 6.4)

•
x = Ax + BO (3.44)

wh ere A

f = x + blO (3.45)

and
A = -aa (3.46)

B = ba - aa b1 (3.47}

3.6
 (3.48)

(3.49)

-1
b 1 = (uQ + 'ra ) (3.50)

In the present case the value of w2 was taken to be

uQ = 12.5 rjs ( 3.51 )

which places the break frequency of Equation 3.42 well above the other
characteristic frequencies of the otolith system.

3.3 Threshold Values

It should be noted that the threshold values reported in Sections 3.1

and 3.2 represent the lower end of the possible spectrum of data. This is
because they were measured in isol ation with the subjects devoting their
full attention to the detection process. As pointed out by the \'Klrk
reported in Reference 3.5, distractions such as mental tasks andjor tracking
tasks can raise thresholds by as much as threefold. Thus the actual val ues
of threshold levels that should be used in assessing flight simulator
perfonnance will vary during the course of a fl ight test in a manner which
cannot be precisely specified at the present time. In this report these low
threshold values will be employed in evaluating all of the motion-base drive
al gori thrns • Thi s wi 11 generate a consi stent but conservati ve est irnate of
the simulator performance in all cases.

References

3.1 Gum, [). R., "Modeling of the H.Jman Force and M:ltion-Sensing
Mechanisms", AFHRL-TR-72-54, June 1973.

3.2 Zacharias, G. L., "~·1otion Cue ~bdels for Pilot-Vehicle Analysis",

AMRL-TR-78-2, May 1978.

3.3 Peters, R. A., "Dynamics of the Vestibular System and Their Relation
to tvbtion Perception, Spatial Disorientation, .and Illusions", NASA
CR-1309, April 1969.

3.4 Meiry, ll. L., "The Vestibular System and Human Oynamic Space
Orientation", NASA CR-628, October 1966.

3.5 Hosman, R. J. A. W. and van der Vaart, J. C., "Vestibular tvbdels and
Thresholds of tvbtion Perception. Resul ts of Tests in a Fl ight
Simulator", Delft University of Technology, Dept. of .Aerospace
Engineering, Report LR-265, April 1978.

3.7
 Table 3.1
Rotational Motion Sensation Model Parameters
(from Reference 3.2)

Pitch Roll Yaw

(about y-axis) (about x-axis) (about z-axi s)
TL(s) 5.3 6.1 10.2
TS(s) 0.1 0.1 0.1
Ta(s) 30 30 30
óTH (deg/s) 3.6 3.0 2.6

Table 3.2
Specific Force Sensation Model Parameters
(from Reference 3.2)

Surge Sway Heave

(along x-axis) (along y-axis) (along z-axis)
'tL ( s) 5.33 5.33 5.33
'tS(s) 0.66 0.66 0.66
'ta (s) 13.2 13.2 13.2
K 0.4 0.4 0.4
dTH (m/s 2 ) 0.17 0.17 0.28

3.8
 Fronto' Pion.

Saoi ttal
AKis

Z Verticol AKis

FIGURE 3.1 HEAD AXIS SYSTEM

(Forward)
X

Left Superior

Left Horizontol Right Horizontal

--1----- Y (Right Side)

Left Post.rior

FIGURE 3.2 ORIENTATION OF SEMICIRCULAR CANALS

 TL·s 8 ~ Toos
w -
~
,.
~

/
V ,. -w
-~
"-

s
( TL S + 1){ T S + 1)
8TH
T S+1
a
--

FIGURE 3.3 MODEL OF THE SEMICIRCULAR CANAL SYSTEM

 o
o
O'-----------------------------~~~----~----~~~----------------------------

o
o
lf)

0
0
0
(Y")
I

~o
CO o
0
~lf)

::1'
I

a:
(IO
0
ei
(!)
I

0
0
lf)
r--
I

0
0
ei
m
I

10- 3
s • 5 e 7 eI
10- 2
S
• 5 e 7 I'
10- 1
S • 5." I.
10 0
, • 5 "" , • S818a

10 2 10 3
FREQ IRISJ
0
0
(!)
(Y")

0
0
::1' Q
N ~
~

0
0 J
'ö" N N
0_

lI:

0
~O
~ . 0
w'T
0

WO
(f)~ ~
(IN ~
:1:-
CL'

0
0
O
:d-
N , ..:r
0
0
(!)
(Y")
, S .., .'" S
• 5 • '7" S • 5. 7" S
• I • ., • •
, '5'7"
10-3 10- 2 la-I 10 0 101 10 3
FREQ IRIS)

FIGURE 3.4 AMPLITUDE RATIO AND PHASE LAG OF THE SEMICIRCULAR CANAL MODEL
 T T-'-r,-1
-- KaL S d _lL 0
-
f
(s + T - 1
) (s + 15- 1
)
--- 7 I
(5 + -ç-1)
--- '"f
L dTH

FIGURE 3.5 MODEL OF THE OTOLITH SYSTEM

 a
a
a ,---------------------------~~~~~------------------------------------

a ~~€HT8~~TB~T8~et~-~
a
lil

a
a
a
(T')
,
~a
CDa
O·
~lIl
::l'
,
cr:
a:g
a
CD
,
a
a
lil
r--
,
a
a
a
(])
, S Ij, 515789 3 Il 515789 3" 515788 S "515709 31l557ag 3 .. 515189
10- 3 10- 2 10- 1 100 101 10 2 10 3
FREO (RISJ
a
a
CD
(T')

a
a
::l'
N

a
a
b~
)I(

a
~a
C) •
W~
0

Wa
c.nC:::
a:N
I-
(L'

a
a
::l'
N ,
a
a
CD
,
(T')
, , 5 I ag , " 5 15
10-3 10- 2 10°
FREO (RISJ

FIGURE 3.6 AMPLITUDE RATIO AND PHASE LAG OF THE OTOLITH MODEL
 8 8

: :~
7
- Model Prediction - Model Prediction
o
:6 I Experimental Dato wlth
One Standard Deviation
I Experimenta I Dato with
One Standard Deviat ion
cP 5 I cP 5
E
..... 4 \ E
~ 4~ I
1\ \
~3
c::
cP
;2 'I'1'1-1_ î 3~ \'I
o 2
....J
-l
1 1 "'-I - -I-r I
O~, __~______~______~______~____~
Î
0
.01 .03 .05 .07 .09 .01 .03 .05 .07 .09
Acceleration (g) Acceleration (g)

Latency Times (or Percept ion of Latency Times tor Percept ion ot
Horizo ntal linear Acceleration, Horizontol linear Acceleration,
Supine Upng ht

FIGURE 3.7 LATENCY TIMES FOR THE OTOLITH (REFERENCE 3.4)

 Cl)
• I 0

I -
Q).!!?
I -
_Cf)
J:: . -

I
e: ><
.- 0
0., I 0'><
._
.... ,0
N
~N
CJ)_
I :3-~ r-
d "'-
(J)

E
I) I) 0
I
I I <.0
•I •I ö

I I
til
~
:;J
H
L.()
~
I I ö I=l

• I • I
H
0
::c
til

~
::c
I I ~
0
H
::c
H
H

I • I H
0
H
0

I I CO
.
I •
t<) ct)

I· Ö ~
:;J
c.!l

• I I H
rx..

I .1 C\J
• I I
ö

I· I• -
I· I• ö
I I·
0
t0 C\J -
Ö C\I d d
~~
E
 4.0 FAST INTEGRATION METHOOS
4.1 Introduction
The quick and accurate numeri cal solution of ordinary differential
equations assumes great importance in moving-base simul ator design. The
vehicle equations of motion and the washout filter equations are largely in
the form of 1 i near and nonl i near simul taneous differenti al equati ons whi ch
must be integrated to obtain the simulator motion. Although good accuracy
is desirable, it is sacrificed to a certain extent, in order to obtain the
fast running times required in real-time operation, and any improvement in
the accuracy of the solution must come at a small cost in running time.
Euler's method (Reference 4.1), which is the simplest, fastest, and
least accurate of the integration methods, is still widely used in
commercial simulators (Reference 4.2) due to its speed and programming
simpl icity. However, it introduces solution errors which are either
corrected by the pilot (an undesirable situation as the pilot is then flying
the errors rather than the aircraft), or accumul ate with time in the
open-loop case and are exhibited as a deviation from the trimmed flying
condition by a nominally trimmed aircraft. An investigation of fa st
i nteg rati on methods was therefore undertaken in an attempt to reduce these
detrimental effects with the minimum increase in running time.
Since an exhaustive treatment of integration methorls is beyond the
scope of th~s work, the reader is directed to any good numerical methorls
text (References 4.1 and 4.3) for more detailed information on the methods
presented herein. The emphasis in these texts is usually placed on
high-order methods which increase running time by at least a factor of two
over Euler's (first-order) methode This investigation concentrated on
methods which increased running time by less than a factor of two while
still achieving substantial improvements in accuracy. Tests were conducted
on two systems; the fi rst was a series of equati ons representi ng a typi cal
fifth-order optimal washout filter; the second was a set of equations
model 1 ing a ty'pi cal cl assi cal second-order hi gh-pass fi 1ter. The test
inputs used (Xl) included sinusoidal, step, ramp, and random functions of
time. The transfer functions of the filters were given as:
Case 1:
~o = a 5 s 5 + a 4 s'+ + a 3s 3 + a 2s 2 + als + a O ~i
(4.1 )
s 5 + e 4S 4 + e 3S 3 + e 2S 2 + els + e o

Case 2:
X 0 = _ _ _ _s_2_ _ _ ~ i
(4.2)
S2 + 2 Lt"
.' nS + t. :'.
n

4.1
The values used for a 5 to ao' e 4 to e o' Uh and ç are given in Table 4.1.
These transfer functions were reduced to the fol1owing sets of
simultaneous first-order differentia1 equations (using Section 6.4):
Case 1:
•
xl = -e 4 x I + x 2 + 1 4 xi (4.3)
•
x 2 = -e 3x I + x 3 + 1 3xi (4.4)
•
x 3 = -e 2x I + x 4 + 1 2 xi (4.5)
•
x4 = -e IX I + x 5 + 1 I xi (4.6)
•
x5 = -e OX I + 1 xi (4.7)
where °
1; = a·1 - ei as (4.8)
and
x ° = XI + X!>xi (4.9)

Case 2:
•
XI = -2 Çwn (x I + xi ) + x 2 ( 4.10)

;2 = -un 2 (x l + xi) (4.11)

and
x ° = x I + xi ( 4.12)
These equations were then integrated to obtain the output (xO) as a function
of time.
The .exact solutions of these equations to sinusoidal, step and ramp
inputs (xl) were found using the J-eaviside Expansion lheorem (Reference
4.4), and compared to the numerical solution to obtain the instantaneous
error at each step. lhe root-mean-square of these errors was eva1 uated and
used as an indication of the re1ative accuracy of each methode
Computer programs were deve10ped to integrate Equations 4.3 to 4.9 or
4.10 to 4.12 using the following methods:
1. Eu1 er (Reference 4.1)
2. Improved Euler (Appendix A)
3. Tustin (Reference 4.2)
4. 2nd Order Mams-Bashforth (Reference 4.1)
5. 2nd Order Runge-Kutta (Reference 4.1)
6. 4th Order Runge-Kutta (Reference 4.1)
7. ABC Method -- Case 2 on1y (Appendix A)
The instantaneous and Rt~$ errors were eva1 uated at each step, as we11 as
the time to integrate over a SOs tota1 simu1ation time. The integration
method, step si ze, and si nusoida1 input frequency (where appl icabl e) were

4.2
user-variable. The equations which were programmed for each method are
given in Appendix A. The running time ratios ([time for given method]/[time
for Euler's method]) are listed in Tables 4.2 and 4.3. The running times
are plotted in Figures 4.2 and 4.12 as a function of the time step size ~t.

4.2 Results - Case 1: Fifth-Order Optimal Washout Filter

As expected, the hi gher order methods were more accurate and sl ower;
although some were more efficient than others. The methods yielding the
lowest time rati os (Improved Eul er, 2nd order Jldams-Bashforth and 2nd order
Runge-Kutta) were chosen for further investigation. For these and Eul er l s
method, plots wre obtained of the input, calculated and exact outputs, and
instantaneous and RMS errors for various inputs with an integration step
size of 0.05s (since a sampling frequency of 20 Hz was the projected
capability of the present simulator installation, this step size was
considered most relevant). $ome representative plots are shown in Figures
4.1 and 4.6 for sinusoidal and step inputs. As can be seen from plots 4.1b
and 4.6b, even the least accurate method (Euler) yielded solutions virtually
identical to the exact ones.

It was apparent that, for thi s system, none of the methods

investigated was clearly best in both accuracy and time. Rather the
performance varied with the input and the final choice was an exercise in
compromi se.

4.2.1 Running Time - Case 1

The total time to integrate over a 50s simulation time was plotted
against step size for each method (Figure 4.2) and yielded the expected
results. The running times remained in the direct ratio given in Table 4.2,
and increased inversely with step size. Since the step size was fixed once
chosen, the total time was unaffected by the input.

4.2.2 Sinusoidal Input Performance - Case 1

Four different sinusoidal input frequencies were used (w = 0.8, 3, 10,

30 rls). At the lowest frequency (0.8 rls), the results were as expected
(Figure 4.3) with Euler performing the worst and 4th order Runge-Kutta the
best. As the input frequency was increased (Figures 4.3a to 4.3d) , all
methods tended to the same accuracy except for Eul er which was consi stently
the worst tand the 2nd order Jldams-Bashforth which was roughly midway
between Eul er and all other methods. Typi cal var i ati ons of accumul ated RMS
error with time are shown in Figure 4.4 for an integration step si ze of
0.05s and an input frequency of 0.8 rls. The RMS error at t = 50s shown in
Figure 4.4 would then be used to obtain points on Figure 4.3a.

Another basi s for eval uati ng the · integrati on methods was the plotti ng
of the RMS error versus running time, which gave an indication of the
relative efficiency of each methode Figures 4.5a and 4.5b are
representati ve plots showi ng how the effi ciency of the methods changes in
going from the lowest to the highest input frequencies.

4.3
 Considering its speed, the Improved Euler flethod yielded the best
overall perfomance, si nce its resul ts i n campari son to the hi gher order
methods deteriorated only at the lowest input frequency.

4.2.3 Step Input Performance - Case 1

The instantaneous error for the step input solution was largest at tl~
step, and then damped out to zero more or less quickly depending on the
methode The RMS error after a SOs simulation time (Figure 4.7) was not, by
itsel f, a good eval uati on of the performance. Instead, the instantaneous
variation of error with time was also considered (Figures 4.8a and 4.8b).
Judging from peak error, time to reach zero error, and RMS error, it was
felt that all methods were roughly equally accurate. The fastest methods
(Eul er and Improved Eul er) were therefore the best overall performers for
this input.
4.2.4 Ramp Input Performance - Case 1
The results with a ramp input wre simil ar to those wi th a low
frequency si nusoidal input, as shown in Fi gure 4.9. The accuracy improved
with increasing order of the methode At a step size of 0.05s, all methods
of order higher than one gave similar accuracies, and performed much better
than the Euler and Improved Euler Methods. Since the 2nd order
Adams-Bashforth algorithm was the fastest of these higher order methods, it
would be the best compromise solution for this type of input. If
integration time were critical, the Improved Euler flethod might be the best
choice.

4.2.5 Filtered Random Noise Input Performance - Case 1

The accuracy for this type of input could not be quantitatively
evaluated ·, as the exact solution was not known. Visually, differences in
the plots of the output obained with the various methods were
indistinguishable. Figures 4.10, 4.11a and 4.11b show the input and outputs
from Eul er IS and the 4th order Runge-Kutta flethods (the 1east and most
accurate methods respectively). A brief comparison was made of the
tabul ated numeri cal output val ues for each method as compared to the 4th
order Runge-Kutta which was the most accurate of the methods eval uated. It
wa s found that the Tust in and . 2nd order Runge-Kutta tended to show
deviations in the 5th significant figure, the Improved Euler and 2nd order
Adams-Bashforth Methods deviated in the 4th significant figure, while
Euler's flethod deviated in the 3rd significant figure. Considering
computing speed, the best compromise performances in this test were obtained
with the Improved Euler and the 2nd order Runge-Kutta flethods.

4.3 Results - Case 2: Second-Order Classical Washout Filter

As in the previous test case, the higher order methods tended to yield
greater accuracy at the cost of increased running time. The relative
performance of each method was simi 1 ar to the previous secti on with some
notable exceptions:

4.4
 1. The Tustin Method maintained its very low error, comparable to
a fourth-order Runge-Kutta, but its runn i ng time rati 0 was
reduced to 1.22 (from 2.37 in the previous section). This was
due to the derivati on of the Tust in equati ons whi ch tends to
be much 1ess campl ex for low-order transfer functi ons. Thi s
radical reduction in running time makes the Tustin Method
eminently suitable for the fast mmerical solution of linear
low-order differential equations, and it clearly surpassed the
other methods in most of the following tests.
2. The Improved Euler t-'ethod, while still fast, gave the ~rst
results for the ramp input wi th thi s fi lter. It was found
that this method reacted . bad1y to these specific conditions
due to its form, and \'tOu1d yie1d even ~rse resu1ts for a
first-order system with a ramp input. However, its
performance for other inputs to 1ow-order systems was
commendab 1e •
Fewer test s were run for thi s fi 1ter case si nce the general resu1 ts
were similar to the previous section. The running times and time ratios are
shown in Figure 4.12 and Tab1e 4.3, whi1e the RMS error plots for various
inputs are shown in Figures 4.13 to 4.15.
Contrary to the previous case, these tests clearly pointed to a best
compromise in the Tustin t-'ethod which yie1ded much better accuracy than the
Eu1er Method with on1y a 22% increase in running time.

4.3.1 Running Time - Case 2

The tota 1 time to integate over a 50s simul ati on time is shown as a
function of step size in Figure 4.12, and these times stayed in the direct
ratios given in Tab1 e 4.3. It shou1d be noted that the ABC method is faster
than the Eul er Method, that the improved Eu1 er t-'ethod has a 1 arger time
ratio than in the previous section, whi1e the Tustin method has a much
reduced time ratio. For this second-order system, most methods yie1ded time
ratios lower than two, thereby making more choices avai1ab1e for a fina1
sol ut i on.

4.3.2 Sinusoida1 Input Performance - Case 2

Three different sinusoidal input frequencies were used (w = 0.8, 5, 30
rls). The variation in results from lowest to highest input frequency,
shown in Fi gures 4.13a to 4.13c, showed the same trends as in the previous
fi 1ter case. At the 10west input frequency, the Eu1 er flethod gave the \'tOrst
resu1 ts, along wi th the ABC Method. The other methods i n order of
i nc reasi ng accuracy were: Improved Eu1 er, 2nd order Jldams-Bashforth, 2nd
order Runge-Kutta, Tustin, and 4th order Runge-Kutta. At the higher
frequenc ies, the Eu1 er and ABC Methods ·gave the same 10w accuracy, whi 1 e all
other methods yie1ded similar, much higher accuracy, except for the 2nd
order Adams-Bashforth t-'ethod which 1ay between the two groups.
Considering both speed and perfomance, the Tustin t-'ethod is c1ear1y
the best solution for sinusoida1 inputs. The next best methods are the 2nd

4.5
order Pdams-Bashforth rv-ethod whi ch i s a lmost as fast but not as accurate as
the Tustin t-'ethod, and the 2nd order Runge-Kutta which is as accurate but
somewhat slower.

4.3.3 Step Input Performance - Case 2

As in the previous section, it was found that all methods yielded

similar ,a ccuracies for the step input, as shown in Figure 4.14. O1ce again,
the error was at a maximum at the step, and decayed to zero with time. lhe
best perfo rmers in thi s category were the fastest al gorithms (i n order of
increasing speed): Tustin, Improved Euler, Euler, and ABC.

4.3.4 Ramp Input Performance - Case 2

The results with a ramp input, shown in Figure 4.15, were again
simil ar to those with a low frequency si nusoidal input, except that for thi s
fi lter, the Improved Eul er t-'ethod fared appreciably worse. The results
coul d be grouped into two categories: the Improved Eul er, Eul er, and ABC
Methods all yielded similar results, while all other methods gave much
better results • In thi s category, the Tust in t-'ethod aga i n gave the best
compromi se with excell ent accuracy, and only a 22% increase in running time
over the Eul er ~thod. lhe next best compromi se was the 2nd order
Pdams-Bashforth t1ethod.

4.3.5 Filtered Random Noise Input Performance - Case 2

The accuracy for this type of input could not be evaluated accurately
due to the absence of an analytical solution. However, a comparison of the
tabul ated resul ts for each method against the 4th order Runge-Kutta t-'ethod
(the most accurate of the methods eval uated) indicated that: the Eul er and
ABC Methods showed deviations in the 3rd significant figure; the Improved
Eul er and 2nd order Pdams-Bashforth tv'ethods showed deviations in the 4th
si gnifi cant fi gure; the Tusti n and 2nd order Runge-Kutta tv'ethods showed
deviations in the 5th significant figure. From these results , the Tustin
Method cl early appeared to be the best numerical method for thi s type of
i n put.

4.4 Conclusions and Recommendations

4.4.1 Case 1 - Fifth-Order Optimal Washout Filter

Assuming comput i ng time to be of greatest priority, the Improved Eul er

Method is recommended as the best compromise of the methods investigated for
thi s fi 1ter. In most cases, i t gave much improved accuracy over the Eul er
Method at only a small (13%) cost in time. If, however, excess computing
time were available, the 2nd order Runge-Kutta Method is recommended as it
yields excellent results at a 90% increase in running time.

4.6
4.4.2 Case 2 - Second-Order C1assica1 Washout Filter
Under the assllTlption that computing time is critica1, the Tustin
t-1ethod is recommended as the best compromise of the methods investigated.
In all cases, it gave exce11 ent accuracy at a minimal (22%) time increase
over Eu1 er' s tlethod. None of the other methods i nvest i gated \\Ou1 d have
significant1y reduced the error, regard1ess of computing time.

4.4.3 General Conc1usions

It shou1d be emphasized that these resu1ts are specific to the
equati ons sol ved and the inputs se1 ected. Other equati ons or inputs mi ght
best be solved with other methods. However, the following suggestions are
cautious1yextrapo1ated:
1. The Improved Eu1 er tlethod shou1 d not be used for fi rst or
second order systems.
2. The Improved Eu1 er tlethod shou1 d be used for hi gher order
systems, whether 1inear or non1inear.
3. The Tust in tlethod shou1 d be used for fi rst, second, or
thi rd order 1 inear systems. Thi s method app1 ies on1y to
1 inear equati ons, and its equati ons are not as easy to
derive as thase for other methods due to the necessity of
computi ng forward and inverse z-transfonns. However,
once the equations are derived and programmed, the
resu1ts are outstanding.
4. The 2nd order Mams-Bashforth or Runge-Kutta shou1d be
used for 1ow-order non1 inear equations. The former
yi el ds good results at a 42% inc rease in time over the
Eu1er tlethod, but has the disadvantage that it is not
se1f-starting, thereby a1most doub1ing the nllTlber of
1ines of computer code. The latter method is
se1 f-starting and as accurate as the Tustin tlethod for
1 inear equations, but takes 70% more time than the Eu1 er
Method.

References
4.1 Conte, S. D., and de Boor, C., "E1ementary J'umerica1 Jlna1ysis",
Me Graw-Hi 11, New Vo rk, 1980.

4.2 University of Dayton, Short Course on F1 ight Simu1 ation rt>tes, Dayton,
Ohi 0, 1980.
4.3 James, M. L., 9nith, G. M., and \o.b1ford, J. C., "App1ied ~merica1
Methods for Digital Computation with FORTRAN", International Textbook
Company, rëw Vork, 1967.
4.4 Etkin, B., "Dynamics of Atmospheric F1ight", John Wiley & Sans, ~w
Vork, 1972.

4.7
 Table 4.1
Filter Parameter Values

Parameter Val ue

a5 0.69354

a4 0.4373326

a3 0.07578681

a2 0.00367443

al 2.15801x10- 13

ao 8.99722x10- 14

e4 1.129427

e3 0.6191589

e2 0.1798585

el 0.02315727

eo 9.787999x10- 4

UTt 0.3
ç 1.0

4.8
 Table 4.2
Ratio of Running Times - Case 1

Method Measuredft Calculated 11 Notes

Improved Euler 1.13 1.16

2nd Order Adams-Bashforth 1.47 1.50 NSS
2nd Order Runge-Kutta 1. 90 2.01
Tustin 2.37 2.40
4th Order Runge-Kutta 4.06 4.20

Notes: Ratios are (time for method)j(time for Euler's Method)

NSS = Not Self-Starting
ft Measured from actual running times for 50s simulation run.
11 Calculated from number and type of operations performed and
average time per operation for a Perkin-Elmer 3250.

4.9
 Table 4.3
Ratio of Running Times - Case 2

Method Measuredtl flbtes

ABC 0.94
Improved Eul er 1.21

Tustin 1.22
2nd Order Adams-Bashforth 1.42 NSS

2nd Order Runge-Kutta 1.70

4th Order Runge-Kutta 3.31

Notes: Ratios are (time for method)/(time for Euler's flethod)

NSS = Not Sel f-Starting

(j) Measured from actual running times for 50s simul ation run.

4.10
 o
o
.-.

r--
CD
o

en
en
o

o
1-0
~~
z
t---f
'T\---r--iM-+-+-+-W---L-L
r

en
en
o ,

r-
CD
o ,

o
o
.-. I ~J \/ \I I \I \I . I

11. 11 16.67 22.22 27.78 33.33 38.89 llll.llll 50.00

'0.00 5.56
TIME (S E C)

FIGURE 4.1A SINE WAVE INPUT TO THE FILTER EQUATIONS (w = 0.8 r / s or 0.13 Hz)
 1.00
Calculated using Euler's Method

0.67
Output
0.33

o I I

-0.33

-0.67
g- Exocf

' - Calculated using Euler's Method

-1.00
o 5.56 11.11 16.67 22.22 27.78 33.33 38.89 44.44 50.00
Time (sec)

FIGURE 4.1B COMPUTED AND EXACT OUTPUTS FOR EULER'S METHOD: CASE 1 (w = 0.8 rls or 0.13 Hz)

!J.t = 50 ms
 10

Total ...........
............
Time
6~"""""""""",
(5) v.~ ............ ............
~.
~':::::-
~'" "
"';~.:~~"~
- ..............
.......... .....
......... '
~ 6~
.....
...0 .............. " ~ ............ ...... ~ ......

.......
' ~<>~6~ ........
.....
.~
. . . . . . . . . .. . . ~-~:::::::::::~0"
'"
-.........~ "'v
.........
. . . . . . . . . ..........
0.1 IO.~""'''''..................~ ...............

o,~ . . . . . . . . . . . . ~~~:
• 4th Order Runge - Kutto ..... ..... .....
6 Tustin
v 2nd Order Runge-Kutto
<> 2 nd Order Adoms - Boshforth '~~~v
•o Improved Euler o~
............ ...... ....... <>
Euler '~
.01

.005 .05 0.5

~t (5)

FIGURE 4.2 RUNNING TIME FOR 50s SIMULATION: CASE 1

 10-2
0,/°/
RMS /e/e
Error

10- 3 o
/0 e/e

10-4
4 th Order Runge - Kutta
b. Tustin
v 2nd Order Runge-Kutta
<> 2 nd Order Adams - Bashforth
• Improved Euler
o Euler
5
10- 1
.005 .05 0.5
~t (5)
FIGURE 4.3A RMS ERROR AT t 50s: CASE 1 (w = 0.8 rls or 0.13 Hz)
 ,/0"'-
2
10-
%/"o
RMS ~.
~
Error ~
~

IO-3~ 0/°
.4
s>' ..,/-;::/
.~ Jr../

~
~7
10- 4 /
o 2 nd Order Adams - Bashforth
. ' ,, /
", / /
/
•o Improved Euler
Euler
'\'- /{
/ .. All Others
'- --~~/

10-5 1

.005 .05 0.5

6t (S)
FIGURE 4.3B RMS ERROR AT t = 50s: CASE 1 (w = 3 rls or 0.48 Hz)
 10-2

RMS
Error f:)

./.
10- 3
o
/0

10-4
o 2 nd Order Adams - Bashforth
o Euler
À All Others

10- 5 I

.005 .05 0.5

6t (S)
FIGURE 4.3C RMS ERROR AT t SOs: CASE 1 (w 10 rls or 1.59 Hz)
 10-2 . 0'

~.
o /
RMS A

Error
/
.&

10-3

o
/0
A'

10-4 <> 2 nd Order Adams - Bashforth

A'
o Euler
À All Others

10- 5 I

.005 .05 0.5

~t (5)
FIGURE 4.3D RMS ERROR AT t 50s: CASE 1 (w 30 rls or 4.77 Hz)
 .010

RMS
Error

.005

Improved Euler

2nd Order Adams - Bashforth

2nd Order Runge -Kutta

o
o 5 .56 11.11 16.67 22.22 27 .78 33.33 38.89 44.44 50.00
Time (sec)

FIGURE 4.4 RMS ERROR vs. t: CASE 1 (w = 0.8 rls or 0.13 Hz, ~t = 50 ros)
 • 4th Order Runge- Kutto

'" . 0"
. , 0\ ""
b. Tustin

'\0\,,-.
, 0"0"",, / 50
" 'lI.
v 2nd Order Runge-Kutto
10- 2 <> 2 nd Order Adoms - Boshforth
" • Improved Euler
o Euler
RMS -\', 50
Error "\ \',/
., °""-
o DGEAR
"" 50 Results tor 1:1 t = 50 ms
\\,
~
- \\, "'°
10- 3
v "'. '.
\\
\\
~\~
",
"' . - -
. .
_~v
\~V
50 '" ,
//:/
v
'b~O\~_"~o-....jO/b., /
~ \

~~"'~./
10- 4

50'/50 o
50

10- 2 10- 1 10° lOl 10 2

Toto I Time (5)
FIGURE 4.5A EFFICIENCY: CASE 1 (w = 0.8 rls or 0.13 Hz)
 0,'" \ o

\/.
50 •
t:.
4th Order Runge - Kutta
Tustin
~~\\
2
10- v 2nd Order Runge-Kutta
e
o 2 nd Order Adams - Bashforth
• Improved Euler
RMS
\ o~,
o Euler
Error
"~ \0~
o
/e DGEAR

"'0 "" 50 Results for À t = 50 ms

.
10-3 50 \ \\
50 \ \
50 ~ \ t:. \.
"'\
0,
\ \
\ \
\ \ o
\ \
\ \
10- 4
\ 6<> .~
.\~.
10- 2 10- 1 10° lOl 102
Tata I Time (5)
FIGURE 4.5B EFFICIENCY: CASE 1 (w = 30 rls or 4.77 Hz)
 o
o
.--<

r--
(.D

en
en
o

o
1-0
=> .
~ ~-r-L-------------------------------------------------------------------------------------------------------------------

f--I

en
en
o_
I

r--
(.D

o_
I

o
o
~ I I I I 1 I I I I I
0 . 00 5.56 11.11 16.67 22.22 27.78 33.33 38.89 44.44 50 . 00
T I ME (SECj

FIGURE 4.6A UNIT STEP INPUT TO THE FILTER EQUATIONS

 1.00

0.67

Output
0.33

o I I \ ;p--

, Calculated using Euler's Method

-0.33

-0.67

-1.00
o 5.56 11.11 16.67 22.22 27.78 33.33 38.89 44.44 50.00
Time (sec)

FIGURE 4.6B COMPUTED AND EXACT OUTPUTS FOR EULER'S METHOD: CASE 1 (UNIT STEP INPUT, ~t 50 ms)
 L()
ö
L()
~ ..
,
I
\

Q
o
en
Lf"\
11
\ \
\ -\
\ \
\ \
\ \
\ \
\ \
\ \
\ \
o • \
\ \
\ \
\ \
\ \
\ \ L()
. ~. \. o
Q
C\J rt) 10
I
b o b
 .010

RMS
Error

Euler
.005
2nd Order Runge -Kutta 8 Adams - Bashforth
Improved Euler

--- --
o v
o 5.56 11.11 16.67 22.22 27.78 33.33 38 .89 44.44 50.00

Time (sec)

FIGURE 4.8A RMS ERROR vs. t: CASE 1 (UNIT STEP INPUT, 6t = 50 ms)
 .012

.008

Error Euler

.004 / 2nd Order Runge - K utta B Adams - Bashforth

IV~edEUler
o " <7'~

-.004

~008

-.012
o 5 .56 11.11 16.67 22 .22 27 .78 33.33 38.89 44.44 50.00
Time (sec)

FIGURE 4.8B INSTANTANEOUS ERROR: CASE l - (UNIT STEP INPUT, Ót 50 ms)

 • 4th Order Runge- Kutto o

10-2 ll.

v
Tustin
2nd Order Runge-Kutto
•
o 2 nd Order Adoms - Boshforth
RMS • Improved Euler o
Error o Euler
• 0
1
10- 3 /
/
/

<:~,
/ _v
/
/
/
/
...... "- /
•
"
//
4
10- //

- . > --x
,0"'"
äL.. _ _ _ _ _

10-5 1 ~. ......... •
~,

.005 .05 0.5

~t (S)

FIGURE 4.9 RMS ERROR AT t = 50s: CASE 1 (UNIT RAMP INPUT)

 o
o
.....

r-
(!)

Cr)
Cr)

j
r-~
~o
0-
Z
o

f--t
1

Cr)
Cr)
lAl

:\l t - -t~ I1~- - - ~- ~\r~;1~ ~ I t\~-!~ ~ ~1f- - ~ ~\l;J/~k~rl'-1:;'- \

111.1 , I I I . 1111 IllM 11 / I /1 11 V If 11 I ,

o
I

r-
(!)

o
I

o
o
..... ~I-------------r------------~------------_,--------------r_------------~------------~------------,_-------------r------------,
10.00 5.56 11. 11 16.67 22.22 27.78 33.33 38.89 l!l!.l!l! 50.00
TIME (S E C)
FIGURE 4.10 FILTERED RANDOM NOISE INPUT TO THE FILTER EQUATIONS
 o
o
...-<

r--
CD

en
en·
o

~
rO
:::JO
e:~ lAl
0 A I * -·~ '\t/l·\/~Yf~1~Î\ft~l~j~·~'~ ~ ft11\r~~~~~\î~~4r~~--~--~~~~~~,
" VI ,.\I - j 11 I I \ I '\ I'" I1 r 1J \I \1',I"II

en
en
o
I

r--
CD
o
I

o
o
"'-<~I~-----------r------------r------------r------------r------------r------------r------------r------------,-----------,
10.00 5.56 11. 11 16.67 22.22 27.78 33.33 38.89 llll.llll 50.00
TIME (SE C)

FIGURE 4.11A COMPUTED RESPONSE TO NOISE INPUT, EULER'S METHOD: CASE 1 (~t = 50 ms)
 o
o

r-
C.D
o

(Y")
(Y")

~o
=>0
CL ~ 'FIt \
~ I "
I
1'1
AI V\ rVJU\"A
V INt~ A j" V· I 11 \ ". r
1 I \I AA.i"\ I \I \ A
J \J
J
\\ I!\IV \1' \ (f ~ 1I1 I/IA
V \(\ rA\
=>
o
(Y")
(Y")

o
I

r-
C.D

o
I

o
o
-I
'0.00 16.67 22.22 27.78
I
33.33
I
38.89
I
44w44
I
50.00
5.56 11. i 1
TIME (SEC)

FIGURE 4.11B COMPUTED RESPONSE TO NOISE INPUT, 4TH ORDER RUNGE-KUTTA: CASE 1 (~t 50 ros)
Toto I • 4th Order Runge - Kutfa
Time ó. Tustin
v 2 nd Order Runge - Kutta
(5)
<> 2 nd Order Adams - Bashforth
• Improved Euler
10° • o Euler and ABC

10- 1

10-2

.005 .0 5 0.5
~t (5)
FIGURE 4.12 RUNNING TIME FOR 50s SIMULATION: CASE 2
 tfII""/ 09
6'-..,.-
• 4th Order Runge - Kutta
6'-..,.-
6 Tustin ~-"'­
~-..,.­
•
v 2nd Order Runge - Kutto ~o

<> 2 nd Order Adams - Bashforth ~-""'"

10- 2
• Improved Euler ~--'
. 9/ 0

""'"

(.)
~-..,.­
o Euler ~-..,.-
RMS T ABC ~o'"
~o'"
Error ~ . .",. IV
9~-t9' / ~
10- 3 • /
/
/
/
/
0 ~

0/ ~
-,v'; ~
10- 4 : /

0005 .05 0.5

~t (5)
FIGURE 4.13A RMS ERROR AT t = 50s: CASE 2 (w = 0.8 rls or 0.13 Hz)
 10-2 •
•
RMS
Error

10- 3

• 4th Order Runge - Kutto ond Tustin

o 2 nd Order Adoms - Boshforth
10-4 • Improved Euler
o Euler ond ABC

.005 .05 0.5

~t (5)

FIGURE 4.13B RMS ERROR AT t = SOs: CASE 2 (w =5 rls or 0.80 Hz)

 10- 2 ..
Y':
8

RMS
Error
o

10-3

o • 4th ond 2nd Order Runge-Kutto

ond Tustin
o 2 nd Order Adoms - Boshforth
.• Improved Euler
o Euler end ABC
4
10-
•

.005 .05 0.5

~t (5)

FIGURE 4.13C RMS ERROR AT t = 50s: CASE 2 (w = 30 rls or 4.77 Hz)

 ."".-.
10- 2 /-/. &

/- /
./ /// _0

RMS
Error
,.
.
./ &//
/./
/
/
///
.

//
~:// .-0
.~
10-3
o
///.
/'7'
,,"" ./
~./
• Improved Euler
o Euler
& All Others

10-4

.005 .05 0.5

~t (5)
FIGURE 4.14 RMS ERROR AT t = 50s: CASE 2 (UNIT STEP INPUT)
 10-2 •
.~ 0
~.
RMS ~ ~O
Error
•
10- 3 ~ • • 4th Order Runge - Kutto

0
l::.. Tustin
• v 2nd Order Runge - Kutto
b.
<> 2 nd Order Adoms - Boshforfh
• Improved Euler /0
o Euler
• ABC ,/
10-4 _._._<>_._.--./
-. ~
~
:::::=--...-
ei
-------b. re::: .

10- 5 I
.01 .05 0.1
~t (s)
FIGURE 4.15 RMS ERROR AT t 50s: CASE 2 (UNIT RAMP INPUT)
5.0 CLASSICAL WASHOUT FILTER ALGORITHM

In developing the underlying equations for the classical washout

filter algorithm it is assumed that the basic task is one of creating at the
aircrew's location in the simulator a specific force vector and an angular
velocity vector approximating those that they would experience in an actual
aircraft. Specifically for the captain one \\Ould like the components of
these translational and rotational motion effects to obey

(5.1)

(5.2)

Because there may be several aircrew in the simulator cab, the

location at which Equations 5.1 and 5.2 should hold may not be unique. For
this reason it is convenient to designate a single reference point in the
cockpit as the location for which the motion drive algorithms apply and then
attempt to achi eve the rel ati onshi ps expressed by these equati ons. The
actual specific forces at the aircrew locations can then be found from
Equations 2.2 to 2.5. Thus the algorithms attempt to achieve

iss "'.!AA (5.3)

(5.4 )

(see Secti on 2.1 for the 1ocati on and orientati on of reference frames FA
and FS)'

The val ues of .!..AA and !!!AA come from the sol ution of the fl ight
equations with

(5.5)

where from Equation ' 2.2

.!AB = ~AB - .9.B (5.6)

and ~B is given by Equations 2.3 to 2.5 with the subscript repl aced
by A. lhus from Equations 5.5 and 5.6

.!AA = .!::AB ~AB - .9.A (5.7)

Simil arly
(5.R)
!...AA and ~A are the inputs to the · cl ass i cal washout al gorithm. lhey
are scalea-, limited and filtered in order to restrict the motion-base travel
to remain within the physical system constraints.

5.1
5.1 Translational Motion Channel

As shown in Fi gure 5.1 !.AA is fi rst passed through the scal ing and
limiting bloek represented by HP SCALE. This process is described in
Appendix Band acts as a nonlinear attenuator depending only on the
ampl itude of the !.AA components. In the present case the attenuati on
takes the form:

(5.9 )

( 5.10)

where Ki(X) is the limiter shown in Figure B.l. As outlined in Appendix B

the scal ing is carried out on cab frame components in order to avoid
cross-coupl i ng of the sensed moti on st imul i components. f1~ represents the
cab frame specific force components that we wi sh to simul ate in FS.

It is now convenient to generate an intermediate acceleration variable

al as

(5.11)

From Equation 2.2 it can be seen that al represents the acceleration of the
origin of FS, in FS components, that-wil1 generate a specific force of
f1 at the origin of FS, in FS components.
The signal ~S is given by

(5.12)

where LSI is given by the transpose of Equation 2.10 with FB replaced

by FS and 1. repl aced by l.S. Thus

(5.13)

The values for ~S come from the rotational motion channel algorithm.

Because the mot i on-base hardware cannot reproduce much of the low
frequency content of al it is necessary to pass it through the hi gh-pass
filter represented by HP FIlT #1 in Figure 5.1 generating the signal aH. As
ex p1 a i n e din Ap pen d i xC, alt hou 9 h th i s h i 9 h - pas s f i1 ter i n 9 -Of FS
components wil 1 not introduce cross-coupl ing of the sensed motion stimlJl i,
unfortunately it wil 1 not prevent the motion hardware from being driven
against the stops in all cases. Thus aH will also be treated as an

5.2
 k,.",J '"'~
intermediate acceleration signal. In generating motion-base actuator
commands i t is most convenient to work with FI cOfTlponents of the vectors
involved. lhus aH, representing acceleration components in FS, is
tranformed into ~,-rëpresenting acceleration components in FI' by

~ = .hlS aH (5.14)

where .hlS is given by Equation ?10 with frame FB replaced by frame

FS and f replaced by ~S.

As implied above, further high-pass filtering of the acceleration

signal is required in order to successfully restrict motion-base actuator
travel • By high-pass filtering a2, which represents acceleration components
in FI' it is possible to ensureadequate motion travel restriction at the /
expense of introducing some degree of cross-coupl ing of sensed f11otion
stimul i back in FS (see Appendix C). As shown in Figure 5.1, a2 is J
processed by a high-pass filter represented by HP FILT #2 to producethe
final form of the ~otion base acceleration command signal ~SI.

The signal aSI along with the rotation matrix LIS is passed to
the actuator extenSlon algorithm described in Appendix 0 wnich generates the
commanded jack lengths .Roi. lhe block ACCN EST in Figure 5.1 corresponds
to Equation 0.6 in Appendix 0 and is used to generate an estimate of jack
acceleration ti t which is required by the UTIAS motion hase
system. These sr?nals are in turn fed to the motion base controller to
generate the simulator motion response.

5.2 Rotational Motion Channel

As shown in Figure 5.2 there are two c~ponents that go into

generati ng the simul ator Eul er angl es ~S. The cOl"1ponent ~SH comes frofll
hi gh-pass fi 1teri ng !!:AA and i s i ntended to simul ate ffi'ë"" rotati onal
velocity of the aircraft. The second component, ~SL, comes frol'l a
tilt-coordination algorithm and is intended to tilt the----slmulator in order
to simul ate susta ined spec ifi c force cues. Thus

(5.15)

Both processes are described bel ow.

5.?1 Simulation of Aircraft Rotational Velocity

!::!AA is passed through the scaling and limiting block represented

by HP SCALE in Figure 5.2. This process has the same features outlined in
Section 5.1 for the translational channel HP SCALE block. The output from
thi s process is wl and it represents the scal ed angul ar velocity written in
FS components siiTiUIating the aircraft's angular velocity.

Because the motion base hardware cannot reproduce the low frequency
content of w1 it is necessary to pass i t through the hi gh-pass fi lter
represented by HP FILT #1 in Figure 5.2 thereby generating the signal uH.
The si gnal wH has the same properti es as aH desc ribed in Secti on 5.1. The
outputs fromthe rotat; onal mot; on channelrequi red by the actuator 1 ength

5.3
algorithm are based on the simulator Euler angles and thus it is necessary
•
to convert wH to a simulator Euler angle rate ~H. Following Equation 2.11
the appropriate relationship is

(5.16)

where TS is based on ~s. That this fonn leads to suitable siMulator

angular-response can be demonstrated as follows.
Consider the case where the second hi gh-pass fil ter represented by
HP FILT #2 is absent from Figure 5.2. Then
•
~SH =f ~H dt (5.17)

and
(5.18)

In Section 5.2.7 it is shown that

~SL = [ :~~ ] (5.19)

Thus from Equations 5.18 and 5.19

• • • • • • • •
~S = ~H + ~SL, eS = eH + eSL, <\IS = 4A-I ( 5.20)

From Equati ons 2.13 and 2.14 (with FB replaced by FS and ..ê. replaced by
--~S) •
( 5.21)

~
Q o-s·

(5.22)

Substituting Equations 5.20 and 5.22 into Equation 5.21 produces

•
pH + ~L
[ PSS]
qss
rSS
=
[
qH +
rH -
~SLCOS~S
eSLsin~S
] (5.23)

5.4
 • •
Thus if the angular rates .SL and aSL are maintained bel ow the threshold of
sensation as explained in Section 5.2.2, then the simulator pilot will sense
only wH as desi reet si nce Fps i s para" el to FS. If the hi gh-pass
filter HP FILT #2 is included as shown in Figure 5.2, the symmetry of the
above solution will be altered resulting in slight cross-coupling of the
rotati onal moti on sensati ons i n a manner simil ar to that descri bed in
Appendix C for translational filtering in FI •
•
The signal ~H in Figure 5.2 suffers from the same shortcomings as the
~ignal a2 in Figure 5.1. That is, if fed directly to the motion system
~H canl elld to excess i ve angul ar travel • To prevent thi s from
nappening ~H is passed through a high-pass filter, HP FILT #2 in
Figure 5.2,iand then integrated to produce ~SH, the high frequency component
of !s. -

5.2.2 Tilt-Coordination

Sustained translational inertial acceleration is sensed by the pilot

as a long-term change in the magnitude and direction of the specific force
in the absence of rotational motion. This cannot in general be simulated by
transl ati onal moti on due to moti on-base travel 1 imi ts. However it is
possible to alter the direction of the steady-state specific force
experienced by the pilot in the simulator by tilting the cab. It has become
common practi·ce in fl i ght simul ators to employ cab ti 1t to simul ate the
effect of sustained translational inertial acce1eration. Because this ti1t-
coordination process cannot alter the long term magnitude of the specific
force vector, it is at best an approximation to the desired effect. Two
algorithms for generating ti1t-coordination are described bel ow.

In one approach the simulator Euler angles for tilt-coordination (~SL)

are computed in the absence of all other simulator displacements as shownl"n
Figure 5.2 and then added to ~SH to produce the final simulator Euler angles
~S. In this process the timerate of change of ~SL is restricted so that
Tts contribution to the rotationa1 velocity of t~simulator is be10w the
pilotls threshold of detection.

The formulation of the tilt-coordination Eu1er angles ~SL starts with

fL as shown in Figure 5.2.The details concerning the calCülation of fL
wi" be discus sed 1ater. r
Fo the present fL can be cons i dered to be the
specific force FS components that are to Oe simul ated through simul ator
cab tilt. In the absence of other cab rotational displacement and
translational motion it follows from Figure 5.2 and Equations 2.2 and 5.13
that

.!.ss = -J!s
9 sinas ]
= -g cosassin·s (5.24)
[
-g cos as cos.s
and
(5.25)

5.5
 From Equation 5.24 it follows that

l1.ss I = 9 ( 5.26)

and thus it is not possible in general to simulate exactly IfLl by using

tilt-coordination. Although the magnitude of fL cannot be simUTated it is
possible to simul ate its direction of appl icatlon by tilting the simul ator
cab. lhis is achieved by setting

where ·
1s s = K .!.!:. ( 5.27)

K = g/ I.!!:. I (5.28)
It is found that this change in the direction of the sensed sustained
specific force is quite useful in the si mul ation of low frequency aircraft
acceleration components. From Equations 5.24 to 5.28 it follows that in the
present case
x x Z Z
fSS = (fL /fL ) f SS
y Z Z
!ss = (fL /fL ) fSS (5.29)
Z Z
fSS = gfL /I.!!:.I

$SL = tan-1(fLY/fL Z)

eSL = -tan-1{(fLx/fLZ)cos$SL} (5.30)

</ISL =0
Thus Equation 5.30 is used to represent ~SL. lhe simulator Euler angles are
then fo rmed as

~ = ~SL + ~SH (5.31)

Since the tilt-coordination is intended to simulate the low frequency
specific force, fL is formed by low-pass filtering !..AA as shown in
Figure 5.2. If tffi-coordination is to successfully simulate translational
motion, then the angular motionrequired to achieve this tilt must take
place at levels below the simulator pilotls rotational motion threshold.
Thi s goal can be assi sted by i ncorporati ng a nonl inear angul ar rate 1 imiter
within the LP FILT block of Figure 5.2 (see Section 5.5.2). It should be
noted that any steady-state ti 1ti ng of the ai rc raft wi 11 al so be simul ated
vi a the ti lt-coord i nati on channel. If it is desi red to further restrict the
motion base travel by scaling and limiting the tilt-coordination, it can be
seen from Figure 5.2 and Equation 5.30 that this can best be accomplished by
I modifying only 1.~A and !JA while leaving fÄA untouched. lhus the LP SCALE
.. block of Figure 5.2 is applied only to f~A and fÄA' otherwise it has the
same characteristics as described in Section 5.1 for the HP SCALE block.

5.6
 In the extreme cases of !AA containing only high or only low
frequency components it can be se en that the above formul ation leads to the
desi red resul ts. In the more common case of both hi gh and low frequency
content being present it is found that in general a reasonable approximation
to the desired results is obtained. This can be seen from the example
described in Appendix E.

In cases where low performance ai rc raft are to be simul ated it is

possible to simplify the tilt-coordination equations by applying the
following approximations:

(5.31)

si nx '" x (5.32)

cosx '" 1 (5.33)

The tilt Euler angles of Equation 5.30 then become

4>SL '" -fLY/g, <\ISL =0 (5.34)

This latter formulation will be used in the present report.

A second approach to the generation of tilt-coordination is depicted

in Fi gure 5.3 Here the underl yi ng concept is the generati on of a low
frequency simul ator angul ar vel oc ity component wSL that tends to orient the
simulator cab so as to align the negative of the gravity vector in the long
term with a specific force vector having FS components.!I. Here f2 is a
scal ed vers i on of !AA.

This process is based on the following. f2 can be considered to be

the components in FS of a "desi red 11 specific force vector which we wish
the simulator pilot to experience as a result of controlled tilting of the
simul ator al igning f2 and -gS. Because the components of f2 come from
scaling aircraft vmables they are independent of the simurator Euler
angl es Is. Thus the orientati on of f2 wi th respect to 9 can
be altered by rotati ng FS with respect to FI. Now consider the vector
~ given by

r = 9 x f2 (5.35)
-+ -+ -+
r is perpendicular to both 9 and f2. If FS is rotated with respect to FI
-+ -+ -+
about r with angular velocity
-+
w2 = Kr (5.36)
-+ -+
this will tend to move the vector f2 into alignment with -~.
Because the cross product ~ tends -+to zero as t2 approaches
alignment with -~ overshooting should be absent.

The components of r in FS can be found from (based on a form

developed in Reference 5.1t

5.7
 r = 9 x f2
+ + +
(5.37)
= -f2 x 9
+

thus

(5.38)

where J!.S comes from Equation 5.13.

As before, the tilt rotational motion must be kept below the simulator
pilot's detection threshold. This is aided by passing we through a low-pass
fil ter (LP FILTER in Fi gure 5.3) to produce wSL \'ttITch has the desi red
properties. As with the previous tilt-coordination case, nonlinear rate
1 imiting can be appl ied to wSL and the LP SCALE algorithm should not be
a p1 i ed to f Z •
AA
The second half of Appendix E contains an example indicating the type
of result generated by the above algorithm.

5.3 Selecting a Location for FS \

In order to apply the motion drive algorithms derived thus far it is

necessary to sel ect alocation for the origin of the simul ator frame FS.
In simul ati ng the hi gh frequency components of the ai rcraft' s transl ati onal
and rotational motion it makes 1 ittle difference wher:-e this reference point
is located. Since this involves rigid motion of the system in six
deg rees-of-freedom usi ng a moti on-base wi th only six deg rees-of-freedom
itself, there is a unique set of simulator actuator time histories that will
be generated no matter where the reference poi nt is sel ected • However the
same is not true for the low frequency tilt-coordination process. Here the
simul ator is driven in only two degrees of freedom represented by (~SL, eSL)
and the location of the origin of the reference frame FS can be selected
to reduce the associated actuatortravels. Consider the following case of
tilt-coordination in the absence of other simulator motions. From Equations
D.3 and D.4 of Appendix 0 it foll ows that the i-th actuator vector is given
by (see also Figure 0.1)

~iI = lIS Ai S + ll(O) - .!!iI (5.37)

where .hlS is based exclusively on ~SL. If ~SL is zero then

~i I :: lli I (5.38)
where
(5.39)

5.8
From Equati ons 5.37 to 5.39 it foll ows that

in = [hlS - l~i S + i~ I (5.40)

From Equation 5.40 it is seen that generally speaking ~i gets
1 arger as ~i gets 1 arger. In fact it can be easi 1y shown that if
FS is rotated through an angle Q about an axis defined by a unit vector
i then the corresponding vector change in !i is given by
(see Figures 5.4 and 0.1)

(5.41)

Thus a common approach to sel ecti ng the ori gin of FS is to choose the
centroid of the upper simul ator frame which is the point approximately
equidi stant from all six upper frame actuator bearings al so giving the
shortest common distance from the origin to the bearings (based on the fact
that the bearings are approximately at the vertices of an equi 1 ateral
tri angl e, see Fi gure 0.1). Thi sis the 1 ocati on sel ected in the present
study.

5.4 Inputs for the Vestibular Response Models

In the case of the simul ator, the inputsto the vestibul ar model s
wSS and !ps are calculated from the signals ~ and ~SI with
the assumptlon of perfect simul ato r response. Thi s requl res the add iti onal
estimation of ~S and ~S. The following steps are used in this
process. ~ is fi rst passed through a hi gh-pass fi lter with the transfer
function
TF = u2s (5.42)
s + w2
•
(with .w2 = 12.5 rls). Thi s forms an estimate of ~S whi ch looks
1 i ke ~S p~ssed through a fi rst-order 1 ow-pass fil ter: Ca 11 thi s
estimate ~ES. Next an estimate of ~SS is formed by using
Equation 5.21. Call this estimate ufSS
•
(5.43)

where Bs is based on~.

An estimate of wSS is formed by passing wESS through the

high-pass filter of Equatlon 5.42 and this is cal led WESS.

Since Fps and FPa are displaced from FS and FA

respectively by RS (see Flgure 2.1 and Equation 2.1) it follows
from Equations 2.2 to 2.5 that

.!PA = ~PA - ..9.A (5.44)

.!pS = ~PS - J!S (5.45)

5.9
where
x x x y • z •
api = aii RSS (q i i 2 ..
+ r 11 2) + RSS (Pi iq i i - rii) + RSS(Piirii + qii) (5.46)

Y Y x • y 2 Z •
api = aii + RSS ( Pi i q i i + ri i ) RSS{Pii + ri i 2) + RSS (q i i r i i Pi i ) ( 5.47)

z z x • y • z
api = ai i + RSS(Piirii qii) + RSS{qiirii + Pii) RSS (p i i 2 + q i i 2) (5.48 )

with i = A or S (note th at from Equation 2.1 ~SS = BAA) and

~S = .!:.sI ~SI (5.49)

In the case of the simulator the values for ~S and ~S in

Equations 5.46 to 5.48 are replaced by wESS and ~ESS when
generating the estimate for !PS, called FE1pS.- The inputs to the
vestibular models of Section 3 in this case are wESS and fElpS.
In order to ensure that the vestibular response estimates for the
ai rcraft are exactly comparable to those for the simulator the identical
steps employed in generating wESS and fElpS are used to produce
wEAA and fE1PA even though the exact~ariables of ~ and ~
are directly available. The inputs to the vestibular models of ~ection 3 1n
thi s case are wEAA and .!!lPA. Thi s is necessary because the
difference between the vestil5iJlar responses in the simulator and in the
aircraft is used as part of the motion base algorithm assessment process.

5.5 Filter Equations

In order to implement the motion-base drive algorithms both high-pass
and low-pass fil ters must be se 1ected. Thei r general forms are di scussed
bel ow.

5.5.1 High-Pass Filters

The general form of the high-pass filters employed in the classical
washout algorithm is given by the transfer function

~ = _____________s_N_____________ (5.50)
y sN + a N-1 s N-1 + ••• + als + aO
In the time domain this becomes
N N N-1 •
x = Y - aN_1 x - ••• - a1x - aOx (5.51)
p
(where x represents P successive differentiations of x if P is positive and
P successive integrations if P is negative).

5.10
 By integrating Equation 5.51 M times (where 0 ( M ( N and assuming
zero initial conditions) it can be transformed into
N-M N-M N-l-M l-M-M
x = Y - aN-l x - . •• - a 1 x - aO x (5.52)
Now Equa t i on 5.52 can be written as

(5.53)

• N-M
xl = Xo = x
• N-l-M
x2 = xl = x

•
xN-M+l = xN_M = x

• l-M
xN = xN_l = x

• -M
xN+l = xN = x

Now, the input to the filter is y and the primary output is x, although when
~mplemented by the ccxnputer all of xo to xN are available ~:~resenting
x for (-M ( P ( N-M). In order to employ Equation 5.53, y must be
available to the algorithm. In the present report M was always selected as
t1 = N. For example, if it is desired to pass an acceleration signal a =
z through a second order high-pass fil ter and then integrate the filter
output twice to find the filtered version of z, called zH, then one can
select in the above

N =M=2 (5.54)
y =a (5.55)
and it follows that

zH = Xo = x (5.56)

zH = x2 (5.57)

5.11
the same answer and sim; 1 ar equati ons woul d al so resul t i f we took
N =2 (5.58)

M =0 (5.59)

y =z (5.60)

(the algorithm would then require y = z = a) from which it follows that

(5.61)

5.5.2 Low-Pass Filters

The general form of the low-pass fil ters employed in the cl assi cal
washout algorithm is given by the transfer function

x aO (5.62)
-y = s N + aN_l sN-l + ••• + als+aO
In the time domain th;s becomes
N
x = aoy - aN-l
N-l
x - ... -
•
alx - aOx (5.63)

Equation 5.63 can be wr; tten as

Xo = aoy - aN_lxl - aN_2 x2 - ... - -al xN-l - aOxN (5.64)

• N
xl = Xo = x
• N-l
x2 = xl = x

• •

•
XN+l = xN = x
The input to the filter is y and the primary output is x, although when
~mplemented by the computer all of xo to xN are available representing
x for (0 ( P ( N). If, as in the case of the tilt-coordination routine, it
is des i red to 1 imi t the rate of change of x,
thi scan be done in the
computer program by limiting xN-l when it is calculated.

5.12
5.6 Test Runs

The cl assi cal washout al gorithm described above has been impl emented
in FORTRAN and run on a PE 3250 digital comuter. The test runs were
designed to demonstrate some of the features of the algorithm and to
determine in a preliminary manner its performance. The program is described
briefly in Appendix F and a listing is included. A number of different
washout fil ter confi gurati ons were employed in the current test s. The
program listing provided is for the general case of filtering in both frames
FS and FI. In this instance there is second-order high-pass filtering
in FS and fi rst-order high-pass filtering in FI for both transl ational
and rotati onal moti on. In additi on to the simul ator moti on-drive al gorithms
this program includes an input generation section which starts with aircraft
val ue s f 0 r a AA a n d ~A A a n d pro d u ces th e fA A a n d wA A ( se e
Figures 5.1 and5.2) required by the algorithm. Alsolncluded are routines
to predict the subjective sensation of motion at the pilotls head in the
aircraft and in the simulator and plotting routines to record the system
variabl es. In these plots, .!PA and ~A are the true val ues whi 1e
.!.PS and ~S are based on fEl pS and ui SS of Section 5.4.

5.6.1 Algorithm Configuration

As described in the above sections there are several items in the

cl assi cal al gorithm for whi ch a parti cul ar form must be sel ected. For the
present test runs the foll owi ng choices have been made. Thi s has resulted
in a general form simil ar to that curently employed by many commercial
simul ators.

The ti lt-coord i nati on al gorithm employed is the 1 ineari zed one

described by Equations 5.34 in Section 5.2.2. This choice is . appropriate
for the si mul ation of commercial jet transports which normally have modest
acceleration capabilty.

As poi nted out in Secti on 5.3 the best choice for the 1ocati on of the
origin of FS is the centroid of the upper simul ator frame in the pl ane of
the upper actuator bearings. Thi s has been adopted in the present tests
(thus FS :: FO and FA :: Fe) except for the runs intended to
demonstate the effect of choosing an alternative location. The FS x-axis
is parallel to the floor of the simulator cab and lies in its vertical plane
of symmetry. The frame FI is located at the centroid of the lower
simulator frame and is parallel to FS when the simulator is in its neutral
positi on.

A number of parameter values must be selected before the algorithm can

be run. The vestibular model parameters are contained in Tables 3.1 and
3.2. The other parameters whi ch are common to all of the test runs are
either contained in Table 5.1 or outlined bel ow. The geometric parameters
represent the UTIAS Fl ight Research Simul ator .•

(1) The iteration rate has been selected to be 20 Hz, thus ~t = 50 ms.

(2) All the scale factors have been taken to be unity.

5.13
(3) The 1 imit val ues for blocks HP SCALE and LP SCALE in Fi gures 5.1 and
5.2 have been set 1 arge enough that they are never reached in the
present tests.
•
(4) The tilt-coordination rate limit on fL (see Figure ~.2) has been
selected to be 1.0 m/s 2 which corresponds to a ~SL limit of
5.8° Is, somewhat above the sensati on detecti on threshol d . -

(5) The sensed motions are evaluated at the location of the captain's head
(the captain is seated in the left-hand seat of the UTIAS Flight
Research Simul ator) •
(6) The filter break frequency w2 used in estimating derivatives in
Equations 3.42, 5.42 and D.6 has been selected as w2 = 12.5 rls.

(7) The notation used in Figures 5.5 to 5.67 of this section is intended to
be self-explanatory. lts relation to the notation used in the text is
given in Table 5.2.
(8) Where appropriate, the sine and cosine approximations of Appendix G
have been used to speed up the running of the algorithm.

5.6.2 Input Test Signals

In order to demonstrate the performance of the moti on-drive al gorithms

i t is necessary to fi rst generate test si gnal s to represent the al gorithm
inputs fAA and wAA (see Figures 5.1 and 5.2). Rather than
generati ng campl ex test si gnal s representi ng canpl ete ai rc raft maneuvers it
was decided to enploy ideal ized representations that approximate typical
aircraft .!AA and ::!AA. This makes it easier to interpret the
responses OT the simuTàtor and vestibular system models. More realistic
cases wil 1 be run later during the detailed algorithm parameter selection
ph a se of the study.
x
For longitudinal translation motion tests an acceleration pulse (aAA)
was used, representi ng in an ideal i zed manner the response of the ai rcran
to a thrust pul se such as during takeoff or during an airspeed change whil e
flying straight and level (see Figure 5.5 for such an 2.AA). In this
instance
(5.65)
z
For vertical translation motion tests an acceleration pulse (aAA) was
used, representing an idealized pull-up maneuver (although the corresponding
pi tch attitude changes are absent). See Fi gure 5.56 for such an 2.AA.
In thi s case
(5.66)

Two rotati onal moti on test si gnal s were employed separatel y, one about
the roll axis and one about the yaw axis. In these tests !AA was set to

5.14
zero. A roll acceleration doublet was employed, representing the transition
into a steady-state banked attitude (with no heading change present). The
corresponding .:!:AA and bA can be seen in Figure 5.33. In the yawing
case a steady-state yaw rate is generated following a constant yaw
accel erati on. Thi scan be seen in Fi gure 5.45. Thi s has the general
rotational motion features of a flat or skidding turn although the lateral
acceleration aY corresponding to a curved flight path trajecory is absent.
AA
The above test si gnal s for transl ati on and yaw represent the worst
cases for which the current washout filters were designed as described in
Secti on 2.5. Thus they shoul d provide a good test of the effecti veness of
the washout algorithms.

5.6.3 The Effects of Filtering Frame Selection

As pointed out in Appendix C, if the washout process is carried out in

FS it can result in undesirable offsets in the motion-base hydraulic
actuators. On the other hand, i f the washout process is carried out in
FI, it can resul tin incorrect moti on cues for the pil ot. Thi s secti on
demonstrates the extent to whi dh these effects are present when the current
algorithms and test signals are employed.

In the first set of runs represented by Figures 5.5 to 5.8 all the
washout filtering was carried out in FS' In Fïgures 5.1 and 5.2 this
corresponds to dropping HP FILT #2 completely. In this case the
translational channel filters, HP FILT #1 in Figure 5.1 were third-order
with the filter parameter set A in Table 5.3. The tilt-coordination filters
LP FILT in Figure 5.2 were second-order with filter parameter set B in Table
5.3. The longitudinal translation test signal was employed (see Figure
5.5). The rotational channel filters HP FILT #1 in Figure 5.2 were not used
in the present instance since ~A was identically zero.

The outputs from the high-pass and low-pass fil ters (aH and fL) are
presented in Figure 5.6. The characteristic spikes at theoutputof the
high-pass filter aW are due to the switching on and off of the input
pulse. The aHz signal is nonzero due to the g~ resulting from the
simulator pitch (see Figure 5.8) caused by the tilt-coordination process.
The resulting undesirable simulator offsets can be seen in Figure 5.8 where
both the longitudinal and vertical displacements do not return to zero as
time increases beyond t = las. A similar effect can be demonstrated in the
rotational degrees of freedom for an angular rate input pul se.

In the next set of runs represented by Figures 5.5 and 5.9 to 5.11 all
the washout fil teri ng was carried out in FI' In Fi gures 5.1 and 5.2 thi s
corresponds to dropping HP FILT #1 completely. In this case the
translation channel filters HP FILT #2 in Figure 5.1 were third order with
filter parameter set A in Table 5.3. The tilt-coordination filters LP FILT
in Figure 5.2 were second order withfilter parameter set B in Table 5.3.
The longitudinal translation test signal was employed (see Figure 5.5). The
rotational channel filters HP FILT #2 in Figure 5.2 were not used in the
present instance since ~A was identically zero.

5.15
 The responses shown in Figures 5.9 to 5.11 are similar to those in
Figures 5.6 to 5.8 except for the vertieal motion variables and the
longitudinal simulator displacement. As seen in Figure 5.11 the washout
process is completely successful in returning all displacements to zero as
time increases. If the translational acceleration plots of Figures 5.5, 5.7
and 5.10 are compared along with the angular displacement plots of Figures
5.8 and 5.11 (recalling that lAA = 0) it is found that the motion cues
generated by filtering in FI appear to be no worse than those generated by
filtering in FS' This was the same conclusion reached in Reference 5.2.

A third alternative, namely some filtering in FS and some in FIt

is used in some commercial simulators. Figures 5.5 and 5.12 to 5.14
represent the results for such a case. In this instance the translational
channel filters HP FILT #1 in Figure 5.1 are second-order using filter
parameter set C in Table 5.3 and HP FILT #2 in Figure 5.1 are first-order
using filter parameter set D in Table 5.3. The tilt-coordination filters LP
FILT in Figure 5.2 are second-order using filter parameter set E in Table
5.3. The longitudinal translation test signal is employed (see Figure 5.5).
The rotational channel filters HP FILT #1 and HP FILT #2 in Figure 5.2 are
not used in the present intance since .!!!AA is identically zero. The
detailed response shown in Figures 5.12 to 5.14 differs from that of the
previous two cases mainly because the filter parameters employed are
different. The most significant effects to note are the similarities in the
general form of the transl ati onal accel erati on and angul ar di spl acement of
the simulator in all three filtering cases. As seen in Figure 5.14 the
present case results in modest 1 inear displacement offsets and thus in this
respect 1 ies somewhere between filtering in FS alone and filtering in FI
al one.

It was noted in further tests that the vertical steady-state offset

noted in Figure 5.14 eaused by longitudinal aircraft acceleration was of a
positive polarity for both 'positive and negative aÄA values. Thus
it will tend to accumul ate throughout the period of operation of the fl ight
simul ator.

From the above it appears that the adverse effects of filtering in

FS are more severe than those associated with filtering in FI. It
should be noted that in Reference 5.2 test subjects did not experience any
improvement in the qual ity of moti on sensati on in going from fi ltering in
FJ alone to filtering in FS alone. It was also reported that these
pllots experienced difficulies in ·controlling angular oscillations (Le., a
tendency towards pil ot i nduced ose i 11 ati ons or PIO) when a thi rd-order
high-pass filter was employed in the rotational ehannels.

5.6.4 The Effects of the Location of FS

As pointed out in Section 5.3 the loeation of the orlgln of FS

within the simulator ean have a significant effect on the motion base
actuator lengths during tilt-coordination. A test case to demonstrate this
effect was run using filtering in FI (as in Section 5.6.3) and parameter
sets F and E in Table 5.3 for the translational channel filters HP FILT #2
and the tilt-coordination filters LP FILT respectively. A longitudinal
transl ati on test si gnal was employed (see Fi gure 5.15). The resulti ng
simul ator moti ons are simi 1 ar to those previousl y reported. The resul ti ng
actuator time histories are shown in Figures 5.18 and 5.19 for two locations

5.16
of the orl gl n of FS' lhe data in Fi gure 5.18 corresponds to the case with
the origin of FS located at the centroid of the upper fame bearings. lhe
data in Fi gure 5.19 correspondes to the case with the origin located at the
pilot's (captain's) head position. lhe relative position of these two
locations is given by RS in Figure 2.1 with the components RSS
1 isted in Table 5.1. The focation of the upper frame centroid is given-by
Figure D.1 and the ~iS data in Table 5.1.

It can be seen from Fi gures 5.18 and 5.19 that the 1ocati on of the
origin of FS has a profound effect on the shape and magnitudes of the
actuator lengths. In the present case most of this motion is in response to
the tilt-coordination generated by ElS' It is clear that the case with
FS located at the upper frame centroid (Fi gure 5.18) produces a small er
peak actuator extension than the case with FS located at the pilot' s head
(Figure 5.19).

5.6.5 Results

The four cases presented in thi s secti on were run to demonstrate the
response of the motion sensation models to simulator motions. Each data set
(Figures 5.20 to 5.31 9 Figures 5.32 to 5.43, Figures 5.44 to 5.55, and
Figures 5.56 to 5.67) contains data on aircraft motions, aircraft pilot
sensations, simulator motions, simulator pilot sensations, the difference
between aircraft pilot and simulator pilot sensations 9 and motion base
actuator time hi stories. lhe washout fi ltering was carried out wi th some
filtering in FS and some in FI' As in Section 5.6.3, the translational
channel filters HP FILT #1 in Figure 5.1 were second-order using filter
parameter set C in Table 5.3 and HP FILT #2 in Figure 5.1 were first-order
using filter parameter set n in Table 5.3. lhe tilt-coordination filters LP
FILT in Figure 5.2 were second-order using filter parameter set E in Table
5.3. To this was added the rotational channel filters HP FILT #1 in Figure
5.2 using a second-order version and filter parameter set G in Table 5.3,
and filters HP FILT #2 in Figure 5.2 using a first-order version and filter
parameter set Din Table 5.3. lhis configuration represents a typical
commercial motion-base drive algorithm.

lhe first test case represents a longitudinal acceleration of the

aircraft as shown in Figure 5.20. Because there is no rotational motion of
the aircraft (see Figure 5.21) the specific force shown in Figure 5.20 is
identical to the inertial acceleration minus 9I. lhe aircraft pilot's
sensation of motion as predicted by the vestibü1ar models is presented in
Figures 5.22 and 5.23. lhe components of sensed specific force in FA or
Fpa are given by [SFX SFY SFZ] for the case where the thresholds dTH
(see Figure 3.5) are taken to be zero and by [SFXTH SFYTH SFZTH] for the
case where the threshold values of Table 3.2 apply. lhe components of
sensed angul ar velocity in FA or Fpa are given by [SP SO SR] for the
case where the thresholds óTH (see Figure 3.3) are taken to be zero and by
[SPTH SQTH SR TH] for the case where the threshold values of Table 3.1
apply. In the present case, in the aircraft there is no rotational motion
sensation and the sensed specific force is affected only sl ightly by the
presence of thresholds.

lhe resulting simul ator motion is presented in Figures 5.24 to 5.26.

The simulator pitch resulting from the tilt-coordination process can be seen

5.17
in Figure 5.26. In the same figure it can be seen that the
tilt-coordination rate limit algorithm has clipped the QS value at 5.8°/s.

The responses of the vestibular models of the simulator pilot both

without and with sensation thresholds are given in Figures 5.27 and 5.28.
The errors represented by the predicted sensations in the simul ator minus
those in the aircraft are presented in Figures 5.29 and 5.30. In Figures
5.27 and 5.28 it can be seen that the presence of sensation thresholds has
had the greatest effect on rotational motion sensation in the simulator. In
the present case all the rotational motion is due to the tilt-coordination
and thus represents an unwanted st imul us whi ch hopefully wi 11 go undetected
by the pil ot.

Based on Equation 5.30 the exact value of the tilt-coordination Euler

angle eSL is 11.52° for the present maneuver when f~A = 2 m/s 2 • This is
i nd icated on Fi gure 5.26 where it is seen that the present al gorithm does a
good job in generating an approximation to this value.

It can be seen from Figures 5.23 and 5.28 that the ~ error in
Figure 5.30 is to a large extent the result of an approximately Is lag in
the simulator sensation behind that in the aircraft. The waveform is
otherwise quite good in the simulator.

The actuator 1 ength time hi stories shown in Fi gure 5.31 represent

reasonabl e extensions except for the unwanted long-term offsets of about
6 cm.

The second test case represents a roll accel erati on doubl et as shown
in Figures 5.32 and 5.33. The specific force shown in Figure 5.32 results
from the offset of Fpa with respect to FA and from the bank angle.
The aircraft pilot' s sensation of motion as predicted by the vestibul ar
models is presented in Figures 5.34 and 5.35.

The resulting simulator motion is presented in Figures 5.36 to 5.38.

Note that tilt-coordination is present due to the sustained f~A
val ue caused by the sustained <l>A val ue of Fi gure 5.33. The rotati onal
response of the simulator in this case begins with a rolling motion due to
the high-passed signal ~SH (see Figures 5.2 and 5.38) and is soon followed
b~ the contributi on due-u, ti lt-coordinati on, ~SL. Although the val ue of
<l>SL is limited to 5.8°/s by the cOOIputer algorithm the second peak in
PSS in Figure 5.38 exceeds this value due to the contribution from
~ which has not yet died out at this point in time.

The responses of the vestibular models of the simulator pilot are

given in Figures 5.39 and 5.40. The error represented by the predicted
sensations in the simul ator minus those in the aircraft are presented in
Figures 5.41 and 5.42. Here the specific force error is again to a large
extent due to thesimul ator 1agging behi nd the ai rcraft. In the case of
rotational roll sensation the simul ator onset cue is actually sl ight1y
1eading that in the aircraft but with a significant1y reduced amp1 itude.
However, after the initial h>A peak, the simulator PPS waveform does
not follow that of the aircraft at all. In fact the second PPS cue is
of the wrong po1arity. This is due to the transient behaviour of the

5.18
high-pass filters HP FILT #1 and HP FILT #2 (see Figure 5.2) coupled with
the onset of tilt-coordination.

The actuator length time histories shown in Figure 5.43 display

reasonable extensions. However it can be noted in Figure 5.37 that unwanted
residual simulator offsets remain in the yand z displacements.

The third test case represents a yaw rate build-up as shown in Figures
5.44 and 5.45. The specific force shown in Figure 5.44 results from the
offset of Fpa with respect to FA' The aircraft pilot's sensation of
motion as predicted by the vestibular models is presented in Figures 5.46
and 5.47.

The resulting simulator motion is presented in Figures 5.48 to 5.50.

In this case there is no f~A and f~A present and hence
tilt-coordination is absent. The orily simulator motion is ab out the yaw
axis (see rSS in Figure 5.50). By comparing Figures 5.45 and 5.50 it
can be seen that the simul ator motion is highly attenuated with respect to
that of the aircraft and that the washout process has introduced a large
spurious negative rSS pul se.

The response of the vestibular models of the simulatQr pilot are given
i n Fi 9 ure s 5 • 51 a n d 5 • 52 • No t e t ha t f ~ S a n d f~ Sar e ze r 0 i n
Figure 5.52 because d is always below the corresponding threshold level (see
Figure 3.5). The error represented by the predicted sensations in the
simulator minus those in the aircraft are presented in Figures 5.53 and
5.54. The most significant error is in rp as expected.

The actuator length time histories in Figure 5.55 show very small peak
extensions (less than 10 cm). This coupled with the large rp error
suggests that the hi gh-pass angul ar rate fi 1 ters in the al gori thm shoul d be
modified to allow a greater simulator response.

The fourth test case represents a vertical acceleration of the

aircraft as shown in Figure 5.56. Because there is no rotational motion of
the aircraft (see Figure 5.57) the specific force shown in Figure 5.56 is
identical to the inertial acceleration minus 9I' The aircraft pilot's
sensation of motion as predicted by the vestibü1ar models is presented in
Figures 5.58 and 5.59. In the aircraft the sensed specific force is not
significantly affected by the presence of thresholds.

The resulting simul ator motion is presented in Figures 5.60 to 5.62.

Only vertical motion response is present. The response of the vestibular
models of the simulator pilot are given in Figures 5.63 and 5.64 and they
are minimal • The sensation error shown in Figures 5.65 and 5.66 indicates a
rather poor simulation in this case.

The modest actuator travel shown in Figure 5.67 would seem to indicate
that the washout in the vertical channel has been overly restrictive.

It i s of interest to note that in the present case, with 15 = 0

reference frames FS and FI are parall el and thus fi ltering 1n F'S
becomes the same as filtering in FI' As expected this will lead to the
successful washing out of an aÄA step input when the current washout
filters are used. This is confirmed by Figure 5.67.

5.19
 References

5.1 Etkin, B., "Dynamics of Atmospheric Flight", Wiley, New Vork, 1972.

5.2 Bitner, M. E., 11 Investigation of fIotion Base Drive Techniques",

NADC-77306-20, March 1978.

5.20
 Table 5.1
Simulator Geometrie Parameters for FS :: FD
(UTIAS Fl ight Research Simul ator)
(all values in m)

AIS = [-1.47 9 -0.10 9 o]T

[-1.47 9 0.10 9 o]T
~S =

[0.65 9 1. 32 9 O]T
bs =

[0.82 9 1.22 9 O]T

~S =

[0.82 9 -1.22 9 O]T

fu>s =

= [0.65 9 -1.32 9 O]T

&S
[-1.22 9 -1. 52 9 O]T
~1I =

[-1. 229 1.52 9 O]T

~I =

[-0.71 9 1.82 9 O]T

~I =

[1.93 9 0.29 9 O]T

!!.41 =

[1.93 9 -0.29 9 O]T

~I =

[-0.71 9 -1.82 9 O]T

~I =

~S = [-0.02 9 -0.47 9 -1.78]T

~I (0) = [0 9 09 -1.84] T
L = 2.34

5.21
 Table 5.2

Notation Used in Fi gures

Text
Vari abl e Notation Used in Figures

~A (same as ~A in cl assical al gorithm)

(AXe Aye AZe) Ale

.!pA (FXP FYP FZP) Ale

lA ( PHIA THETAA PSIA) Ale

::!c A 0 r .!:!.AA ( PA QA -RA) A/C

"
.!PA (with thresholds set to zero)

( SFX SFY SFZ) A/C

"
~A (with thresholds set to zero)
(SP SQ SR) A/C
" ( SFXTH SFYTH SFZTH) A/C
.!pA
" (SPTH SQTH SRTH) A/C
~A

.!sI (AXSI AYSI AZSI) SIM

.!pS ( FXP FYP FZP) SIM

~I (VXSI VYSI VZSI) SIM

[~I ( t) - ~I ( 0 ) ] ( XSI YSI ZSI) SIM

is ( PHI THETA PS I) SIM

~S ( PS QS RS) SIM
"
.!.pS (with thresholds set to zero)

( SFX SFY SFZ) SIM

~S (with thresholds set to zero)

(SP SQ SR) SIM
" ( SFXTH SFYTH SFZTH) SIM
.!pS
" ( SPTH SQTH SRTH) SIM
~S

5.22
 Table 5.2 - Continued

Notation Used in Figures

A A

(.!pS - .!.PA) (with thresholds set to zero)

( SFX SFY SFZ) ERR

(;,S - ~PA) (with thresholds set to zero)

(SP SQ SR) ERR

A

(!PS - .!.PA) ( SFXTH SFYTH SFZTH) ERR

(frs - @PA) (SPTH SQTH SRTH) ERR

ii ( JACKL i)

Pxi (PXi)

Pyi (PYi)

Pz ( PZ)

P<\I (PPS!)

5.23
 Table 5.3
Filter Parameters
s2 s
(TF = ---=----- · - - - fo r hi gh-pa ss )
S 2 + 2 çl1ln s + utl 2 s + ut

2
(TF = l1ln for low-pass)
s2 + 2tur,S + utl 2

Set*
~s ~s
A 1.25 0.707 0.125
B 1.25 0.707
C 3.10 1.4
D 0.• 2
E 3.10 1.4
F 3.10 1.4 0.2
G 1.02
t·o
*A and B from Reference 5.2.
C to G, typi cal commercial simulator.

5.24
 -- - - - -- - - - - - - - -

Lr S
,~ ,--------, /',/
f1 s a1 s aHs a 2I Sr
!A - HP
A-- SCALE ~
HP
FILT --
~
Lrsx 1-
I"
HP
FILT
#2
aS!
--- I
5 2
...
I"
,t.I - Ij
+"
#1 I I
L ________ ..J
ACCN
~S ~
EST ~ iiest

FIGURE 5.1 TRANSLATIONAL CHANNEL FOR TEE CLASSICAL ALGORITHM

fA
LP
~ SCALE
f2
-
-
... LP
FILT
--
fL
p
TILT
COORD
~SL
gs ~ gs
J'
r-------l
w1.~ wHs ~~ ~SJ~,+~S
HP
WA~ SCALE - HP
FILT
#1 -
... TSx ,r_

I
...
HP
FILT
#2
....
--
I
-S ~ Sin
COS -... LIS ---. LIS
I
h L ________ J
Ts
.. Ts ~ Ts

FIGURE 5.2 ROTATIONAL CHANNEL FOR THE CLASSICAL ALGORITHM

 ~S ~ 9s
A~
gs
,It
w2 {3SL ~S
f2
- ..
- . LP
WSL
..
-
{3SL
- -I
-
-+ -- Sin
- --- LIS
LP
-- -.;,;,;
LIS ~

. fA~ SC ALE g)(f ~

FILT TSx S ~ cos I"

-"7 -"7 ~

J~
{3SH
Ts , r - - - - - -I--l ~ Ts -Ts
~

w1 wH {3H HP I
-. HP - .. - I
HP
WA ~ SCALE -- FILT
#1 -- TSx I ..
I"

I
FILT
#2
L ________ J
I
~
-
S I

FIGURE 5.3 ROTATIONAL CHANNEL EMPLOYING CROSS-PRODUCT TILT-COORDINATION

 .ui

A·
,I /
/
/
/
/
/
/
I.

FIGURE 5.4 ól
~
DUE TO TILT-COORDINATION
 ~ r-----~----~----------, ~ r_----------,-----------, ~ ,-------,-------.

~
w
o g~----------~----------1
,u ~ 0
'"u
x
'" ..'"
0-----------::'''o."''o----------,''''0.0
:."' ~. o 20.0 "0.0 ~."'o-----------""".-0----------,..Jo., ;. ~o-----------:';;0."0------------:,.:
T IH E (SEC) TlHE (SEC) T IHE (SEC) TI HE (SEC)

g r---------~------_. g.-------,-------, g r---------r--------,

~
~ ~r-----------~----------~
~_ 1
g~--------r_------_1
I ...,
~-- ~r-----------+-----------~
'",,
'"
~ ~r-----------1-----------1
~ ei ! ;:
I
'"'"
~....o -----------::'''0."''0----------'' '0. 0 :.~o-----------:,::'o.-o=----------,c:lo. 0 , -----------"".-
:.L. 0----------,..Jo., :.':-0-----------::,.:-o."o----------,,::!o.o
T [H E (SEC) TIHE (SEC) T IHE (SECJ T IHE (SEC)

I ~
~
w

I o

~
g ~----------_r----------1
ci r
u

'"
~. ~
o ------..""..
o."..o------,,~o.o :.'0:------------='.:-0.""0----------,',.:0.0 ~.o 20.0
TIME (SECl T I ME ISECl TIH E (S ECJ

FIGURE 5.5 INPUT SIGNALS FOR FILTERING FRAME EFFECTS TESTS FIGURE 5.6 FILTERING IN FS' FILTER OUTPUTS

~ r_----------,_----------, g r---------,--------,
g'----""""--I-----,
I
~ gr_----------+_----------~
'" 0

'"x> '"x

:.~o------'------=,.:-o.o"----------",.:o. 0 g~. ~----~----:~-----------:'"

o 20.0 1/,0.0 •• 0 20.0
TIME ISEC) TIHE (SECJ TIHE (SEC)

~ r------------r----------~
~
~
~ ~~----------~----------1 ~ ~~----------~----------~ ~ ~r----------1-----------1

'">
1

o
:.~o-----------:,,,Jo.-o=----------c:l
l/,O.O ~'------------:~----------~
I

Ii.~o-----------='.:-0."0------------:',.:0. 0 ~.o 20.0 "0.0 ';:0 20.0 "0.0

TIHE (SEC ) TlHE (SEC) TIME (SEC J T I HE (SEC)

- ,-------,------,

~ gr_----------+_----------~
'" ö

'"
N

'"
..'"
- ~----~~----....".:
11.0 20.0 1/,0.0 ~.':-o-----------:''' 0.-0::----------'-'0.0 ; ':-o----------~~~. o"..----------:.,.:o.o ~...., -----------,,"".-,----------,~,. ,
TIHE (SEC) TIHE (SECJ TI HE (SEC) TIHE (SEC )

FIGURE 5.7 FILTERI NG IN FS' SIMULATOR RESPONSE FIGURE 5.8 FILTERING IN FS' SIMULATOR RESPONSE
 ...

:C\
~
~
,on

r~
'",, ,,on "'r

= ~fj- A"
'"!;
on
..on .;g
.; ~

on
"'" '">
~
~ ~
~.O 20 . 0 \10.0 ~.O 20.0 1110.0 ~. O 20.0 1110.0
TIME ISEC) TIME ISEC) TIME ISEC)

- ~ ~
,on
,on ~
'", r !;
~ g g
=~
!; .;
'"=
.; on
...
~
~

,..
v.
~

,..
on
>
~ '" g ~
~.O 20.0 \10 . 0 ~.O 20.0 1110 .0 ~.O 20.0 \10 .0
TIME ISEC) TIME ISE C) TIME ISE CI

~
.; ~

~
!~
'",r.
r
!J" JI
on V =
on

-.
v·
-
~
'" - g
B.O 20.0 1110.0 ~.O 20.0 "0.0
TIME ISEC) TIME ISEC)

FIGURE 5.9 FILTERING IN FI' FILTER OUTPUTS FIGURE 5.10 FILTERING IN FI' SI MlJLA TOR RESPONSE

g
~

'" g~
è.
~
"
on .; L
~
-
on
" Îj ~
"
i1:
g
~.O 20 .0 1(0.0 ~.O 20.0 110.0
TIME . ISEC) TIME ISEC)

.
;n
g

~
~
g

-,..=
r
on .;
on
on
'"r
w

.,
i.a 20.0 1110.0 'D.o 20.0 110.0
TIME ISEC) TIME ISEC)

~ ~

g
! ~I-- ~
9

-=
on zo g

on
on
N

aV ~.o
TIME
20.0
ISEC)
1,10.0
t
~
~.o 20 .0
TIME
I
ISEC)
\10 . 0

FIGURE 5.11 FILTERING IN FI ' SIMULATOR RESPONSE

 r

"Ir
m

~
,
~ ~Ir-- - ~
rt
g
.;
\ ;;;
m
!
~ .;
,
m

'"
~

m
!tTi{\
m x
x >
I g '" g
r.o
~
a=.D
TIME
20.0
[SEC)
~.o
TIME
20 . 0
[SEC)
1.10.0 20 . 0 1.10.0 ti.D 20.0
TIM E [SEC) TIHE [SEC)

~ ~
-
,m
,
m

'",!
,
m m

'"
~
~ g
.; ~ '" .;
'"
" '2 '">
'"
~
~

~ g '" g
~
~. o 20.0 ~.o 20.0 \1 0.0 ~.o 20.0 "0.0 ~,o 20.0 110.0
TI HE [SEC) TIME [SEC) TIHE [SEC) T IHE [SE C)

! -
.; 0

~
g
m
'",,
'"
" !Ir-v
!
" .;
'"
~
g('--I\
"0 '"
'"
N '">
N

'"
~,o 20.0 \10.0 8.0 20. 0 1.10.0 ~.o 20.0 1.10.0
TIME [SEC) TIME [S EC) TIM E [SEC)

FI GURE 5.12 FILTERI NG IN FS AND FI ' F I LTER OUTPUTS FIGURE 5.13 FILTERING IN FS AND FI' SIMULATOR RESPONSE

:lv-
'-'
-
w
0 '" ~
'" '"
~
i '"
;;; g I L

w ci
"'-'
W

,
w 0

on i:: '" "

w

"
.
'"
&.0 20.0 ~.o 20.0 ~.o 25.0 ~.o 25.0
TIME [SEC[ TIME [SEC) T [ME [5) T IME [5)

g
'"m
m
~
-

I
~~
'"
-
m
"
w
"
'"

g
0

.;
n I
,
m
,m
'"
,w
"
W
~
tt
.;
~
w
0

w .;

'"
tt
co

~.o
TIME
20.0
[SEC[
1.10.0 'ti.a
" TIME
20.0
[SEC)
~,o 25.0
T [ME [SI
50.0 ~,o
TIME
25.0
[SJ
50.0

;kC
0 g
-
:1:
~
'" e m

" "w
'" g
0

'" ,
w
m
N
.
m
.;
tt
W
0
~ Q

" '"'"
~ g
ti.0 20.0 "0.0 ~.o 20 . 0 1.10.0 ~. o 25.0 50.0 ~,o 25.0 50.0
TIME [SEC) TIME [SEC) TIME [SI TIME [5)

FIGURE 5.1l! FILTERING IN FS AND FI' SIMULATOR RESPONSE FIGURE 5.15 INPUT SIGNALS FOR FS LOCATION TESTS
 /

.
8 g
,;
,
..,
~ ~

,
~

,
~ I, I w
9 ~
!; 8 I g ! u
. . g. 9g
,
~ ~ , rr 0
~

1
ei I ei à

~
i ~
,I
I ."
rr
..
rr
rr 8
W.a 20. 0
11HE IS )
1110 . 0
g
~ . 20.0
11"E
I
IS)
'-10.0
~
G".' 20. 0
TI"E IS)
1110.0
~
G".' ....
TlHE IS)
110.0

~ ~ ~ , ~

~
I

...
~
i ..,
w
II ~.
I
c ..,
~
I
~

:': g ~ g ~ g
rr ö ,
Bg I
~ ei ~ ei ~ I
rr

~
! .
~
I rr

~ i ~
ei
I
I
... ~ I
....
~
g , ~ ! ! ~ I
· ~.o zo., "0.0 G".' 20.' 110.0 G".' 20.0 1110. 0 G".' 110.0
11HE IS ) 11HE IS) 11HE ISI TlHE ISI

~ ~ ~ ~
i !I
I i
~
,
~
,
on
I I ,..,
on

'"~ ,
on
! ,I
r i w

~ ei
g
, -

~
0

ei
u
;X
g
ei
eg
~ ei i
rr
. , rr
..
on

1
u rr
N N

'" ~
G"•• 20.'IS) 110 . 0
~ ~

~.o lO.D
i 110.0
~
G".• 20 . 0 "0.0
'"
~
G".' 20.0 ,0.0
TIME 11HE IS I 11HE IS) TlHE IS)

FIGURE 5.20 LONG I TUD I NAL ACCELERATlON TEST CASE FIGURE 5.2i LONGITUDINAL ACCELERATION TEST CASE

~~I~
~ ~ ~

~
K
~
g
ei
ili g
~

~
..
on
ei

,
L~
;x
E
r

on
ei

I
!
i
u
;;0

~
~
0
,

:
'" ~ :-: ..l
Ul ~ ,
G". • 20.0 110 . 0 G".' 20.0 '0. 0 G".' 20.' IS) 110.0 G". ' ZO., 110.0
11HE IS) TlHE IS) 1 Jt1E TI"E ISI

~ , ~ I
~ ~
I I I
i '",, ,'"
~8
!;
i
:
~
eg
I
I
'"
:r
;:;8
II
I ~ g
II
~ ei ! ~ ö
: ;X 0
I ~ ö
,."...
rr
~ 0 I i ;!:
I g
....I
"- on
'" ~ ~ ~ ~ ~
G".• 20.' IS ) 110.0 G".' 20.0
TIME
"0.0 G". ' 110.0 G".' 20.0 110.0
11"E ISI TI"E IS) 11HE IS)

~ ~ ~ ~
,'"
I
~
~
..,~ ,
'"
:r
! ~
~ i
!; g ~ g ;:; g
9
g
~ci ~ ei ~ ei ~ ci

'"iJi Ë ~
N
"- i ::;8 on
"'8
tr.G ....
11"E IS)
,,0.0
~
G".' 20.0
TI"E IS)
110. 0 ni.a
TI"E
.... IS)
110.0
8
~. o ....
TI"E IS)
110.0

FIGURE 5.22 LONGITUDINAL ACCELERATION TEST CASE FIGURE 5.23 LONGITUDINAL ACCELERATION TEST CASE
 ~ ~

~:n
~ I ,; ,;

Yi
'""- '""-
'""- '"
'"-
~o
g
v--- I
'" ,;
I '"
~~
'" 0
v-f\ '" g
;:;; ~ -
'"
'"><
..
'"
>< '">>< '"
><

.
~

'" g ~
rf.• 20 . 0 110.0 rf.• 20. 0 110.0 B.• 20.0 11 0 . 0 ~ 20. 0 110.0
TIME ISI TI ME ISI TI ME ISJ TI ME 151

g
~ ~ ~

'""-
'""-~ '""-
~
~ g
~ '"
g
~~ ;:;; ~
'" 0
~o '" ,;
-'"
",0

'" '" -'"

>- >-
'">- t >
I
'" :s
~ .. 20.0
TI ME ISI
110.0 ~ .. 20. 0
TIME ISI
11 0 . 0
:s
~ .. 20 . 0
TI ME ISJ
110. 0
g
~. , 20. 0
TI ME 151
110.0

~ ~ ~
,; - ,; I
'""-
'""-~ 0 j '""-
'"~"-
'"
"-

:'" ~ J g I
II
g 0 '" g
~ o r ,; ;:;; ei ~ ~
'"
-
.... ..'" -'" I '"
N
'"
I
N
~~ I
'" ~
B.• 20.0 110 . 0
~

~. O 20 . 0 \1 0.0 B.,
I
2 0.0 11 0 . 0 B.,
i
20.0 110. 0
TIME ISI TIME ISI TIME ISI TI ME ISI

FIGU RE 5.24 LONG!TUDINAL ACCELERATlON TEST CASE FI GURE 5. 25 LONGlTUDI NAL ACCELERATl ON TEST CASE

g g g

~~'
~:
'"cw '""- '""-
'"eg
w '"ew g
'" 8 0
'"
:1:

'"
-'"
'" ó

1::
:s
~., 20.0 110.0
'""-
~ . I
20 .0 110 . 0 ~. ,
1
20.0 110. 0
:);
g
~ . 20 .0 11 0.0
TI ME ISJ TIME IS I TIM E 15J TIME IS)

_. :s

n-
~: ILvl
8
~ m

- - EXACT
- -TILT
- --
'""-
~g
Lv
'"cw
~
" 0
eg ~ I
", 0
'" ,; '" ,;
;;;
'"w
~
'"g; >-
;= ~

~
~. o
:s
w.•
VI
w.,
20.0 110.0 20.'
TI ME 15J
110. 0 ~. ,
20.'
TIME 15J
110 . 0 20.'
TIME 15J
110 .0
TIM E 15J

~ :s ~ :s

~ ~
9 g
'""-on
"-
~ g .....
'""-
'"w~g
'" g -
;r; ei '"
;;;
,; '" ,; r
;;;
Ó

-...'" '"
'"'"
N
'"a: ~

:s
~., 20.0
TI ME 151
110.0
:s
~ .. 20.'
TI ME 15J
110.0
'" g
~. o 20 . 0
TIME 15J
110.0 ~ .. 20.'
TI ME 15J
110.0

FI GURE 5.26 LONG!TUD I NAL ACCELERA Tl ON TEST CASE FIGURE 5.27 LONG!TUDINAL ACCELERATlON TEST CASE
 ~,--------.--------, ~r----------r---------' ~r-------_'--------,

~
~
! eigr---------~--------~
ea: 8Ó I-----------+----------~

Ê
'"
~
&;
...>< ..'"
!:j

~~------~~----~~ ~~------~~------~ '" Ir~.•~--------~..~.=-.---------:,,:O.O ~

rr. !:-.---------~..~.•~---------::,:!•.•
tr. O 20.0 '0.0 rr. O 20.0 .0.0
TIME IS) TIME IS) TI"E ISI TI"E IS)

~r-------~r-------_' ~r---------.---------~ 8,---------,-------_,

~
~ gr-----------+-----------~
I ~
~ g I-----------+
I
I------------i

r ....
TIME
I
. ;~.----------~~~--------~.~
IS)
... 20.0
TI ME IS)
'0.0
'" a
!:j
...>-
'" g
Ir.!:-.-----------::..,:..=-.---------:,,:•.•
TI"E ISI
20 .0
TIME IS)
110.0

~r----------r---------, ~r---------~------_. ~,--------,--------~

,'"
~
! eigr---------~--------~
~9 r------------r------------j
8
I
'" ~ei !
"
~
'" I
N

~ gr-----------~----------~
~
<r ...
N !!;
.. !
If.~.----------~
..~.•:----------:,,:•.•
'" g7-------~~------~ "'.":".-----------,,='•.• g7-------~~------=
!i.O 20.0 110. 0
Ln ;.'-.__________-::
..
tr. O 20. 0 "0. 0
TIME IS) TIME IS) T I"E IS) TIME IS)

FIGURE 5.28 LONGlTUDINAL ACCELERATlON TEST CASE FIGURE 5.29 LONGlTUDINAL ACCELERATlON TEST CASE

~
~ ~ Ó

..,..,'",
~ <> '" g
'"
;;: g
a:
g - g
..,0• ~ ei ~ Ó

~
Ê
'"
.... tO. O
..'"'"
~

g
tr. O 20.0 "0.0
~IU
fi.o 20.0 tO.O
~
fi.o .... .0.0
TIME IS) TIME IS) TIME IS) TIME IS)

~ ~ ~
'", ,'"
'"
;;; ~ g g
;;: g
e N il
~ei ;t ei

~
E ~ e
"'~
Ir.' ....
TI"E IS)
110. 0 110.0 ....
TI"E IS)
tO.O ~ .. ...
TI"E IS)
110 . 0

~ ~
~ Ó

~ g
a: .
g
'"~ ilÓ
g
;!
!:ja
..,'"~
U

..."...
~

~ '"..,
"'~
ri.• 20.0
TI"E IS)
110.0
'" g
ri.O 20. 0
TI"E IS)
110.0
~
~. ....
TI"E ISI
'0.0
~
~ .. 110.0

FIGURE 5.30 LONGlTUDINAL ACCELERATlON TEST CASE FIGURE 5.31 LONGlTUDINAL ACCELERATlON TEST CASE
 ~r-------_,r_------_. ~ r-------------------.

~
~ g~--------~----------_1
1
1

i :~
_ ~I_---; ~
~ ~ 1----------,-----------1
I

r
~ 0
1
.."'" ~
" 0 1
.."'"
I ~ ~ ~
~~.--------~~~.7.--------~.~
TIME 151
•.• ~.L.__________:-:~'-:
.•,---------.~•.•
T IHE 151
~.L...----------"",.
•..,...----------::-!"o,0
TIME 15 1
~.L.o----------:,~
•.-::.----------::-!
TIME 151
.. o.o

~r----------,-----------, ~r---------r-------_. ~r------'-------' ~r----------r---------.

'""- '""- '""- '""-

'""- '"w '""- '"w
'" "
~ ~~ -
9 g ~----------t---------_1
I ~ ei Ug~
;;; ei ~ ei

~ I V----'\- ----i '"'"

Ê --~~--------~ ~

~ ;~.--------~~~.7.--------~.~O.O ;O7---------~,~.~.•--------~'O.O Ul :.~.--------~,~•.7.--------~lIo.o U'l ;.~.----------:,~-:.--------~D.O

• • ..

TIME 151 TIHE 151 TIHE 151 TIHE 151

~ ~'-------'-I---'
~r----------,-----------, ~ r-----------'----------' ~r----------'-----------'

~
h
'""-
'"

~ 'lJ---.---
9 g 1---------.....1----------1 u ~ 1----------+-----------1
~ ;;, 0

--1
ei
!'-----,f-----l "
~

!:iJ) "'"
V) ~.L.----------:,~•.-::.----------.~•.• :,':-.----------,"'.'-:
.•----------::-!
.. o.o ~.L...----------:,.:-.:-.----------::-!
.. o.o ~.~.--------~,~•.7.--------~"o.o
TIME 151 TIHE 151 TIME 151 TIME 151

FIGURE 5.34 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 5.35 ROLL ACCELERATlON DOUBLET TEST CASE
 ~ !! I! ~

'""-
'""-~
i ~ '"
g
~ '" ~
'" ~
E Ö U>
'"
'"
'"
"'"
.."
;;;
U>

"> "'"
'" ~
~ .. 20.0
TlHE IS)
110.0
"-
~
~ . ....
TI HE IS)
1110.0
~
~. 20.0
TlHE IS)
110.0 ..
~
~ 20. 0
TI HE IS)
\10.0

~
U>
"-
~ '"
_....-- .. -

~~I\
"-:;:
,..>
U>
~

110 . 0 ~.O 20.0 110.0 20.0 110.0 110.0

T IHE 15J T IHE IS)

'"
No
U>
N
>
'"
N

« ~.L.--'-----=,":-----...,.c:•.• 20.0 110.0

!!
1i.'="o------=,,-=-.7o-----:,c:o.o ~.~O------:.,:.O.70-------:.,.:O.O
TlHE 15J TlHE 15J TlHE Is J

FIGURE 5.36 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 5.37 ROLL ACCELERATION DOUBLET TEST CASE

~r---------._--------~

'""-
C>
'""-
W
C '""-
-
zo
0
Ö
~
E
gr------i-------
Ö
;;;
..'"
'"
'"
"-
'" !!
W.a ".0 110 . 0 G!.o 20.0 ... ~o------:.c:o.7o----,---.-:;!o.o ~~o------:.~•.~.-------:,~•.•
TlHE 15J TIHE 15J TlHE 15J TlHE 15J

~r--------.--------, !!,-------,------, ~r---------.---------, ~r-------.--------'

~
e
~ ~~-----+------;
'""-
C>
w
9 gl-_____+ - - - - - - - I i 81--_ _-+1_ _--\
'"

~
;=
0

g
i'.~o-----="L.o:------:.:!D. O
TIHE 15 J
'" Ö

'"co
'"
~
1f.~O------:.,.:o.-=o----,---.::;!o.o
TlHE 15J
,..
"-

'" g':---~~---:-:-.,,-------:~
ia ~. O
r 1
~~.------:.~o.7o-------:.~o. o
TlHE 15J

~r--------'--------, gr------~-----, ~r------,------, I!r-------.-----~_,

....
'"
c
'""-
C>
w
I '""-
....
'"e 81-_____1-____-1
~ ~~-----+------; !2 g~-----+-------I
'" 0 '" Ö
'"
.'" ~g~--r-----~
~.L.-----.,."J...O:-----..."c:O.O :.~o------:.::o':.O,-------.::;!O.o ti.a ".0IS) 8~------~------~
~.o 20.0 'lD. 0
TlHE 15J TlHE (5) TIME TlHE IS)

FIGURE 5.38 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 5.39 ROLL ACCELERATION DOUBLET TEST CASE
 ~ ~ ~

on on
"-
on
"-
zo
~ ~ I
"-
L>
W
c ~
!: g
~w
c
- 0
(\
::;'" ~
on 0 .; a

I
'"x
~

"-
I
'""-
~ x
"-
"-
on
V
..
on on on
~
~ 20.0
T IHE ISJ
"0 . 0
g
rf.a
TIME
20.0
ISI
110. 0
~
~ .. ....
TlHE ISI
1,10 .0 ~.D 20 . 0
TIME ISI
110.0

~ ~ ~
on on
"- "- on
on L> "-
"- W L>
ZO 9 w
~ g g 9 g
~ ei
~ Q
~ ei

~
~g >- CO
"- "- on

..
on on
g
W.O 20.0
TIHE ISI
"0 .0 ~
TIME
20.0
ISI
110 .0
g
tr.O 20.0
TIHE ISI
11 0 . 0
~
~ .. 20.0
TlHE ISI
110.0

~ g
~ 8
,,;
on on

~!: g
"- "- on
on L> "-
"- W L>

eg
rv
ZO 9
ZO 0
g
~ ei ~ ei
~ ei ~
w I w
0

...'"
~
~ ... I
'"on
"-
on
g
11'•• 20.0 110.0
ili
~
~ .. 20.0 110.0
"-
on
g
11'.•
!
20.0 110 . 0
g
~ .. 20.0
I 110.0
TI HE ISI TIME ISI TlHE ISI TIME ISI

FIGURE 5.40 ROLL ACCELERATJON DOUBLET TEST CASE FIGURE 5.41 ROLL ACCELERATION DOUBLET TEST CASE

~,---------.---------, ~r-----------._----------, ~ r--------.--------, 8,---------.---------,

on
"-
on
"-
ZO

~~ eigr-----------~----------~ '" 8

'"
~
X
"- r~----+---I
VI ~
~.L. -----------:..-='.-=.-----------"...J
T IHE
-
ISI
•.•
TIME
.... ISJ
110. 0 !:---------::,.;;.."'.---------,,:-!o.o
TlHE ISI
:.L.------:,-•.='-=.-
- -- ----,,7,.•
TIME IS J

g , - - - - - - - - -r - - -- - - - - , ~ r-------r------. 8,----------,---------,
on
"-
L>
W
C

'"~ ~
"- CO
Vl g ~~
II'.!:-.-----------;..:::.-=,-----------,.oo.o
:! :::.-=.-----------"o:!•. ,
ff.!:-.-----------;.. 110.0 ;!:-.------;..:::.-=.--------,,:-!,.•
TI HE ISI TI HE ISJ TIME ISI

~r--------.--------, ~ r------,------, ~r------._-----,

on
"-
on
on
"-
L>
I
"- W
ZO C
I
~ !I~~I----t
- -- '"
riV
VI ~.~,------:,7,.-=,-------"...J,.•
TIHE ISJ
::;

'"
~

'"on
~r_----------~----------,
0

~.!:-.-----------:,:::•.-:.-------,.o:!•.•
T IHE ISJ
~I ~
~.o 20.0
TI HE
I

ISI
~. , 20.0
I
TlHE 15J
11 0. 0

FIGURE 5.42 ROLL ACCELERATJON DOUBLET TEST CASE FIGURE 5.43 ROLL ACCELERATION DOUBLET TEST CASE
 SF Z A/C IH/S/S) sn A/C I H/S/S) SFX A/C IH/S/S) AZC A/C IH/S/S) AYC A/C IH/S/S) AXC A/C IH/S/S)
~ . OO
o. . ,
5.00 fU·0600
0
0.0000 0.0600 !=O .O750 0.0000 0.0750 ~.OO 0.00
, ,
La\) ~.oo
o•
0.00
. ,
1.00 f'l.OD
0
0. 00 1. 00

O
l
(
------------ r l

::0 "W
~

"'"c:
~?
~
n1?

~
. %N I
m~

~
%N
m? f

~
- 1- -1 ~ ~ r----I---, ~-
;;;
"
'"
c:
'"'" '"
'"
:.,
'" :E:
,.-<
OE
0

,.-< "
SR A/C IDEG/S) SO A/ C IDEG/S) SP A/C IDEG/S) OE: FZP Al e IH/S/S J FlP Ale IH/S/SJ FX P Ale IH/S/S)
~ ;; ,00
--i

'"
--i
~o . o 0.0 20.0 ~.OO 0.00 1.00 0 . 00 1. 00
~
--i

'"
O
l
~1 . S 0.0 17.5
O
l
~ . O750 0.0000

~
0 . 0750
O
l
fO·IOO 0 . 000

r==:J
0 . 100

'"
(I)
--i
--i
'"
(I)

,.n --i

,.n
~NI
(I)

'" %N (I)

rn~ ~N I
m? m ~. '" %N!
mo %
NI
m? -- -r- -
%N
j'TI ~ .-

~ ~ ~ ;;; ! ~ -I ~

" '-----'--------'
,,: l ______ ~ ______ ~

SF ZTH A/C IH/S/S) SFYTH Ale [HISIS) SFXTH A/C IH/S/S) PSIA A/C IDEG) THETAA A/C IDEG) PH I A A/C ID EG)
;;.00 0.00 5.00 !=O.osoo ;;.00 0.00 1.00 ~ooo.o ~.OO 0.00 1.00 ?I.OD 0.00 1. 00
,-- r

"'"c:
:~!! ! %N!
m?

~ ( !
%N
m~

~
";;:;c:
~N I
m?
;;; \ 1 :~I -I' ---I %N
mo

~ "
Rl '"
'"
'"
~
~
'"
" '"
~ SR TH A/C IDEG/S) SQTH Ale IDEG /SJ SP TH Ale IDEG/SJ
~ AR R/C IDEG /SJ OA Ale IDEG/S)
" PR Rle IDEG/S)

S
'"
;;,.5 0. 0 ~.OO ;;,00 0.00 1.00 ~
--i
'"
~s.o

Ol
0.0 25.0

-=:::::::::::::1
?I.OO 0 . 00
1 ,
1. 00 ;;,00 . '"
~
--i
~ --i
--i

~
(I)
,.
n
(I)

'"
~

'" %N
:~! ( -I iNI
m? ~? ~?
:~ m~
- - 1- - - -
§c
I I ~ ~
~

" IL._ _ _' - -_ _--' " 1'--_ _-'--_ _- '

D'
 ~
:; ~ ~
,; I I
I I

I ~
li1
.. f
§
8 I i I
'" 8 ;;; ~
'" 0
~ ei ;;:; ei
'" 0
I
on
on
><
I
I
on

x 0
on
><
>
I on
><

~
'"
. ....
LL
~ ~ ~
~.O 20.0 "0.0 &.• iO . O _0.0 ~ 20.0 \10.0 ~.O 1110.0
TIME ISI TI ME ISI TIM E ISI TIME ISI

~
~ ~ ~

!:ril
I I

~ I
I I .-
,on
'" I §

~ ei ~ ~
on 0 ~~
on
I
'"
;;;
0
1
-
on
; I
J
on >-
>- ~ g
g
'"
.
~
~
~. O 20 . 0
TIME ISI
1110.0 ~. O 20.0
TI ME ISI
\1 0.0
~
~ .. 20.0
TIME ISI
\10 . 0
~
~ 20.0
TIME ISI
\10 . 0

8 8 ~
-
,on
,on
'"
~ ei
8 I L
'"
- ,;
,on
'"
'" 8
;;:; c:i
'"
;;:; ~
'" 0

..
on
on
N
I on
N
I
~
on
N

'" 8
~

~ ~
~., 20.0 1110.0 ~. o 20.0 \10.0 ~. , 20.0 110 . 0 ~. , 20.0 1110.0
T IME ISI TIME ISI TIME ISI TIME 151

FIGURE 5.48 YAW RATE TEST CASE FIGURE 5.49 YAW RATE TEST CASE

~,---------------- ~ r------~----- ~ r------------, ~ r-----r----'

~ ~~ {\
,
on
!
I
~ ~f------------
on
'"w
9 gf------~----­
" ,;
: 0
Lr----+---
~,;
"I --------I

~ ~ ~
:.~,------:,,:-.:-,-------:::.. o.o :.~O------:,:-O.7,-------:::"O.O &.~o--'------:,,:-.:-,- - - - - :,::!o.o U'. ~o------:,;;,.7,-------;,:;!o. o
TIME 151 TIME 151 TIME ISI TI ME 151

!",o~:.I ~~------------~ ~
8 , - - -- -- - - - , ~r----~----, ~ r----~---'I

~ ~ 1
1

=~f_--------- --__I ~
I
gf-- - -- -...,_-----.,
:;:

:,; i
r
" 8 f - -- --...-;- -- - - - j

I
:.0':----- --::""-.,C------:::!
TIME ISI
1I0. 0
r TIME
!
:. ~,------:,:-,.7,-----:,::!o.,
ISI
'"r
~
VI

§
&.~,- - - - -::,,:.-,: ------:'=
TIME ISI
0.0
~
~
Il'.!:-
, - - - - --:;,,;;-.;;-,- - - ---;,.;,.,
TIME 151

8r-------------, ~ r------...,------, ~ r------------------. ~ r----~----,

;:~~:!...-----i
.'" on
a:
I
I N
.I
~ ~W'--
'"
on
a:
on
0

~ gr-------;------i g~. ~--_;:.

l:7_--_:;;!
~. L,------"-.O-----..,.,.
,,O.O tr.G
1 _ _--:::!
8 '---_ _--::'-:--
20.0 \1 0.0 ~~
, ------:; ,o;;-.;;-
, -------;,~,.o O 20 . 0 "0.0
TIME ISI TIME ISI TI ME 151 TI ME ISI

FIGURE 5.50 YAW RATE TEST CASE FIGURE 5. 51 YAW RATE TEST CASE
 - -----

SFZTH ERR IH/S/SI SFYTH ERR IH/S/SI SFXTH ERA IHISISJ SFZTH SIM fH/S/SJ SFlTH SIM [H/S/SJ SFXTH SIM fHISISJ
~.OD 0.00 1. 00 I.U:'UV u.uuuu 0.0500 ~.oo f'S. 00 !"I.OO 0.00 1. 00 ;;.00 0 . 00
r- I i I o r-

."

ê...,
'"
\n
\n
zo

..,
-<
l>
"W
m?
~

:LlJ "OW "'?

~ ."

ê
...,
'"
\n
\n
N

-<
l>
~' I
m?

~~J
1
~' I
m?
;;;
~,
m?
§

~
SRTH ERR tOEG/SI
0.0 11 . 5 PI·aa
saTH ERR
0.00
IOEG/SI
1.00 !"I.OO
SPTH ERR
0 . 00
IOEG/SI
1.00
""
~ !=Q.1S0
SRTH SIH
0.000
IOEGISI
0.750 PI.OO
saTH SIH IO EG/SI
0.00 1.00 PI·OO
SPTH SIH
0.00
IOEG/SI
1.00
...,
--< 0 0 --<
!Tl
Ol ======r- O l I I 0

...,
--<
en
--<
!Tl
en
--< --<
n n
l> ".
...,
en en
!Tl
"'rn ,~
~
~?
~
~~
~
~- I
m?
~ ""I
m~
~
-- 1- -I ~~ I
~
1--

"

JRCKL 3 IHI JRCKL 2 IHI JRCKL 1 IHI SFl ERR IH/S/SI SFY ERR IHISISI SFX ERR IHISISI
0.100 ~.IOO 0 . 000 0.100 ~.100 0 . 000 0 .1 00 PI·DO 0.00 1.00 ~.O600 0.000t' 0.0600 !=V. 0150 0.0000 0.0150

":~
I I

"W "W '"W ~' I

~,
m? rn~

~
mo

~
.. m~
§
m?
~
m?
~
." ~ ."
c;:;
c::: ê
'"
!Tl '"
!Tl
\n \n
\n
'"
\n VI
?I ~I
0
-<
..,~
0 0 0
".

~
J ACKL 6 IH) JACKL 5 IHJ JRCKL ~ IHI '"'~ SR ERR IOEG/SI sa ERR IO EG/SI SP ERR tOEG/SI

,:m
--i
~ . 100
0
0.100 ~.IOO 0.000 0 . 100
--i ~~.o 0.0
~
,
25.0 PI· OD ,
0.00
,
1.00 !=,!.OO
0
0.00 1. 00
!Tl !Tl
--i --i
!Tl !Tl
en en
--i --i
n n

i_I \
l> l>
en ...,
en ....
~~I I ~ -~
!Tl

"'.
mo ~~ m~ . m? ~? I --
§o
~ " ~ ii: ~
ii:

" LI______-L______~ ,, ' - - - - - ' - - - - '

,, ' - - - - - ' - - - - '
 ~r-----------------------, ~r_--------~-------. ~r-------------------~

~ o f------------;------j o
"'
~ g~----------~----------1 !; g ~ ~I------------------------, 9 8 f------------------------I
~
u
0
.
~ ö

..
.. Q

.
~ ei

g:
x x
cr ~ "'- 8
~.!:-,-----------:,,:-.:-,-----------;:;!'o. 0 0".1,-, ----------..,,0:,':.,----------,:::',., ~.'-,-----------"":-."",---------:,,.:o.o :.I,-,-----------::,"",."",---------...,,:-:l,.,
TIME IS) TIME IS) TIME IS ) TIME 151

gr-----------------_, g
~r_------------------~ ~r_------~--------~

~
,"'
"' o
!: 81---____; - . - - - --1 !: g
.
~ .;g f------------------------I :! g f-----------..;--------
~ è
..
~ Q

i
I
!
~
..
ei

U ..
, o
~ ~ ~ g
i
~.,L-----------.,.,,=-.=-
, ---------,-J.o.o Ir. 0
~.L,___________",,:_."",________-:,.: ;!:_,----------::,f,.7,---------:,:-:l,.,
lO . D \10.0
TIME IS) TIME IS) TIME IS) TIME IS)

~ r----,---------------, ~r-----------------------, ~r_------~--------~

on
~8 ~ ;;;
~ ~ ~ 1------------------------, 9 : -----------1
8 f-----------+_

.
~ ei
.
.. Q
~ d
a:

.. ;,L,----------.,.,,=-.=-,---------="0.0
U
N .."' ~

20.0 "0.0
:.L,__________. ,,.,=-.",________-:,,.:,., ~.'-,-----------::,"",.7,---------:,,.:,.,
TIME IS) TIME IS) TIME IS) TIME IS)

FIGURE 5.56 VERTICAL ACCELERATION TEST CASE FIGURE 5.57 VERTICAL ACCELERATION TEST CASE

~ g
~ ~

"',, "':;: 0 "',

"' "''!: "'<>w
~ g eg ug
.
..."'
~ ~ 0 ~ ei ~ ei
0
! Ê~ ~~
x
~

"' ~
g Ion

~., 20.0 "0.0 ~., 20 . 0 110.0 ~., 20 . 0 .0.0 ~., 20.0 110.0
TIME IS) TIME IS) TIME IS) TIME IS)

~ ~ ~ ~
I I
0
~
I
"',, "', :;:
"'
~ g
"'eg
w
! ~ g
:!
u ~
~ ei ~ ei
;x ct
I ..
' Q

~
~
0

"'
I ~
~ 8
;!:
0
en .~
"' ~ ~
~. , 20.0 110.0 ~., 20 . 0
TIME IS)
110.0 Ii.a 20 .0 _0. 0 ~., 20.0 \10.0
TIME IS) TIME IS) TIME IS)

g g
~ ~
~

0 ,"'
~ 0 "'w
~ 8
:;:
~ g ~
u g
03
8
~ ei
..
~ ei ;r ei
.~ ei

Ë
N

~ ~
r.'
fL 20.0
TIME IS)
\10.0
"''" g
~., 20.0
TIME IS)
,.0.0
~

IJ') ~
r.'
fL 20.0
TIME IS)
"0 . 0
;!:
"''" ~
~. ,
I
20.0
TIME IS)
"0.0

FIGURE 5.58 VERTICAL ACCELERATION TEST CASE FIGURE 5.59 VERTICAL ACCELERATION TEST CASE
 g
!l ~

~ ,
on
on
,
on
E
E
g !: g E g r Q
~ ei ;;; ~
on I E
;;;
.; ;; ei

~
on
x ~ B
'" ~

~ ~ ~
~. O 20 . 0 110.0 ~.O 20.0 110.0 ~.O 20.0 110.0 ~.O 20.0 ~o.o
TIME 151 TIME 151 TIME 151 TIME 151

~ g
I
~ g ~ g
,on
E

E g E
E

Q
~ ei
on ~
.;
;;; ei i ;;; ~
on Q.
on on
~ ~
~ ~
co ~
>
g g g
5.0 20.0 110.0 ~.O
~
20.0 \10.0 ~.O 20 . 0 \10 . 0 ~. O 20.0 IIQ.O
TIME 151 TIME 151 TIME 151 , IH E 151

g
.;
- ~ Q

Iv
on
,
on
r
~, ~ §
~ ei
g "
~
.; " g
;;; ei 1"-- - ;;; ~
" Q

'"co
N
Q.
N '">
N '"
N

g
~ ~
~. O 20.0 \10.0 ~.O 20.0 \10 .0 ~.O 20 . 0 110.0 ~.O 20.0 \10.0
TIME 151 T I ME 151 TIME 151 TIME 151

FIGURE 5.60 VERTI CAL ACCELERATION TEST CASE FI GURE 5.61 VERTI CAL ACCELERATlON TEST CASE

~ g
~

~
"w '" '"
0
"
0
g
"
0
g
"on Q .; .;
"
Q

" "
;;;
or on
Q. X Q.
Q.
~ on
~ '" g
~.O 20.0 'h).a ~.O 20 . 0 ~.O 20.0 \1 0.0 ~.O 20.0 110.0
T IHE 151 T I HE 151 TIME 151 , IME 151

g g g

~ '"
"
0
,'"
,'"
,'"
"0w
g
" ~ .; " g
.;
g
.;
'" " " ~
'"w
~

or
~
'"
0 ~ 0

'"
~ g '" g
~
~.O 20.0 110 . 0 ~.O 20.0 11 0.0 ~.O 20.0 \10 . 0 ~.O 20.0 110.0
, I ME 151 TIME 151 TIHE 151 TIME I SJ

~ g g
~
~
I
"w
0
~ '",,
'"
i i
g
e~ E

"on .; ~ ~
r
" "
;;; " i
'"
<C
::;
or
on
g
~ g
~
~.o 20.0 110.0 ~. O 20.0 \10.0 W.O 20.0 110.0 ~.O 20. 0 110.0
TIME 151 Tl HE 151 TIME 151 TIME 151

FIGURE 5.62 VERTICAL ACCELERATION TEST CASE FIGURE 5.63 VERTICAL ACCELERATION TEST CASE
 SFZTH ERR IHISISI SFYTH ERR IH /S/ SI SFXTH ERR IH/S/SI SFZTH SIH IH/S/SI SFYTH SI H IH /S/SI SFXTH SIH IH/S/SI
~.15
0
0.00 1. 15 !='t .DO
ol
0 .00
I
1.00
I
!='t.OO
° i
0 . 00
i
1.00
i
~.OO
0' ;
0.00
,
5.00 !='t.OO
0 '
0 . 00
,
1.00 !='t.OO
0
0.00 1.00

." ." ~

;;:;
:~I
,,~
~~
en
~~

I "~I ~~
c:: mo c::
m~
'" 0"
TT1
V1 0" 0 " '"'"
V1 0 0
'" ~
'"
<
TT1
<
TT1
....
'" ....
'"
,..
;:;
r-
,..n
0 0 . ,..
;:;
,..nr-
.
~

SRTH ERR lOoG/SI saTH ERR IDEG /SI SPTH ERR iDEG/Si SATH SJH tOêG/SJ saTH S IH IDEG /SI SPTH SIH IDEG/SI
n n
f'i.OO 0.;)0 1.00 ?I.OD 0.00 1.00 f'I.OO 0.00 1.00 PI..OO 0.00 1.00 ?I . OO 0.00 1.00 ;; .00 0.00 1.00
TT1
r- 0 0 0
'"
r- 0 ' i i i i

'"....~ '"
~
;j
(5 0
:z :z
.... ....
'"....en ~ ~
'"....en
i~1
,..
n
en
'"
§
.
i:i~ ~~
§
~2:
§
. ,..en mo
n

'" §
. r "~I
mo
~o ----c
1- -1 ~~I
§o

"
JRCKl 3 (Hi JRCKl 2 IH i JRCKl I IHi sn ERR IH/S/Si SFY ERR IH/ S/ SI SFX ERR IH/S/Si
~ .I7S 0 . 000 0.175 ~.I7S 0.000 0.175 ~.175 0.000 0.175 0.00 1.75 PI. .OO 1.00 PI..DO 0.00 1.00
Ol I Ol I

."
;;:;
c::

'"'" '"
V1

"'"
<
~

~~ ~~
'"
~~
§
.
."
;;:;
c::

'"'"v , 0
'"
V1

<
,,~

m~

7 I
~

i~1
.
mo

0
~~

0
..

'"=l '"=l
;:;
,..r- .. 0 .. ,..
;:;
r-
,..n J RCKl 6 IHi JRCKl S IHi JRCKl Y IHI ,..
n
SR ERR IDEGISi sa ERR IDEG ISI SP ERR IDEC/Si
n n
'"
r-
~ .175 0.000 0 . 115 ~.175 0 . 000 0.175 ;;',175 0.000 0.115
'"
r-
?I.DO
0 '
0.00
i
1.00
i
PI..OO
0 '
0.00
i ,
1.00 PI..DO
0
0.00 1.00

'"....~ '"~
....
(5 (5
:z :z
.... ....
'"
....en "N
~

'"en
....
,..nen m~
'"
§
~~
§
. ---
,,~

m~
§
"~ I
,..nen m~
'"
§ I· "~I.
mo
§. -1--- I ,,~

m~
§

.. ~I_ _ _~------~
. '-----'------' .
 6.0 OPTIMAL CONTROLLER ALGORITHM
The development in this section closely follows that of Reference 6.1
which in turn is based on Reference 6.2. The simulator problem is
formulated in the following manner. Consider motion inputs required by the
vesti bul ar model s of Section 3 to be represented by ua in the aircraft and
by uS in the simulator. Let any other motion variables in the problem be
represented by 1Ë a in the aircraft and by 1E.s in the simulator (see Figure
6.1). In the aircraft the motion ua would be processed by the aircraft
pilotls motion sensing system (part of ~a) to produce motion sensations va.
In the simul ator uS is processed by the vestibular subsystem within S~to
produce motion sensations Xs • -
The optimal control problem is to select uS subject to the dynamical
constraints of the problem so as to minimize the cost functional
J = E@T Q! + p[~s T ~ ~s + 1is T ~ 1E.S]} (6.1 )
where
a
! = 1..s - 1.. (6.2)

Q ) 0, .&t ) 0 (6.3)

R >0 (6.4)
Q, .&t and ~ are symmetric and p is a positive scalar. The first term on the
ri ght hand si de of Equati on 6.1 forces the contro" er to attempt to create
the same motion sensation in the si mul ator as in the simul ated aircraft
while the remaining terms restrict the simulator motion by penalizing such
motion.

6.1 System Dynamics

In order to apply conventional optimal control it is necessary to
formulate the dynamical constraints as 1 inear differential equations. Thus
it will be assumed that systems Sa and SS can be represented by

(6.5)

(6.6)

(6.7)

(6.8)
where it has been assumed that the same motion sensing system dynamics can
be applied both în the aircraft and in the simulator and all system matrices
are taken to be time-invariant.
In deriving an optimal controller it is necessary to specify the
nature of the aircraft motion ua since this influences the details of the
controller. Typically one would specify ~a either as a deterministic signal

6.1
or a random signa1 and the resu1ting controller wou1d be optima1 for either
that particu1ar deterministic signalor for that class of random signa1.
Because we des i re our control to cover a wi de range of simu1 ator scenarios
it appears that sel ecting a random signal representation for ua wou1d lead
to usefu1 results more qui ckly than attempt; ng to di scover the best si ng1 e
deterministic signal to use during the design process. For this reason the
ua used in designing the optimal controller was taken to be filtered white
no;se represented by ~n where

(6.9)

and n is white no;se. let

(6.10)

In order to s;mp1 i fy the resu1 ti ng formul ati on and thereby reduce the
computation required to implement the controller it will be assumed that the
bandwidth of the motion base is sufficiently wide that it can be represented
by a pure gain of unity in the foll owi ng analysi s. Thus uS represents both
the input command to the moti on-base and its response. Also we wi 11 take

(6.11 )

6.2 Optimal Controller Eguations

The simul ator controll er dynamical constraint is formu1 ated by

adjoining the systems represented by Equations 6.5 to 6.10 to form

!T = liaT (6.12)

yT = C~T (6.13)
.X = A X '" '"
+ B US + Hn (6.14)

y = CX + 0 uS (6.15)

and from Equation 6.1

J = E trT §. r + pus T ~.!!S} (6.16)

where

[~ ~J
0
'"
A= (6.17)
-A
0

[~J
'"
B = (6.18)

6.2
 c
c= [ -~ St -~J (6.19)

o = [~J (6.20)

G= [~ (6.21)

~= [tJ (6.22)

In the above it can be shown that G :> 0 and is symmetrie. It is shown in

Appendix H that J of Equation 6.16 is minimized by seleeting

US = _~;l[~ f. + ~T2].!
(6.23)
= -F X
where
R
-2 =pR+DTGD
- - __ (6.24)
;:r ~

~12 = f §. Q (6.25)

and P is the solution to the algebraie Riccati equation

B.~ - 1: i B.; ? f. + ~ T 1: + f. ~
1
Q = I I (6.26 )
where
B.~ = E.l - B.12 E.ï 1 B.T2 (6.27)

B.l = cr §. f (6.28)
'" ~ -1 T
AI =A - ~ E.2 E.12 (6.29)

Now partition Equation 6.23 and rewrite it as

(6.30)
The bloek diagram corresponding to equation 6.30 is given in Figure
6.2. This is the so-called closed-loop form of the optimal controller
(based on the loop containing I2)'

6.3
 An open-loop form of the optimal controller (corresponding to the
classical washout filter of Section 5) can be generated in the following
manner. Take the Laplace transform of Equations 6.5 and 6.7 to produce

;a = (sI _ A)-1 ~ ~a (6.31)

;s = (sI _ A)-1 ~ ~s (6.32 )

From Equations 6.10 and 6.30 to 6.32 it follows that

~s = _ I.1(sl- A)-1 ~ ~a

- I.2 (sI - A)-1 ~ ~s (6.33)

or
(6.34 )
or
~s = .!! (s) ~a (6.35)

where W(s) is the optimal washout filter. Now in Reference 6.3 (page 40) it
is shown that

(6.36 )
Thus using Equation 6.36 to alter the first factor in Equation 6.34 it
foll ows that

(6.37)

Note that the number of elements in the transfer function matrix W(s) (and
hence the number of individual filters that must be implemented) is-given by
{the number of elements in up. In general the order of each filter v«>uld
appear to be 2x dim(A). However in practice it is found that to the
precision allowed by numerical calculations there is a cancellation of poles
and zeros leading to W(s) having a~ order of only dim(A). It is interesting
to note that the eXlstence of thi s effect has not yet been demonstrated
analytically for this form of the simulator controller problem. Equation
6.35 represents an open-loop process linking uS to ua. This is the form of
the controller implemented in the present report. -

Before proceeding with the development it is necessary to specify the

variables to be represented by ~s in the cost functional J of Equation 6.1.
An obvious choice would be the actuator lengths 1i of the motion base. As
shown in Appendix 0 the linearized form of the equations for the 1; of each
actuator involve all of the simulator Euler angles J!s. This means that it
would not be possible to uncouple the system equations and thereby realize a
reduction in the real-time computation required to implement the controller
algorithm. On the other hand, if ~s is taken to represent J!s and il and

6.4
their integrals and derivatives an uncoupling of the system equations
follows as shown in Section 6.3.-2. However in this case the limiting of
actuator motion occurs only indirectly and thus may not be as effective. In
the present report thi slatter approach i 5 taken in order to reduce the
real-time computing load. yd has been selected to consist of the elements
xd found by solving the equation contained in Table 6.1. xd contains
lntegrals of ~.

6.3 The Uncoupled System Eguations

Before writing out the system equations it is necessary to establ ish

the location and orientation of the reference frames employed. The motion
sensations y are given with respect to the pilot' 5 head reference frames
F Ps a n d F P (5 e e Fi 9 ure 5 2. 1 a n d 3 • 1). Th e 5 i mul at 0 r mot ion 5 ~ 5 are
represented ~y those of the simul ator reference frame FS. As pointed out
in Section 5.3 the location of FS can have a considerable influence on the
motion-base actuator extensions when tilt-coordination takes place. (It is
found that tilt-coordination is present in the optimal controller due to the
off-diagonal transfer function blocks in W(s) of Equation 6.35.) On the
other hand i f FS and Fps are displaced from one another by an
aS of finite magnitude as shown in Figure 2.1 then the
transl ati onal accel erati on at the pil ot' 5 head when expressed in terms of
FS motion variables introduces unwanted cross-coupling among these
variables (even under the assumption of small motion perturbations) and
prevents the desired uncoupling (see Equations 5.46 to 5.48). For this
reason it was decided to select FS to coincide with Fps and FA to
coincide with Fpa in the present development.

As discussed in Section 5.6.3, in general the washout filtering

process can be applied in FS' Flor in both frames with different
results. In the present optimal control case, once the linearization
process has been carried out, it can be seen from Equations 6.42, 6.43,
6.44, 6.48, 6.49 and 6.50 that there is no significant difference between
frame FS and FI for the present probl em and thus at thi 5 stage in the
process a choice between filtering frames is not required. When the optimal
controller is appied to the actual simulator (rather than its linearized
model) in Section 6.6.1 this question will arise again and will be dealt
with there. The required linearized system equations are developed in the
following subsections.

Foll owi ng the development in Reference 6.1 the control inputs will be
defined to be

uaT = [pJ ~IrJ (6.38 )

and
usT = [!! ~!rJ (6.39 )

6.3.1 Vestibular Models

Following the approach taken in Reference 6.1 reduced order

approximations to the vestibular models of Section 3 are employed for Sa

6.5
and ~s in order to reduce the real-time computing load generated by the
optimal controller. In addition the models are linearized by deleting the
threshold feature. Thus the transfer function of Equation 3.1 for the
sensation of rotational motion is replaced by

(6.40)

The transfer function of Equation 3.23 for the sensation of specific force
is repl aced by

f = Go (s + Ao) (6.41)
f (s + Bo)
In the case of rotational motion sensation the dummy variable w in Equation
6.40 takes on one of the three values PpS' qps rp~. In the case of
translational motion sensation the dummy varlal>le Tin "Equation 6.41 takes
on one of the three val ues f~S' f~S' f~S.
Now ~ can be related to ~S through Equations 2.11 and 2.12.
In keeping with the need to employ linearized equations in developing the
optimal controller these equations will be linearized about ~S = Q,
~ = O. This results in
•
..ês = Is ~S (6.42 )
with
0
.Is =
[~ 1
4>S
-;s
eS ]
(6.43 )

Thus from Equations 6.42 and 6.43 and linearizing by retaining only
first-order terms in the perturbation quantities
•
..ês = ~S (6.44 )

From Equations 6.40 and 6.44 one can derive (where ~S = ~Pps)

(6.45)

qp S = ~_ _G..:;:.S..:J.q-:-s~2_ _ _~ (6.46)
9 . ( "'Iq s + 1)( "'2q s + 1)
S

6.6
 (6.47)

A similar set of equations holds for the aircraft.

Since the pilotls head is assumed to be located at the origin of FS
with Fps parallel to FS, it follows that

!PPs = .!SS
= ls I .!S I (6.48)
Assuming a linearization carried out about .!SI = 0 and ..ês = Q, it
follows from Equations 2.9 and 2.10 that

(6.49)

From Equations 6.48 and 6.49, keeping only first-order terms in the
perturbation quantities, it follows that

.!PPs = .!S I (6.50 )

Also from Equation 2.2

.!PPs = .!PPs - iSI [ OO

g
J (6.51 )

From Equations 6.49 to 6.51

.!PPs = .!SI - 9 [ -1: 'I'

SS
J (6.52)

A similar equation holds for the aircraft with

(6.53 )

From Equation 6.41 one can write (where .!PS = .!PPs):

(6.54 )

6.7
 Goy(s + Aoy)
(6.55)
(s + BOY)

(6.56)

A similar set of equations holds for the aircraft.

In the time domain the rotational motion sensation c~annel represented
by Equat i ons 6.45 to 6.47 has the general form

.. •
~ + T1w~ + T3w~ = T2wf3 (6.57)
where
(6.58)

(6.59)

(6.60)

and ~ represents a component of ,;S 0 r !!J'A, w ta kes on the

correspondi ng val ue of p, q or rand f3 represents . a component of 15 or
1A. Equation 6.57 can be shown to lead to (following Section 6.4):

(6.61)

(6.62)

~ repre~ents ~PS or ~PA and ~, ~, ~ can have either ( )a or ( )s

superscrlpts where

(6.63)

(6.64)

6.8
 -T lp 1 0 0 0 0
-T 3p 0 0 0 0 0
0 0 -T 1q 1 0 0
Al = 0 0 0 0 0
(6.65 )
-T 3q
0 0 0 0 -T Ir 1
0 0 0 0 -T 3r 0

-TIp T2p 0 0 0 0 0
-T 2p T3p 0 0 0 0 0
0 -T Iq T2q 0 0 0 0
.!!.l =
0 0 0 0 0
(6.66)
-T 2q T3q
0 0 -T 1r T2r 0 0 0
0 0 -T 2r T3r 0 0 0

o o o o
fl = [ 01

o
o
o
1
o
o
o
o
1 ~J (6.67)

T~p
0 0 0 (j 0
QI = [ T2q
0
0
T2r
0
0
0
0
0
0
] (6.68 )

Now under the assumption of a reference equilibrium condition in which ~I'

arI' 4fc and ~S are all zero, the equilibrium values for.!pA and.!PS are
T rom quations 6.52 and 6.53)

(!PA)e = (.!PS)e
= [0 0 _g]T (6.69)
A A

and the corresponding equilibrium values for .!pA and .!PS are (from Equations
6.54 to 6.56)

6.9
 (!PA)e = apS)e

= [0 0 (6.70)

The following equations deal with the perturbations about these equilibrium
values.
In the time domain the equation of the translational motion sensation
channel represented by Equations 6.54 to 6.56 has the general fonn
•
ti + BOifi = Goifi + GOiAoifi (6.71)

where r} represents the i th (i = x, y, z) component of the perturbation in

!PS or !PA and f1 represents the corresponding component of the perturbation
1n !PS ?r !PA. Equati?ns 6.52, 6.53 and 6.69 can be shown to lead to (based
on Sèct10n 6.4 and not1 ng that !PPs = !PS and !ppa = !PA)
•
~ = A~ + ~~ (6.72)

1.. = f.~ + Q~ (6.73)

where y represents perturbations in f ps or t pA and ~'1..., ~ can have either

( )a or ( )S superscripts and

usT = Lcl ~!rJ (6.74)

uaT = LêÀ ~ArJ (6.75)

J
0 0

[
Box
A2 = - 0 Boy 0 (6.76)
0 0 Boz

0 0

l
Gox (Aox -Box)
~2 = 0
0
Goy ( Aoy-Boy)
0
0
Goz (Aoz-Boz)
Jy (6.77)

~J
0 0 1 0

[
9
V= -g 0 0 0 1 (6.78)
0 0 0 0 0
(6.79)
f.2 = I

6.10
 o
Goz
J
V (6.80)

The total specific force sensation is found by adding the equilibrium value
of Equation 6.70 to the perturbation values represented by 1...a and yS,
thus
(6.81)

(6 .82)

6.3.2 Collected System Equations

Collecting the system equations of Section 6.3.1 and Table 6.1 it is
found that they can be represented by
•
x = A x + B u ( 6 .83 )

Y = f. ~ + Q~ ( f) .84 )
yd = ~ (6.85)

where~, y, yd,.!! can have either ( )a or )S superscripts.

A= [ i ;] 1 o
~2
o
(6.86)

(6.87)

~l ~J
0
C= [
(6.88)
f. 2

o=
-
[ .Q.l
0
-2
] (6.89)

Si = [Q 0 .!J (6.90)

u sT = Lê! ~!d (6.91)

6.11
 ~T aT]
rL-t:A (6.92)
-AI
and ~ is partitioned as follows:

{xl to X6} represent rotational state variables

{X7 to Xg} represent translational state variables (6.93)
T
xd = [x10 to x 24 ] represent the additional motion
- state variables yd

The additional motion state variables represented by xd obey t he

differential equation shown in Table 6.1. The sel ected generaT form for K
and i results in a ES (from Equations 6.85 and 6.90) that contains
elements of the form

which can be used in the cost functional J of Equation 6.1. From the above
equati ons representi ng the system it can be seen that they can be broken
down into four independent sets of equations involving the following groups
of variabl es:

Set 1 - Pitch/Surge

X = [x 3 X 4 X 7 X 12 X 13 X 16 X 17 X IS] T (6.94)

(6.95)

(6.96)
Set 2 - Ro 11 / Sway

(6.97)

u
-
= [u 1 u]T
5 (6.98)

(6.99)
Set 3 - Yaw

(6.100)

(6.101)

(6.102)

6.12
Set 4 - Heave
(6.103 )

(6.104 )

(6.105 )
Now i f the matrices .9" B. and B.d in the cost functional J of Equation 6.1
are selected so that variables from the above four sets of equations are
kept separate (i .e., do not occur in the same term) and if ~ and §J of
Equation 6.9 are taken to be diagonal, then it is possible to solve the
optimal control problem as four isolated lower-order problems thereby
significantly reducing the amount of computation required as well as
reducing the order of the W(s) filters. This is the approach taken in the
present report. -
The system equations and co st functional J for each of the four
reduced order optimal control problems have the form of Equations 6.1 and
6.83 to 6.85 with the appropriate variables included. The optimal control
solution of Section 6.2 is then applied to each of the four in turn.
The resulting sets of system equations are summarized in Appendix I.
Here choices have been made for K and E which correspond to the final forms
adopted in Secti on 6.5. These -either fo" ow Tabl e 6.1 or represent the
removal of some of the state variables (xlO to x24) from the problem.

6.4 Filter Equations

The fi lters represented by the blocks of W(s) in Equati on 6.35 can be
represented in the time domain by the differential equation

~Na d kx - (6.106)
L k --
k=O dt k
where y is the input and x is the output and zero initial conditions are
assumed. Now arrange thi ngs so that aN = 1 by suitabl e mani pul ati ons.
Thus Equation 6.106 can be written as

a dx + aox
I dt

dNv dN- 1y dv (6.107)

= bN ..::......l.. + bN_1 - + ••• + bI ...::.l.. + boY
dt N dtN-1 dt
Now integrate Equation 6.107 N times with respect to t to generate

6.13
 t tN t 2 [ dNy(t) d N- 1y(t ) dN- 1 x(t )
x(t) =f f f bN I + (b
N-1 N1
I a
- N-1 N1
I )
o 0 o dt N
I dt I - dt 1 -

+ ••• + (b 1 dy ( tI) _ a I dx ( tI)) + (b oY (t 1) - a OX (t 1)) ] dt 1 ••• dt N

1 l
dt dt for all ti ... t (6.108)

From Equation 6.108 it follows that

t t N
x(t) = bNy(t) + f {(bN_1y(tN) - aN_1 x(t N)) + f {(b N_2y(t N_1 )
o 0

t 2
- aN_2 x(t N_2)) + ••• + f (boY(td - aOx(t1))dtd ••• }dt N (6.109)
o
for all ti ... t

Equati on 6.109 can be represented schemati cally by Fi gure 6.3. Fro!11 the
notation of this figure it follows that Equation 6.109 can be written as:
x = xl + bNy (6.110)
where
•
x =~~+!!.y (6.111)
-
with
-aN_l 1
0
-aN-2 0 1
A = ••• •• •• (6.112)
• •
-a 1 0 1
0
-a 0 0

bN~l - aN_l bN
b N_2 - aN_2 bN
B= • (6.113)
••

6.14
These equati ons were sol ved as part of the eva1 uati on program descri bed in
Appendix F using the Improved Eu1 er integration scheme described in Appendix
A.

6.5 Optimal Controller Design

6.5.1 Expressions for ~

The generation of the transfer function matrix W(s) of Equation 6.37

requires a number of matrix manipulations. The computer programs uti1ized
in this process (except for the 7th order pitch/surge case) are described in
Appendix J. In carrying out this process it is found that only a subset of
the elements of the solution P to the algebraic Riccati equation of Equation
6.26 is required. This can be seen from the following.

From Equations 6.23, 6.25 and 6.30

(6.114)

where Il' f2 and f3 have as many col umns as there are el ements i n ~a, xs
and ~n respectively. Now partition f as

(6.115)

where fl~ and f22 are square matrices Qf the same dimension as the system A
matrix te.g., the sub-block Ain A
of Equation 6.17) and all the
fij wi th the same i val ue have the same number of rows and all the
Pij with the same j value have the same number of columns. When
tquation 6.114 is expanded using Equations 6.18 to 6.21 and 6.115 it is
found that

(6.116)

(6.117)

(6.118)

Thus only ~21' f22 and f23 are needed.

6.5.2 Vestibular Model Parameters

The model parameters for the mot i on sensati on transfer functi ons of
Equat i ons 6.40 and 6.41 were se 1 ected so that the frequency response of
these model s matched closely those of the hi gher order model s of Equati ons
3.1 and 3.23 in the frequency band 0.1 to 1. 0 rls. It was found that thi s
could be achieved by taking

(6.119)

6.15
 ';2 = TS (6.120)

Ao = ,;-1
a (6.121 )

Bo = ,;-1
L (6.122 )
and selecting Gs and Go to obtain the desired match. The resulting
parameter values are listed in Table 6.2. The corresponding Bode ploi:s
(with sensation measured in rls or m/s2) for the yaw rate sensation and the
speci fi c force sensati on (for all channel s) model s are given in Fi gures 6.4
and 6.5 as typi cal exampl es. These can be di rectly cOOlpared wi th Fi gures
3.4 and 3.6 which are the corresponding plots of the higher order models
used in evaluating the pilot's sensation of motion in the aircraft and in
the simul ator. The match is judged to be adequate for the present purposes •
When emp 1 oyi ng the out put of t he mot i on sen sat i on model sin the cost
functional J of Equation 6.1 the question arises as to what units should be
used for the rotational and transl ational variabl es in order to give them
suitably similar impact on J. It is suggested in Reference 6.1 that
expressing these variables as a multiple of the corresponding threshold
value for their sensation is a reasonable approach. The resulting variables
are said to be expressed in "threshold units". Thi s has been adopted in the
present report. The threshold values employed are listed in Tables 3.1, 3.2
and 6.2 and the corresonding Gs and Go val ues are presented in parentheses
in Table 6.2.

6.5.3 Random Input Parameters

The spectrum of the random input for which the simul ator control
system is optimized is determined by the matrix ~ in Equation 6.9. In
the present report An is taken to be diagonal and the elements Anij
represent first-order low-pass filter break frequencies for eacn
degree-of-freedom. The val ues sel ected are given in Tabl e 6.2. These are
within the range of values used in Reference 6.1.

6.5.4 Selecting the Order of the Filters

Once the vestibular models are specified the order of the filters
represented by W(s) is only influenced by the number of elements in xd given
by Equati on 6.93. In Reference 6.1 the foll owi ng xd s were empl oyed:-

Pitch/Surge: xé = [JuSIdt usd T

Roll/Sway: xd s = [fvSIdt vSrJT

Thi s resul ted in fifth order fil ters. It was found in thei r exampl e that
the transfer function on the diagonal of W(s) representing direct rotational
to rotational filtering had the form -

(6.123)

and the transfer function representing direct translational to translational

filtering had the form

6.16
 W ( 5) = 5 2N ( 5 ) (6.124)
22 0(5)
where N(s) and 0(5) are polynomials in 5 with non-zero constant terms. This
results in the simul ator Eul er angl es generally tracki ng those of the
aircraft and as outlined in Section 2.5 the washout filter represented by
Equati on 6.124 wi 11 produce steady-state di spl acement offsets i n response to
steady a i rcraft accel erati ons. Thi 5 was i ndeed verifi ed through simul ated
test runs. For exampl e Fi gure 6.6 shows the resul ts of a test run of a
response to a 1 m/s 2 longitudinal acceleration step (following the
procedures outlined in Section 6.6) which employed the optimal controller
that generated the results of Figure 12 in Reference 6.1. The large
steady-state lateral displacement is apparent.

In studying the properties of W(s) it was found that each additional

hi gher order integ ral of s imul ator transl ati onal vel oc ity employed in xd s
added a factor 5 to the nLlllerator of Equation 6.124 (and to the nurneratorof
W (s), e.g., a;rcraft pitch to simulator surge, as well) and increased the
2J
or er of each transfer function element of W(s) by one. The addition of the
integral of the Euler angle to xd s in the-pitch/surge and roll/sway cases
caused Wll(s) to take the form -
= sN(s) ( 6.125)
0(5)
where N(s) and 0(5) are polynomials in s with non-zero constant terms and
added a factor s to the numerator of W12 (e.g., aircraft surge to simulator
pitch) as wel 1 • Thi 5 step also increased the order of each transfer
function element of ..!i(s) by one.

Based on the arguments of Section 2.5 it was decided to ensure that an

S3 factor appeared in the numerator of all the translational motion to
transl ati on al moti on transfer functi ons and to ensure that an 5 2 factor
appeared in the numerator of the yaw motion to yaw motion transfer function.
The need to incl ude sn factors in the nUl'lerator of the remaining
rotational motion to rotational motion transfer functions was not
immediately apparent. In the present section a pitch/surge case is
described which incl uded f9Sdt in xd s in order to demonstrate the
process.

The following test cases resulted from the above decision process:

Nominal Filters

Pitch/Surge 6th order:

(6.126)

Roll/Sway 6th order:

(6.127)

6.17
Yaw 4th order:

(6.128)

Heave 4th order:

(6.129)

Special Filters

Pitch/Surge lth order:

xd s = [J0sdt fIuSIdt2 JUsr dt uSrJT (6.130)

The val ues employed for R, Rrl, 0 and p in the cost functi onal J (see
Equation 6.1) are listed in TabTe b.2.- These values resulted from trial and
error tests starting from the optimal controller values of Reference 6.1.
They represent a reasonable set for demonstration purposes.

6.5.5 The Generation of ~(s)

This portion of the optimal controller design is carried out with the
aid of several computer programs described in Appendix J. The design
process contains three steps:

(1) Gi ven the system parameters of Tab 1e 6.2 generate the sol ut i on to the
appropriate algebraic Riccati equation (see Equation 6.26).

(2) Using the results from (1) solve Equations 6.116 to 6.118 for i!' i2
and i~. Then solve for ~(s) by using Equation 6.37. !!(s) is found as
a ratlo of polynomials

W.. (s) = Nij(S) (6.131 )

1J Di j (s)

(3) Find the roots of the polynomials Nij(S) and Di.1(s) of (2) above and
cancel any roots common to both. Expand the remalning terms to obtain
the final fonn of !!( s) •

Step (2) above is carried out with the aid of a symbolic manipulator
compiler called FORMAC. This allows the solving of Equation 6.37 for W(s)
in terms of the Laplace variable s, an otherwise overwhelming process.- In
Step (3) above it is found that hal f the numerator roots can be cancell ed
out in all cases.

Thi s process has been carried out for the system parameters of Tabl e
6.2. Table 6.3 gives the solution P of the algebraic Riccati equation for
the pitch/surge 6th order case. Taol es 6.4 to 6.7 give the corresponding
numerator and denominator roots for the transfer functions represented by

6.18
W(s). In these tables roots 7 to 12 were cancelled out from the numerator
N(s) and the denominator O(s). Of the remaining roots 1 to 6 those elements
marked with an asterisk were sufficiently small that they were replaced by
zero. This resulted in the following optimal controller transfer
functions for the pitch/surge 6th order case:

W11(s) = (s + 0.188)(s + 0.076)(s2 + 1.385s + 0.548)

x (s2 + 0.766s + 0.652)/0(s)

W12 (S) = O.Oll(s + 9.989)(s + 1.000)(s + 0.192)(s + 0.072)

x (s2 + 0.781s + 0.305)/0(s) (6.133)

W21 (S) = -0.032(s + 0.191)(s - 0.070)(s + 0.076)s3/0(s) (6.134)

W22 (S) = 0.826(s + 0.825)(s + 0.190)(s + 0.076)s3/0(S) (6.135)

O(s) = (s + 0.187)(s + 0.076)(S2 + 1.385s + 0.548)

x (S2 + 0.766s + 0.651) (6.136)

Similar results were found for the other three sub-cases. The corresponding
Bode plots for all the current filters are presented in Figures 6.7 to 6.11.
The plots for the 6th order pi tch/surge case are presented in Fi gure 6.7.
It can be seen that Wll has almost identically a unity transfer function.
The cross-feed from aircraft surge to simulator pitch (Wl.~) has the form of
a low-pass filter which (as can be seen from Section 5.2.2) is the same fonn
used for the tilt-coordination process in the classical washout filter
al gorithm. W22 has the form of a hi gh-pass fi lter as expected. The
cross-feed from ai rcraft pi tch to simul ator sur geW al so has the fonn of a
high-pass filter but it will have very little inf2t uence on the simulator
response due to its low gain over the frequency band of interest. There is
no equivalent filter in the classical algorithm of Section 5. This
cross-feed allows some sensed specific force which in the aircraft was due
to ~ to be generated in the simulator by aSI. In the present case
this is not very useful. It is anticipated that W21 would be deleted in
practice.

The present form of W1 may cause some prob 1ems duri ng the operat i on
of the flight simulator by a~lowing large aircraft pitch angles (and roll
angl es in the roll /sway case) to become simul ator Eul er angl es whi ch may
exceed the physical limits of the simulator. However as demonstrated by the
coordinated t~rn example in Reference 6.1 this may not occur in general. If
the filtering is carried out in FI it is possible that the W12 cross-feed
may help to reduce the simulator pitch and roll angles. Further study is
required. Nonetheless tests were carried out in an attempt to modify the
fonn of the ai rcraft pi tch to simul ator pi tch transfer functi on (W 11) in
order to introduce some high-pass filtering. The first test involved the
6th order pitch/surge fonnulation. In an effort to penalize the generation
of eS the penalty weighting on eS in J (R ll ) was increased by a factor

6.19
of 100 and the penalty on (CIPS - CIM) in J (Qu) was reduced by a factor of
0.01. With these changes it was fóund that WIt remained a pure gain but
with reduced magnitude. ~ high-pass filtering e fect was achieved for Wu .
The test with the 7th order pitch/surge formu1 ation was more successful.
This emp10yed the expanded xd s of Equation 6.130 to penalize f9sdt in J.
The weights emp10yed are listed in Table 6.2. The resulting Bode plots are
presented in Figure 6.8. It can be seen that the desi red 10w frequency
roll-off has been introduced into WIl' although the effect begins at a very
10w frequency. There al so appears to be an undesi rab1 e low frequency
ro11-off introduced into the tilt-coordination channe1 W12 • The utility of
this 7th order form is therefore not c1ear.

The reason why Wu tends to be a unity transfer function is based on

the form of the present problem. The on1y way for aircraft tilt (e.g., ~A
= 0, ~A = constant) to be passed through to the simu1 ator as simu1 ator tilt
i s Vla the Wll rotati ona1 channe1 (un1 i ke the cl assi cal al gorithm whi ch
because . it is based on !..AA as an input rat her than ~ I, can use the
tilt-coordination channe1 for this purpose) • Thus if "W ll were to b10ck 10w
frequency signa1 s it cou1d lead to large fpS errors due to
incorrect tilt matching. For this reason a high-past filter form for W11
tends to be resi sted by the al gorithm. It shou1 d be noted that the above
ho1ds for pitch and roll (aS and 4»s) but not for yaw (q,S). In the
case of yaw, since there is no tilt-coordination involved in the
formulation, a high-pass version of WIl is found (see Figure 6.10).

6.6 Test Runs

6.6.1 A1gorithm Configuration

As indicated in Section 6.3 the 1inearized optimal controller ana1ysis
makes no distinction between reference frames FS and FI. However when the
optima1 controller is app1 ied to an actua1 simu1 ator the governing equations
are non1 inear. Thi s 1 eads to two rel ated effects :

(1) The resu1ting control will be sub-optimal.

(2) A choice must be made between app1ying the filters in

FS or in FI.

As a result of the findings in Section 5.6.3 it was decided to carry out the
filtering in FI in order to avoid simulator displacement offsets. One
possib1e configuration of severa1 that wil1 achieve this is depicted by the
b10ck diagram given in Figure 6.12. The input is the motion of aircraft
frame FA written in FA components (recall from Section 6.3 that FA and
FPa are coincident for the optima1 controller algorithm). These inputs
represent the motion of the aircraft pilot's head. B10ck LIA represents
the app1 ication of the transformation matrix .!:.IA to create the FI
components ~I from 2AA. B10ck OPT F ILT #1 represents the three W11
filters for surge, sway and heave. B10ck OPT FILT #3 represents the two Wl2
cross-coup1ing filters that provide the pitch/surge and roll/sway
tilt-coordination. B10ck OPT FILT #2 represents the two W21 cross-coup1ing
filters for the pitch/surge and roll/sway channe1s. Blod OPT FILT #4
represents the three W22 filters for pitch, ro" and yaw. The rotationa1
and translationa1 slJllmations are then performed to generate .51 and 15

6.20
which are then fed into the same actuator extension algorithm employed in
Section 5 and described in Appendix D.

It should be emphasi zed that the cross-feed from aircraft surge and
sway to simulator pitch and roll is fundamentally different from the
tilt-coordination of Section 5.2.2. For example, in the present roll/sway
case the W12 cross-coupling filter creates ~S responses to a~I
inputs while the tilt-coordination in the classical washout algoritRm
creates ~S responses to f~~ inputs. Thi s means that i f the
aircraft is perfonning a coorainated turn (in which case a~I "* 0 but
f~A = 0) the contri but i on to simul ator ~S response wi 1 f' be quite
different for the classical and optimal algorithm cases. In this sense it
can be seen that W12 cannot strictly be described as a tilt-coordination
channel.

The appropri ate moti on base geometri cal parameters are conta i ned in
Table 6.8.

6.6.2 Inputs for the Vestibular Response Models

In the present case the inputs to the vestibular models representing

the simulator pilot are determined from ~S and aSI with the assumption
of perfect simulator response. Because FSis located at the pilotls head
the vestibul ar input .!PS can be computed directly (using Equation 2.2)
from ~S and ~SI (with ~SS = .!:.sI~sd without using any angular ra te
approximations. However since the vestibular input ~s is not available
it must be est imated as ufSS i norder to compute- -fPS. The process
employed is described in Seëfion 5.4 (with IJJ2 = 12.5 rls). For consistency
the same procedures are employed in calculating the vestibular inputs for
the aircraft pilot used to compute iPA.

6.6.3 Input Test Signals

Two types of test signal were employed.In the first the aircraft
motion was represented by filtered white noise of the type used in the
optimal controller design phase (see Equation 6.9). The second type
consisted of the discrete transients described in Section 5.6.2. As in
Section 5 these inputs were taken to be motions 'of the aircraft frame Fe
(see Figure 2.1). (Note that in Section 5 F~ coincides with Fe while in the
present section FA is taken to coincide wlth Fpa .) The input signals ~AA
were generated by using Equations 5.46 to 5.48 wlth i = e and Bssreplaced
by Ree (see Fi gure 2.1 and note that FA == FPa). The val ues of tfle el ements
in Bte are contained in Table 6.8.

6.6.4 Results

The cOOlputer program used to test the optimal controllers contains

modules similar or identical to those employed in Section 5 for the
cl assi cal washout al gorithm. Of course the fil ters and confi gurati on now
reflect the optimal controller equations. The listing and program
description are contained in Appendix F for 6th order pitch/surge and

6.21
rolljsway, and 4th order yaw and heave cases. Unless otherwise noted this
was the system used in the following test runs.

The notation used in the data plots is contained in Table 5.2. The
only estimated variables plotted are wss' wPA and wpS as explained in
Section 6.6.2.

In the first trial documented in Figures 6.13 to 6.24 the input was a
random pitch angle signal created by passing white noise through a
fi rst-order fi 1ter wi th a 0.2 rjs break frequency. Thi s was the form of the
random input used in the pitchjsurge controller design. In this case only,
the pilotls head was assumed to be located at the origin of Fç in the
aircraft (and at the origin of FD in the simulator). This simpl1fies the
input signals to the simulator system and allows us to look at pure
rotational effects at Fps • The values of ~A (or $:A) and hare shown in
Figures 6.13 and 6.14. The aircraft pilotls sensation of motion as
predicted by the vestibular models (the models of Section 3 used previously
in Secti on 5.6.5) is presented in Fi gures 6.15 and 6.16. The sensed
specifi c force ari ses mainly from the til ti ng of the ai rcraft cockpit wi th
respect to g.
-+

The resulting simulator motion is presented in Figures 6.17 to 6.19.

It can be seen that the primary response is eS which is almost identical to
e A as expected. The a~I in Figure 6.17 is due to the cross-coupling filter
W21 and its effect is seen to be minimal • The reason that the qS$ of Figure
6.19 is slightly smaller in general than the qAA of Figure 6.14 lS that due
to the process used to generate eA, qAA is avail abl e in an exact form whil e
in the case of the simulator, qss is only estimated, using the process
described in Section 6.6.2.

The response of the vestibular models of the simulator pilot are given
in Figures 6.20 and 6.21. The error represented by the predicted sensations
in the simul ator minus those in the aircraft are presented in Figures 6.22
and 6.23. In thi s case it is seen that an excell ent simul ati on has been
achieved. The corresponding actuator extensions of ' Figure 6.24 also seem
reasonabl e.

The second trial run is documented in Figures 6.25 to 6.36. The input
was a random surge acceleration signal created by passing white noise
through a first-order filter with a 0.2 rjs break frequency. The
corresponding aircraft motions ar~ shown in Figures 6.25 and 6.26. The
aircraft pilotls predicted sensation of motion is plotted in Figures 6.27
and 6.28. The simulator motion response ,is given in Figures 6.29 to 6.31.
The rotational motion is present due to the W 12 cross-coupling and is quite
large as seen in Figure 6.31. The corresponding predicted simulator pilot
motion sensation is presented in Figures 6.32 and 6.33. The sensation error
is plotted in Figures 6.34 and 6.35. It can be seen that the simulation has
been reasonably successful. The double spike at t = 27s in SFXTH ERR is due
to the slight lead of the simulator with respect to the aircraft motion (see
Figure 6.7) at the time when the vestibular threshold limiting takesAeffect.
The eS caused by the W12 cross-coupl ing creates errors in f~S and
QSS. The actuator extensions of Figure 6.36 are quite reasonable.

6.22
 The thi rd case is covered in Fi gures 6.37 to 6.48. The i nput was a
surge pulse of the aircraft as shown in Figures 6.37 and 6.38. The aircraft
pilotls sensation of motion is predicted in Figures 6.39 and 6.40. The
simulator motion response is plotted in Figures 6.41 to 6.43. The specific
force at the location of the pilotls head in the simulator shown in Figure
6.41 is amazingly close to that in the aircraft as shown in Figure 6.37.
However, this is the result of unacceptably high simulator tilt pitch rates
(from W1Z cross-coupling) as depicted in Figure 6.43 and this could never be
allowed 1n an actual simulator application. Tilt rate limiting may h3ve to
be incorporated into the algorithm before it is employed in a piloted
simulation. The simulator pilotls motion sensation is given in Figures 6.44
and 6.45. The sensation error of Figures 6.46 and 6.47 indicates that the
specific force sensati ons are quite well represented but the rotati onal
velocity error is probably unacceptable. The actuator length responses of
Figure 6.48 are also a bit excessive.

The fourth trial is presented in Figures 6.49 to 6.60. The aircraft

motion in this case was a roll acceleration doublet as shown in Figures 6.49
and 6.50. The aircraft pilotls predicted motion sensation is given in
Figures 6.51 and 6.52. The resulting simulator motion is given in Figures
6.53 to 6.55. By comparing the aircraft rotational motion of Figure 6.50
wi th that of the simul ator in Fi gure 6.55 it is seen that the simul ator
motion exceeds that of the aircraft. This is caused by the W.l 2
cross-coupling effect augmenting the direct Wll effect (which lS
approximately one-to-one as seen in the fi rst trial run results) • The
simulator pilotls motion sensation is predicted in Figures 6.56 and 6.57.
The sensation error shown in Figures 6.58 and 6.59 indicates reasonable
specific force sensations and probably acceptable rotational velocity
sensation. The actuator lengths of Figure 6.60 are unacceptable. The peak
values are too high as are the sustained levels.

The fifth test case is a yaw rate build-up as shown in Figures 6.61 to
6.72. The aircraft motion is given in Figures 6.61 and 6.62. Because </IA
becomes qui te 1arge it was necessary to employ the exact rather than the
series approximation for cos</lA and sin</lA in the lIA block of Figure 6.12.
The aircraft pilotls sensation of motion is predicted in Figures 6.63 and
6.64. The resulting simulator motions are shown in Figures 6.65 to 6.67.
The oscillatory nature of many of these motions is due to the nonlinearities
caused by the linearly increasing value of </IA with time. In FI canponents
this results in periodically varying ~N elements (see Figure 6.12) which
then feed through the optimal control f1 ters to drive the simul ator. It
should be noted that in the present case a~A and aAA are non-zero. This can
be seen from the plots of f~k and f.~A in Fi gure 6.61, whi ch because <!lA and
SA are zero, represent exactly a~A and aÄA' Because </IS is restricted to
much smaller values than <VA by the filters, the acceleration components in
F S retain the periodic nature of the ~kL elements. All this is the
consequence of employing an optimal controller designed to handle small
perturbations, in a situation \'A1ere large variations exist. From the
simul ator pilot ' s vestibul ar response shown in Figures 6.68 and 6.69 it can
be seen that the unwanted periodic motions were too small to be detected in
thi s case. In the sensati on error plots of Fi gures 6.70 and 6.71 onl y the
yaw rate sensation appears to be poor. Thi s results from the early
reduction in rSS caused by the high-pass filter in order to limit the

6.23
simulator travel. The onset of this motion cue however appears to be quite
reasonable (see Figures 6.63 and 6.68). The associated actuator extensions
are plotted in Figure 6.72 and they are a little excessive.

The si xth test run represents an attempt to remove the unwanted

osci1lations experienced in the fifth run. The results (for the same input
as run five) are presented in Figures 6.73 to 6.84. Since the oscillations
were caused by <VA and <vs being out of synchronization and since it is not
possible to increase <\IS to match <\IA' it was decided to equate <VA in LIA with
<vs (i.e., restrict tne value of <\IA employed). This produces the required
synchronization and also reduces <\IA sufficiently to allow the approximation
to sin<\lA and cos<VA to be used in the computer program. From Figures 6.77 to
6.79 it can be seen that the oscillations have been successfully removed.
The corresponding vestibular response of the simulator pilot is about the
same as in run five (see Figures 6.68 and 6.80), as are the actuator
extensions (see Figures 6.72 and 6.84).

The seventh run (Figures 6.85 to 6.96) represents a vertical

acceleration pulse of the aircraft as seen in Figures 6.85 and 6.86. The
aircraft pilotls motion sensation is one of pure heave as shown in Figures
6.87 and 6.88. The simulator response plotted in Figures 6.89 to 6.91 is
quite mild. This is reflected in the simulator pilotls motion sensation of
Figures 6.92 and 6.93 and the sensation error of Figures 6.94 and 6.95. The
actuator extensions of Figure 6.96 are reasonable.

Tests carried out with the 7th order pitch/surge form of the optimal
controller were somewhat disappointing. When the random input eA of the
first test case (Figures 6.13 to 6.24) was rerun using the 7th order form,
the simulator eS was not significantly reduced. This can be seen by
comparing Figure 6.97 with Figure 6.19. In addition, since the W12
cross-coupling filter is also altered by the order of the controller, tests
were carried out in which a 2 m/s 2 step in aar was used as the input. The
resulting tilt-coordination simulator pitch angle for the 6th order case is
shown in Figure 6.98 and for the 7th order case in Figure 6.99. It is seen
that the 6th order form produces a sustained tilt cue while there is a slow
decay of thi s cue in the 7th order case. Thi slatter decay was expected
because the W12 frequency response plot of Figure 6.8 indicates a high-pass
type roll-off at low input frequencies. The benefits to be achieved by
usi ng the 7th order form of the optimal controll er are thus seen to be
somewhat disappointing.

Based on the probl ems encountered in the fifth test run (yaw rate
buil d-up) further anal ys i s of the confi gurati on represented by Fi gure 6.12
was carried out. It was concluded that this particular structure, although
it represents the optimal contro" er for small motions , does not perform
very well for large motions in cases where ~S differs significantly from lA
when simultaneous appl icaJion of aAI occurs. This can result in very bad
cross-coupling within the ..!.PS components. It is felt that a structure more
like that used for the classlcal algorithm (see Figures 5.1 and 5.2) would
be better , provided that it reduced to the exact optimal control form for
sma 11 mot i ons.

6.24
 References
6.1 Sivan, R., Ish-Shalom, J. and ijJang, J-K., "An Optimal Control
Approach to the Design of Moving Flight Simulators". IEEE
Transactions on Systems, Man, and Cybernetics, Vol. SMC-12, No. 6,
Nov./Oec. 1982, pp. 818-827.

6.2 Ish-Shalom, J., "Design of Optimal Motion for Flight Simulators".

Ph .0. Thesi s, Man-Vehi cl e Laboratory , Center for Space Research, MIT,
Dec. 1982.

6.3 Kwakernaak, H. and Sivan, R., "Linear Optimal Control Systems".

Wiley-Interscience, New Vork, 1972.

6.25
 Table 6.1
Additional Motion State Variables
•
xd = K xd + E u

Roll x10 Ol 0
Xu 00 1
Pitch x12 Ol 0 0 0
x13 00 o1
Yaw x 14 Ol o0
x1S 00 o0 1
x- Tran s x 16 Ol xd + 000 -u
xl? Ol o00
x 18 00 000 1

V-Trans x 19 Ol 000 0
x20 0 Ol 000 0
x 21 00 o0 0 0 1
Z-Trans x22 Ol o0 0 0 0
x 23 Ol o0 0 0 0
X 24 00 . o0 0 0 0 1
lSx1S lSx6

6 . 26
 Table 6.2
S~stem Parameters - O~timal Control A19orithm

Parameter Pi tch/Surge Roll/Sway Yaw Heave

6th Order** 6th Order 4th Order 4th Order
Vestibular
Semici rcul ar Canal
Threshold (rls) 0.063 0.052 0.045
'tI (s) 5.3 6.1 10.2
't2 (s) 0.1 0.1 0.1
Gs 5.41 6.21 10.3
(86.31 )* (118.55 )* (226.87 )*
Tl 10.19 10.16 10.10
T2 162.86* 194.30* 222.33*
T3 1.89 1.64 0.98
Otolith
Threshold (m/s 2 ) 0.17 0.17 0.28
Ao (S-1 ) 0.076 0.076 0.076
Bo (s-l ) 0.19 0.19 0.19
Go 1.0 1.0 1.0
(5.86)* (5.86)* (3.56)*
Random Noise Break
Freguenc~ x(-l)
Anll (rls) -0.2 -0.2 ' -1.0 -1.0
An22 (rls) -1.0 -1.0
Penalt~ Weights
Rll 0.9 0.9 0.577 0.5
R22 0.158 0.158
Rdll 0.158 0.158 0.577 0.5
Rd22 0.158 0.158 0.577 0.5
Rd33 0.158 0.158 0.5
p 1.0 1.0 1.0 1.0
Qll 0.447 0.447 1.0 1.0
Q22 0.894 0.894
Miscellaneous
W2. (rls) 12.5 12.5 12.5 12.5
*With gains expressed in threshold units.
**The 7th order version weighs f9 Sdt by 0.9 in J with the other weights as
1i sted.

6.27
 Table 6.3
P For PitchjSurge 6th Order

*{ 6.95893E-03
-1 . 45963E-03
1.45963E-03
1.92554E-Q2
-1. 92554E-02
-1 . 97005E-02
1. 49837E-11
-7 . 28182E-02
-1. 34853E-11
-0. 12455
1. 36352E-ll
5.19046E-04
-6 . 95893E-03
-7. 14280E-02

1.45963E-Q3 1. 14812E-02 -2 . 39383E-03 -1. 49836E-12 1. 34B59E-12 -1. 36316E-12 -1 . 45963E-03

-1. 14812E-Q2 2 . 39383E-03 1 . 9471BE-03 6. 53839E-03 6 . 37960E-Q3 5 . 18786E-05 1. 78839E-03

-1. 92554E-Q2 -2 . 39383E-03 5 . 54156E-02 -3 . 59614E-12 -1 . 53305E-11 -8. 42845E-ll 1 . 92554E-02

2 . 39383E-Q3 -5. 541:56E-02 5 . 62010E-02 O. 20760 O. 35492 5. 52322E-02 0.20352

1. 49B37E-11 -1. 49836E-12 -3 . 59614E-12 O. 00000 0 . 00000 0 . 00000 -3 . 14201E-17

5 . 71274E-18 9 . 14038E-17 -1. 98042E-16 -2 . 74211E-16 -4. 11317E-16 -1. 21883E-07 1. 44921E-13

-1. 34853E-11 1 . 34859E-12 -1. 53305E-11 0 . 00000 0.00000 0.00000 -3 . 02401E-17

1. 84748E-17 2 . 84257E-16 -5. 86659E-16 -1. 25799E-15 -8. 70916E-16 -4 . 99111E-07 1. 26434E-11

1.36352E-ll -1. 36316E-12 -8 . 42845E-11 0 . 00000 0 . 00000 0 . 00000 -1.03276E-16

9 . 12402E-17 2 . 66221E-15 -1. 34832E-15 -3. 58022E-15 -3. 39662E-15 -2 . 60381E-Q6 6 . 66886E-11

-6. 95893E-03 -1. 45963E-03 1. 92554E-02 -3 . 14201E-17 -3.02401E-17 -1. 03276E-16 6 . 95893E-03
0"1 1. 45963E-03 -1. 92554E-02 1 . 97005E-02 7. :i!8182E-02 O. 1:i!455 -5. 19046E-04 7 . 14:i!80E-02

N -1. 45963E-03 -1 . 14812E-02 2. 39383E-03 5.71274E-18 1. 84748E-17 9 . 12402E-17 1. 45963E-03

00 1. 14812E-02 -2. 39383E-03 -1. 94718E-03 -6. 53839E-03 -6. 37960E-03 -5. 18786E-05 -1. 78839E-03

1.92554E-02 2 . 39383E-03 -5. 54156E-02 9 . 14038E-17 2 . 84257E-16 2 . 66221E-15 -1. 92554E-02

-2. 39383E-03 5 . 541:56E-02 -5 . 62010E-02 -0. :i!0760 -0. 35492 -5. 52322E-02 -0.20352

-1. 97oo5E-02 1 . 94718E ·-03 5 . 62010E-02 -1 . 98042E-16 -5. 86659E-16 -1. 34832E-15 1. 97005E-02
-1. 94718E-03 -5. 62010E-02 O. 55867 0.90971 0 . 811:i!7 4 . 6262óE-04 0 . 30212

-7. 28182E-02 ó . 53839E-03 0 . 20760 -2. 74211E-16 -1. :i!5799E-15 -3. 58022E-15 7 . 28182E-02
-6. 53839E-03 -0. :i!0760 O. 90971 2 . 4159 2 . 5615 1.0ó751E-02 1.0553

-0. 12455 6 . 37960E-03 O. 35492 -4 . 11317E-16 -8. 7091óE-1ó -3. 39662E-15 O. 1:i!455
-6. 37960E-03 -0. 3549:i! 0.81127 2 . :5615 3 . 5398 8 . 0928:i!E-02 1. 6315

5. 1904óE-04 5 . 18786E-05 5. 52322E-02 -1. 21883E-07 -4. 99111E-07 -2 . ó0381E-Q6 -5. 19046E-04
-5. 18786E-05 -5 . 52322E-02 4. 62626E-04 1. 06751E-02 8 . 09282E-02 2.2487 7 . 07135E-02

-7. 14280E-02 1. 78839E-03 0 . 20352 1 . 44921E-13 1. 2ó434E-ll ó. óó886E-ll 7 . 14280E-02

-1. 78B39E-03 -0.20352 O. 30212 1.0553 1. 6315 7 . 07135E-02 0.87838

{PIJ
PI (J+7)
 Table 6.4
!!u (s) Polynomial Roots, Pitch/Surge 6th Order
Aircraft Pitch + Simulator Pitch
Root No. N{s) O{s)
Real Imaginary Real Imaginary
1 -0.69228 0.26168 -0.69259 0.26153
2 -0.69228 -0.26168 -0.69259 -0.26153
3 -0.38280 0.71070 -0.38283 0.71053
4 -0.38280 -0.71070 -0.38283 -0.71053
5 -0.18778 -2.18593x10- 3* -0.18681 -3.86197x10- 3*
6 -7. 60202x10- 2 o -7.60671x10- 2 o
7 -10.000 0 -10.000 0
8 -0.19179 0 -0.19375 0
9 -0.18778 2.18593x10- 3 -0.18681 3.86197 x10- 3
10 0 0 0 0
11 0 0 0 0
12 0 0 0 0
*Set to zero before forming the final filter.

Table 6.5
!!12{S) Polynomial Roots, Pitch/Surge 6th Order
Aircraft Surge + Simulator Pitch
Root No. N{s} O{s}
Real Imaginary Real Imaginary
1 -9.9894 0 -0.69259 0.26153
2 -1. 0001 0 -0.69259 -0.26153
3 -0.39032 0.39086 -0.38283 0.71053
4 -0.39032 -0.39086 -0.38283 -0.71053
5 -0.19223 -5.19045x10- 3* -0.18681 -3.86197x10- 3*
6 -7.18229x10- 2 0 -7.60671x10- 2 0
7 -10.011 0 -10.000 0
8 -0.19223 5.19045x1 0- 3 -0.19375 0
9 -0.18289 0 -0.18681 3.86197 xlQ- 3
10 3.83831x10- 6 0 0 0
11 -1.91915x10- 6 3.32415x10- 6 0 0
12 -1.91915x10- 6 -3.32415x10- 6 0 0
*Set to zero before forming the final filter.

6.29
 Table 6.6

~2ds) Pol~nomial Roots, Pitch/Surge 6th Order

Ai rcraft Pitch + Simulator Surge

Root No. N(s) D(s)

Real Imagi nary Real Imaginary

1 -0.19085 -2. 76090x10- 3* -0.69259 0.26153

2 6.95373x10- 2 0 -0.69259 -0.26153
3 -7.60524x10- 2 0 -0.38283 0.71053
4 4.88642x10- 5* 8. 48942x10- 5* -0.38283 -0.71053
5 4.88642x10- 5* -8.48942x10- 5* -0.18681 -3.86197 xlQ-3*
6 -9.77287x10- 5* 0 -7.60671 x10- 2 0

7 -10.000 0 -10.000 0
8 -0.19085 2. 76090x10- 3 -0.19375 0
9 -0.18565 0 -0.18681 3. 86197 xl 0- 3
10 0 0 0 0
11 0 0 0 0
12 0 0 0 0

*Set to zero before forming the final filter.

Table 6.7

~22(s) Pol~nomial Roots, Pitch/Surge 6th Order

Aircraft Surge + Simulator Surge

Root No. N(s) D(s)

Real Imaginary Real Imagi nary

1 -0.82543 0 -0.69259 0.26153

2 -0.19022 -1. 76239x10- 3 -0.69259 -0.26153
3 -7.60691x10- 2 0 -0.38283 0.71053
4 4.39323 x 10- 6* 7. 60935x10- 6* -0.38283 -0.71053
5 4. 39323x10- 6* -7.60935x10- 6* -0.18681 -3.86197x10- 3*
6 -8.78646 x10- 6* 0 -7.60671 x10- 2 0

7 -10.000 0 -10.000 0
8 -0.19022 1. 76239x10- 3 -0.19375 0
9 -0.18697 0 -0.18681 3.86197 x10- 3
10 -1.69953x10- 6 0 0 0
11 8.49763x10- 7 1. 47176x10- 6 0 0
12 8.49763xl0- 7 -1.47176xl0- 6 0 0

*Set to zero before forming the final filter.

6.30
 Table 6.8
Simulator Geometrie Parameters for FS = Fps
(UTIAS Flight Research Simulator)
(all values in m)

~lS = [-1.45, 0.37, 1. 78]T

~lS = [-1.45, 0.56, 1. 78]T

A~S = [ 0.67, 1. 78, 1. 78]T

~4S = [ 0.84, 1. 69, 1.78]T

~5S = [ 0.84, -0.76, 1.78]T

AóS = [ 0.67, -0.85, 1.78]T

!!.l I = [-1.22, -1. 52, o ]T

!!.2I = [-1.22, 1.52, o ]T
!!.31 = [-0.71, 1.82, o ]T
!!.4 I = [ 1. 93, -0.29, o ]T
!!.sI = [ 1.93, -0.29, o ]T
!!.6I = [-0.71, -1. 82, o ]T
!kc = [-0.02, -0.47, -1.78]T

11(0) = [ 0, 0, -1. 84]T

. L = 2.34

6.31
 Ua

e
US
yd S

FIGURE 6.1 THE SIMULATOR PROBLEM

US
ua
( xn )
-F
_3
+ +'(
-
-- MOTION
BASE --
... US

a
---
S
X X
Sa -F
_1
- -F
_2 "
SS ~
'-

FIGURE 6.2 CLOSED-LOOP FORMULATION OF THE OPTIMAL SIMULATOR PROBLEM

y
,r ,r
" " "
bo b1 b2 bN-, bN

1r + XN "+ X ,r + X ,r + X 111+
)----. -
S
I
~)---. -
S
I
~ ... -
S
I
~)---. -I
S ~ -x
'"
A.. - ,~ - A.. -
"

°0 0, °2 °N-1

~~ ~~ ~~ J~

FIGURE 6.3 GENERAL FILTER BLoeK DIAGRAM

 0
0

0
0
Lf)

0
0

0
en ,
'"",
,
""-"
~o
aJ o
D·
~Lf)

,
:::t'

CC

CCD
0

0
lD
,
0
0
Lf)
r-- ,

0
0

0
Ol
,
10- 3
• • 5 6 78 sd
10- 2
• • 5 8 7 e lil
10- 1
3 11 5 8 7 BliJ

10 0
3
• 567 UI
101
3
• 5 6 1 a lil
10 2
3
• 5 6 7 SIiJ
10 3
FREQ (RISJ
0
0

lD
en

0
0

:::t'
ru

0
0

b~
;t:

0
~O
C) •
w9
0

WO
(f)~
CC ru
:1:-
0...'

0
0

:::t'
ru ,
0
0

lD
en ,
10- 3
• • 5 8 7 lUI
10- 2
• • S878S1
10- 1
• , 587 eSil
10 0
• • 5 IS 7851
101
• • 5 871.
10 2
3
• S87es•
10 3
FREQ (RISJ

FIGURE 6.4 AMPLITUDE RATIO AND PHASE LAG OF REDUCED ORDER SEMICIRCULAR
CANAL MODEL
 0
Cl

0
0
Lf'

0
0
0
cr· ,
~D
CD o
0
~Lf)

::1'
,
cr:
a: 0
o
0
(I)
,
0
0
Lf'
,.....
,
C'
c·
0
cr>
, • '7" • • '7' •
10- 3
3
• 5 B .,.~

10- 2
3 5 •
10- 1
3 11 5 • ., ...

10 0
3 5 • '7 a.
lOl
3 5 •
10 2
S
• 5 1'7 I I
10 3
FREQ (RIS)
0
0
lD
en

0
0
::1'
N

0
0

b~
Ä

0
~ O
C) •
w~
0

WC
(f) ~
a:N
I-
CL'

0
0
::1'
N
,
0
0
(I)
en ,
10- 3
3 • 5 IS '71S1
10- 2
3
• 5 • ., ••
10- 1
3 11 S • '7 I .

10 0
3
• •• S • ., "
10 3
FREQ (RIS)

FIGURE 6.5 AMPLITUDE RATIO AND PHASE LAG OF REDUCED ORDER OTOLITH MODE~ ·
 0
0
R/C DI5PLRCEMENT 0
0
IJ")
r-- m

~ o
0
. ............................ ....
~

~
......
lf)

......
0
0
0
IL- . ... .......... •...........

lf)
cr: x
x

0
0
0
IJ")
0 :
0.0 17.5 35.0 tr.O 17.5 35 . 0
TIME (SEC) TIME (SEC)

0 R/C VELDCITY 0
0
0
IJ")
-
lf)
.......
~

lf) ~ 0
0
.......
lf) 0
~

lf)

cr: x
~
>
0 0
0
0
tr.O 17. 5 35.0 0. 0 17.5 35 . 0
Tnl E (S EC) T If1E (SEC)

0 0
0 0

en

t.:l en
UJ .......
Cf)

I~
0
.......
~
~
0 0
Cf) 0 ~ 0
0 0
cr: Cf)
t-
UJ
I Cf)
t- X
cr:
0
0
0
0 I
5'.0 17. 5 35.0 0.0 17.5 35 . 0
TI ME (S EC) TIME (SEC)

FIGURE 6.6 FIFTH ORDER ROLL/SWAY FILTER (FOLLOWING REFERENCE 6.1)

 -30
A.R.
(db)
-60

-90

-120

-150

10'
Freq. (rls)
36 x 10

24
Phase
(deg)
12

-12

-24

-36+------~~~~~~--~~~~~n_--~~~~~~
10- 2 10'
Freq. (rls)

FIGURE 6.7 AMPLITUDE AND PHASE OF ~ (jw) FOR 6TH ORDER PITCH/SURGE FILTER
 Wil
0
W'2
-30
A.R.
(db) W22
-60
W2;
-90

-120

-150

Freq. (r Is)
36 x 10

24
Phase
(deg)
12

-12

-24

-36+-__~~~~~~--~~~~~~--~~~-r~~
10- 2 lOl
Freq. (r Is)

FIGURE 6.8 AMPLITUDE AND PRASE OF ~ (jw) FOR 7TH ORDER PITCH/SURGE FILTER
 -30
A.R.
(db)
-60

-90

-120

-150

Freq. (rls)
36 x 10

24
Phase
(deg)
. 12

-12

-24

10'
Freq. (r Is)

FIGURE 6.9 AMPLITUDE AND PRASE OF W (jw) FOR 6TH ORDER ROLL/SWAY FILTER
 -30
A.R.
(db)
-60

-90

-120

-150

Freq. (rls)
36 x 10

24
Phase
(deg)
12

O+-----------------------~~~~~~~~

-12

-24

-36+-__~~~~~~--~--~~~~----~~~~~
10- 2 10'
Freq. (r Is)

FIGURE 6.10 AMPLITUDE AND PRASE OF ~ (jw) FOR 4TH ORDER YAW FILTER
 0,-------------------------------------_
o 000 0 0 0 0 0 0 0

-30
A.R.
(db)
-60

-90

-120

-150

-180 +-----r--.,..--r-.,.....,.-r"T...,..,.---,---,---r--r-T""r"rT"ï----,~_r__r_T...,._,._rT'1
10- 2 la'
Freq. (rls)
36 x 10

24
Phase
(deg)
12

oL-------------------------~~~~~

-12

-24

Freq. (rls)

FIGURE 6.11 AMPLITUDE AND PRASE OF ~ (jw) FOR 4TH ORDER HEAVE FILTER
 ---
GA A .. L1Ax
aAI
---.
-
OPT
FILT
#1
-+A
,. y
°SI
-
,.
I
S2 -.. i,.I ..... ii
J h

OPT
-. • ACCN
~A
"'\
,/ FILT
#2
EST r-. t iest

OPT
.... FILT
#3

+,if ~S
-
~
OPT
FILT
#4
-
,. "'\
..- LIS

FIGURE 6.12 OPTIMAL CONTROLLER

 ~r-----------~----------~

~ ~ 81--_ _ _-'-1------1
C>
w
!"!

r
~ 81---- - - ; . - ------4

TI ME
I
151
.
><

:.~.-----------;,-=-•.7,---------,,-::!•.• ~ ~.!:-,----------_:..f.7.----------:,:-}•.•
T I ME 151
~ ~1-----------7-----------4

~
~. ~
. ----------7.,,'f-.:---------"',-::!
TIME
•
15 1
•.•
!"!
~ ei
<JO
<JO
~

8
~. ~,
I ...
----------..,,,,...,..----------,-'
TIME 151

~ r-----------------, ~r-----------------------,

~ ~
~
~ 81 - - - _....;-_
I - ---1
~
~ 81---_ _ - ' -_ _~
~a I ~a
'"u
~
'"
~
~

'" ~
~
~. ~
. -----------;,-•.=-7.----------,,-::!o.o
TIME 151
Ir.!:- ••7.----------:,:-}•.•
, -----------::,:-
TIME 151

~ r-----------------, ~r-----------------------, ~r----------~

,"' ,"'
~
C>
W
,"' "',
C>

r
~ g 1-----------------------4 ~ ~f------------_l 9 g 1-----------7_----------4
I
- 0

~ ci ~ ci
<JO 0

~ a
~
'"
~
~
~.O~----------::20:.-:---------"',:-}
'" 0

~.~
. -----------;,7,.-:.----------,:-:!o. o •.
~. ~.-"""====..,,~..,.-..c:---=,.J
•.• :.!:-.----------::20".•:----------:,:-}O.O • •.•
TIME 151 TIME 151 TIME 151 TIME 151

FIGURE 6.13 RANDOM PITCH ANGLE TEST CASE FIGURE 6.14 RANDOM PITCH ANGLE TEST CASE

"',
,"'
:;:
eg
~ C>
W
C>

~ö ~ ei
;:
><
~
~ ><
~
...."
~

"' "' "'

~ .. 20 . 0
TI HE 151
~o .o
"'
~ .. 20.0
T IHE 151
"0. 0

~ ~ !i
N I
,"'
~g
~
~ ö
"'x
~ g i i
:
;;: ei
x
'"
~
~ ~
~
~
x
0

"'
~.. 20.'151
TIME
~o. o P.• 20.0
TI HE 151
"0.0
"' 8
~ . 20.0
TI ME 151
..o.e
"' ~
W.G 20. 0
T I HE 151
~o . o

8 8
~

,"'
~g
~
"',
C>
W
eg
~ C>

C>

ua ~ ~ u ~
~ ei , '" 0 ;;0 0

N
'"g; "
N ....X
"''"
~ ~

"' 8
n'. o 20.0
TI ME 151
"0 .0
8
~ .. 20. 0
TIME 15 1
"0. 0
"'
~ . TIME
T
20.0
151
"0 .0 ~ . 20.'15 1
TIME
'0.0

FIGURE 6.15 RANDOM PITCH ANGLE TEST CASE FIGURE 6.16 RANDOM PITCH ANGLE TEST CASE
 ~.----------------. ~r-------------------~ !.--------,

A-
U>
~
~ 8 '">
'" 8
~ .. 20.0 '0. 0
'"
~
8
~ .. 20.0 '0.0
I.
~ 20. 0 '0.0 2G . D "0.0
TIME IS) TIME IS) TIME IS) r IME IS)

~ ~ ~
~. ~
U>
~
!: ~
g ~ g
~ ei E Ó
~ ~
U> a ~~
U>
;;;
U>
~ e: ; U>
~

'"
.. .. .. ..
~
g
~ ~ ~
~ 20.0 '0.0 ~ 20.0 "0. 0 ~ 20.0 '0.0 ~ 20.0 1110. 0
TIME IS) TIME IS) TIME IS) TIME IS)

8 ~
-
~
,
U>
~
,
U>
!: ~
ZO
g ~ ~
~ c:i
- a
ZO Ó
~ ~
U> a U> a
U>

... U> ...

U>

'"
N
A- N
>
'" ~

~
~ .. 20 . 0
TIME IS)
'0.0
Ó
~ .. 20.0
TIME IS)
'0.0 ..
~
~ 20.0
TIME IS)
'0.0
~
~ .. ....
TIME IS)
'0.0

FIGURE 6.17 RANDOM PITCH ANGLE TEST CASE FIGURE 6.18 RANDOM PITCH ANGLE TEST CASE

~ ~ ~
~
i
G
w
0>

~ d
g
Ó
e" 8
ZO Ó
'" !:
'"a. '"
'"a. '"
~
A-

'"
..
~
~ lO.D
TIME IS)
tO.O
g
~ .. 20 . 0
TIME IS)
110.0
'" ~
~ .. ZO . D
TIME IS)
"0.0

"
W
0> ~
~ g
ZO Ó

'"
~
W U> ~
g
~
0 ~

~
U>

..
~ JO.O
TIME IS)
.0.0

~ ~ ~ ~
I
~
,~
,'"
"e
W

~~
"w
e 8
I
I !: 8
I "98
U>
E Ó E Ó
Î ~ ei
U>
a.
U>

~
i N
~
Ir
U>

..
~
~ 20.0
TIME IS)
'0. 0 ..
~
~ 20.0
TIME IS)
.0.0
U>

..
8
~ 010.0
TIME ISI
'0.0
8
~.O ".0
TI"E ISI
'0.0

FIGURE 6.19 RANDOM PITCH ANGLE TEST CASE FIGURE 6.20 RANDOM PITCH ANGLE TEST CASE
 5FZTH ERR IHI51S1 SFYTH ERR IHISISI SFXTH EAA IH/5/S1 SFZTH 5 J H IH/S/SI SFYTH 5 J H IH/S/SI SFnH SIM IHISISJ
?I.OO 0 . 00 1. . . ~.115 0.000 0.175 ~ . oo D.DO 5.00 ~.DO 0 . 00 Z.50
i '"I I

::~I :1 r : 1 -- 1-- 1
"Tl

~-I ~- "Tl
c;:; me me ~-
me
c:
'"
(Tl
~o
~
;,
'"
c::
'"
(Tl I ~
;,

'"
N N
'" .....
~ ~ ' L'"'--_ _ _----'
~ '=' ~
o
' - - I_ _- - - '_ _- - - - '

~! c=::
3

"
;::j ;;,00
SATH EAA IOEGISI SQTH ERR IDEG/SJ SPTH ERA ' DEG I SI
"
;::j
SATH SIM iDEG/SI SQTH 5 I H IDEG/SI SPTH SIM IOEG/SI
:t:
~.OO
°l
0.00
I
1.00
:t:
~I.OO ~o.o 20. 0
-, ;; .00 0.00 1.00

".
:z s;:
'"
f;; '"
f;;
--< --<
[;J (Tl

--< ~
'"--< ~

~
'"
(Tl ::1 - I ::1 =ê I ~f:
~
n
".

'"
(Tl
:~I ::I~ I. ~-
rTl~ 1

~
_

" L'_--==-"------'

JACKL 3 IHI JRCKL 2 IHI JRCKL J IHI SFZ ERA IH/S/SI SFY EAR IH/S/SI SFX ERA IM/S/SI
fO.00 2S0 0.00000 O.oozso ~ . oo 0 . 00 1.00 ~'WW'WW 0 . 00750
fa. zoo 0.000 D.lOO ~.soo ~ .500
w.wwwww

°l ~ i ° l I i I 12 I

"Tl

lij
~~I
I
"Tl

) :~I
c;:; ~? ~-
m,. m,. ;;:;
c: c:
;;; ~
'"
(Tl ~ ~ ~ '"
(Tl

'"s:
N
'"
N
N

~ ~
:z
:z
'=' pi .c ?
0'=' 0
e =<
"
3:

::;
" JAC KL 6 1111 JRCKl 5 1111 JRCKL " IHI "
::; SA ERA IOEG/SI sa EAR IOEG/SI SP ERA IOEG/SI
n
n
:t:
0.200 0.200
:t:
~.OO 0 . 00
,
1.00 ~: OSOO 0 . 0000
,
0 . 0500 ~ . oo 0.00 1.00

". ".
:z :z
'"r
(Tl
'"r
(Tl

--< --<
(Tl
(Tl

'"
--<
n
".

'"
(Tl

~ ~ ~
'"
--<
n
".
'"
(Tl
i_I
~

me

~
;, ~- I
me

~ " -? ~-
m,.

~ LI ______~______~
 ~ I! I!

~
e
u 8 8
;r: ei
~ ei

u
x
a:
.~ ~ .. I!
a:
%
a:
f
. ~ 110. 0 ..~ >0.'
TIME IS)
110. 0 ~.
I!
>0.'
TIME IS)
110. 0

I! ~ I! I!
.,w
~ '"
'"
e ,..,'"
~ 8 ~ 8 ~ 8
a: 0
I
e8
~ ei ~ ei ~ ei
a:

~
.
~
Ë
%
a:
a:
" I!
I! .... I!
~
~. lO . D
TIME IS)
110 . 0
~
~.D 20, 0
TIME IS)
110.0 ~. D lO . D
TIME IS)
110.0 U lO.'
TIME IS)
'0. 0

I! I! I!
-
..,
~ '",
'",!;
w
<>
~8
~ ~ 0
~ 0~ e
~ ei ~ 0
a: ~ ei

u
N
..
a:
..'"
~ a:
a:
a:

.
a:

. .
~

~ I! 8
~ lO.D
TIME IS)
110 . 0 ~ 20 . 0
TIME IS)
110. 0 ~ >O.D
TIME IS)
1110.0 r.a lO.'
TIME 'IS)
110.0

FIGURE 6.25 RANDO/l SURGE ACCELERATION TEST CASE FIGURE 6.26 RANDOM SURGE ACCELERATION TEST CASE

~
I! ~ I!
;;; ~
.., ~
'"
I:! 8
e
~ ei
u
,0
I!
a:
or
x
~
e;
~ .'" I!
~
'" ~
~.o 20 . 0
TIME IS)
110.0
I!
~ . 20 . 0
TIME 151
110 . 0
~
~.o lO. D
TIME 151
tO.O .. ~ lO.'
TIME IS)
ftO . O

I! I! ~

,'" ~ ~

!'" e
~8 ~
a::
~
e
~
8
ei ~ ct
;:; 8
;Z ei ~
a:
..
8

>-
~ ~
a:
51
8
~
~ g
.
~
"'g
~. D lO.'
TIME IS)
tO . O ~.o 20 . 0
TIME 151
110. 0 r.a lO . D
TIME 151
1110 . 0 r.o >0.'
TIME IS)
110.0

8
~
8
~ ~
~ ~
~
~
..,~ ~ ~
~ 8 9g a:: e
~ ei
~ 0~ ~
I
~ ei a:
a:
ct
a: a:
~ ~
N
~ ffi ~ a:
., I!
'" 8
W.• lO.' (5)
TIME
110. 0
I!
~ . lO.'(5)
TIME
110. 0
., I
W•• >0.'(5)
TIME
tO. O U ....
TIME (5)
110.0

FIGURE 6.27 RANDOM SURGE ACCELERATION TEST CASE FIGURE 6.28 RANDOII SURGE ACCELERATION TEST CASE
 ~

;;;
'"crx ~ '">x '"x
~
..
~

~ '0.0 "
TIME
.. ISI
'0.0
~
~ .. 20.0
TIME ISI
'0.0

~ ~ ~ ~
I
i
'",
'", ~ ~ ~
'" g ~ 8
=~ ~~
!
~ ö '" ,; ",0

'"
..'" ! '"
'"<r
~ ~
'">
~
~

..
~

~
~ .. 20.0
TIME ISI
~O.O
~
~.O 20.0
TIME ISI
\10 . 0
~
~.. " ..
TIME ISI
160.0
~
~ 20.0
TIME ISI
.. CI.O

g g
~
-
'",,
'" ~ ~ ~
'" g x
- 0
'" g
x 0

~ ei '" ,; ;:;:; ei ;; ~
'"
'"<r
N
..
N ~ '"
N

..
~

~
~ 20.0
TIME ISI
'0.0
0

~.O " ..
TIME ISI
'0.0
~
~.. ....
TIMÈ ISI
110 .0
~
~ .. 20.'
TIME ISI
'0.0

FIGURE 6.29 RANDOM SURGE ACCELERATION TEST CASE FIGURE 6.30 RANDOM SURGE ACCELERATION TEST CASE

~ ~ ~ ~

~
~ ~
~ ~
0
~
~

'"
g
,;
'f y ~

'"
g
,;

iE :c x
~
..
'"
~
~ .. 20.0
TIME ISI
'10.0
g
''''. 0 20.0
TIME ISI
"0 .0
'"
~ .. 20.0
TIME ISI
"0.0
g
~.O 20.0
TIME ISI
110. 0

~
,'"
'"~ g
'" ,;

~ '"
0
~
~

~
:;;
~
~.Co " ..
TIME ISI
V)

~ .. 20.0
TIME ISI
" ..
TIME ISI
..0.0

g g
~ ~
~

~
§ ,..,'" '"
e
0 9g '"~g eg
~ ~ '" ,;
'" ó '" ó
'" ;;;
.'" :g
:;, '"'"
~
~ .. 20.0
TIME ISI
110. 0
~
~ .. 20 .0
TIME ISI
\10.0
g
~ . "
TIME
.. ISI
tO.O
~
~ .. " ..
T IHE ISI
110.0

FIGURE 6.31 RANDOM SURGE ACCELERATION TEST CASE FIGURE 6.32 RANDOM SURGE ACCELERATION TEST CASE
 ~ !! ~ !!
'",
'"w ~
0

~ ö
8 e8
U'> ffi 0

Ê ;= I >< :l; I
0-
~ ! i
.. !!
.. .. ..
U'>
~
U'>
! !!
~ lO.O 1110.0 ~ 20 . 0 1110.0 ~ 110.0 ~ 20.0 \10.0
TIME (SI TIME (SI TIME (SI

!! ~ ~
~ ,
U'>

~
~ '"ew U'>

~ ~ 8 ~ g
'" 0
~ ö
U'>
'" .;
'" '"
W

...>- ~ ...>- 0

!! '"
'" ~
Jr.. 20.0
TIME (SI
"0. 0
U'> ~
W:O 20 . 0
TIME (SI
.. 0.0
U'>

~ .. 20.0
TIME (SI
..0.0 20.0
TIME (SI
1110.0

8 ~.; !!

~ i ~
~
";: g 8 ~I eg

::;'"
~ ei ~ ei
~ ei .;
'" '"
...
N
'"
lli ~ ~ lli
'" g
~ .. 20.'
TIME (SI
1110.0
g
~ .. 20 . 0
TIME (SI
110.0
U'> S
~ . 20.'
TIME (SI
110. 0
g
~.o 20.0
TIME (SI
110.0

FIGURE 6.33 RANDOM SURGE ACCELERATI ON TEST CASE FIGURE 6.3Q RANDOM SURGE ACCELERATION TEST CASE

~ !!
,'" U'>

,'" g
~ "
~~
WO r ~
~
~
g
Ó

~ '"
~
0-
:;; ~
..
U'>
§ §
!!
ri.• lO.O 110.0 ~ 20.0 110.0 ri.• 20. 0 110 . 0 ti.D 20 . 0 1110 . 0
TIME (SI TIME (SI TIME (SI TIME (SI

~
N

~ .n
~ ~ g g
";;:
g
e
~ 0 '"~ !! 0

~ !!
;=
~
~ .. 20.'
TIME (SI
'0.0
~
~ .. 20.0
TIME (SI
1110.0
§
~ .. 20.'
TIME (SI
110.0
§
~ .. 20.0
TIME (SI
1110.0

5 !!
,
U'>

~
U'>
g g
'"~ I.
e
WO '" !!
::; 0

...'"
~
;=
~ ~ lli
ri.• ....
TIME ISI
"0 . 0
!!
~ .. 20.'
TIME (SI
"0.0
~
~ .. ....
TIME (SI
1110. 0
~
~ . ....
TIME ISI
1110.0

FIGURE 6.35 RANDOM SURGE ACCELERATION TEST CASE FIGURE 6.36 RANDOM SURGE ACCELERATION TEST CASE
 ~ ~ ~

0
zo
u 0
~ ö
g
'"cw
,
u
co 0
i~
g
ö
co
u
x
co
.
co
x
~
.."
~
.
co
g g
~ ~
ei.o lO.O "0.0 W., lO. O ~ o.o ~., zo . o "0 . 0 ~. , lO. O "0 . 0
IIHE 151 TIME 15 1 T IHE 151 T1HE 151

g
~ ~

,;;; ,on '"cw

,'" ,on ~
w
'!; ~ g ,
u g 9 g
,u ~
co .;
~ ci ~ 0
co
co
u
~
..
co co
w
co
co
co
co
~
~
"
~
g
~
~. , zo. O "0. 0 ~., 20 . 0 "0 . 0 ~. , l O. O "0.0 ~. O 20.0 "0.0
T IHE 151 T IHE 151 TIHE 151 T1HE 151

g g
~

,
~
i
on
'",
'"zo ,
on
zo g I
~ ~ 0~
u
~ 0
co
~ 0
!
'"u :!;
I ~
co
N
co
~ a::
~ ~
~ ~ ~
~., lO.O "0. 0 ~.o 20.0 "0 .0 ~. , 20 . 0 "0 .0 ~. , ~O . O "0 .0
T I HE 151 TIME 15 1 T IHE 151 T IHE 151

FIGURE 6.37 SURGE ACCELERATlON TEST CASE FIGURE 6.38 SURGE ACCELERATlON TEST CASE

~
~

~ "~
~
!:: g
~ 0
0
'"w~ g
~
I ~'~
~ 0~
'",
~
~ ci
or V-- .
ci

I
co

"x
co

~
~
x
"- on ~

'" g V>
~
~.o .!o.o "0 . 0 ~., l O.O "0. 0 ~. , 20 . 0 "0. 0 ~. , 20. 0 "0 . 0
T IHE 151 T1HE 151 TIHE 151 T IHE 15 1

~ ~ g
-

~
,
V>

~
!:: g
~
eg g
'"cw
u
u ~
~ ö ~ 0
;;; 0 ;;: 0

>-
"-
co
V>
~
"-
"
co
V>
g ~ ~
~ Vl V>
~., 20 . 0 ~O.O ~., 20.0 ~., lO . D "n. o ~., l O. D ~D . O
"0,0
T1HE 151 T IHE 151 T1 HE 151 T IHE 151

g
~
8
~

I
~
,onon ,
on
~
w
'"ew
!: g eg
~ ~ u ~
~ 0 ~ ei I co 0 ;;: 0

co
N
"
~
"a::
~
N ~
"-
on on on
g g g
~
tr.o 20. 0 "0.0 ~., 20 . 0 ~O . O !i. o l O. D "0 . 0 ~., 20 . 0 "0 . 0
T1 HE 151 T IHE 15 1 TIME 151 TIME 151

FIGURE 6.39 SURGE ACCELERATlON TEST CASE FIGURE 6.40 SURGE ACCELERATlON TEST CASE
 ,

~ g ~ ~

°A
"~.
~

I I
I
-

vv~
V>

~!: g
,
V> ,
V>

'"- g i
'"
~ c~ ~~
~ ei
V> \J '" 0
-
I
V> V> c
---'

~~
V>
V>
x ~ ~ i
<r
~
~.o 20.0
I 110.0
~
g
Ii.o 20.0
I "0.0
~
~.o 20.0 -'0.0 ~. o 20.0 1f0.0
TIME ISJ TIME ISJ TIME ISJ TIME ISJ

g
~ ~ ~
I
I I I

~
,
V>
,
V> I
~

=
g
V>

!: g
"
" g
;J; ö
I ~~
I
0 '" 0 i
V>

V> I
V>

~
-
V>
>-
I V>
>-
>-
<r
~
~.o 20.0 1f0 . 0
~

~
~.O 20.0
I
1f0.0
>
g
~.o 20.0 110.0
8
~.o 20.0
! "0.0
TIME ISJ TIME ISJ TIME ISJ TIME ISJ

g 8 8
- - i i
- I
0'" ~ ,
V>
r ~

~ ei
8 '"-r c
0
'" 8
;;; ei
'" c
;;; ~
V>

-
V> .;;; -
V>
N
I V>
N
N N >
<r ~
8 0 ~ ,
~.o 20.0 "0 . 0 ~.o 20.0 \10.0 ~.o 20.0 1f0. 0 ~. o 20 .0 1f0 . 0
TIME ISJ TIME ISJ TIME ISJ TIME ISJ

FIGURE 6.41 SURGE ACCELERATION TEST CASE FIGURE 6.42 SURGE ACCELERATION TEST CASE

~ r-----------------------, ~r-----------------------, 8 r---------~------_.

i:~
i
'"
'"w ~
9gr---_____+-
I ______--I
~ ~ 1------------1
V>
9 81---- - - - - - -- -1
r 0 '" ° ~------t '" 0
'" V>

I ..;;;
~. LO----------~,,:'1-70----------,~o. -----------:,70'.~0-----------'
x
V>
V>

0 l(' :.L
o .. O.O ;Lo----------~,O~.70----------~,O. O ~LO----------7,O~.~0----------,~O.0
TIME ISJ TIME ISJ TIME ISI TIME IS J

~ .---------------, ~ r---------------, ~r---------T

i ---------.

LI~'- -liI-------t
~

I
LI1r"-"
'" c

'"
"'" c
~
'" 0

>-
~
V>
g 1-------------'------------1
2:

~
Q I
I
~. !:-o----------""":".:-0---------,,:::!o.o e:.!:-o-----------;,:-o."O
"----------:::!
. o. 0 :.!:-o----------""":".:-0----------:::!
"o.o ~.!:-O-----------:;;
,,-:-.o;:----------,,:::!o. 0
r IME ISJ TIME ISJ TIME ISJ TI ME ISI

~ r_----------~----------, ~ r-------~--------, ~ r_-----------' ~r-------~--------,

~ I
I ~ I I
,
V>

'"'"
dl------'-· - ---I 9 8 f-------------...:I-------------j 9 g f-------------+------------l
V> I ~° I i :z: ei ~

.. ~ I
:. ':-0----------7,,:'I.-:o ----------,~o.o
V>
a;
i
I
~
:,LO-----------::,.,-
O...,.O----------,..JO.O ~.':-o-----------:-",..1.,-0----------..J"o.o !
~':-O----------~'O~.O:----------:,..JO.O
TIME ISJ TIME ISJ TIME ISJ TIME ISJ

FIGURE 6.43 SURGE ACCELERATION TEST CASE FIGURE 6.44 SURGE ACCELERATION TEST CASE
 g ~ ~

,
on ,V> !
,
on
r
<>
"'0
I
~g
~
~
~ g
r 0
;;; 0 :":
V>
0
~ ei '" ,;
:f.
'"
~g
~
X X Q.
"- "- on
V> In Ë
~
~.o lO.O ~O.O ~.O 20. 0 11 0 . 0 ~.O 20 . 0 110.0 ~.O 20 . 0 110.0
TIME 151 TIM E 151 TIME ISJ TIM E 151

~ ~ ~

,
i
V>

,
V> ,
on
r ,'"
;; g !; 8
;;; 0
'" ,;
:f.
'">- I
~
~
>-
"- "- '"'"
'" g '" l!
~. O 10 . 0 110.0 20 . 0 110.0 ~.O 2<>.0 110 . 0 ~.O 20.0' 110 . 0
IIME 151 TI ME ISJ TI ME ISJ IIME 151

~
~ ,; ~

0 ,'" , ~
on
<>
~ '"0 ,
,n
;; g g !: g 98
Vi ~ ei
'"
Q
,; ~ ei

Ë '" :f.
'"
~
...'"
N
"-
'" g '"
V>
g ~ '"'"
l!
~.o 20 . 0 '110.0 ~.O l O. D 110 .0 ~.o 20.0 11 0.0 ~.O 20. 0 110.0
TIM E ISJ TI ME 151 TIME 151 TIME ISJ

FIGURE 6.45 SURGE ACCELERATlON TEST CASE FIGURE 6.46 SURGE ACCELERATlON TEST CASE

,; ~

,'" ,
on

'",!; ~ ~
0 ~
'" .
'" 0 a: ~
",0
:f. 0

~
"-
~
Ë :l;
I/')

~ ~ ~
~.o 20.0 110.0 ~. O 20.0 \10.0 Fi.D 20.0 110.0 ~.O ZO.D 110.0
TIME 151 TIME ISJ TIME 151 TI ME 151

~
-
~
,on
<> ~ ~
"'
0

'"'" g.
",0

Ë
... '"
~

V> '"
V>
~ ~
~.O 20.0 \10.0 20.0 110 . 0 ~. o lO . D "10.0 ~.O lO.O
TIME (SJ TI ME (SI TIME ISJ TI ME ISJ

~ g ~
,;

,'" I
,
V>
r
g i I
~ r

r
'" g "' 8
'"
:f. ,; a::
:f. 0 ~ ö
'"
~
N
'"
~
"i
"-
VI ~ '"
V>
8 ~ ~
~.o 20.0 110. 0 ~.o 20 .0 110.0 ~.o 2<>.0 110.0 ~.o 20 .0 110. 0
TIM E 151 TIME IS) TIME ISJ TIME ISJ

FIGURE 6.47 SURGE ACCELERATlON TEST CASE FIGURE 6.48 SURGE ACCELERATlON TEST CASE
 ~:~
~ ~

'""-
'""-
!: 2
~ ei
~g
!:
~ ei
~:v
"-
a: 0
'
~ ei

u
><
..
><
~
a:
f :1:
'" ~ ~
0

IT" 20.0 11 0.0 IT.' 20.0 loIO.O ~.o 20.0 "0. 0 ~., 20 .0 110.0
TlHE ISJ TlHE ISJ TlHE ISJ 11HE ISJ

~ ~ ~
-
~
'""-
'"!!ö"- g
~
z:
- 0
e
~ 8
'""-
9'" g
a: .;
~ ~ ~ö

u
>-
'" g
ci

IT. ' 20.0

1 IHE ISJ
~
~
ei

.;
IT"
\ 20.0
11HE ISJ
"0. 0
~
w
"
~
IT" 20.0
TlHE ISJ
"0.0
'"a:
0

~
IT. ' 20.0
11HE IS J
110.0

g 0
~
~ ~

~
~ g
~! ~
'"wc ~g
~
~ \ g - - - - - -- --
~ 0
u.;

. .
;:[ ei
~ ei

u
N
a:
8
'"
N
~

f ..'" '"
0<

~
~
IT., 20 . 0 110.0 ~.o 20.0 110.0 IT" 20.0 1,10 . 0 IT" 20 . 0 110.0
T1HE ISJ 11HE ISJ 11HE ISJ 11HE ISJ

FIGURE 6.49 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 6.50 ROLL ACCELERATlON DOUBLET TEST CASE

~ ~

~
'""-~ g
'"
"-
'""-
"
;:;8
i
~ ei
;;: ei

><
~
.'" ~
~ ..'""
'" ~ '" ~ ~
Ir., 20 . 0 ~O.D ~. O 20.0 110.0 IT" 20.0 '0.0 i.o 20.0
ISI
110. 0
11"E ISI TlHE ISI TI"E ISI 11"E

~ ~ ~ ~

~
'""-
'"~
~
:z '" i
~.~
8 e 8 8

.
~ ct ~
a:
Ö .. 0 ~ ei

>-
~
:;: Ê :;:"
>-

'" ~ '" :;: ~

20 , ' 110.0 IT" 20 . 0 '0.0 fi.D 20.0 '0.0 IT" 20.0 110. 0
1 I"E ISJ 11HE ISJ TlHE ISJ 11"E ISJ

~ ~
'""- '""- '""- '""-
'" '"eg '""- '"cw
"
~!j-- ë'j-- .
u ~
~ ei
.. 0 "-
a:
a: ~ "
a:
'" 8 '" ~
:;;8 '" ~
~, ' 20 , ' 110.0 IT" 20.0
ISJ
1110 . 0 ~,' 20,'
TlHE ISJ
110.0 IT" 20 . 0
TlHE ISJ
110. 0
TI"E ISJ 11HE

FIGURE 6.51 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 6.52 ROLL ACCELERATl ON DOUBLET TEST CASE
 8r------------------, ~ r-----------------__, ~ r----------'I---------'
!
g~------~
' ----- --1 ~ ~ f----------,------------I ~ ~ 1-"---------------------1
~ 0
;;; '" 0 I
~ ~
~
'">><
I
~. L..---------:-:,,--:.•---------c:"o.o
TlHE 151
. -----------,,,="..,-.--------,-="
:.'-
TIH E 151
0. 0 :.!:-
. ---------,:.1"..:---------="..
TlHE IS)
:.!:-
. ----------;;,,!-:I.•' ,---------=\lO.O
TlH E 151

!.--------------------. - ~ r--------'--------, ~ .----'

I · -----,
I
I
"'
, I
!

~~ ~ V\ 0v I

~!r\r
I

i
1 I
,,-------------1
I,
==:::1:c:::===:j
: 0

~ ~

: L::::=
\ - >~ V ~
.•----------,-J
« :.'-.-'------ -- -- ,.,."""' IIO.O ~.o 20.0 110.0 8.0 20 . 0 110.0 ri.o! :---------::
" c:.•,---------=\lo.0
[IHE IS) T IHE 151 T I HE IS) T IHE 151

~o r---------~,----------, ~ .-------~-----.

!~I I~__ i~jv~

'
'" ,"'
,
"'
--__j -" 0

" 0
'" "'
·n
~ ~
~.':-.
V ~ J
----------="--:.•---------,,,J~o. o ~ ~."'.<----------:,.,.-•..,.-.- - - - - -- --llo.a ~ !:--------~, ~------~
TIHE 151 TIME IS)
l
e.o TI HE IS)
20.0 11 0.0

FIGURE 6. 53 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 6.54 ROLL ACCELERATION DOUBLET TEST CASE

i:~
:ë ~

'"cw
~ ~
'"
I L~
" ö
;;;
'",,
'"~
" ö
g
I
" 0 I
" I '">< '"
0.

~.o
I
>0.' 110. 0
"'
0.

~. o 20.0
I
110.0
~~
~ .. I
20.0 110.0
:l;

~ .. 1
20.0
Tl HE IS)
110.0
T IHE IS) T IH E IS) T IHE IS)

l! ~ ~ ~

, "
'"cw ~ I
~ I
! ~
'"

'"w
0
I

,
9'" g
" ö
;;;
~
I
L\;- JI
~
~
0

I
e8
" ö
"'co
I
I,
;:
i I "'
l!
~ .. 20.0
TlH E IS)
110.0
~
~ .. 20.0
T IHE IS)
\10.0
'" ~
ri.D 20.0
TI HE IS)
11 0 . 0
~
~ .. lO.D
T IHE 15 1
110. 0

~ ~ ~
~
I
,'" ~
~ ~ ,'" I '"
w
II
~ ~
'"
'"
0 9g
~ ö
;;; ;lj
'" .
I
, 98
"
;;;
Ö

!
0.
"'"
N
~ '"
V>
I
g
~ .. ilO.O
TIME IS)
110.0
~
~ .. >0.'
T IHE 151
11 0 .0
'" g
~ .. TIME
20 . 0
IS)
\10 . 0
~
~ .. I
>0.'
TI ME IS)
110. 0

FIGURE 6.55 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 6.56 ROLL ACCELERATION DOUBLET TEST CASE
 g ~

,;;;
~
~ ~ g
" C
~ Q
~ Ö

~
~ ..'"
~ x
~
.'"
'"
'" !!
.... 20.0
llME ISJ
"0. 0
'" C

~ .. 20.0
TIME ISJ
"0 . 0
~
.... 10.0
TIME IS)
1t0. 0
g
tr.a 20 . 0
llME IS)
"0 . 0

~
~
8 !!
,"'
"e
W

g
~eg
!: ö '" .;
"' ~
Ë '"
0
0
~

"' "' "'

lO.O
TIME ISJ
110 . 0 ....~ lO.O
llME IS)
110.0 20.0
llME IS)
110.0 ....~ 20.0
llHE 151
110 . 0

g ~
~ .;

0"
;
~ ei

'"
~
N
~

j-
~
~
e
~ ei

"'
'"
g -- !.(
~
W

N
C
W

~
"',
"e 8
a

'"'"
~

"''"
~

.. .... .... . ..
"' g "'
~
C g
~ 20.0 110.0 lO.O 1(0.0 ti.• 110 . 0 20.0 'tO . O
TIME ISJ TIME 151 llHE IS) llME 151

FIGURE 6.57 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 6.58 ROLL ACCELERATlON DOUBLET TEST CASE

g ~
;;; ;;.
,"' ..,
W ~ ~
r 0

'" .
'"
W
C
C

'"x
~

"'
."''"
~

g ~
~
.... 20.0
1 )ME ISJ
11 0.0 tr.G 20 . 0
TlHE ISJ
110 .0 f1.o .!O.O
llHE 151
!t0. 0 20.0
llME 151
"0. 0

~
~
e
'"
'"
W

;:
5l
g
C
.
~
N

r
g
~

g
r ~

r
"' 8
8
~

8
j
11 0 .0 . .. 20 . 0
TlHE IS)
110 . 0 · n:.O 20.0
TIME IS)
\10 . 0 W
.a 20.0
llHE IS)
1i0. 0

~ ~

,
o>

~
'"~ . .
8
!'
m

~
g
Q
J- ::!
~ci
"'!i
~ '!i
~
~g
"'~
ti.• ....
llHE 151
1l0.0 .... ....
llHE IS)
.0. 0 ..~. 20.0
llHE IS)
110.0
~
~. O ....
TIME IS)
110. 0

FIGURE 6.59 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 6.60 ROLL ACCELERATlON DOUBLET TEST CASE
 SFZ R/C IHISISJ sn R/C IHISISJ SF X R/C IHISISJ RZC R/C IH ISISJ AYC A/e IH/S/SI RXC R/C IM/S/SJ
?S .Da 0. 00 5. 00 fO ·oooo fO·01SD 0.0000 0. 0150 ~ .OO 0. 00 1.00 ?l.OO 0 . 00 loOD ?l.OO 0 .00 1. 00

Ol ç==s
....

"W "'~. I
"'~
~~ - ~~
. mo %~I
mo
. %~
1"Tl~

~: ~
--+-
.." § § .."
§ § §
;;:; ;;:;
c:: c::
'"'"
Ol
'"'"
Ol
Ol Ol
.....
VI

~
>C
0 . ~
. .
$! ;;0. 0
SR Ale
0.0
IOEG/SI
lO.D :".00
SO Ale (OE G/SJ
?l. OO
SP Ale IDEG/SJ
1.00
'"$! FZP Ale
~=I7.S 0.0
IH/S/SJ
n.S !=O.07S0
FYP Rle IH / S/S)
0.0000 0.0750 ~. I OO
FXP Ale IHISISJ
0 . 000 0. 100
..... .....
'".....
0

'"
.....
Ol I Ol c:::::::::::::: Ol ;==J
'".....en '"
en
.....
(") (")
". ".
en en
'" '"
"' o~
m
.. ~~ - - ---- ~~ %~
m~ I %~I ~~
§ § § § I·· -I m?
§ ·1 §

.
0
.. .. .. ..
SFZl H R/C IH ISISJ SF YTH Rle IHISISJ SFXTH R/e IHISISJ PSIR R/e IOEGJ THET RR R/e IOEGJ PHI R R/ C WEGJ
~; DO , 0.00
,
5. 00 ~ .O500 0.0500
~ i OO ,
0.00 1.00
I
~~OO.D 0.0
I ,
1000. 0 ?l.OO
,
0 . 00 1.00
I
;;,00 0. 0' 1.00

.... ....
~~I
.."
;;:;
c::
~~
§
~~
§
~~
§ .."
;;:;
c::
m~
§ \ I i~ 1
m~
§ I ~ ~I--
§

'"'"
Ol '"'"
Ol

~ Ol
N

-<
".
0 0 0
-<
".
~'
) ,
..
>C

~
..... ?l 1.S
SRTH Rle
0. 0
IOEG /S J
PI · OO
SoTH Rle
0 .00
IOEG /SJ
1.00 PI.OO
SPTH R/C
0.00
IOE G/SJ
1.00
'"~ ~s.o
RR R/C
0.0
IOEG/SJ
15.0 ?l . OO
OR Rle IOEG /SJ
0 . 00 1. 00 PI.OD
PR R/C IOEG/SJ
0.00 LOD
..... Ol --=:::::::::::: 1
0 0 0 0

'"..... '".....
'".....en '".....en
n (")
". ".
en en ....
'"
.-
"' o
m ~ %~
mo
. %~
mo
. '"
~~I
rn~ i~l
.
mo
-I ~~
§ § § § 11 § in

.. L - _ - . J . ._ _- '
 ~ ~
.;
~
.;
~

~
'!ö
~

"
V> '"
;;;
!
on on
x x
"-
~ x
~
~
>

"10. 0 20.0 "0 .0

TIME ISJ

;;;
V>
~
"- ;
'" ~
~
~,O 20.0 ~O.O
TIME ISJ

8 8 8r-----------------,
V>
,'" ,'"
V>

'!ö ."
, "
~ .;
8 '" 0
0 ;;;"
8
0
~ 8f-------------j
";;; '"

on
0

V> "-
N
'">
N
N

'" g
~

.;
Ir.• 20.0 "0.0 Ir.' 20.0 "10.0 Ir. • 20.0 ,,0 . 0 •.-;:.----~,-::!•.•
Ir.:-.------:,:;:-
TIME ISJ TIME ISJ TIME ISJ TIME ISJ

FIGURE 6.65 YAW RATE TEST CASE FIGURE 6.66 YAW RATE TEST CASE

~r------------ö ~r-------------------~
;;,
'"w .;
"
'"
V>

..."
;;; V
'"
"-
~L-_ _ _ _~_ _ _ _~
20.0 "0.0 20.0 \10.0 fi.D 20 . 0 1110.0
TIME ISJ TIME ISJ TI ME ISJ

~
9

;;;
;;;
Ë '" "'"
;= co
~ ~
lO.O "0.0 Ir.• 20.0 "10.0 fi.D 20 . 0 "0.0 20.0 '0.0
TIME ISJ TIME ISJ TIME ISJ TlHE ISJ

~
, ,on
~
V·

c' ,on
0
"'c
- 0 '!ö g
"
on
.;
" .; " .;
;;;
on '"
"- '"'" N
~
on
'"on
~. O 20.0
TIHE ISJ
\10 . 0 i.o 20.0
TIME ISJ
"10.0
g
~ .. 20.0
T IHE ISI
'0.0 ~.O lO.O
TIME ISJ
1110.0

FIGURE 6.67 YAW RATE TEST CASE FIGURE 6.68 YAW RATE TEST CASE
 SFZTH ERR IH/S /SI SFYTH ERR IH/S/SI SFXTH ERA [HISIS) SFZTH SIH IH/S/SI SFfT H SIH IH/S/SI SFXTH SIH IH/S/SI
~.00075 f't.OO
o r-
0.00 !='lS.OO
r-
0.00
0,
~.OD 0.00 1.00
0,
PI · OO 0.00 1.00

::u ::uu ':w

~

i~1
:~II I·
x~
~~ "' ? "'?
§ §
";;:;<= "
;;:;
<=
'"'" '"
'"
...."'" ën
U)

,.-< ;. ,.-<
'"
~
SRTH ERA IOEG /S J SUlH ERA IDEG /SJ SP1H ERR IO EGISJ '"~ SATH SIM IDEGISJ SDTH 5 I H IDEG/SI SP TH S IM IDEGISJ
~17 .s 0. 0 17.5 ~.OO 0.00 1.00 f'I.OD ~7 . 5 17.5 ~ .. oo ;;.00 0."0 1.00

,:F;l
--< 0.00 1. 00 --<

"'m
'"--< '"--<
'"
en
--< '"en--<
,.nen ,.nen
'" ~
'" ~

x.
m~
x~
"'0
;. - ~---- -
x~
m~
~~ "'~ · - -1 - mr
,, ~

_0

'" '" § § " § §

;. ;. ;. ;. ;. ;.

J RCK L 3 IHI JRCK L 2 IHI JRCKL I IHI SFZ ERR IH /S/S I sn ERR IH / S/SI SF X ERR IH/S/SI
0.000 0 . 500 ~.OO 0.00 1.00 ~.500 O.DJO 0.500 fU·OOOSO ~.I1S 0.000 0 .115
Ol
:=E
0

1.___

";;:;<=
'"
'"
'"j:j "W
"' ~
§o
x~
m~
§
~~
;;;
( I
";;:;
<=
'"
'"
'"
"
0
~~
§
;.
x~
"'0

§"I /' §

~
'"~
0

JAC KL 6 (MI
0

J RCK L 5 I HI JRCKL 4 IHI

~
'"~
" SR ERR IDE G/SI SD ERR IDEG /SI SP ERR IDEG / SI
--< ~ .500 0.000 O.SOO ;;>.500 0.000 0.500 ~.7S O.Ol
--< ~~ . o 0.0
, l O. D
I
~iOO 0.00 1.00
I
~.soo 0.000 0.500

'"--< '"--<
'"en--<
,.en
n

'"
~~
§
;.
~

~?
§
x~
m~
§
( '"en--<
,.nen
'"
;~I î i~ 1
m~
§.
-y I ~~
§ "

;.LI______-L______-' ;.LI______ ~~ ____- ' c

 8,-________________'
~,-----------------ï ~r_----------------~ ~,-----------------~
,
'n
'"cw ~
'", :;:
9 g I--____________________~
'" g
~ cr-----------------------, ~ 0~ r-----------------------~
~
~ c:i
~

~ ~ ~
~
a: .. !
:.~.-----------::..".•:--------·-:,c:!o.o fl.~.-----------:,=-•.-=.----------,-::•.• ;.~o-----------:,,=-.=-.-----------,',:0.0 ;.~o-----------:,-fol.-=-o-----------,,:-!o.•
TIHE 151 TlHE 151 TIHE 151 TIME ISJ

~,-----------------, ~r_----------------, ~r_----------------,

'"
~ ~
w
e
'" :
~ gr---------------------~
~ ei
~ ~ r------------------------j e~ 8 r-----------..,.-------------j
ei

u ~
~
w
~

cr
~
.. ~
0
;= '"
;. ~.-----------;:..=-.=-.-----------,',0:.0 ~.~o-----------:,:::-o-:.o~---------:::!
. o.o ~. ~o-----------::,,=-.=-.-----------"c:!o.o :.~o-----------:,:::-o.-:.-----------::.o. o
TIHE 151 TIME 151 TIHE 151 TIME 151

~ r--------------------. ~ ~----------------~

~
~ gr-----------------------~
~ 0
~
u ~
~
~
~
~
'"
~

:.L.o-----------:-":-.0"..----------'-'0,0 1r,L.o----------"',"0-:.o----------....Jijo.0 ~~. L.o-----------:,-:-o.-:-o----------,,.,.o. 0 ~. L.o----------.."-:-o.-:-o----------,.,.,,.o

TIHE 151 T1HE 151 TIME 151 1 IME 151

FIGURE 6.73 YAW RATE TEST CASE WITH 'A -.S FIGURE 6.74 YAW RATE TEST CASE WITH 'A - 's

~r-----------------------, ~ ,-----------------, ~,-----------------, ~r-------------------~

0:(l
~ g
u.; ------------------~
:;:'"
9 g 1-----------------------1
~ 0
'",,
'"
"
;:; g
~ ei
r------------------------l
~
u
,
~
~ 1------------------------1
0
~
;= %
x ~ x ~
~
~ 0
'" ~

'"
U1 ;.~o-----------=lO=-.: -O-----------,',:0.0 :.~.----------_:,=-•.-=.----------,::!•.• ;.~o-----------::,,=-.=-o-----------,',:0.0
'" 8 '---________-,.,-,-__________..j
~. o 20.0 '0. 0
llHE ISJ llHE 151 TIME 151 llHE 151

~r-----------------------, ~r-------------------,

~ '",
'" '"w ,

e~ gr------------------------j
ei
9'
u
~ 0
l
~ 1------------;------------1
!
~

~
;I
ri.~o-----------:lO=-.::-.-----------,.,:O. O :.~.-----------:,:::-o-:.•-----------:::!~o.o
TIME 151 llHE 151

~r-----------------~ ~r---------------------,
';'If~
~ I1 ~
w
c
- 0
'"~ ~
~ci - - -____--., ~ 0

...
~
::i ~ ;=
::; 8 r-------------------------j ::i
'" ~.~.-----------::,,=-.=-o-----------,',:0.0 o
~.L.o----------..,,"o-:
.•----------.,,-!
. o.o ti.L.o-----------:,,-:-."'"o----------,'-'•.o ~~.-----------:"-f~-=-o-----------"c:!o.o
TIHE ISJ T !HE 151 TIME 151 TIME 151

FIGURE 6.75 YAW RATE TEST CASE WITH 'A - 's FIGURE 6.76 YAW RATE TEST CASE WITH 'A • • s
 ~ ~-----------,
g
~ .------------------. ~ .--------------. ~ r------------,

0..
>< 0

L -_ _ _ __ _ ~------~
20 . 0 ~o.o
~~~------~~----~
8.0 20.0 "0 .0
lO.O toIO.o
TIME ISI TIME ISJ TI ME ISJ
lO.O "0.0
TI ME ISI

'">>- 0

L~-----:.,..,.----___,,~O. O 8. ~o------:,::O-:
.O-----;;t,jO. O
TI ME ISJ
~.0~----~~----~
8 20.0 110.0
~~----~~----~
t?O 20.0' "0 .0
TI ME ISI TI ME ISJ

~ .--------------. ~ .--------------. ~ r--------------- __,

,'" I
'",
'" 8~ __________________ ~
~
~ ei x ei
;:;; g~-----------~ ~ ~I-------------I

'" '">
N '"
~~
N

~. LO------:''''O.70-------''~O.O :.~O------:,:-:O.70-----=t,jO.O
TIME ISI
~.o 20 . 0
TI ME ISI
110.0 :~O------,~O-:.O-----:::!
t,jO. O
TIME ISJ TIM E ISJ

FI GURE 6.77 YAW RATE TEST CASE WITH ' A = 's FI GURE 6.78 YAW RATE TEST CASE WITH ' A = .s

!r-----------_~ ~ r------------------. ~ .--------------.

og
~
Q

" ó
;;;
><
:::; ~

toIO.O 20 . 0 "0.0 ri. ~O------:,:-:O.70-----,,.JO.o

TI ME ISJ TI ME ISJ "0.0

~ r------------~ ~ .------------------.

LI\ (
"

'"
0

'"
o ~
>-
0

!:---- ---:,,:-..,.O-------,,~O.O
TIME ISI
~.~O..L-----:,:::O-:.O-----,:::!O.O
TIME ISJ
~ ~LO------:,O.,...,.O-----.~IO.O
TI ME ISJ

8.-----------------,
,'"u ,
w ,'"
9 0 !: g
" ó '" ó
;;;
'"
0..
'"'"
'" g l - - -- - - - - - - - - - I
~. LO------:..:-..,.O-----.~O.O ~.~o------:,::-o-:.o-----,:::!o. o
TIME ISJ TIME ISJ
~.o 20.0
TIME ISJ
110.0 ~.o 20.0
TIM E ISI

FI GURE 6.79 YAW RATE TEST CASE WITH ' A = 's FIGURE 6.80 YAW RATE TEST CASE WITH ' A = 's
 ~ ~ I.; ~
~ ~
~I
~ ~
'" . ~
g
ö i ei cr: .;
('v -
§

'"
~

"
~
'"~
e;
~
..'" ~
~

§
'" ~
~ .. 20. 0
TIME (S(
'10.0
8
r.o TIME
20.0
(5)
"0.0 20. 0
TIME (5 )
'0.0 ,;. 20.0
TIME (5)
.0.0

~
.; ~ ~ ~
~ '"""-
~ ~

~~
9

'" . ~ ei
'"
8

'"
~
~ co
~~ '"" ~I '"
ti.D lO .O
TIME (5)
.. 0 . 0
~
~. TIME
10.0
(5)
"0.0 ri.D 20. 0
TIME (5)
1&0.0 20. 0
TIME (5)
1110.0

8 i
.;
0

Ii
v;

~ I~
'"""- '" ...'"
'" '"c
'" '"9
'"
;; 8 '" 0
;;; ei cr: .; cr: .;
~ ~
'"
~
N ~ cr:
~ ~i
~

'"
'" g
!i. 0 20.0
TIME (5)
'0 . 0
~

i.a 20.0
TIME (SI
"0 . 0 .
,; ....
TIME (5)
1110.0 ....
TIME (5)
.0,0

FJGURE 6,81 VAW RATE TEST CASE WJTH ~A • ~S FJGURE 6,82 VAW RATE TEST CASE WJTH .A •• S

~
~ ~
~ , '"""-
~ ~ '" ~
;;; g g - §
cr: •
~ è ::i • ~ö
u
'"
~

'" ~
"
~

~
'" ~
~ . ....
TIME (5)
'0 . 0
g
~.. TIME
20.0
(5)
!tO . G ....
TIME (5)
'0 .0
~
~ . 10.0
TIME (5)
1110, 0

! ~
§
.;

~ ...'"
'" '"9'" ~ ~
~I
cr: • 8
'" 0 ~ ei

'"
~

~
~~
/i.• ....
TIME (5)
'0.0
'" 8
r.G
TIME
20.0
(5)
1110 . 0
8
r.o 20.0
TIME (5)
'0 . 0 1110 . 0

~ §
.;

~
~ ~ ~
9

~
~
~ ~ cr:
"'§
,; . 20. 0
TIME (SI
1110.0
'"
TIME
10.0
(5)
• 0. 0 ....
TI"E (5)
1110. 0
§
,; . ....
TI"E (SI
1110.0

FJGURE 6.83 VAW RATE TEST CASE WJTH .A •• s FJGURE 6.84 VAW RATE TEST CASE WITH .A •• s
 g
~ ~
-
,
~
~ ~

~
'"wc g
'" g
.;
, ~
u
,u ~ ei
<r ~
<r
0

<r ~
u
X
"-
X "
"- ~
<r ~

~ ~ ~ ~
~. o 2<>.0 110,0 ~.o 20 .0 110 .0 ~.o lO.O 11 0.0 ~. o ZO. O 11 0 .0
r IH E 151 T IHE 151 T I HE 151 T IHE 151

~ ~

,
~

,
~ ~ ~ ~w
'" g ~ g ~ g c g
u .;
<r .;
~ o ~ ei <r
<r
,
<r
<r
u "- w <r
>-
<r
>-
~ ~ " g
g ~
~
~.o ".0 _0.0 ~.o 20. 0 110 .0 ~. o 20.0 110.0 ~.o 20 . 0 . 110. 0
TIHE 151 T IHE 151 T IHE 151 T IHE 151

~
~

~
~
,
,~
m

'!;
'"r g
8 ~
.; u.;
~
0
u <r ei
<r
'" ~
u
N
"-
N '""- '"or
<r
g ~
~ r-----, ~ g
W.o lO.O \10 . 0 ~.o 'zo.o 110.0 ~. o 20 . 0 11 0 . 0 ~.o 20. 0 101 0 . 0
T IHE 151 T IHE 15 1 T I HE IsJ T IHE 15 1

FIGURE 6.85 HEAVE ACCELERATlON TEST CASE FIGURE 6.86 HEAVE ACCELERATlON TEST CASE

g
~ ~

,in ,
m
~
~ .;
:;:
9 g
~ <:>
~
,u 0
~
~
,
u
'"r 0
<r
x '"
:J;
x
~
r
~

~
~ :J;
'" g
~ ~
~.o 20.0 110.0 ~.o 20.0 11 0. 0 ~. o 20.0 ,", 0 . 0 ~.o 20. 0 \1 0. 0
II HE 151 r IHE 151 T I HE 151 T IHE 151

~
- - m in

~ ~
~
'"cw
'"
'!; 9 g Uo
~
,u ~
u
;;: ei
~o ;;0 0

'">-
"m
t:
~
"
~

~
~
~
"
~

~
g ~ ~
~. o 2<>.0 110 . 0 ~. o 20.0 11 0.0 ~.o 20 . 0 110.0 ~.o 20.0 1,10 . 0
T I HE 151 T I HE 151 TI HE 151 T IHE 151

g
~ ~ ~

,
~
~

'",
.n ~ ~

,'"<ru
9
r eg g
g
.;
u ~
;;0
~
0
u ei
0

cc
Ë "or
~ "'-
N or ~
~ m
~ ~ ~

~
IJ)
tri

If.o 20 . 0 '0.0 ~.o 20 .0 110. 0 If. o 20. 0 110.0 ~. o 20.0 110.0

TlHE 151 T IHE 151 TIHE 15 1 T IHE 151

FIGURE 6.87 HEAVE ACCELERATlON TEST CASE FIGURE 6.88 HEAVE ACCELERATlON TEST CASE
 ~ ~ ~
U>

...
U> 0!: !i
'" g g
~ ~
~ ei !: ei U> 0 ~~
.."
U>
U>
U>
U>
'""
"'" ~
">-
~ ~ ~
~. o l O. O \10. 0 ~. o 20 . 0 ~O . 0 ~. o 20 . 0 ' 0. 0 ~. o 2U . O 110. 0
TI ME 151 TIME 15 1 TIME 151 TIME 151

~ ~ ~ ~
...
U>

...
U>

'" g
~
~ g
0
!: !i
~ ei '" .;
~ ~
U> 0 ~~
U> U>
U>
>-
Ë' E >-

'" ~ ~ ~ ~
~. o 20. 0 \10.0 ~. o 20.0 \1 0 . 0 ~. o 2<1. 0 '0. 0 ~. o 20. 0 \10.0
TIME 151 TIME 151 TIME 151 TIME 151

~ ~
.; - ~ .;
;;;

~,
...
...~

î'lVf
U>
~
r '"
~
0
- 0
'" g 8
! ei
'" .; ;; ei U> 0

..
U>
U>
U> N
U> N
N N >-
'" ~
~

~ ~
~.o 20.0 1f0.0 ~.o 20. 0 "0.0 ~.O 20 . 0 1110 .0 ~. o 20. 0 110. 0
TIM E 151 TI ME 15 1 TIME 151 TIME IS)

FIGURE 6.89 HEAVE ACCELERATION TEST CASE FIGURE 6.90 HEAVE ACCELERATION TEST CASE

~ g ~
~

... ...
~ ~g
U> U>

"''"9 g ~
'"ew g
! ~ '" .; r .; '" .;
U>

il: ."'
;;;
~"
U>
..
U>
g g
~ ~
~.o 20.0 110 . 0 ~.o l O. D 1110.0 ~. o 20. 0 \1 0 . 0 ~. O 20.0 1f0 . 0
TI ME 151 TIME IS ) lI "E IS) TIME IS)

g g
~

'"
'"
e
...
U>
~ ...
U>

'"9 g ~ g
'"
98
!U> ~ 0
'" .; ,-'" .; '" .; i
~
U> U>
U> >-
;= "
" i
~ U>
U>
~ ~ ~ 8
~. o 20. 0 .0.0 ~. o 20 . 0 \10.0 ~. o 20. 0 '0 . 0 ~. o 20 . 0 '0. 0
TI ME IS) TIM E IS ) TI ME IS) TIME IS)

8 ~ 8
~

'"'" ...
U>
~
'" "'"9 R ~ 8
ieli 8
'" 8
U> '" .; '" .;
r .;

.
;;;
"''"
U>
N
~
;;;
'"
U>
i
~ g U>
8 8 i
~. o 20. 0 \10.0 ~. o 20 . 0 1110 . 0 tr.o 2<1.0 110 . 0 ~. o 20.0 110.0
TIME IS) TIME IS) TIME IS) TIME 151

FIGURE 6.91 HEAVE ACCELERATION TEST CASE FIGURE 6,92 HEAVE ACCELERATION TEST CASE
 ~ ~ ~

~
x g
0 ~ ei
~ g
.;
~
eg
.;
on
'"~ '"~
X
X
U.
on
'"
'""- X
U.
.
on
g
~ .. 20.'
TIHE 151
110.0
'" ~
~ . 20.'
T IHE 151
"0.0
'" ~
~ .. lO.D
TIHE 151
110.0
~
~ .. 20.0
T IHE 15 1
110.0

~ ~ ~ ~

,'"
~
,'"
x
on
.;
~
9
~ ei
g
,'"
x

'"~
~
i
'"'"w
~
x
........ '"
'".... ....u.
u.
on "'" on "on
g
~ . 20.0
TlHE 151
110 . 0
~
~ .. 20.0
TlHE 151
110 .0
~
~ .. 20 . 0
TlHE 151
110 . 0
~
~ .. 20.0
TlHE 151
110.0

g ~ g
1<0
~

on
~

~~ ~
,'on. ~
x 9 w
g g c g
r
.; ~ ei .;
on
'" '"~ '"~
....N-"
on
~ N
U.
on '"on
g
~ .. 20.0
TlHE 151
110.0
'" ~
~ .. lO.D
TlHE 151
.0 .0 ..
1<0
~ >0.' IS)
TlHE
110.0
~
~ .. 20 .0
TlHE IS)
110 . 0

FIGURE 6.93 HEAVE ACCELERATION TEST CASE FIGURE 6.94 HEAVE ACCELERATION TEST CASE

~ ~ ~
.;

,on ,'" I
,on ~
~
x
r

'"'"w
I -~ ~

~
0 0 ~ 0
I ~
><
'-'
><
'-'
....x '"
-, Cl:
-,
X
u. '""-....
on

~ .. 20.0
TlHE IS)
110.0
'"
~ .. 20.0
TlHE 151
110 . 0
~
ri.D 20.0
TlHE 151
110.0 ~ .. 20.0
TlHE IS)
110.0

~
: ~ ~

~
~ I
~ I r x
9 g g

r
g on
'" ~
N
.; .;
~ '"'"w ~
><
x '-'
........ ~
u. '"....
on ~
g
~ .. 20 . 0
TlHE IS)
.0.0 ~ .. 20.0
TIME 151
~
11.0 ZO.O
TlHE 151
~
6'.0 20.0
TIME 151
~O .D

- ~ ~
.; ~

~
'"
g
.;
~
9
g
.;
~

'" !
r

"' g
.;
'"w '"~ ~
><
~
><
'-' '-'
~
u. '".... '"-, '"-,
on '"'" ~ ~
20.0
TlHE 151
"0.0
~
~ . 20 . 0
TlHE 151
,,0 .0 ~ .. 20.0
TlH E 151
"10.0 fi.o 20.0
TI ME IS)
110.0

FIGURE 6.95 HEAVE ACCELERATION TEST CASE FIGURE 6.96 HEAVE ACCELERATION TEST CASE
 .l ! . - - - - - - - - - - - - , ~ r---------------------,

~ ~~--------------------~ " ~ /------- -------- -------- t

'"
..'"'"
8L-____ __~~--------~
~. o ~o . o
11 0.0
~.O~---------:lO::"•..,.0------ ---:::!
. O. O

11HE 151 T ) HE 15 1

~
w '"
0
~
'"
110 . 0
TIHE
"'.0 151 lO.O
TIHE
2'0.0
151

g .~

'"'ew
0
~8
e
~ ~ " .;
'" ;;;
..
'" '"'"
~
~
' .0 20. 0 110 . 0 ".0 lO.0
1 ) HE 15 1
110 . 0

11 HE 151

FIGURE 6.97 RANDOM PITCH ANGLE AND 7TH ORDER PITCH/SURVE FILTER

8 ~

~ ~ ~
e w
~ g
8
~ ~ "
"
;:;; d
0
" .;
.; ;;;
'"
.
% .'"
'" ..
%
~
~
l! ~
~ ".0 20.0 110.0 ".0 20 . CI 110. 0

".0 ·1 )HE
20. 0 110. 0 ".0 lO . 0
11HE 151
'0 . 0
TI HE 151 T IHE 151
151

~
0
~
g
~
~10-- ~ ~ ~

~
'"'wc ~
e ~
e
I~
o
e o
;;;0
r 0

" ö
"
;;; 0
0 " .;
'" '"w
~
:g'"
Ë '"
0 '"
~ 0
;!:
~.O ~.O 20 . 0 110. 0
~. O 20.0 110.0
20. 0 110 . 0 T )HE 151
".0 TI HE
20. 0
151
110 . 0
TI HE 151 11HE 151

8
l!
~ ~
~
~
e .~
~
~ ~
e 8
~~
e 8 ~ ~ " ö
" .; '"
'" '"
.
;;; '"
~
~ '"'"
~
~
".0 20. 0 110. 0
~
".0 T I HE
lO . 0
15 1
110.0
~
".0 lO.0
TIHE 151
110.0 ". 20 . 0
T )HE 151
110.0

T )HE 151

FIGURE 6.99 SURGE ACCELERATION TEST CASE AND 7TH ORDER PITCH/SURGE FILTER
FIGURE 6.98 SURGE ACCELERATION TEST CASE AND 6TH ORDER PITCH/SURGE FILTER
7.0 COOROINATEO AOAPTIVE WASHOUT ALGORITHM
This section is based on the development presented in References 7.1
to 7.3. The intent of the algorithm is to adapt the severity of the
simulator washout filters according to the current state of the simulator.
In this way it should be possible to make full use of the simulator's motion
system at all times. The other consi derati on of course concerns how the
simulator pilot' will react to varying simulator motion-base dynamics.
Following the findings of Section 5 it was decided to apply the adaptive
washout to F1 components in order to avoid residual motion base
displacements. The axis systems were selected in the same manner as
described in Section 5 (i.e., FA = FC and FS = FO).

7.1 System Eguations

The general philosophy concerning the development of the system
equations follows that employed for the classical washout algorithm of
Section 5. The block diagram is given in Figure 7.1. This figure is almost
identical to Figures 5.1 and 5.2 if the HP SCALE and HP FILT #1 blocks are
removed. The mai n difference is that the ti lt-coordi nati on process now
takes place inside the block ADAPT FILT #2. Note that the],s input to the
translational channel of Figure 5.1 has now been moved downstream of the lIS
block and changed to ~I in Figure 7.1. The signals aC and ~ represent tne
inertial frame acceleration components and Euler angTëS which if applied to
the simulator frame FS will produce specific force and rotational velocity
components in FS identlcal to the corresponding FA components of the origin
of FA in the aircraft. From Figure 7.1

aC = !:..IS !..AA + ],1 (7.1 )

•
~ = Is .!!!AA (7.2)

From Equations 2.10, 2.12, 7.1 and 7.2 it follows that

(7.3)

7. 1
 (7.4 )

(7.5)

•
~C = PAA + (qAA sin4>S + rAA cos~s)tanes (7.6)

•
SC = qAA cos ~s - r AA sin ~S (7.7)

•
(7.8)

As with the optimal control al gorithm it is found in the present case

that a considerable reduction in the complexity of the formulation can be
achieved if the motion equations are separated and thèn dealt with as four
parallel sub-problems following Section 6.3.2. This is carried out in the
following subsections.

7.1.1 PitchjSurge
As in the optimal control formulation, the pitchjsurge problem is
assumed to deal primari ly with eS and a~I. Foll owi ng Reference 7.1 the two
adaptive filters shown in Figure 7.1 are taken to have the form

(7.9)

7.2
 • •
es = LIM(px2 aC x ) + Px3 ec (7.10)

•
where LIM( x) 1imits the magnitude of x to be 1ess than eLIM. The parameters
k xi are fixed while the parameters Pxi are continually adjusted in an
attempt to minimize the instantaneous value of the cost function

x ..x x·
This te~ds to make the simulator x-acceleration aSI (or SI) track aC and eS
track ec while restricting the actual motion generated due to the presence
of penalty weighting of s~, S~, êS and eS. Including aC x as an input in
Equation 7.10 allows for tllt-coordlnation similar in effect to the type
described in Section 5.2.2. The tilt-coordination takes place because the
aC x dependence on eS (see Equation 7.3) allows .the first term in J x (see
~quation 7.11) to be reduced by using changes in eS (tilting) while keeping
s~ and s~ and their associated penalties at modest levels. The tilt rate
llmiting included in Equation 7.10 (i.e., limiting on Px2aCX) was added to
the formul ati on of Reference 7.2 because it was found fhat thi s 1 imiti ng
could not be achieved satisfactorily otherwise. The penalties of the form
Wx ' (Pxi - PxiO) 2 in J x are incl uded in order to cause the parameters Pxi to
return to thelr original values PxiO af ter a maneuver is completed. If this
is not done it is found that the Pxi can drift to become unacceptable values
as time goes on. In fact in Reference 7.2 bounds are placed on the Pxi as
well.
As explained in Reference 7.1 the instantaneous application of the
continuous steepest descent technique leads to the use of

•
p . (7.12)
XJ

as a reasonable equation for the adaptive parameters where Gx ' is a

constant. It now becomes necessary to eval uate the right hand s1de of
Equation 7.12. From Equation 7.11 it follows that

7.3
 ··x •
OJ x x ··x :""C x oSI • • ) (OOC oas)
-- = (aC - sIH-ua- - --) + Wx1 (SC as -- -- -
OPxj OPxj Opxj Opxj OPxj

OP x 3
Px30) ~
ut'xJ
(7.13 )

In the present derivation it is assumed that

(7.14)

and that all the Pxj are independent. From Equation 7.3

(7.15)

where

(7.16 )

z
+ f cos <PS cos aS cos <Vs
AA

where it has been assumed to a first approximation that

(7.17)

and
(7.18)

7.4
From Equations 7.7 and 7.17

•
Me =0 (7.19)
Opxj
Now let
(7.20)

(7.21 )

•
(7.22)

(7.23)

•
(7.24)

Then in combination with Equation 7.16 the system equations become

•
( 7.25)

• (7.26 )
X2 = Px1 aCx - kx1 x 1 - kx2 x 2

•
x 1P = x (j = 1, 2, 3) (7.27)
xj 2Pxj

• °Px
x 2p . = __1 aCx + Px1 fn x aSPxj kx1 Xlp . - k x2 x 2P . (j = 1, 2, 3)
XJ OPxj XJ XJ (7.28)

• •
as = LIM(p x2aCX ) + Px3 ac (7.29)

7.5
 Op2 Op3·
= OUM·[_X_ aC x + Px2 fn x esp.] + _x_ ec (j = 1, 2, 3)
~Pxj XJ ~Pxj (7.30)

where
OUM = 1 for

=0 otherwi se

•
Pxj = -G xj [{(1 - pxl)aC X + kx1 x 1 + kx2 x 2 }{(1 - PXl)fnxeSPXj

(j = 1, 2, 3) (7.31)

Because all the Pxj are assumed to be independent it foll ows that

~Pxi = 0 i "f j
Opxj

=1 i =j (7.32)

Thus fran Equations 7.30 and 7.32

•
(7.33)

If we assume zero initial conditions then it follows from Equation 7.33

that

(7.34)

7.6
Thus Equation 7.30 need only be evaluated for j = 2, 3.

7. 1. 2 Ro 11 / Sway
The roll/sway case is solved in a manner similar to that employed for
the pitch/surge case in Section 7.1.1. The present formulation allows for
tilt-coordination. The filter equations are

(7.35)

• •
'1>S = LIM( Py2 acY ) + Py 3 '1>C (7.36)

•
where LlM(x) limits the magnitude of x to be less than '1>LlM. The cost
functi on is

The adaptive parameters obey

• OJ
pYJ· = -G·
YJ~
Y (7.38)
YJ

where from Equation 7.37

7.7
From Equation 7.4

(7.40)

(7.41)

where to a first approximation it is assumed that

oas = 0
(7.42)
OPyj
and
ö~S
--= 0 (7.43)
OPyj

From Equations 7.6 and 7.42

•
ö~ = (7.44)
Opyj
where
(7.45)

Now 1et
:: . Ö$S (7.46)
$Sp .
YJ OPyj

Y
Yl :: Sr (7.47)

•
Y2 :: Y1 (7.48)

7.8
 - ~Y 1
Yl p . = - - (7.49)
YJ ~yj

= ~Y 2
Y2p . - - -
YJ ~yj

•
= Ylp . (7.50)
YJ

Then in combination with Equations 7.41 and 7.45 the system equations
become
•
(7.51)

•
Y2 = Py1 acY - ky1 Y l - ky2 Y2 (7.52 )

•
Y lp . = Y2p . (j = 1, 2, 3) (7.53)
YJ YJ

(7.54 )

• •
<Ps = LIM( Py2 aCY) + Py 3 <pC (7.55)

• ~2 ~P3·
<PSp . = our~.[-Y- aCY + Py2 fny <PSp.] + -y- <pC
YJ ~Pyj YJ ~yj

+ Py3 hny <PSp . (j = 1, 2, 3) (7.56)

YJ

7.9
where
DUM =1 for

= 0 otherwise

•
Pyj = -G yj [y{(1 - pYl)acY + ky1Y 1 + ky2Y2}{(1 - Py l) fny ~SPyj

• •
+ Py {Wy2Y~ 2P yj + WyJY lY lPyj + Wy4 ~S~SPyj + WY5~S~SPyj}

(j = 1, 2, 3) (7.57)

Opyi
=0 i of j
Opyj

=1 i =j (7.58)

For zero initial conditions it follows from Equations 7.56 and 7.58

Thus Equation 7.56 need only be eval uated for j = 2, 3.

7.1. 3 Yaw

The yaw case is solved in a manner similar to that employed for the
pitch/surge case of Section 7.1.1. However no tilt-coordination is
involved. Note that the second and third terms on the right-hand side of
Equation 7.60 have been added to enhance the high-pass filter effect.

7.10
 • •
cjJs = PcjJ c\iC - kcjJ1 f cjJSdt - kcjJ2cjJS (7.60)

The cost function is

(7.61 )

The adaptive parameter obeys

(7.62 )

where from Equation 7.61

(7.63 )

To a first approximation it is assumed that

(7.64 )

and
(7.65 )

Thus from Equations 7.8, 7.64 and 7.65 it follows that

(7.66 )

Now 1et
(7.67)

•
(7.68)

7.11
 (7 .69)

(7.70)

The system equations become

•
(7.71)

• •
<Ji 2 = P<Jiq.C - k <Ji1 <Ji 1 - k <Ji2<Ji 2 (7.72)

•
(7.73)

• •
<Ji 2P <Ji = <j.C - k<Ji1 <Ji1P<Ji - k<Ji2<Ji 2P <Ji (7.74)

• ••• • •
P<Ji = -G<Ji[( q.c - <Ji2) (-<Ji2P <Ji) + P<Ji(W<Ji1 <Ji 2 <Ji 2P <Ji + W<Ji2 <Ji 2 <Ji 2P <Ji)

( 7.75)

7.1.4 Heave

The heave case is solved in a manner similar to that employed for the
pitchjsurge case of Section 7.1.1. However no tilt-coordination is
i nvol ved. The second term on the right-hand si de of Equati on 7.76 has been
added to enhance the high-pass filter effect.

(7. 76)

The cost function is

7.12
The adaptive parameter obeys

(7.78)

where from Equation 7.77

•• Z ·Z z
~Jz = .. z)(~aCz ~SI) ·Z ~S Z ~SI
(aC Z - SI - - - - + Pz (W ZI SI ~z + Wz2 SI -
~pz ~z ar>z ~z

+ Wz3 (pz - pzO) (7 .79)

To a first approximation it is assumed that

~cj>S =0 (7.80)
~z

(7.81)

Thus from Equation 7.5

(7.82)

Now 1et
z
Z 1 :: J SI dt (7.83)

•
(7.84)

•
(7.85)

(7. 86)

7.13
 (7.87)

(7.88 )

The system equations become

•
(7.89)

•
(7.90)

•
z3 = pzaC z - kz1 z l - kz2z 2 - kz3 z 3 (7.91) .

•
(7.92)

•
(7.93)

•
z3Pz = aC z - kZ1zlPz - kZ2z2Pz - kZ3z3Pz (7.94 )

•
Pz = -G z [{(l - pz)aC Z + kz1 z l + kz2 z 2 + kz3 z 3}{-aC Z + kz1zlPz

+ kZ2z2Pz + kZ3Z3Pz} + PZ(WZ1Z3Z3Pz + WZ2Z2Z2Pz) + Wz3 (pz - pZO)]

(7.95)

7.1.5 Including Vestibular Terms in J y

It is suggested in Reference 7.3 that in the case of roll response a

• •
y1 (4)C - 4>S)2 in Equation 7.37 is
better simulation js achieved if the term W
replaced by W OIPP.AI(PPA - PQS)2. This would tend to cause the vestibular
sensation of rbll in the simUlator to match that in the aircraft with the
•
IpPAI factor tending to enhance this match when there is a rapid sensation
change present in the aircraft. It is argued that it is at just such times

7.14
that a simulator motion-base can make the greatest contribution to fidelity.
However, as can be seen bel ow, this approach does add a considerable number
of differential equations to the formulation. The filter equations are

(7.96)

(7.97)

o
where UM(x) limits the magnitude of x to be less than 4>UM. The cost
function becomes

J y = O.5[y(aCY - S~)2 + WYOI~PAI(PPA - PPS)2

~y2 y2 02 2
+ py (\.Jy2 5 1 + W
y3 S1 + W
y4 4>S + W
y5 4>S)

The adaptive parameters obey

o OJ
Pyj = -G . -y- (7.99 )
YJ Opyj
where from Equation 7.98

7.15
From Equations 7.40 and 7.41

{7 .101)

where to a first approximation it is assumed that

(7.102)

and
(7.103 )

From Equations 7.44 and 7.45

{7 .104)

Let 4>Sp .J Yl' Y2J Yl p .J Y2p . be defined by Equations 7.46 to 7.50 and
let YJ YJ YJ

{7.105)

The vestibular res~onse is found using Equations 6.57 to 6.60 with ~•

replaced by pand w by xl" This results in (after first integrating
Equation 6.57 with respect to t):

•
Xl = -TlpX I Tjpx:l + T:lpp (7.106 )

•
x2 = Xl (7.107)

P = Xl {7.108)

7 . 16
Substituting Equation 7.108 into 7.106 and 7.107 leads to

(7.109)

•
(7.110)

where pis the input and p the output.

A
From Equati ons 2.13 and 2.14

• • (7.111)
PpS = $S - <)issi neS

and from Equations 7.102, 7.103 and 7.111

• (7.112)

Then in combination with Equations 7.41 and 7.45 the system equations become
Equati ons 7.51 to 7.56 and the simulator pilotls semicircular canal model is
(from Equations 7.109 and 7.110)

(7.113)

(7.114)

and the aircraft pilotls semicircular canal model ;s

• A
(7.115)
y,+ = PpA

•
A
(7.116)
PPA = - Tlp PPA - T3p y,+ + T2p PpA
and •
•
A
(j=1,2,3)
(PPS)p . = -T lp(PPS)p YJ. - T3p Y3Pyj + T2p $Sp YJ. (7.117)
YJ

7.17
 (j = 1, 2, 3) (7.118)

•
and Pyj is given by Equation 7.57 with

(7.119)

and with the addition of the following term to the right-hand side

•
(7.120)

For testing purposes it was necessary to develop a formulation of the

problem which excluded tilt-coordination in order to allow direct
comparisons to be made with the results reported in Reference 7.3. This was
achieved by setting

(7.121)

The lateral translation in this case is generated by using the fonn of the
heave channel of Section 7.1.4 with the appropriate changes in sub- and
superscripts. Here it is assumed that

aacY = 0 (7.122)
~Py

7.1.6 Including IpPAI(PPA - PPS)2 in J y

•
It was seen in Section 7.1.5 that the incorporation of PPA in J
introduced a considerable increase in the number of differential equatio~s
that had to be solved. For this reason it was decided to attempt the same
J formulation with PPA replaced by PpA. This removes all the additional
d1fferential equations yet still retains the feature of emphasizing the
fidel ity of the simul ator when sudden rapid roll rates occur. The desi red
formulation is achieved by replacing PPA by PpA and PPS by PpS in Section
7.1.5 and dropping all the equations related to the vestibular model,
Equations 7.113 to 7.118.

7.18
7.2 Test Runs

7.2.1 A1gorithm Configuration

The formu1ations of Sections 7.1.1 to 7.1.4 are solved simu1taneous1y

to produce a six degrees-of-freedom a1gorithm. The corresponding b10ck
diagram is given in Figure 7.1. The input is based on the motion of the
aircraft frame FA written in FA components. B10ck ADAPT FILT #1 represents
the three trans1ationa1 filters for surge, sway and heave. B10ck ADAPT FILT
#2 represents the three rotationa1 filters for pitch, ro11 and yaw inc1uding
the tilt-coordination for pitch/surge and roll/sway. The b10ck LIMIT is
used to represent the 1 imiting of the tilt-coordination rotationa1 rates to
be10w thresho1d levels. The adaptive parameters P which a1so feed into the
fil ter b10cks are formul ated by sol ving thei r corresponding differenti al
equations. ~Sl and is are then fed into the same actuator extension
algorithm emp10yea in Sections 5 and 6 and described in Appendix D.

A number of parameter val ues must be sel ected before the a1gorithm is
run. The vestibular model parameters are contained in Tables 3.1 and 3.2.
The other parameters whi ch are common to most test runs (unl ess otherwi se
specified) are either contained in Tables 5.1 and 7.1 or out1ined bel ow.
The geometric parameters represent the UTIAS F1ight Research Simulator. The
adaptive a1gorithm parameters are loosely based on those of Reference 7.2.

(1) The iterati on rate is 20 Hz, thus t:,t = 50 ms.

• •
(2) The rate limit on tilt-coordination is 9LIM = 4>LIM = 5.8°/s.

(3) The sensed motions are eva1uated at the 10cation of the captain's head
(seated in the 1eft-hand seat of the UTIAS F1ight Research
Simul ator) •

(4) The filter break frequency w2 used in estimating derivatives with a

first-order high-pass filter is w2 = 12.5 rls.

(5) The notation used in Figures 7.2 to 7.101 · is related to the notation
used in the text in Tab1e 5.2.

7.2.2 Inputs for the Vestibu1ar Response Mode1s

In the present case the axes employed are the same as in Section 5.
The ca1cu1ation of the inputs to the vestibular models is the same as in
•
Section 5.4 with the exception that ls is known exact1y and th us from
Equation 2.13

•
~s = B.s is (7.123)

• •
and an estimate of wSS (i .e., uf SS ) is obtained by passing ~S through the
filter of Equation 5.42.

7.19
7.2.3 Input Test Signals

The primary test signals for the adaptive algorithm are identical to
those described in Section 5.6.2.

An additional test signal has been included in this section in order

to all ow some cOl'1pari sons to be made with results presented in Reference
7.3. Thi s maneuver approximates that of a coordi nated turn in whi ch

(7.124)

(7.125)

with PAA as given in Figure 7.59. It is seen that to the extent that there
is no yawing in this test case, it only approximates the c~Täïnäted turn.

7.2.4 Results
'" '"
In the plots presentedin this section only those for !iA' .!.PS and .!.PS
are based on any approximations. In this case ~A and ~SS were used in
generating the estimated specific force signals fE2 pA and fE2pS.

The computer program used to test the adaptive algorithm is described

in Appendi x F. The test cases were run in a manner simil ar to that used in
Sections 5 and 6.

The fi rst case represents a longitudinal accel erati on of the ai rcraft

as shown in Figure 7.2. The complete set of results is presented in Figures
7.2 to 7.15. The aircraft pilotls sensation of motion is predicted in
Figures 7.4 and 7.5. The simulator motion resp~nse is plotted in Figures
7.6 to 7.8. The adaptive nature of the algorithm can be seen by comparing
x
the aSl of Figure 7.6 with that for the classical algorithm in Figure 5.24.
In the 1 atter case the fi rst and second accel erati on pul ses are of the same
magnitude whil e the adapt i ve al gorithm has attenuated the second pul se
significantly. In this instance
x the
. attenuation results in an unwanted
specific force spike in f pS at t = lOs for the adaptive case. The
successful washi ng out of the transl ati onal di spl acements i s evident in
Figure 7.7. The tilt-coordination and its rate limiting are shown in Figure
7.8. The simulator pilotls motion sensation is given in Figures 7.9 and
",x
7.10. The fpS plot in Figure 7.9 is se en to begin to rise quite rapidly
before the longitudinal travel of the simulator is used up. The sensation
error is shown in Figures 7.11 and 7.12. It is seen to be similar to that
for the classical algorithm shown in Figure 5.29 for example. The actuator
length responses of Figure 7.13 are a bit excessive. As seen from Figures
7.14 and 7.15 the only adaptive parameter showing a variation with time is

7.20
Px1' lts response after the initial transient drop is seen to be sluggish.
In fact it j ust happens to be near zero at t = lOs where the second input
transient occurs. Based on J x of Equation 7.11 and the data presented in
the figures one woul d not expect the adapt i ve al gorithm to attempt to behave
any differently at t = lOs than it did when t = Os since the only simul ator
state variable which is large at t = lOs is eSand from Table 7.1 it is seen
that its weighting Wx5 = O. This suggests that further tuning of the
algorithm should be carried out in an attempt to speed up its response. It
may al so be necessary to employ a nonzero val ue for Wx5 i norder to all ow
the algorithm to take tilt-coordination angular displacements into account.

The second case represents a roll acceleration doublet. The

corresponding plots are contained in Figures 7.16 to 7.29. The aircraft
motion is presented in Figures 7.16 and 7.17. The aircraft pilotls
predicted motion sensation is given in Figures 7.18 and 7.19. The simulator
motion is shown in Figures 7.20 to 7.22. The peak simulator roll ra te is
somewhat less than that in the aircraft but the final bank angles are almost
identical. The simulator pilotls motion sensation is predicted in Figures
7.23 and 7.24. The sensation errors shown in Figures 7.25 and 7.26 are less
than those found for the cl assi cal al gorithm (see Fi gures 5.41 and 5.42).
The actuator extensions of Figure 7.27 are a little excessive but are
comparabl e to those for the cl assical algorithm shown in Figure 5.43. As
seen in Figure 7.28 all of the adaptive parameters associated with the
roll/sway channel responded to the maneuver.

The third case is a yaw rate build-up and is presented in Figures 7.30
to 7.43. The aircraft motion is given in Figures 7.30 and 7.31. The
aircraft pilotls sensation of motion is predicted in Figures 7.32 and 7.33.
The simulator motions are shown in Figures 7.34 to 7.36. It is seen in
Figure 7.36 that <jJS is washed-out smoothly to zero. The predicted simulator
pilotls sensation of motion is plotted in Figures 7.37 and 7.38. From the
motion sensation error plots of Figures 7.39 and 7.40 it appears that the
rpS error is about the same as that experienced with the optimal algorithm
and shown in Figures 6.70 and 6.71. The actuator extensions of Figure 7.41
are reasonable. Onlya small response in P<jJ resulted, as seen in Figure
7.43.
The fourth test case (Figures 7.44 to 7.57) represents a vertical
acceleration pulse of the aircraft as shown in Figures 7.44 and 7.45. The
aircraft pilotls motion sensation is one of pure heave as shown in Figures
7.46 and 7.47. The simulator motion is plotted in Figures 7.48 to 7.50.
The simulator pilotls predicted motion sensation is given in Figures 7.51
and 7.52. The large sensation error relative to the aircraft pilotls
sensation is evident in Figures 7.53 and 7.54. It is quite impossible to
AZ
simulate sustained fpS with the limited motion-base travel available
al though the onset and decay of the stimul us are represented by modest
bumps. The actuator travel of Figure 7.55 is quite acceptable. The
variation of Pz with time is shown in Figure 7.57.

7.21
 7.2.5 Comparisons with Reference 7.3
In Reference 7.3 it was suggested that improved simulation of
coordinated turns could be achieved by using an adaptive algorithm
incorporating vestibular sensation as outlined in Section 7.1.5. In this
formulation the intention is to have the simulator produce accurate roll
motion cues whenever the roll motion cue in the aircraft is changing
rapidly. This of course means that large simulator motions wil 1 be
generated at the same time which could result in hitting the limits. The
vestibular responses are calculated for a head located at the origin of FA
and FS in order to duplicate the conditions used in Reference 7.3.
The input used was that described in Section 7.2.3 (an approximate
coordinated turn). The adaptive algorithm configuration used in the first
test run, designed to duplicate that of Reference 7.3, is the one described
in Section 7.1.5 which excluded tilt-coordination. In order to exactly
• •
dupl icate the reference case the signal ~C in ~C of Figure 7.1 was passed
through a first-order high-pass filter with the transfer function

TF = _s__ (7.126)
s + Wo

and then fed i nto the ADAPT FIL T #2 bl oek. Thi s was necessary because the
removal of the ti lt-coordi nati on al so removed some · of the hi gh-pass
filtering properties of the algorithm.
The roll channel parameter values are contained in Table 7.2. This
case is labelled as SC#l. Any roll/sway parameter not listed in Table 7.2
was taken to be zero. The corresponding lateral transl ational response is
found by using the heave algorithm of Section 7.1.4 with z replaced by y and
the z parameter values of Table 7.1. The pitch/surge, heave and yaw filter
equations have been deleted. The parameter values given in Table 7.2
correspond to those values listed in Reference 7.3 except for:
(i ) G 3 which was listed as 1.0. It was found that this value \'K)uld not
re~roduce the responses plotted in the reference.
(i i ) Py30 was not listed in the reference.
It was found that a much higher iteration rate than 20 Hz was required
to reproduce the rapid responses shown in Reference 7.3. The value used was
not specified but it was found that 100 Hz was sufficient.
The results of test run SC#l are gi ven in Fi gu res 7.58 to 7.71. The
aircraft motion is shown in Figures 7.58 and 7.59. The predicted aircraft
pilotls motion sensation is presented in Figures 7.60 and 7.61. The
simulator motion is given in Figures 7.62 to 7.64. The lateral motion of
the simul ator is an attempt to cancel the lateral specific force caused by
the bank angle ~S. The simulator pilotls predicted motion sensation is
plotted in Figures 7.65 and 7.66. The rotational sensation is quite good
and matches quite closely that of Figure 7a in Reference 7.3, but the

)
7.22
 "y
nonzero f pS is not acceptable. This is reflected in the sensation error
plots of Figures 7.67 and 7.68. The actuator extensions of Figure 7.69 are
quite reasonable. The adaptive parameter variations are shown in Figure
7.70. t-Jote the rapid response of Py3' This is largely responsible for the
excellent simulation of the negative PPS transient.
The second case run in this series, designated SC#2, was intended to
demonstrate the effects of including tilt-coordination and reduced iteration
rates on the results obtained in SC#1. The full adaptive algorithm was
employed as outlined in Sections 7.1.1,7.1.3,7.1.4 and 7.1.5 (roll/sway
.
including tilt-coordination). The high-pass filter represented by Equation
7.126 was included but it only acted on ~C. The parameter values presented
in Table 7.3 were selected by a trial and error process in an attempt to
duplicate the results of SC#l while still retaining a significant level of
tilt-coordination. The iteration ra te employed was 20 Hz. The vestibular
model used inside the algorithm now employs the (dimensional) parameter
values employed in the optimal control algorithm and were obtained from
Table 6.2. The value of Wo has been changed to 0.5 rls. The results are
plotted in Figures 7.72 to 7.85. From Figure 7.78 it is seen that the
simul ator ~S now returns to zero more rapidly which is a good "feature since
it quickly reduces the unwanted specific force sensation 1ps as seen in
Figure 7.79. However the PPS suffers as a result as shown in the same
figure. The less rapid variation of p 3 with time seen in Figure 7.84 is
partly due to the reduced iteration rate.Y It appears that additional tuning
of the algorithm is necessary in order to reduce the unwanted ~S.
As suggested in Section 7.1.6 it may be possible to replace P by P in
the formulation of Section 7.1.5 with little effect on system performance
and a savi ng in computer 1oad. To test thi s the roll channel was
implell1ented in isolation with no tilt-coordination (as in SC#l) and all
other channels deleted. The system parameters used were the same as in
Table 7.2 but with W 0 = 9257 to reflect the change from p in threshold
units to p in rls. (~ne threshold unit was taken to be 0.025 rls in this
instance, following Reference 7.3.) The vestibular parameters are not used
of course. The iteration rate was set at 100 Hz.
Thi s test case was 1abell ed SC#3 and the pertinent resul ts are shown
in Figures 7.86 and 7.87. From the plots of PPS in Figures 7.86 and 7.65 it
can be seen that this newly proposed approach 1S able to match the response
of the more complex formulation.
For comparison purposes case SC#4 was run with the coordinated turn
inputs using the standard form of the adaptive algorithm outlined in
Sections 7.1.1 to 7.1.4. The parameter values used are the ones contained
in Table 7.1 with the following changes made in order to improve the
response:

Gy 3 = 0.5

7.23
The iterati on rate was kept at 20 Hz. The results are plotted in Fi gures
7.88 to 7.101. The predicted simulator pilotls motion sensation is given in
Figure 7.95. When compared with the SC#1 results of Figure 7.65 it is seen
that the present case has a better ~s response but a poorer PPS response.

References
7.1 Parrish, R. V., Dieudonne, J. E., Bowles, R. L. and Martin, D. J. Jr.,
"Coord i nated Pdapti ve Washout for t-'oti on Simul ators" • Journalof
Aircraft, Vol. 12, No. 1, January 1975, pp. 44-50.
7.2 Parrish, R. V. and Martin, D. J. Jr., "Comparison of a Linear and a
Nonlinear Washout for Motion Simulators Utilizing Objective and
Subjective Data from CTOL Transport Landing Approaches". NASA TN
D-8157, June 1976.

7.3 Ariel, D. and Sivan, R., "False Cue Reduction in Moving Flight
Simulators". IEEE Transactions on System, Man, and Cybernetics, Vol.
SMC-14, No. 4, July/August 1984, pp. 665-671.

7.24
 Table 7.1
Adaptive Algorithm Parameters (SIU )

(i = x or y)

y = 1.0
Pi = 1.0 P<\I = 1.0 Pz = 1.0
Wil = 0.00929 W<\I1 = 0.0 Wz1 = 5.0
Wi2 = 0.2 W<\I2 = 1.0 Wz2 = 0.5
Wi3 = 5.0 W<\I3 = 1.0 Wz 3 = 0.3333
Wi4 = 1.0

WiS = 0.0
Wi 6 0.194
Wi7 = 17.26
Wi8 = 46.45
kil = 1.0 k<\ll = 0.09 kz1 = 0.5
ki2 = 2.828 k<\l2 = 0.4242 kz 2 = 3.0

kz 3 = 4.5
Gil = 0.517 G<\I = 1.0 Gz = 0.3
Gi2 = 0.00579
Gi3 = 0.0108
PilO = 1.0 P<fiO = 1.0 PzO = 1.0

Px20 = 0.164
Py 20 = -0.164

Pi 30 = 0.5

7.25
 Table 7.2
SC#! Adaptive Algorithm Parameters (SI U)

Y = 0.0

Py = 1.0

*W yO = 0.15

W
y1 = 0.0

y2 =
W 0.0

y3 =
W 0.0

y4 =
W 0.0

y5 =
W 1.0

Wy6 = 0.0

W
y7 = 0.0

W
y8 = 1.0

Gy 3 = 0.1

Py30 = 1.0

Wo = 0.7 rls

Tlp = 200.17

*T 2p = 7559
T3p = 34.0
A

*for p expressed in threshold units as in Reference 7.3.

7.26
 Table 7.3
SC#2 Adaptive Algorithm Parameters (SIU )

Y = 1.0

Px = 1.0 Py = 1.0 P</I = 1.0 Pz = 1.0

WyO =500.0
Wx1 = 0.00929 Wy1 0.0 W</I1 = 0.0 Wz1 5.0

Wx2 = 0.2 y 2 = 0.2

W W</I2 = 1.0 Wz 2 = 0.5
= 1.0 Wz 3 = 0.33~3
Wx3 = 5.0 y 3 = 5.0
W W</I3
Wx4 = 1.0 y4 = 0.0
W
Wx5 = 0.0 Wy 5 = 1.0
Wx6 = 0.194 W
y6 = 0.194
Wx7 = 17.26 W
y7 = 17.26
Wx8 = 46.45 Wy 8 = 1.0
kx1 = 1.0 ky1 = 1.0 k</ll = 0.09 kz1 = 0.5

kx2 = 2.828 ky 2 = 2.828 k</l2 = 0.4242 kz2 = 3.0

kz 3 = 4.5
Gx1 = 0.517 Gy1 = 0.517 G</I = 1.0 Gz = 0.3
Gx2 = 0.00579 Gy2 = 0.0

Gx3 = 0.0108 Gy 3 = 0.075

Pxl0 = 1.0 Py 10 = 1.0 P</IO = 1.0 PzO = 1.0

Px20 = 0.164 Py 20 = -0.0328

Px30 = 0.5 Py 30 = 1.0

Wo = 0.5 rls
Tlp = 10.16
T2p = 10.18

T3p = 1.64

7.27
 p
21
'r
f3 oe -,. ti
fA A
~

~ LISx
- -
--..
"
1'" - ""
ADAPT
FILT
#1
OS!
-
~
I
S2 -
... J,.I
..
L IS • p~
-
"
LIMIT
~" ~
ACCN
EST
...
~
Liest

~A A
...
~ Tsx
f3"C ... ADAPT ~S...
-
~

,.
FILT
#2
~ -I
S
ês
-.. LIS -
~

LIS

"
p

FIGURE 7.1 COORDINATED ADAPTlVE WASHOUT SCHEME

 ~ ~ I ~ ~
I ! I
! i ~ I ,
g g g
~ ~
~ ei I ~ ei co 0
~ ei
co
u "- co
x
co
g
~.. i
20.0
TIHE ISI
,"0.0
x
~
g
~. "
TlHE
I
.. IS)
'"0 .0
iE
~
~.. 20.0 ,"0.0
"-
~
~.. ....I ,"0. 0
TlHE IS) TIME ISI

g 8
~ ~

~ I
~
, ~
,'"'" I! t::
~ 8 !; g I ~
co
8
.;
I eg
~ ~ ei ~
u
ei
co ! co
~
j

co
ei

....I I..
>- :<
co
I " l! I
..
l!
~
T IHE ISI
,"0.0
I,L~
~.. 20.0
TlHE IS)
,"0.0
I- ~
~.. "
TIHE ISI
,"0. 0 ~.. 20.0
T IHE ISI
foIO.O

~
l! l! ~

~!:
,'"
~ ~..,
,'"
~
~ ei
g I
I
~ ei
Cl _
~ 0~
9 g I
co
I co
co
~ ei
co
u
N
co

..
I
"-
N
, :c co
'" l! i
..
~ ~
~
~
TlHE
20.0
ISI
f,j0.0 ~. o 20.0
TIME ISI
'0 . 0
~
~.. "
TIHE
.. IS)
foIO . O ~ 20.0
TIHE IS I
'0 . 0

FIGURE 7.2 LONGlTUDINAL ACCELERATlON TEST CASE FIGURE 7.3 LONGlTUDINAL ACCELERATlON TEST CASE

/r
k , ~r---------~---------'
! ~
t::
e
~~--------r--------'
I

r
~g _------t u ~ I------------:-----------j

I
;.~.----------::,,'-:.•,-----------::!.o.o
TIHE IS)
~ I
~.L.----------:-,,='.•-,---------:'
TIHE IS)
.0. 0 ~.!:-.---------..
::I-.•:----------::!'D.O
TIHE IS)
r I
:.L.--------~"L.•,----------:'
TIHE 151
.O. D

l!r------------------, ~ ~--------~-------, ~r--------------------' ~r--------------------'

~ ~
I I
i I i i I

r
~ 81-----------r------------1

!
:.~.----------::,,'-:.•,-----------::!.o.0
TI HE IS)
e 81-----------..,----------j

;' :.!:-.
!
----------;:,,~.•;----------::!.o.
TIHE 151
0
~I--------~i----------j

•
i!:-.---------.::.~
TIME
I
.•----------::!
ISI
.D.O
u~I----------r---------j

;:
~!:-.----------::
I
..~.•,-----------::!'O. D
TlHE IS)

~. r--------~-------, 8r --------,-------_, ~.-----r----, ~.---___rl----,

~ _. ~ . I
~ g 1-________+ __________1 e g I------------+----------j ~
I
I- ---------+-----------t
i I
u ~ 1-----------r------------1
~ ~ I
:.~. .~ i": ----~.. ~ ;~. ~I.,- - - - - -: !.D.O
Q ei

~ ~.L.1-----------:----------1
g
---------,,=.•---------::-!.o.o
~~ I
Ir.!:-.---------:,,~.•;----------::!.o. D _____ .. ---------::..
TlHE 151 T IHE 151 TIME ISI TIME 151

FIGURE 7.4 LONGITUDINAL ACCELERATlON TEST CASE FIGURE 7.5 LONGlTUDINAL ACCELERATlON TEST CASE
 ~ I .
g
! I
!l.;
0
~
'"-
g
~ 0 '{-V
t

I
'"
g
.;
r-\ i ~
~ ~
U> a
'vîJ
I
I
'"
~! ~
!
!
I

U>

..
U>

:l>
U>
x

I
U>
x
<r g
0'. 0
I
20 . 0 1110.0
x
... g
~.O 20.0
IS)
1110.0
§
~.O 20.0
I 1110.0
§
~.O 20.0
IS)
'0.0
llHE IS) llHE llH E IS) llHE

~ ~ ~ ~
t
~. II
~
!;
g I
,
U>

U>

~ 8
I

I
,
U>

!;
I
'"
~ ~
I
'" 8 I
'"
~ 0
I ~ ei
I u>a
U>

.....
U>
.;
I: I
U>

, ~
U>
~
> i U>
~

I
~
<r i
~ ~ ~ I ~
~.O 20.0 1110 .0 ~.O 20.0 1110.0 ~. O >0.0 1110 . 0 ~.O 20.0 1110.0
T1HE ISI llHE IS) T1H E IS) llHE IS)

,!~
! - I
:,;';IV~G
i
0 ,
,
U>
II
~ ,
U>
!; a i !; a

~ ei
g
V--v '"
-

'"
a T ---- ~~ ~ - ~
U>

U> .....
U>

~
.;
I
I
U>
~
I
I
~
I
I
~
> ~
I a: ~
~.O 20. 0
T1HE IS)
"0.0 ~.O 20.0
llHE IS)
"0.0 fi.o 20.0
T1HE IS)
"0.0
~
~.O 20.0
T1HE ISi
1110. 0

FIGURE 7.6 LONGITUDINAL ACCELERATION TEST CASE FIGURE 7.7 LONGITUDI NAL ACCELERATION TEST CASE

I g g , ~ g
-
I ~
!:~
<.> ,
U>
~
C ~
:;:
9 g - a
9 g
~ ~ '" '" '"
, U> .; .; .;
I
~ I
U>

~
I
U>

...
x
, ..;;;
U>

~
~.o >0.0 'l0.0
~
~.o >0.0
!
"0.0
U>
~
~. o >0.0
I
1110.0
~
~. O 20.0
I "0.0
I llHE IS) llHE IS) T1HE IS) llHE IS)
I

I
8 ~ 8
- - 11 -
n I
I I

~
9
I
I ~
I
,
U>

~
I ,
U>

<.>
~
I
~!t
98 _ !: g 9 g
! '"
~ ei
a
'" '" '"
( I
.; .; .;
;;; I
I <r
~
I

I
U>

... I
U>

"
U>

"
~

U>
"
U>

g g
I ~
~.O 20.0 .. 0 . 0 W
.O 20.0 1110.0 ~.O >0. 0 '0.0 W.O >0.0 1110.0
T1HE IS) llHE IS) T1HE IS) llHE IS)

~ ~ ~ ~
I I
<.>
~ ~ I, ,
U>
I
I :;:
~
C

~ ~
I eg ~ g 9 g
U> '" .; x: .;
I ZO .;

I
U> U> U>

~ U>
CE: I ...
~

U>
g
!
I
~
g
~ I ~
~.o 20.0 1110.0 ~. o 20.0 1110.0 ~.o 20.0 1110.0 ~.o 20. 0 ".0
T1'HE IS) llHE IS) I T1HE IS) _. . llHE IS)

FIGURE 7.8 LONGITUDINAL ACCELERATION TEST CASE FIGURE 7.9 LONGITUDINAL ACCELERATION TEST CASE

. -
,

-
 l!

g
.;
'"on
"
~
"-
on
; Lu--------~~--
20.0 ------~
"0.0
l!
~.u
TIME IS)

~r---------._------_. ~
.; ~r---------~--------__, g

,on r\
'"w ,
on
I

v-i
'" ~
g~--------~--------~
.; ~ g ~--------~------------I I
a: .;
::;
"
0 ~
on ~
on
~

TIME
".uIS) '0.0 ~. u
~
~':-u---------::>u:'-:.u:----------:-!.o. o
TIME I S)
8
~o
TIME " ..
I
IS)
,DO.

~.-------r------. l!
l!.---- ----.-- ------ -,
,on
'" ~
w
'"
'"
8
.;
I e'" 8 ~---------+--------____I
on
~
;:
a:
I ...
ö

l! :li
:.~u==::::...--"""'!'!
20 . 0 ~---~~
'0.0
on
~.u
'" I.u ~ ~
~.':-.------- --::
~
TIME IS) TIME IS)
..:'.•
-::--------- -:-!'o.o If.!:-.--------- ::":-:.•;-------- -.,,.}
.. o. 0
TIME IS) TIME IS)
FIGURE 7.10 LONGITUDINAL ACCELERATION TEST CASE
FIGURE 7.11 LONGITUDINAL ACCELERATION TEST CASE

~.---------,----------,
~r----------r--------__'
on

~
g ~------------------~ - g

..'"'.u;---------::!'o. 0
;,!:-u-------- -;;;
~

"e;
ei

:.L
I r ~
U--------- -:-,,"-.u:--------- c:'''o. 0 ~. ':-
TI ME IS) TIME IS) ... u '0.0 . - - - - - - - - - - :.: u---------: ;:!"o. 0
."=.
TIME IS) TIME IS)

~.--------------. :.---------------------, ~r---------~--------,

I ,
on

~1------:-I_----1
'"w
'"
I "
~
o

~. u
I
".u tn ~.LU-------- --:-,,:-.u:---------c :'''o.0 E
TIME IS) 20.0 ~LU----------::..~.u---------:;:!
.. o.o
TIME IS) TIME IS) TIME IS)

8.-____ ____~--------__,
~r---------~--------~

~
a:l!~----_+I------~
I §
gf------'t- --::;;......-------- ---t
5
r
(I)
::; 0
ei

~ I !!i V I
;.!:-u---- -----::..:-.u=---------:c:!'o.o
TIME IS)
:,!:-.---- -----::,,:-.•;----------- ;;;!"O.O
TIME IS)
:!:-.---------=..~.u:--------~.o.o
TIME IS)
~L-_______ _~_______ _~
ljI. a
TIME
20.0
IS)
'0 . 0

FIGURE 7.12 LONGITUDINAL ACCELERATION TEST CASE

FIGURE 7.13 LONGITUDINAL ACCELERAHON TEST CASE
 -

~
~
~ I
I II

I~
,I I I
_

~ ~
0
- 8
~ 0
- .
N g
.; - - -
I

:e
~ .. .... tO.O
8
tr.• 20. 0 \10 . 0
~
11'•• .... 110,0
I
TIME 'SJ TIME 'SI TIME 'SI

0
:<! .
8
I
I

t---
I

N
x
..
§.
0

ti.o .... "0.0

N
~ ei
8

~ .
--

20. 0
I
I
I!
I
TIME 'SI
'110. 0
.. .
;;; g
.. 0

g
ai.O
+--
20 .0
TIME
i
---

"0.0
I
I
I
i --- -

TIME 'SI 'SI

~
:<!
! !
i i J
I
I

~ ~
'" 0
-----+-- ---
~ ~
'" 0 - ---- - - - - - -
!
I ! I
~ .. 20.0
TIME 'SJ
110.0
:e
~ . 20.0
TIME 'SI
110.0

FIGURE 7.14 LONGJTUDINAL ACCELERATION TEST CASE FIGURE 7.15 LONGJTUDINAL ACCELERATION TEST CASE

_. -

-
 ~ ~ ~

~=g
~
, ~
~
iS
Ö
~
u
;i;
0
ei
/ i:iA
~
u
x
a:
ei
i .
a:
x
~
I .'"
~
I ~
ei

I
iS ~ I I
~.O 20 .0 "0. 0 ~.O ".0 110.0 ~.O 20.0 110 .0 ~.O 00.0 tO.O
TIME 151 T IHE 151 TlHE 151 TlHE 151

~ ~ ~
- I
,"'
,"'
i ~
,"'
'"9
w
I
! ~
!: g i '" -- - ~ 8 ~ g

~ ~I\
a: ó
~ ei
II I ~ I
a:
u
~
II ~
w
I
a:
ei

a:
~
~ ~ '"
~
"
~ ~
~.O 20.0 110.0 ~.O 00.0 \10.0 ~. O 20. 0 \10.0 ~.O 00.0 '0.0
TIHE 151 T IHE 151 T IHE 151 TIHE 151

~ ~ ~ ~
I
~ I ,"' '"cw ~
~ '"9w
~ g ~
-- , I iS
~ 0~
0

~ ei ~ ei
a:
~
a:
u . I
.."'
~
I
1 a:
Ö

~~
~. O 20.0 \10.0
N
~ 0
ó
~.O
f 20.0 110.0
~
~.O 20, 0 110.0
a:
~
~.o 20.0 \10.0
TIME 151 TIME 15J TI ME 151 TIME 151

FIGURE 7,16 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 7,17 ROLL ACCELERATION DOUBLET TEST CASE

i'~
~ ..

~
in
"', ,
~g "',
"
'"w
~

~ ei ,
-

~
0

ei
~~
a: 0
'"
~
0

ei

I
~~
I i i
::;
~
x
~

"' ~
~ I
~.o 20.0 "0.0 ~.O 20.0 \10.0 ~.O 20.0 \10 . 0 ~.D 20.0 \10.0
TIME 151 T IHE 151 TlHE 151 TlHE 151

~ ~ ~ ~

,"' "',, ,"'

"' ~ "' I
'"cw

~'~ ---- -
:;~ t
9 g
,, u ~
~ 0 I ~ ei ;;: 0

a:
~

""' ~ ~g I
"' ~ ~ "' ~
S=.O 20.0 110.0 ~.O 20.0 \10 . 0 b:.o 20.0 \10.0 ~.O 20.0 \10.0
T IHE 151 TIHE 151 T IHE 151 TIHE ISJ

g
~ iS
~
~
I
,"' "',
~ ,"'
~ '"~ "', I '"cw
!: g I - g " g
;:;
-
~ ~
;;;; ei

1 u ~

J-
,0

j-
ei ei a:
I I '"
N a:
~
N i=
~
"' ~ ~ :li
"' ~ \ ~ i ~
~.o 20. 0 .. 0.0 ~.o 00.0 '0.0 ~.o 20.0 \10.0 ~.O 20.0 110.0
TIME 151 TIHE 151 TlHE 151 TIHE 151

FIGURE 7,18 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 7.19 ROLL ACCELERATION DOUBLET TEST CASE
 ~ ~ I ~ l! I
, i I
:
,
V>
0
,
V> '", !O
zo
zo '"~ g
I

~ ~
zo 0
~ ö zo .; ;;; ei
'"
~
..
V>

x
i
I V>
x
>
i
I
~
iI
.
a: ... g g
l! ~
~.O 20.0 '10.0 ~.O 20 . 0 .. 0. 0 ~. O 20.0 '0.0 ~ 20.0 \60 . 0
TIME ISI TIME ISI TIME ISI TIME ISI

~ ~
§
!i ! .;
,
~~ g I
'",,
'" !I --
0
!O
~ I !

~:I\
--- §

"iV
r
q
~ ei
V>

V>
V
I
i
I §';I~ -'"
'"

>-
0
.
I
>- I
a:
u.. ~ ~ ~
0
~
!i.a 20.0 "0 .0 ~. o 20.0 '0.0 rio >0 • • tO.O W•• 20.0 '0 .0
TIME ISI TIME ISI TIME ISI TIME ISI

g
~

::~
0
~
.; !i

~ §) f\..
~ 0
V>

-
V>

~ ~
~!O
zo .;

.....
'"
N
0

J
!;~v ~
V>

V>
N
0

8
..
>
~
W.• 10.0 tO.O ~.o >0 •• '10.0 ~ 20.0 '10.0 W•• 20.0 '0.0
TIME ISI 0 ___ TIME ISI TIME ISI TIME ISI '

FIGURE 7.20 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 7.21 ROLL ACCElERATION DOUBLET TEST CASE

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

§]!\
r .; '-------------------1
o
V>

=.;g ~-----------------------~
" L~
zo .; - - - -----j
'"
:c
;;;
...x .
V>

V>

~LO----------~'.~.~.----------,~O .• ~. LO----------~,,~.~.----------,~ •.• '" ;.~,------------;:,,~.~o----------,=,., ~. ':-O-----------:,,.,.O ...,.O-----------,,:c!,.,

TIME ISI TIHE ISI TIME ISI T IHE ISI

g,--------------, ~r-----------------------, ~,---------------, ~r-----------~

~ ~
Lç'_
'"
'"e
w

V; ó8r-----------~----------~
r
g I-______________________~ e f------------r-----------I
g
~ci !
r .; zo 0

~
:
... --------------~
I
"'"
~~.----------~>O~.~.--------~,~•.• ~.~.----------~>O~.~.--------~.~•. 0 V> ~~. ~--------_::~----------=
o 20. 0 110. 0 :.~.-----------:,: -,.~O-----------,,=,.,
TIME ISI TI HE ISI TI ME ISI TIME ISI

~r-----------------~ ~r-----------------------, ~r-----------, 8r-----------------,

o
~
,
V>

'" ~ !
9 g r-___________i__________--t
~
I
~ f------------'------------j e 81------+! ____--I ~ 8

~:r_ --
r .;
" .;
'"
.
'" '"a:
V>

ei
:.L.----------~,.~.~.----------,~O.O :.L,----------~>o~.~,----------,~o.• ~., 20. 0 .. 0. 0 :.L,-----------:>o::-.~.----------,:c!,.,
TIME ISI TIME ISI TIME ISI TIME ISI

FIGURE 7.22 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 7.'13 ROLL ACCELERATION DOUBLET TEST CASE
 ~ ~
.-
, f'

~
~

,
0
~

,
~

"" "'w
c
~ 0
~ g =' g
~
;:;; Q
~

I~
c ó
r "'r
x ~ X Q.
Q.

~~
~ ~
~ ~
g g
Ir.o 20. 0 '0 . 0 ~. O 20.0 110 .0 Ir.O 20 . 0 11 0.0 ~.O 20.0 110 .0
T IHE IS) 11ME IS) TIME ISt. 1 1HE IS)

~
~ - ~

~
~ ~
9 ~
"~ ~
g e g
~ ei
~

r
c
~ ei

~
~
>- >- 0
~ ~ ~
~ ~
Sl ~ ~
S=.O 20.0 110. 0 Ir.O 20.0 "0. 0 Ir.O 20 . 0 110.0 Ir. O 20.0 "0.0
11 ME 151 11ME 151 11M E 151 TIME 15 1

~
~ ~

~
~

,
~

" 9
8
~
!: g
~
98
~ ei
~
~ ei '" .;
r
r
w ~
N N
~ or ~ :li
~
"' ~

~ g
20.0 "0.0 Ir. O 20. 0 "0.0 Ir.O 20 . 0 _0.0 Ir. O 20.0 "0.0
1 1ME 151 1 1HE 151 TI ME 15 1 TIME 151

FIGURE 7.24 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 7. 25 ROLL ACCELERATION DOUBLET TEST CASE

~
g g
~ ~.:

,
~
,"'
,
~

~
~
~
~
g - § ~ 8

r~
or
!5 ó ~
~
.;

1\
u
~
~
r
~
Q.
...,
'"
~
g
~
g g ~
Ir. O 20. 0 '0. 0 !i.a 20 .0 "1 0 . 0 ~.O 20.0 '0. 0 Ir.O 20.0 110.0
TIME 151 11 HE IS) TI ME 151 11 ME 151

~
-
,"'
j
!

~
"'cw
g
i "
~ g
~ ei ~ Ó

~ i= !!i
5l
~ ~ ~ ~
Ir. O 20 . 0 '0. 0 Ir.O 20.0 "0 . 0 ~. o 20.0 110.0 Ir. O 20.0 liO. O
TI ME 151 11 ME 15 1 TIME 151 11 HE 151

~
~
,"'
~ ~ " "
V
i~
9
-
IV
~ g
'" . a: ~ "' g .-
we ~ c ~ ei
r
~
N
r
~
!!i
~
~ :li
~ ~
Ir. o 2Q.0 li O. O Ir.O 20.0 "0. 0 20. 0 11 0.0 Ir. ' 20 .0 .0. 0
11 HE 151 11HE 151 TIME 151 TlHE 151

FI GURE 7. 26 ROLL ACCELERATION DOUBLET TEST CASE FIGURE 7.27 ROLL ACCELERATION DOUBLET TEST CASE
 ~r--------------------' ~r-------~--------~ ~r---------~--------~

:
.
.N 8
.;

~r-------~--------~ ~r-------~-----. ~r---------~--------~

~ ~I-------,-------l
.. 0
'" §
:: ct - ---l---- ----
I I
-'-
I
!

~. o '0 . 0

~.------------------, ~.-----------------,

LA-------I
'"
M 8• '" 8
.. 0 ;;: 0

~. !:-o--------:,:-o.:-o----~\lO. D :!:-o----~~~.o--------:,~o. o
TIHE ISI T IHE ISI

FIGURE 7.28 ROLL ACCELERATlON DOUBLET TEST CASE FIGURE 7.29 ROLL ACCELERATlON DOUBLET TEST CASE
 SFl Ale
!:jOO I 0."
'HISIS)
....
I
~.0600
SFY Ale 'HISIS)
fO·ono
SFX Ale ,h/S/SI
0.0000 0.0150
Ol
~.OO
Ale Ale
0.00
I
,HISIS)
1.00
I
~ . OO
Aye Ale
0. "
I
'HISIS)
1. 00
I
!=II.DO
Axe Ale ' hIS IS)
0 . 00 I."

?I °l 0

Ol

.... ....

."
äi
c::
~~
§
"'~
"'~
§
~T
~~
§ " ."
äi
c::
:~I -
1- -I i~1
"'0

§o - I "--~ I
ili~
§ "
'"
'" '"
'"
" "0
'"'" '"
-<
i:
0

SA Ale IOEG/S)
" 50 Ale IDEG /S)
" SP Ale mEG/S)
-<
""
'" " FIP Ale 'HISIS)
" FlP Ale '"/5/S)
" FXP Ale ''''/5/51
~ ~
.... ~.O PI.OO 0." ~ .OO 0." 1.00 .... PI'· s 0.0 11.1 ~ . OlSO 0.0000 O.07SQ !=U. lOCI 0 ....

'"
0 0 0

'"
.... Ol Ol ,----l
-<
'"
....
'" '"
....
'" I °l :--====i
n ~
""
'" '"'"
'" ~~
§°
~~
§
0
"'~
"'?
U> ° "'N"I
"'0

§
1-- ,- -··1 "'~I
m~ _

§ -I ~~
§°

~I
,,'--_---'L.-_ _......J
E~I~~--------~ ~LI------------~

SFZTH Ale 1"/5/51 SFYTH Ale 'HIS IS) SFXTH Ale 'HISIS) PSIA Ale IDEG) THETAA Ale 'DEG) PHIA Ale IOEG)
~iOOi 0.00 5.00 . ~ . 0500
Cli
0. 0000
i
0.0500
i
~.oo 0." I." fllOOO.O
°l
D.D 1000.0
I
fII.OO
°l
0.00
,
1.00 -
I 0,
PI . DO 0.00
, 1.00

~
.,.,
:~II - I "'NI
"'0
§o . ( I "'~
m~
§ ."
~:!t
~o
~~
_0

!:e
~~.­
§
c:: äi
c::
'"
'" il:
::"
i:: ~LI -'--_ _ _ _~
~IL----.L----l ~~I ______~______......J
...'" ;:LI_____--l...J ELI__ __ ~ ___l

~ SAYH Ale IDEG/S) SOYH Ale IDEG/S) SPT.H Ale IDEG/S)

~ AA Ale IDEG/S) aA Ale IDEG/S) PA Ale 'DEGISI

~ ~7.5 0.0 n ,.s ~ . oo 0." 1.00 :-.00 0.00 1.00

~ r· o 0. 0

- Z'S •• PI.DO D.DO 1.00

z
!=Il . GO 0 . 00 1. 00
a::::::::::: i 0l i 0, i
'"
~ E
~ .... ~
'" '"
~~I­ "'~
"'0 ~- -I-- - ~~ "'N
,..,~
"'N
"'? ~:
§o § " § § § §

"IL-____ _____ J
f~'------L------'
~~ ~L...I _ _--'-_ _- - '
 PSI SIM WEGI THETA SIM WEGJ PHI SIH IOEGI AlSI SI H IH/S/SI ArSI SIH IH/S/SI AXSI SIH IHISISJ
~o.o 0.0 50.0 ~iOO 0.00
, ,
1.00 ~.OO 0.00 1.00 ~iOO ,
0.00
.
1.00 ;;.00 0.00 1.00 ~.OO 0 . 00 1.00

.."
;;:;
c:::
'"
JTl

"
'"-<'"
3>
oE:
~~
~

RS 5 I H WEG/S J
~:w
QS SIH IOEGISI
··W
m~

~
_

PS SIH IOEG/S'
.."
;;:;
c:::
'"
JTl

'"
~

-<
3>
oE:
~~I
m"

FlP SI H IM ISISJ
~~I
m"

~
- ·1·

FYP SIM
---I

IH I S ISJ
~~,--
~

FXP StM 01/5/51

~ ~7.S 0.0 ~.DO ~ .OO 0.00 ~ p'7,S 0.0 17.5 ~.Oê!SO 0 . 0000 O.DZ5D fO·075D 0.0000 0.0750
-< 0.00 1.00 1.00
-<
Ol Ol l~---
JTl 0 0
JTl
-< -<
rsl
JTl
(I) JTl
-<
n
3>
(I)
-<
n
I
(I) 3>
JTl (I)
JTl
~~
m~
~
~~
mo

~
;"
~?
~
~~I
fT1~
~ I ~~I
m~
§ - I -I ~E ~--
~

~LI ______~______~ ;"LI______ ~ ______ ~

;"

,
SFl S IH IH I SISJ SFY SIH IH I SISI SFX SIH fM/SIS) vZSJ SIM 111 /51 vrSI SIH IH I SJ VXSI SIM IHISJ
;;;00 , . ;;;00 . . :, .00

r
~ . OO ~.OI7S 0 . 0000 0.0175 0.00 1.00 0.00 1.00 0 . 00 1.00
~~
0.00 5 . 00 0.0500
r- CC
:::::>
I
h
~~ ~~
mo
~~
m~
~~ ~?l ~~ ~._- -
;" ;"
.."
§O ~ ~ .." ~ ~ ~
;;:; ;;:;
c::: c:::
'"
JTl ~
" "'"
'"" ;"LI______~______~
IJl

: SR SIH WEG/SJ SQ SIH IOEGISI SP 5 I H IOEG/SI

: lSI SIH IHI YSI SIH IMJ XSI SIH IHI
~ ~ .

c:r
;;7.5 0.0 11.5 PI.OO
or-
0.00 ?I.DO
or-
0.00 1.00 ~oo ,
0.00 1.00 ~.oo LOO 1.00 ~.oo D. DO I.DO
JTl
-< ;;j
E
n
(I)
-<
~
~ (I)
JTl JTl

~:l:: ~~'---_ -I_- ___ ._

~~
m~I_· __ ~~ ~~
mo
--~----
~~
m~, _ __
;" ;" ;,
~ § ~ ~ ~ ~

;"IL______J -____ ;"LI______ ____ ______ ______

g,~----~------~
~ ~ ~
0~1 ~ ~

;" ;,, ~'------~------~

 sFZTH EAA IH/s/sJ sFYTH EAA IH/s /sJ sFXTH EAA .H/s/ sJ sFZTH 5 I H IH/s/sJ sFYTH 5 I H IH/s/sJ sFXTH 51H iHISISJ
~.oo 0.00 1.00 u. u u"" 0.0500 ~.OO 0.00 l.00 ~.DO 0.00 5 . 00 ~ . oo 0.00 1.00 ~.oo 0.00 1.00
I i

:~I :~ I \. " :~II %~ I

%~
~~
I· m~
~c

."
;;;
c:::
I I ~
;,
."
;;;
I ~ I I ~
:0
c:::
:0
'" '"
"o
..,
~ LI ______-L______ ~
~ ~ L
I -L ____________ ~ ~ LI ______- L______ ~ ~ L
I ______-L______ ~

~ ~
SRTH ERA IOEG /SJ saTH EAA IDEG/SJ SPT H ERA IOEGISJ SATH SIH IDEGISJ saTH 5 I H WEG/SJ SPTH 5 I H IOEG ISJ
~ 0.0 17 . 5 ;;.00 0.00 ~ ;;0.0 ;;.00 0 . 00 1. 00 ;; .00
-i ~.oo '.00 -i
0.00 1.00
::>
'"
-i '"
-i
'"~ ~
-i
,..n
(I) ~
(I)

'" '"
%~ I
m~
~ \ I
%~~ I
~c

~ I I :~I -' I I
%~I
~ c

~
~
r I :~ I -I I ~~
~
~

~ I • I ~ L -_ _--'-_ _- ' ~ IL- ______~______J

JAC Kl 3 IHI JRCKL 2 IHJ JACKl I iHI SFZ ERR I HISISJ SFY EAA IHISlsJ SFX ERA IHISISI
~ . 500
° l :;:>
D.SOO
I
~.600
Ol
0 . 000
c::
0.600
I
0.000 ;;.00 0.00 '.00 !I .L/SDO 0.0000 O.KOO 0.0000
'. j I

:~I-' ~ %~ I %~I
:~I
%~

/
% ~

."
;;;
m~
~
~c

~
;,
."
;;;
m~
~
\ L m~
~
c::: c:::
:0 :0

'" '"
".....
.., VI
W
~ L
' ______-L______ ~

;,
~ ~
'" J ACK L 6 IH ) JRCKL 5 IHJ JA CK l Ij {Hl SA EAA IDEGISJ s a ERR WEG /SJ SP ER R IDEG / SI

~
'"
Rl . SO D
ClI
0.000
CO
0 .600
i
' . :tuu
I
0.000
::;::>
0 . 500
I
!=(I.6OD
°l
0.000
c:
0.600
I ~ ~~.o 0.0
, ,
20. 0
;;:00 ,
0.00 '.00
, ;;,00
-i -i
~
~ -i

~ ~
(I)
(I)

'" '" ~

i~ 1
%~ I
~~

~
~/ - I :~ I - \-- - I ~?
~
~~
§o \ _. I :~I -- I ~~
~

o LI______ ~ ______ ~ c LI________ ~ ____ ~

~ LI------~------~
 ~
8 8
.; .;

, ,
~ ~
- 8
:L 0 ..
No
,;

8
&.• 20 . 0 110.0
8
1'.•
i
20.' 110.0
8
1'.•
i
20.0 \10.0
TJHE IS) TIHE IS) TIHE IS)

:: :
'"
~
i
~ ~ _. '" 80
~
..
. .. - - ;;; 8
t ct

~
~ .. 20.0
TlHE IS)
110 . 0 S••
'" 20.'
TIHE IS)
"0 . 0 ~.
'" 20.'
TIHE IS)
,0 . 0

'"
'"
x g. '" 8.
~
.. 0 .. 0

..
I.
~ I
~ 20.0 110.0 S•• 20.' 110.0
TlHE IS) '" T IHE IS)

FIGURE 7.42 YAW RATE TEST CASE FIGURE 7.43 YAW RATE TEST CASE
 g
~
-
,
~ ~
~

,
~

!: g g
~ 0 ~ Q
~ ~
'" . 9g
~ ei
U
><
'" ~ I
~
><
~

I
'"
f
~
..'"'" ~
~.o ilO.O "0. 0 ~.O >0.0 tO.a ~.o 20 . 0 \10.0 ~. o 010.0 "0 . 0
T IHE 15) T IHE 15) T IHE 15) TIHE 15)

g
~ ~ 8

,
~

,
~
~

~
~ ,
~

'"w
!: g :!O g I ~8 e 8
~ ei ~ö '" à ~ ei

'" '"
,..
u

'" ~ g
I ~~ '"~
~
~.o " .0 .. 0.0 ~.o 20.0 "0.0 ~.o 20.0 110.0 ~.o 20.0 110 . 0
TIHE 15) TIHE 15) TIHE 15) T IHE 15)

g
.; ~ ~

, ,
~ ~
~ ~

, ,
-" .
~ ~

!: g 9'" g
~ ei ~ . ~ ~
'" . ~ ei
'"~ ~
u
N

'" g ~ . ~
..
~
'"'"
à ~ ~
W.G 20 . 0 110 . 0 ~.o 20.0 '0.0 ~.o 20. 0 "0.0 ~.o 20.0 "0.0
TIHE 15) TlHE 15) T IHE 15) TIHE 15)

FIGURE 7.44 VERTlCAL ACCELERATlON TEST CASE FIGURE 7.45 VERTlCAL ACCELERATlON TEST CASE

g
~ ~ ~

'", '", ~

'"~ g
~
eg
~
~ g
'"<>w
g
~ ei ~ ei
~ Cl ~ Cl

'"><
~
~
~
"
~
x
~
"
~
~
~
~
~
g 11'1
~
~. o 20 . 0 110.0 ~. o 20 . 0 "0.0 ~.o 20.0 110.0 ~.o 20. 0 "0.0
T IHE 15) T IHE 15) T IHE 15) T IHE 15)

~ ~ ~ ~
!
, I ~
~
~
'", ,
~
~

,
~

:!O g
"9w I "~ I
~ ei ~ ci
g

I
'" .~ u
'
'"
.
~

,..'"
~ ~
",..
~

~
! ~
~
~

~ 1/'1
~
~.o 20.0 ~.o 20.0 110.0 ~.o 20.0 110.0 ~.o 20.0 .0 . 0
TIHE 15) T IHE 15) T IH E 15) T IHE 15)

8
~ ~ ~

, ,
~
,
~
~
'", ,
~
'"ew
~ ~
"
!: g
~ ö
eg
~ ei
;:; g
;:r ei
-+ u
~ .
~

"
h- rL- "
~
N ~
~ '"on N
~ or
on
~ ~
'" ~ ~
~. o 20.0 tD.O ~. o 20.0 110.0 ~.o 20. 0 ,0.0 ~. o 20.0 "0.0
TIHE 15) T IHE 15) TIHE 15) T IHE 15)

FIGURE 7.46 VERTlCAL ACCELERATlON TEST CASE FIGURE 7.47 VERTlCAL ACCELERATlON TEST CASE
 ~ ~ ~ ~

~
~ ~
~ g
g 8
~ ei
:!:
.; '" g
~ ö ~~
~
x '"x
'"x
'" ..."-x g
>
~ ~ ~
~.O 20.0 \10 .0 ~. O 20.0 110 . 0 ~.O 20 . 0 1110. 0 ~. O 20.0 "0.0
TIME ISt TIME IS) TIME IS) TIME ISt

g
~ ~ ~

,'" ,
~
,'"
,
~

,
~
g
'" g ~ g
'"
'" g ;;; ~
'" 0
~ ei x.; ~ ö
~
~

~
"-
~
~ '"
~

'" ...
~ >
g
~
~.O 20.0 110 . 0 ~.O 20.0 110.0 ~.O 20.0 110.0 ~.O 20.0 \10.0
TIME IS) TIME ISt TI ME IS t TIME ISt

,
,'"
~

'"
~ ei

'"
~
~

'"
~

g
1'.•
IV
1 20 . 0 110. 0
L
'"
..."-
~
-

.;

~. O 20.0 "0.0
'",
'"
'" g
;;; ö

'">
~
!

~
~. O
ILv 20 .0 \10 .0
;:rv
-
~

~
~
~
0

~
~. O
.

20.0 \10.0
TIME IS ) TIME ISt TIME ISt TIME IS) ·

FIGURE 7.48 VERTICAL ACCELERATION TEST CASE FIGURE 7.49 VERTI CAL ACCELERATION TEST CASE

g g
~ ~

~g
,
~
G ~

co
'" ~
S' g ~ g
~ ~ '" .; '" .; '" .;
~

'"
"-
~
"- .~ "-
~

~
~ .. TIME
20 . 0
IS)
~O.O
g
~ .. 20.0
TIME IS)
\10.0
~
~ .. 20.0
TIME IS)
\10 . 0
~
~ .. 20 . 0
TIME IS)
toIO .D

~ ~ ~ ~

~
, ,
~
~ ~

,
~

x g '"!:!'" g !; g !:! g I
;;; ö x.; x.; x.;

~
'"'"
~
~

~
<> ...
~

~
<>
~
I
~ ~ ~ ~
~.o 20.0 '0 .0 ~.o 20.0 11 0 . 0 ~. o 20. 0 \10.0 ~.o 20.0 110 . 0
TIME IS ) T IHE IS) TIME IS) TIME IS)

~ ~ ~

'"'" ,
~

~ ,
~

'" ~
!:!
~~
'"
!:! 8
.;
~ 8 9 g !
x x.; x.;
~
~ ;;:; ;;:;
~
...
::?
I
~
"- '"
~

~
~.. TIME
20 . 0
IS)
\100. 0
~
~.o 20 . 0
TIME ISt
110.0
'" g
1V. 0 20.0
TIME IS)
\100 . 0
g
a=.O ZO.O
TIME IS)
'0.0

FIGURE 7.50 VERTICAL ACCELERATION TEST CASE FIGURE 7.51 VERTICAL ACCELERATION TEST CASE
 g
~

'",
'"
'"'"
i g ~~
'"
'"w
9 ~
;r; 8c .;
= '"w'" '"'"w
~~ ~ x ..
'"
~ ~
~ .. ilO.O
T1HE ISJ
110.0
Cl) ~
~ .. 20.0
T1HE ISJ
\10.0 ~ .. 20.0
T IHE ISJ
110.0
~
~. 20.0
T1HE ISJ
110 . 0

~ ~ ~

'", ,'"
~
:!:
g
'"w
'" 8
~
~ g
'"
'"9w g
'" .; ~ .; .; .;
'"
">- '" '"'"w ~
~
"
~ >-
'"'"
'"'"
~ ~

'"
~ .. 20 .0
T1HE ISJ
110.0 ~ .. 20.0
T1HE ISJ
\10.0
'"
~ .. 20.0
T1HE ISJ
110.0
g
~. 20.0
TIHE ISJ
110.0

~ 8 ~
"
,
,'"'" ~ ,'" ,'"
w
,'" :;:
:!:
g '" g " ~ 9
" .; ~ ei ~
'"
" ~ '"~
~
;:
'"'"
N N
~
:J; ~

'" g
~ .. 20.0
llHE ISJ
IIQ.O ~ . 20.0
T IHE 151
110 . 0
'" g
~.O 20.0
T1HE 151
110.0
~
0'.• ii!0.0
TIHE ISJ
110.0

FIGURE 7.52 VERTlCAL ACCELERATlON TEST CASE FIGURE 7.53 VERTlCAL ACCELERATlON TEST CASE

~
~ ~ .; !
0" '"
'"9w §
'"
g0
~ g g - g ~

~ ei

Eg
'"
::;
w

..'""
~
0

~
~
a:
--,
0

~
r ~
0'.' 20.0 0'.' 20 . 0 110.0 fi.o 20.0 110.0 ri.• ilO.O 110.0
T IHE ISJ T1HE 151 T IHE ISJ TIME 151

~
~ ~ ! .;

,'" '",:;:
~ §
9 '"
"
;; g
0

r
8 '" g '" 11
ffi ei I '" .
~ 0 ~ 0
;: !!i
"
~
~

'"'" ~
8
~.. 2<1.'
T1HE ISJ
110.0 0'.' 20.0
T IHE ISJ
110.0
~
ri.a 20.0
T IHE 151
\10 .0
~
ri.a 20 . 0
TIME 151
1110 . 0

g
~ ~ ~
.;
"
~
,'"
'"w §
"
;; g '" g I (Y) g I.D g
'" .
WO '"~ . 0 5 ei ~ ei

"
~
N
~
~~
I !!i !!i
'" g ~ ~
~. o 2<1.' 110.0 0'.' 20.0 110.0 ri.• zo.o 110.0 ri.• 20.0 110.0
TIHE 151 T IHE ISJ T1HE ISJ T IHE 151

FIGURE 7.54 VERTlCAL ACCELERATlON TEST CASE FIGURE 7.55 VERTlCAL ACCELERATlON TEST CASE
 ~r--------,--------, ~r-------~--------~ ~~-------r--------.

:.~.-----::..L.•:-----..,,;;!•.• 8L-______~~------~
·;.0 20.0 1110.0
~L-________~________~
lr.O ZO.D 110.0
TIME IS) TIME IS) TIME IS )

~ r--------------, ~ r---------~------~ ~r----------------- __,

N gl-------'--------1
~ ë f ..~ of------"=-+---=---j
~

;.L.------".-.
TIME
.,-.------"
IS)
_O,O ~.L.------:,7•.-:0-----.".}~O.O
TI ME IS)
8~

8'.0
__________________~
20.0 11 0.0
TIME IS)

~ r-------~---------. ~ r---------------,

~.L.-----.",.-.•,.------,,-J•.•
TIME IS)
..
~ zo.o
TIME IS )
11 0.0

FIGURE 7.56 VERTICAL ACCELERATION TEST CASE

FIGURE 7.57 VERTICAL ACCELERATION TEST CASE
 ~ ~ -

~
g
ei ~
g
ei
~
9
u
;r
0
ei
V ~ ei
n
.."
« «
u ~ «
~
'"
« '"
~

~
~.o '.0 lil.a ~.o •. 0 Il.O ~.o '.0 ~.o '.0 12.0
T IHE IS) T IHE IS) TIHE IS) T IHE IS)

~ ~ ~

~
~
'"
'"
~ 8
~
!: g I
9
~ g
~
!: g i
----'---- « .;
~
I
~ I I ~
U
~
«
ei

~
~
~
ei
I

I
~
w
;: g;
ei
I
g ~ ~
~.o '.0 12.0 ~.o '.0 12.0 ~.o '.0 12.0 ~.o '.0 12.0
TIHE IS) T IH E IS) T IHE IS) T IHE IS)

g ~
.- ~

L '",
Cl
W

Li\
9 Cl
w
u 0 9 g
~ ~ ei
~ ei
~ ei
i
0
«
u ~ «
N ~
N

'" g
~
'"
~ ~
~.o '.0 12.0 ~. o ' .0 12.0 ~.o •. 0 12.0 ~. o ' .0 12.0
T IHE IS ) T IHE IS) T IHE IS) TIHE IS)

FIGURE 7.58 COORDINATED TURN, SC #1 FIGURE 7.59 COORDINATED TURN, SC #1

~{~ i
'",
'"~
«
x
g
~ ei
'"
Cl
w
9
~
«
~
0

ei
L----!
,
,'"
,'"
r

~ ~
'"

~
0
!:!\
,

'"
"
~
0

'" ~
'" ~

'" g
~

'" ~
~.o '.0 12.0 ~.o '.0 12.0 ~.O '.0 12.0 ~.o •. 0 12.0
T IHE IS) T IHE IS) T IHE IS) TIHE IS)

~ ~ ~ ~
I I
0 ,'" '",, :;:'"
'"
Cl
w
'"
!: g I 9 g '"
~ ~
9
u ~
~ ei ~ 0
I « 0
! ,
a:
0

"
~
~
~ co ~
~

'" ~
'" ~
~

'" ~ '" ~
~.o '.0 I l.0 ~.o '.0 12.0 ~.O •. 0 12.0 ~.O ' .0 12 . 0
TIHE IS) TIHE IS) TIHE IS) TIHE IS)

g g g g
~ ~
I
0 0
~ i
~
!: g
~ ei
-lI :;:
9 g
~ ei
~ ~
'"

"
0
u~
,
a:
0
I

N "
~

'"'"
N

~ Sr---.... g
~
on g r--.. '"
11)

~
!i.o '.0 12 . 0 ~.O •. 0 12.0 If.o '.0 12.0 ~.O '.0 12.0
TIHE IS) TIHE IS) TIHE IS) TIHE IS)

FIGURE 7.60 COORDINATED TURN, SC #1 FIGURE 7.61 COORDINATED TURN, SC #1

 ~ ~ ~
-
,"' ~ ,"'
,"' g
"' '"
'" " g
~ .; : .; '" .; ~ ~
'" 0

'"
;;; "'
"- "'>
X "'
X

"'
X X
~

'" ~
g ~
~.o '. 0 12.0 ~.o 6.0 12.0 ~. o '.0 12.0 ~.o 6.0 12.0
T IHE ISJ I IHE ISJ TIH E ISJ TlHE ISJ

!!.;
~

~
"',,
"' ~ g
'" ~ -~----I _._._-- '" g
~ ~
'"
"'"- "'>
~ "'
~

"'
~ ~

'" g ~
$ ~ !!
ai. 0 '.0 12.0 /i.a 12.0 fi.D '.0 12.0 rio '.0 12.0
T IHE ISJ TlHE ISJ T IH E ISJ

g g
~
~

"',
"''.
'"
"'
"'zo ~ '"
zo g
~ zo
0
0 '" ",0

;;; "'
"- "'>
N
"'
N

"'
N
~

'"
~.o '.0 12.0 ~.o 6.0 12 . 0 ~. o 6.0 12 . 0 ~.o 6.0 12.0
TIHE ISJ TlHE ISJ TIH E ISJ T IHE ISJ

FIGURE 7.62 COORD I NATED TURN, SC #1 FIGURE 7.63 COORDINATED TURN, SC #1

r
~

in
~
~~
<;
"' ~
0
0
"'zo g
w
0
0

'" 0
~
zo 0
zo
.;
"' ;;;
I X "-
"- "'"- "'
"' g
~.o 6.0 12.0 ~.o '.0 12.0 ~.o '.0 12.0 ~.o '.0 12.0
T IHE ISJ T IHE ISJ T IHE ISJ T IHE ISJ

g
- '"
,"' in in
~ ~
w "', ~
w
~ g
'" g !? g
'" ~ 0
~
.;
zo 0

"' '"
'"
~ "' "'
"''" "''"
W ~
I

"' g g
~.o '.0 12.0 ~.o ' .0 12.0 ai. 0 '.0 12.0 ~. o '.0 12.0
T IHE ISJ T IHE ISJ T IHE ISJ TIME ISJ

g g
~

~ ~
<;
w
,"'
~
0 w
" g '" 9 ~
'" .;
zo 0 0

'"
"' ;;; '" ;;;
"'"- l2 N
~
"''"
~ "' g ~
~.o '.0 12.0 ~.o 6.0 12.0 ~. o ' .0 12.0 ~.o '.0 12 . 0
TIM E ISJ TIME IS J TIME ISJ T I HE ISJ

FIGURE 7.64 COORDINATED TURN, SC #l FIGURE 7 .65 COORDINATED TURN, SC #1

 SFZTH ERR IH / S/ S) SFYTH ERR IH / S/ S) SFXTH ERA ' tt / S/SJ SFZlH SIH IH /S/S ) SFYTH 5 I tt IHISISJ SFXTti SI" iH/S/ S}
~.7S 0 . 00 1.75 ~.oo 0.00 2.00 ~.oo 0.00 1.00 ~.oo 0.00 lil. 00 ~ . oo 2.00 ~.oo 0 . 00 l. 00

"Tl
;;:;
c:::
'"
%~ I
m.

~ / I :~I \ I
~~I
~
-- I "Tl

ê
m.
"'m
§
ilo
I "'~I \
m.

§
0
m.
"'m
§

'" '"'"
&: 0>
0>
n
g o lL----L--------~ § o~
!! --------------~ o LI----~--------~ ~ IL-------~------J
El El
Z
~
SRlH ERR tDEG/S) salH ERR tDEG /S) SPTH ERR , DEG /S J ~ SRlH SIH IOEG/S) salH SI H IOEG/S) SP lH SI H IOEG / S)

~
~.oo 0.00 1.00 ~.oo '.00 1.00 ~.OO 0.00 6 . 00
[ij ~.OO 0 . 00 1.00 fIt . DO
O[
0 . 00
i
1. 00
i
~.oo 0 . 00 6.00

ë!
~
ë! ~
'"
f"

..
'"
n

.... ..
'"n
.... ~:
:~ I .I I ~~ I
~
0

I I ~~
~ ~
I ~:
~
~~
~

~ LI ______~____~ ~ LI ______-L____ ~
E LI______-L______~

,...
JRCKl 3 IH) JRCKL 2 [Hl JAC)I;L 1 IKJ SFZ ERR IH/S/S) SFY ERR IH/S/S) SFX ERR 'M/S/SJ
~ . ,oo 0.000 ~.300
° l
0.000
J
0 . '00
I
::',"00 0.000 ~.7S 1. 75 ~ . OO ;;.00 0 . 00 1. 00

"Tl
;;:;
c:::
'"
~ I \ "' mi \
~o ;0"'~ "Tl
;;:;
c:::
'"'"
m.
"'~mi i I I :~I \ t I ~ !"

'" "
fB"
0>

n
"o
n
g
El
o LI-----L--------~ o IL --------------~
~ 0 0
,.z
~ ~ ."OO
° l
J ACKL 6
0 . 000
:e:::::::
I H)
0.1iI00 f(I.I7S
0 l
JACK L S IH)
0 . 000
oc::::::
0.175
i
RI . 175
° l
JACKL Y IH)
0 .000
::;:::>
0.175
I
[ij ;;,.00
SR ERR
,
' .00
IOEG/S)
,
I.DO ;;.00
sa ERR
.... tDEG /S)
1.00 ~o . o ...
SP ERR iOEG / SJ
10 . 0

ë! ë!
'" '"
f"
f"

..
'"
n

....
..
'"
n

....
"'~ I ~~ I
~~ I
~
0

/1 ~~ I
~
0

I \ I ~~
~
m.

~ I I ~
0

I I ~:
~

~ LI __________-L__~ o~ ll-------------L~ o lL-----~L-----~ ~ 'L-------~------J

 g
N - ~

~
g
~ ~ r ~
N
"- .;

g ~
lY.o 6.0 lil.O a".0 6.0 U.O a".0 6.0 12 .0
T IH E ISJ T IHE 151 T IHE 151

~ ~
~

~ ~
'" 0
j - '~
"
"-
0~
0
..
-
V>
..
0
0
0
. -

~ g
a". o ' .0 a".o '.0 12.0 W.O
T IHE ISJ '.0 12.0
T IHE 151 T IHE 151

~
-

..
~
x
0
0
.; .
~
~
0
0
.;
Il-
a".o '.0 lil.O a".o '.0 12.0
TIHE 151 T IHE 151

FIGURE 7.70 COORDINATED TURN, SC #1 FIGURE 7,71 COORDINATED TURN, SC #1

 - - ---- - -

~ ~
g ,:

0on
..,
~
g
ci
~
,u
..'"
g
,;
~
u 0
~ ei
i
~ ei
n
u
x
'"
x
.."
~
.'"'"
~
..,
~

~.,

TIME IS)
12.0
~
~., ..,
TI ME IS)
u. a
0

~., ..,
TIM E IS)
12.0 ~.O ..,
TI ME IS)

8
~ ~ .~
~

~_._--+-
I
on on ..,w
,
on on 0
..,,
on
~ g
--_.- '" l! ~ g
w
~ 0 u, ,;
'" ,; 9 ~
u
~

'"
..'"
~
'"'"
w
~
'"
'"c
,
u
I
I
8
W., ..,
TI ME IS)
12 . 0
~

~
~.,

TI ME IS)
.. , 12 . 0
"~

~
~., .., 12. 0
~
~. , ' 8.0 12.0
TIME IS) TIME IS)

: ~ ~ g

,
on
,
on ..,
,on
,
i
on w
0
'" '"
~
u
N

'" l\
8'., ..,
TI ME IS)
12.0
,u
.
'"
N
~
0

~.O ..,
TIME IS)
12.0
,u ~
'"
'"~
~.,
~
..,
TIME IS)
12.0
, ~
u

'"
'"or
~
~., ..,
TI ME ISJ
12. 0

FIGURE 7.72 COORDINATED TURN, SC #2

FIGURE 7.73 COORD I NATED TURN, SC #2

~
~ ~
~

~
,on
~
1'l
,..,on

~
8
ei
-

~
..
fL 0

ei
~
~~
'" 0 ~
w
'?
ei

~
x
~
on
~
..,
on
.."
~

~.,

TI ME ISJ
12.0 ~.o ..,
TIM E ISJ
12.0
~
~., ..,
TIME ISJ
12.0
on 0
,;
~. , .., 12.0
TI ME IS)

~ ~
I ~
,
on
~
on
~
,..,on
!O eg ~ w

,u ~
I 0

~ ei
~ ~ ,
u
8
,;
'"
~
~ c
'"

"
~
0

'"
on on ~
"c
..,
~
~ on
~.,

TI ME IS)
12.0
~
~., ..,
TIME IS)
12.0
~
~., ..,
TI ME IS)
12 . 0
V) ~
~., .., 12.0
TIME IS)

8
~ ~
~
,in
~
0
,on
'"
~ ei
8
:;:
9 g
~ ei
~ ~
i
,u ~
'"
'" 0

N
~ or "
~ '"
on on N
"
~ .

81-----.
w., ..,
TIME IS)
12.0
8
~. , ..,
TI ME IS)
12.0
~
on
gt-------
~. o .., I l.O
::i
~
~., .., 12 . 0
TIME ISI TI ME IS )
FIGURE 7.74 COORDI NATED TURN, SC #2
FIGURE 7. 75 COORDI NATED TURN, SC #2
 ~ ~ ~ ~
:
,'"
~g
~
!; ~
r
~ =~
0
r
r
.; ." 0 ~ ~
'" ;;;
." ~
V> ~ X
X X >
or ~

~ g
8 ~
' .0 '.D 12 . 0 !i.o ' .0 12 .0 !i.0 6.0 12 .0 ' .0 ' .0 12 . 0
T IHE ISI TIHE ISI TIHE ISI Tl HE ISI

g ~ ~
~

,
."

,
V>
r ~ ~
'" g r g
~ ei V> .;

."
'"
~

~
~ ~
~ >
~
~
'.0 6.0 12.0 Il.O '.0 '.D 12 . 0 '.0 '. 0 12 .0
TI HE 15 1 Tl HE ISI , I ME 15 1

.; ~ .;
."

,
V>
'"on,, ,
."

r
r ~
r ~ " zo
! r
~
V> !:
0
."
'"
V>
'"
~
N
'"
<r
~
N
N
~

e-.o
~

6.D Jl.O ~. O 6.0 12 . 0 ~.O 6.0 12.0 ~.O '.0 12 . 0

Tl HE ISI TIHE 151 T IHE ISI Tl HE 151

FI GURE 7.76 COORD I NATED TURN, SC #2 FI GURE 7.77 COORD INATED TURN , SC #2

~, ~j ~I

---
;; 0'-'
w
0 ,'"'" '"'-'
w
r
0
~
.;
=
r 0

'" !:
"
~
'"
x ~
~ ~

'" g '"
.;
"' .0 '. 0
T I HE 151
Il. O ~.o '. 0
TI ME 151
1.i!. 0 '. 0 '.0
TlH E 15 1
0. 0 ' .0
TI ME 151

g g ~
.,;

;;; ;;;

~
'-'
w
0 '-'
'"
r
w
g g 9 g
~
r
0 0 0
r
'"or '" !: ;;;
~
w
'" lil
" g '"
~

'" g
~
' .0 '.0 Il. O ".0 '.0 1.i! .0 '. 0 1.i!.0 '.0 '.0 Il.O
T IH E ISI TI ME 151 TlHE ISI T I HE 151

g
~
- ,;

'-' ,on ,
<0
,'"
'",
W
c '-' '-'
w w
r
0
I g e~
'" ~
V> !:
0 r
.;
r
;;;
'"
~
~
N or
~

'" g '"
~
' .0 '.D Il. O ".0 ' .D 12.0 ~. o '. D 12.0 ' .0 '.0 12 . 0
TIHE ISI TIHE ISI TlH E ISI TlHE 151

FIGURE 7.78 COORDI NATED TURN, SC #2 FIGURE 7.79 COORD I NATED TURN, SC #2
 ~ ~

~ '",
'-'
'"'" ~
~
" g
;;:; ei ~
'"'"
~ .'"" '"
"
~
..'"
~ '" ~
~.o '.0 12.0 fi.D '.0 12.0 ~.O '.0 12.0 '.0 12 .0
TI ME 15) TIME 15) TIME 15) TIME 15)

g
~ ~
,'" ,
V>

,'" '-' ~
" '"c '-'
'"
g 9 ~
" Q
;;:; ei ~ Ö
'"'"
" :;:'"
~
~
~ ~
~ ~

'" '"
'.0 12. 0 ~.O '.0 12.0 '.0 ~.O '.0 12.0
TI ME 15) TIME 15) TIME 15) TIME 15)

~ ~

~
,..,'"
~ V>

'"c ~ :;:
g g !: g !:?
"
;;; Q
" .; .; ~
;!:
V> '"'" :;;'"
N "
~
'"
~
V> :li
N
~
V>
:li
g
~ ~
ri.D '.0 12.0 ~.O ' .0 12.0 ~.O ' .0 12.0 ~.O '.0 12.0
TIME 15) TIME 15) TIME 15) TIME 15)

FIGURE 7.80 COORDINATED TURN, SC #2 FIGURE 7.81 COORDINATED TURN, SC #2

~ ~I 5
'",,n
,
"
..,,
V>

'"c
;;:
- .;g
/ ~
~ §
'" .
r
'" Q
WO ~

.,'"or
u
;!:
"V>
~ .."
~

V>
~ ~
~.o '.0 12.0 '.0 12.0 fi.D ' .0 12 .0 fi.D '.0 12 . 0
TIME 15) TIME 15) TIM E 15) TIME 15)

~ ~ ~ ~

,
V>
,..,
V>

~
~ '"cg
"
:;;'" ~ '" ~
r
or • "' g
~
:;; Q

'"
u
" '".,
~
~
~
~

,. . ...
V> ~
~ ~ ~
w.o '.0 12.0 ~.o 12.0 ~ 12.0 ~.o ' .0 12.0
TIME 15) IIME 15) TIM E IsJ TIME 15)

~ ~ 2
.;

~
,
V>

"~ ~
i ~
~
'" 8.;
"
'" §
'"

"
N
~

'"
Q
~
~
'"
~

'"
U

~ r !l
~
~.o '.0
TIME 15)
12. 0 ~. o '.0
TIME 15)
12.0
!l
~.o ...
TIME 15)
12.0 ~.o ...
TIME 15)
12.0

FIGURE 7.82 COORDI NATED TURN, SC #2 FIGURE 7.83 COORDINATED TURN, SC #2

 g
,; ~

\I~,-
I

~ ~ ..
~ ~
0 ..
N g
.;

g g
lio '.0 12.0 ~.O '.0 12 . 0 ~.O '.0 U.O
TIHE 151 TlHE ISJ TIH E 151

~ ~
N

.~ ~
NO
><
'"
..
0
0
0 ~ : .. 0 I

~ g
~.O '.0 ~.O
12.0 ' .0 12.0 ~. O ' .0 12.0
TIHE ISJ T IH E 151 TIME ISJ

~ ~
'" 0 '" 0
~ ~

~ ~
~.o ' .0 12 . 0 ~.O
TIME 151 '.0 12.0
TIME ISJ

FIGURE 7 .84 COORDlNATED TURN, SC #2 FIGURE 7.85 COORDINATED TURN, SC#2

gr--;;:::::=:::==========I -
"
:.l! '"cw
'"
V>
0
..'" ~
~
.- .... :- .... _-

..'"
!:
~

•.
i~0----------7~0----------~
12 . 0 ~. O '.0 12 . 0 ~. O '.0
T IH E 151 TIME ISJ TIME ISJ

~r-------------~
- I
I

~, -
,
V>
I
I
I
'"
-t-----··
I
w i
c
-

!;
0 I
-1-
~
..
V>

~
II I! I
!
i~o----------~.~.O----------~
12.0 ~.O ' .0 12. 0
TIHE ISJ TIME ISJ

~r------------~----------~

,
V>

'"cw
~ Q -_.
------+---------
'" .;
V>

:l;
~

~~0----------~.~
. 0~--------~12. 0 ~. O:;---------;.-:.O:------I"J
•. O
TIME ISJ TlHE ISJ

FIGURE 7.86 ROLL TEST, SC #3 FI GURE 7.87 ROLL TEST, SC #3

 ~:I/
~
-
~g
~

~ ei I ~ ei

.
g
I
I
~ ei
~
e
~
o
ei
n
.'" ..
or or or
u or
X

~ ~ ~

I
~.o '.0 12.0 ~. O '.0 12. 0 ~.O ' .0 12.0 ~.O • '.0 J2 . 0
TIME ISJ TIME IsJ TIME ISJ TIME IS'

~ ~ i:!

iV
,
U> '"
W
C ,
U>

U>

~ g
'"~
W
g ~ g g
or Ö
~ ei ~ ei or ~ ei

u
~
or
..
~
~
or
~
w
'"
~
or
0
g g
1!
r.o '.0
TIME IS'
12.0 ~.O '.0
TIME IS'
12. 0 ~. O '.0
TIME 151
12.0 ~. O 6.0 '
TIME IS'
12 . 0

~ ~ ~
- I
,
! ~

~'!\-
U>

, '"w :.;
~
U>

"
- 0

- -f
C

u 8 9 g I
~ 0
~ ei
~ ei ~ ei
or
u
N
or ~
.
N
~
i ..
or
U>
or
~
I
~ ~
W.o '.0 IZ.O '.0 ' .0 " 01. 0 '.0 '.0 12.0 '.0 ' .0 12.0
TIME IS' TIME IS' TIME IS' TIME IS'

FIGURE 7.88 COORDINATED TURN, SC #4 FIGURE 7.89 COORDINATED TURN, SC #4

i:! - I -
~g
~ LrL ,.
~
~
~ ~
i f\
~ ei
0

~
or
~
0
ei 0
II
x
"-
U>
..
U>
I
'"x
~
U>
'"
~
"-
U>
g 1!
~. o '.0 12.0 ~.O '.0 12. 0 ~.O '.0 12.0 ~.o '.0 12. 0
TIME IS' TIME IS' TIME IS' TIME IS'

~ ~ i:!
,on
~
~ 8
,
U>

"'w~ g
i
~ ~
I '"cw
u ~
I

~
or 0
~
, 0

ci ei or

~
"-
on
0
U> E~ '"
~
0
on
~ ~
on
~
~.o '.0 12.0 '.0 ' .0 12.0 '.0 '.0 12.0 '.0 '.0
TIME 15)
12.0
TIME IS' TIME IS' TIME 15)

~
8 ~
~ I ,;

,
U>
~
~
~ , :.;
I
-1-
U>

~ e
~ g - eg '"~ ~ 1--- u ~
I

~
or
~
0 ;;; 0
ei ei

:;;gl----.
or
U>
"
N
~

:;; gt------
'"or
~

on
8
I
i:! r.G 0=.0 '.0
tr.o '.0 12.0 '.0 '.0 12.0 ' .0
TIME IS'
12.0
TIME IS'
12.0
TIME IS) TIME IS'

FIGURE 7.90 COORDINATED TURN, SC #4 FIGURE 7.91 COORDINATED TURN, SC #4

 8 ~ ! !
'", '"~ ~
~
~ ~ 8 zo g ! g
! zo 0 ;:;:; ó

'"
.;;; E '"
...
M M

.. ... .. ...
~

" ~
~
1 IM f IS.
12.0
~
~
1 IMf 15 '
11. 0
~
~ .. 11 M ISI
I l.O
8
r.o ..
1 IMf IS '
al. O

§
.;

'" '"
'",~ § !

!
~
...e:
'"a '">
::
...
1 IMf 151 1I"f
.. 15'
11.0 fiÎ. D ti .a
1I Mf IS I
1&'. 0
!!
!i. • ...
1 IM f IS.
u. a

0 r
~ .;

~
'"
'"
~ §
'"
'"zo ~ !
0 zo § ! §
~ Q zo .0 ;;; D

'"
. ~
on
~

'"a
...
N
r
::
ti., ..
11" ( IS'
U.O
0
~ .. 1 I Mf IS'
11.0 u .a
~
ti.• ...
11ME 15 1 .
U.O

FJGURE 7.92 COORDJNATED TURN, SC #4 FIGURE 7. 93 COORDINATED TURN, SC #4

g g
~ ~ I~

*
1\,
!. ~
.....
<,
.,
eg
" 0
\
\
\y-
'"
'"zo
zo
g
0

'"
.
%
..'" ...on .
on
g
~ .. ...
1 IMf ISI
11.0
8
w.• .. ,
1 IM f IS'
11.0
g
~ .. ..c
11 " [ IS'
U.C
g
r.o ...
1 IMf IS'
11 .0

g g
~ g

<;
~
..,'" ~ ..,'"
:> '"~
9 eg
zo 8
;;; D
!
g
.;
"
! z: "
" '"
~
% '"" "'"
g
~ .. ...
1 IMf 151
12 . 0
~
G"•• •. e
11ME IS '
11.0 B.c ...
11ME IS I
I l. O
g
G"•• ...
llME IS'
Il . O

~ ~ .
g
~

<; '" ~ on
w
c :.; ~ "
~ : :> g
.;
~ g eB
zo .;
zo zo D

'"
.
"' '"
Ir
N

::; g
ffi
g
~ .. ...
1 1ME IS '
U. O
~
G"•• ...
lIM E IS'
12.0 r.o ...
11ME IS'
U .D
~
G"•• I.'
llME 151
12. 0

FIGURE 7.94 COORDINATED TURN, SC #4 FIGURE 7.95 COORDINATED TURN, SC 1/4

 u
~
~ ~
0

'",,
'"
ZO

ZO
.
~g
~
~
e
0 ~ 0 ,; ,;
'" '"'"w '"~
~
~ .
"
~ x
~
..'"
'" '" '"
Ir. 0 ' .0
lIME ISI
U .O
~
1r.0 '.0
l IM E ISJ
12.0
~
1r.0
lIME
... ISJ
12.0 ...
lIME ISJ
12.0

~ 0
g

,'"
~ ~
~
zo g
,;
"cw
zo ,;
g ~! ~ ~
'" '" ~ '"'"w
,...'"
~~
~

~
"'"
'"
fi.D '.0
lIME ISJ
12 . 0 1r.0 '.0 ISJ
TIME TIME
'.0 ISJ 12.0
g
Ir.• ...1
TIME ISJ
12.0

~ g
~ ~

~
,'"
~ i
'" I
'9 g g g I
~ ~
",0 ~ ei ,; ,;
'" '"~ '"~
'"
~
N '"
~ N I
'"'" '"'"
...
~ ~

'" g
~.o '.0
lIME ISI
12.0
~
1r.0 ...
TIME ISJ
12. 0
'" ~
1r.0
lIME
... ISJ
12.0
~
Ir••
TIME ISJ
li.O

FIGURE 7.96 COORDINATED TURN , SC #4 FIGURE 7.97 COORDINATED TURN, SC #4

- ~ ~
,;

,'" '"
,'"
~
zo §

~
zo
g
. -! '" ,;g
'".."'~lJ
,;
~ ~
~
~
~
u u
'"><
~ ...,
'" ~

...
~

...
~

'" ~
Ir.• ...
lIME ISJ
12.0
'" ~
~. O ...
TIME ISJ
12 . 0
!!
~ .. lIME ISJ
12.0
~
~ . TIME ISJ
12.0

! ~
'",, '",
'"~ "cw zo zo
g
~ ~
8
ei
'" ~
~
~
u
'",...
~
~
'" '"...,
~

'" ~
"'" ~ ~
&.0 12.0 Ir•• •• 0
TIME ISJ
12.0
-
&.0 '.0
TIME ISJ
12.0 &.0 ...
lIME ISJ
li.O

~
~ ~ ,;

,
,'"'" '", §
~
"ew zo
g
'"~ ~ -+---- 8
~ó
'"
~
~
,; - +
<D
~ ! - ---
~
u
~'" '"
~ I ...,
'"
...
N

'"'"
~

'" ~
Ir•• ...
lIME ISJ
12.0 1r.0 ...
TIME ISJ
12 . 0
~
fi.D
TIME
... ISJ
12.0
~
fi.D
lIME ISJ
12. 0

FIGURE 7.98 COORDI NATED TURN, SC #4 FIGURE 7.99 COORDINATED TURN, SC #4

 ~ r---------~-------. ~ r--------------.
r---------_ ___- - , ,

~.~.-----7,.. ,.- - - - --,,7,.•

llME 151
~.!:-.-----..,,-;.•,------,.,.:,.•
llME 15J
..
~ ,. 15J
llME
12.0

~ r------------------. ~ r------------, ~ r------------------'

~ .. ...
llME 15J
12.0 :.~:--~::::===:.=:.O:======:I1
TI ME I S)
2. 0 ..
~.~.------::-..,.-----~
llME 15J

~ r--------------.

.
'x" 0
.;

~.~
O -----..,.-;
.•,-------:-!
IO!.O ~~.~o------;.-,.o;-------:-!12.0
TI ME 15J llME 15J

FIGURE 7.100 COORD INATED TURN, SC #4 FI GURE 7.101 COORDINATED TURN , SC #4

8.0 SUMMARY
A number of computer algorithms have been developed and tested. They
have been used to impl ement and invest i gate the perfonnance of three fl i ght
simulator motion-base drive techniques:
(1) classical washout
(2) optimal control
(3) coordinated adaptive washout
These motion-base drive schemes are based on reports appearing in the open
1 i terature. As part of the eval uati on process programs have al so been
developed to:
(1) generate input test signals
(2) predict the pilotls motion sensation
(3) generate the commands to the six hydraulic
actuators employed by the UTIAS Fl i ght Research
Simulator
In the case of the optimal control algorithm, software has been produced to
solve the equations associated with the generation of the optimal
controller.
The proper functioning of all three algorithms has been demonstrated.
At the present time, they have not yet been refined for the best possible
real-time perfonnance. This will take place during the next phase of the
study.

8. 1
 APPENDIX A

EQUATIONS PROGRAMMED FOR FA ST INTEGRATION METHODS

Equations 4.3 to 4.9 and 4.10 to 4.12 were solved using a number of
i ntegrati on methods. The al gorithms used were taken from References 4.1 to
4.3, and we re formul ated with special consideration for computing
efficiency. The ABC method was only used on filter Case 2 since its form
was not adaptable to Case 1. The following sections list the equations
programmed for filter Case 1 (except for the ABC equations which are given
for Case 2); the corresponding equations for Case 2 being analogous. In the
equations listed below, the following apply:

Xi=Xi(t)

I
xi = xi (t - h)

11
xi = xi(t - 2h)

h = step size

= Dot

A.1 Euler's Method (Reference 4.1)

This method i nvol ves a single derivative evaluation at each step.

I I I
x 1 = X1 + h(-e lt x 1+ x 2 + .tI! i I) (A.1 )

I I I
x2 = x2 + h(-e 3x 1+ x 3 + .t xi I ) ( A.2)
3

I I I
x3 = x 3 + h( -e 2x 1+ x It + .t xi I ) ( A.3)
2

I I I
XIt = x lt + h(-e 1x 1+ Xs + .t xi I ) (A.4 )
1

A.1
 I I

X5 = X5 + h ( -e Ox I (A.5)

(A.6)

A.2 Improved Euler's Method

This method uses the same equations as given in Section A.l except
that:

xii is replaced by (xi + xi' )/2 in Equations A.l to A.5

I I
xl is replaced by (xl + xI)/2 in Equations A.2 to A.5

These improvements provide an evaluation of the derivative using an

approximation to the trapezoidal rule (Reference 4.1).

A.3 Tustin Method (Reference 4.2)

Thi s method uses the Tustin approximation to the z-transform and is
valid only for linear equations.

The Laplace operator s is approximated by: 2(1 - z-l) (A.7)

h(1 + z-l)

The pertinent z-transforms are:

Xi (t) is the inverse z-transform of xi (z) (A.8)

Xi (t - h) is the inverse z-transform of z-lxi (z) (A.9)

This yields:

A.2
 ( A.1Q)

(A.ll )

(A.12)

ia ~)( xi + xi ~
I

( A.13)

(A.14)

x0 = Xl + a sX i

A.3
A.4 2nd Order Adams-Bashforth (Reference 4.1)

This method involves one derivative evaluation at each step; but the
two previous derivatives are used to determine the present function value.
This is a second order method which performs much better than Euler's for a
slowly changing function, and only marginally better for a rapidly varying
one since old evaluations are more quickly outdated in the latter case.
This method is not self-starting, so the first step was performed using a
2nd order Runge-Kutta algorithm.

- I
xl = -e4 x I + x2 + J. xi I
4

(A.15)
I
- I
X4 = -elx l + Xs + J. xi I
I

+ J. 0x i
- I
Xs = -eOx I I

xl = Xl + h (1.5 xl
- I
.5 xl)
-..
(A.16)

x4 = x4 + h (1.5 x 2
- I
.5 x 2)
-..
Xs = Xs + h(1.5 Xs - I
-.
.5 xs)

x O = xl + a sxi

A.4
A.5 2nd Order Runge-Kutta (Reference 4.1)

This method is equivalent to an Euler Predictor-Corrector with a

si ngl e corrector step. Two deri vati ve eva 1uati ons are performed at each
step.
I I
P1 = h(-e 4 x1 + x 2 + .tltx i' )

Q1 =
··
Rl = (A.l7)
I I

Sl = h(-e1 x 1 + Xs + .t1Xi I )

Tl = h (-e OX 1 + .tox i I )

P2 = h(-elt(x~ + P 1) + (x 2 +Q1)+ l 4 xi )

Q2 =
·· (A.18)
R2 =
I

S2 = h ( -e 1 (x ~ + P1) + (xs+T 1)+ l 1xi )

T2 = h(-eo(x~ + P 1 ) + lox i )

Xl Xl + .5(P 1 + P 2 )

·· I
(A.19)
xit = xit + .5(Sl + S2)
I

Xs = Xs + .5 (T 1 + T2)

X = Xl + a sX i

A.5
A.6 4th Order Runge-Kutta (Reference 4.1)

These equati ons are simil ar to the 2nd order Runge-Kutta, but i nvol ve
four derivative evaluations at each step.

··· (A.20)
·
+ lel i I )

h(-e4(x~ + P/2) + (x 2 + 0/2) + J.. 4(x i + x i I ) /2 )

I
P2 =

··· ( A. 21)
· I
T2 = h(-eO(x I + P/2) + J.. o( x i + x i I ) /2 )

·· (A.22)
·

h(-elt(x~ + P 3 ) + (x 2 + 0 3 ) +
I
P4 = J.. xi )
4
··
·· I
(A.23)

T 4 = h (-e o( x I + P 3) + J..ox i )

xl = xl + (PI + 2P 2 + 2P 3 + P 4)/6

···· (A. 24)

I

X5 = X 5 + (T I + 2T 2 + 2T 3 + T 4) /6

xo-x
- I taxi
5

A.6
A.7 ABC Method - Case 2 Only
This is a modification of Euler's Method which, in essence, uses some
new values on the right hand si de of the differential equations when they
become available. It is not inherently more accurate than Euler's Method,
although it is slightly faster. The equations programmed were as follows:

x 0 = xi - 2ÇunY' - un 2Z '

Y = y' + hx o ( A. 25)

z = z' + hy

where y and z are intermediate variables.

A.7
 APPENDIX B

SCALING AND LIMITING

Filtering signal s tends to modify their magnitude according to their

frequency content. Sealing modifies the amplitude unifonnly across all
frequencies. Limiting is a nonl inear process that cl ips the signal so that
it is limited to be less than a preselected magnitude. Sealing and limiting
can be used to reduce the motion response of a flight simulator.

B.l HP SCALE
The HP SCALE block in Figures 5.1 and 5.2 represents both sealing and
limiting of the signals input to the block. The corresponding input/output
characteristic for one channel is given in Figure B.l. In the case of an
angular rate input such as ~AA in Figure 5.2 the output wl is given by

( B.l)

where !(~AA) is diagonal with each element having the nonlinear fonn of
Figure B.l. In general all of these elements can be different.
In the case of a specific force input !AA as in Figure 5.1 the output
f1 is given by

..!l = !(!.AA + .9.1 )[.!.AA + .9.IJ - ]:1 (B.2 )

wher~ !(.!.';A + .9.1) has the same properties as K(~A). The fonn of Equation
B.2 lS arr1ved at by noting that under equilibrTum flight conditions

(B.3)

If one were to scal e thi s val ue of !AA and ask the simul ator to match it,
then following a transient it would be found that the simulator was only
able to reproduce the unscaled version with fSS = -gl. Equation B.2
recognizes this fact and only scales and limits that-part of !AA that
departs from the value ~.9.I.

B.2 LP SCALE
The LP SCALE block in Figure 5.2 represents both sealing and limiting
of the signals input to the bloek. The corresponding input/output
characteristic for one channel is given in Figure B.l. The signals

B.l
generated by the output from this block are used for the tilt-coordination
process. As can be se en in Seetion 5.2.2 it is the ratio of the x and y
components to the z component of the si gnal processed by thi s branch that
determines the degree of tilt-eoordination. Thus the sealing and limiting
need only be applied to the x and y signal eomponents. f2 in Figure 5.2 is
given by

(B.4 )

where

(B.5 )

B.3 Reference Frame Effects

Sealing and limiting can produce effects similar to those caused by

filtering as described in Appendix C.

If the sealing and limiting are performed on F I components then

cross-coupling among the motion stimulus components can occur in FS •
Consider a case where the HP FIlT #2 of Figure 5.1 is replaced by sealing
and 1imiti ng of the form

(B.6 )

where K(a2) has the same properties as !C!.AA + JII) and !(~A). Thus from
Equati on 2.2

.!..ss = .!::..sI ~SI - JIs

(B.7)
= !:.sI ~(a2)~ - Jls

Assume that filtering and sealing in FS are absent. Thus from Figure 5.1

From Equations B.7 and B.8 it follows that

(B.9)

8.2
In order to avoid cross-coup1 ing among the components in .!Ss and lAA the
matrix lsI!(~)hIS must be diagona1. This is on1y true in general for

(B.10)

For all other cases .!SS wi 11 suffer from cross-coup1 ing of the components
from 1A.A.

B.3
 OUTPUT

L·I

INPUT

FIGURE B.l SeALING AND LIMITING

 APPENDIX C

FILTERING FRAME EFFECTS

As indicated in Section 2 there are several reference frames involved

in the motion-base drive task. A choice must be made as to which motion
variable components to filter as part of the washout process. The two
logical possibilities are cab frame components and inertial frame
components. Each of these has advantages and di sadvantages in the present
application. If the high-pass filtering is carried out in the cab frame (F S
or FA) then the motion-base travel in the inertial frame may not be
adequately limited. If the high-pass filtering is carried out in the
inertial frame (Fr) then cross-coupling among the motion components to which
the simulator pllot is exposed can occur. These features will be
demonstrated bel ow.

C.l Filtering in FI

In order to illustrate the effects of filtering in FI consider the

linear motion channel represented in Figure 5.1.. Take tne case where
sealing (HP SCALE) and filtering in FS (HP FILT #1) are absent. Thus

~ = 1.1 S[.!AA + .9.s]

(C.l)
= 1.1 S .!AA + ..9.1

and
~I = [a2]h2
(C.2 )
= [1.IS .!AA + ~1]h2

From Equation 2.2 the specific force at the origin of FS is given by

lss = !:.sI ~I - .9.s

(C.3)
= !:.s I [1.IS .!AA + ~I ]h2 - ..9.s

Equation C.2 ensures that the simulator linear travel in FI is properly

washed out. Equation C.3 can be used to demonstrate the cross-coupl ing
effects of this process. In the present case assume that the same high-pass
filter is applied to each of the three components in Equation C.3. · Thus
from the theory of 1inear differenti al equati ons:

C.l
 t
.!.sS(t) = lsI(t) f h2(t - 1:) o:'IS( 1:)!AA( 1:) + ~I )d1: - 1!.S(t)
o ( C.4)
t
=f h2(t - 1:)lsI(t) o:'IS(1:)!AA(1:) + 1!.dd1: - 1!.s(t)
o

where h2(t) is the impulsive response function of the high-pass filter. Now
in order to avojd cross-coupl ing of. specific force sensations it is
I
necessary that f~S be independent of fÄA for i :# j. This in turn requires J

that

t j
f h2(t - -t)LLij(t, 1:)f AA (1:)d't =0 . for i = x, y, z
o
(C.5)

where
(C. 6)

It should be noted that in general LL(t, 1:) is independent of fAAh). Since

E q u a t ion C. 5 must hol d f 0 r aT 1 t a n d ..! AA ( 1:) i t fa 11 0 ws t ha t
h2(t - 1:).!:.!:.(t, 1:) must have off-diagona1 el ements of zero. Now

.!:.!:.(t, t) = .!. (C.7)

However in general fo r t :# 1: LL(t, 1:) wi 11 not be diagona1 si nce .lIS(t)

can be arbitrari1y generated. ThuS-the on1y way to ensure that Equatlon C.5
ho1ds for all t is to require that

h2(t - 1:) = 0, t :# 1: ( C.8)

This gives
h2(t - 1:).!:.!:.(t, 1:) = h2(O).!. for t = 1:
(C.g)
=0 for t :# 1:

Thus from Equation C.4

(C.10)

C.2
un1ess
h(t - 't) = Kö(t - 't) (C.l!)

for which case

(C.12)

Thus it is seen that the on1y nontrivia1 solution satisfying the requirement
of no cross-coup1ing is the case of no filtering at all.
It can be shown that the cross-coup1ing caused by filtering in FI is a
second-order effect in cases where is «1. In this case it can be seen
from Equation 2.10 that

~S
es
hls • [
-es
-<Vs
1
cl>s
-cI>s
1
] (C.13)

and T
hs I = 1.IS (C.14)

From Equations C.6, C.13 and C.14 it fo110ws that to the first-order in the
Eu1er ang1es

(C.15)

and thus Equation C.5 is satisfied.

C.2 Filtering in FS
Consider again the trans.lationa1 motion channel represented in Figure
5.1. Take the case where stal ing (HP SCALE) and filtering in FI (HP FILT
#2) are absent. Thus

(C.16)

and as shown in Section 2.5 the double time integra1 of a2 wil1 exhibit the
washout feature of returning to zero in the long term-if the high-pass
filter is suitab1y selected. But in the frame FI in which the simulator
motion takes p1ace

C.3
 ~I = .hlS ~ (C.Il)

a nd t he element s of ~S I are not the direct output s from a hi gh- pa ss fil ter.
Th usin gen era 1 t he ij 0 ub 1e t i mei nt e 9 ral 0 f ~S I (i. e ., t hes i mul at 0 r
displacement) will not return to zero in the long term. Also from Equations
C.3, C.16 and C.17

.!ss = .!:.sI ~I - ~
=~-.9.s (C.18)
= [lAA + .9.SJh I - ..9.s

and thus there is no cross-coupl ing among the specific force components to
which the simulator pilot is exposed.
This effect of insufficient washout in Fr is second-order in cases
where ..ê.s,
~ «1. In thi s case from Equat i ons C. 3 and C.17

(C.19)

to the first-order in the perturbation quantities ..ê.s and~. From Equations

C.16 and C.19 a2 and thus ~Sl have the feature that their double time
integral wil 1 return to zero 1n the long term if the high-pass filter is
suitably selected. Unfortunately, because time can become quite large
during a simulator run even second-order effects become significant and can
lead to unacceptable simulator displacements.

C.4
 APPENDIX 0

ACTUATOR EXTENSIONS

The geometrical layout of a six degrees-of-freedom synergistic motion

base is given in Figure 0.1. The relf~ant vectors relating the locations of
the upper and lower bearings of the i actuator are shown in Figure 0.2 It
can be seen that the location of the cab frame FS with respect to the
inertial frame FI is given by

(0.1 )

Thus the actuator length vector can be found from

(0.2)

Expressed in FI components Equation 0.2 becomes

lOI=AOI+SI-BO
_1 _1 _ _1 I
(0.3)
= -LIS -1
Ao S + -SI - -1
BOl

where ~i S and j!0 I are geometrical constants and ~I COO1es from the output of
the mot10n-base t1lter algorithms as

tt
~I = ~I ( 0) + JJ ~S I d t 2 (0.4 )
00

~I(O) is selected so as to start the simulator from a desired location.

Usually thi s will be with all the actuators extended to one hal f thei r
stroke although in special cases a bias towards some other locat~on may be
useful. The actuator 1 ength command signal employed for the i actuator
is
T 1/2
~ = ~I ~d -L (0.5)

where Lis the length of the actuator when the simul ator is at its neutral
positi on.

The CAE motion-base employed by the UTIAS Fl ight Simulator Facility

0.1
r.equires both ~i and ~i as input commands. The simplest means of generating
~i signals is to pass ~i through a high-pass filter. The resulting signal
~iest is an estimate of ~i passed through a first-order low-pass filter. In
Laplace transform notation:

__ - w2s 2
~~~ (D. 6)
S + w2

In certain applications it may be of interest to have a linearized

version of Equation D.5. Such a formulation can be achieved by linearizing
about the neutral position represented by

~I = ~I (0) (D.7)

~IS = 1. (D.8)

and
~ =0 (D.9)

Consider motions resulting in small perturbations away from Equations D.7 to

D.9 such that

~I = ~I (0) + !::S (D.I0)

~IS = .!. + &IS (D.ll)

and
I~I « 1 (D.12)

where from Equations D.4 and D.I0

tt
!::S = IJ 2.s Idt 2 (0.13)
00

and from Equations 2.10, D.ll and D.12 (us;ng the small angle approxirn-
ations)

D.2
 0

[ l
-<\IS es

&.IS = <\Is 0 - 4>s (0.14)

-es 4>s 0

Combining Equations 0.3, 0.10 and 0.11 it fo" ows that

~I = AiS + &.IS AiS + ~I(O) + óS - ~iI (0.15)

or
~ I = ~ Ie + ~ I ( 0.16)

where
~Ie = ~iS + ~I(O) - ~il (0.17)

represents conditions when the base is in its neutral position and

Mil=&ISAiS+ óS (0.18)

From Equations 0.5 and 0.16

(0.19)

Keeping only fi rst-order expressions in the perturbation quantity ~.R.i I it

follows that

T T 1/2
1; ... ~i Ie ~i Ie + 2~ Ie ~~ I ) - L
(0.20)
T 1/2 T -1/2 T
... ~i Ie ~i Ie ) + ~ Ie ~ Ie ) ~ Ie ~~ I - L

Now from the definitions of L and ~ I it follows that

1; = 0 ( O. 21)

when
(0.22)

D.3
Thus from Equations 0.20 to 0.22

T 1/2
L = C~iIe .til e ) (0.23)

Substituting Equation 0.23 into Equation 0.20 it follows that

T -1
.ti = &Ie Mi I L (0.24)

Equation 0.24 is the linearized version of Equation 0.5.

0.4
PAYLOAD PLATFORM

FIXED PLATFORM

FIGURE D.I MOT ION-BASE GEOMETRY

 Fs
PAYLOAD PLATFORM

ATTACHMENT POINT
PAYLOAD PLATFORM

S
"-7

F1
FIXED PLATFORM x
ATTACHMENT POINT
S·
~I FIXED PLATFORM
y Z

FIGURE D.2 VECTORS FOR A SINGLE ACTUATOR

 APPENDIX E

TILT-COORDINATION EXAMPLE

Consider a case where the aircraft is undergoing a constant

longitudinal acceleration and a simul taneous hi gh frequency pi tch
oscillation described by

~I = [a 0 O]T (E.1)

h. = [0 e 0] T ( E.2)

where
a = constant (E.3)

e = esinwt (E.4 )

and w represents a high frequency relative to the filters of Figures 5.1 and
5.2. From Equations E.1, 2.2, 2.10, 2.13 and 2.14 it follows that

acose +0 gsine ] .
lAA = [ (E. 5)
as i n e - gcos e

(E.6 )

•
where eis al so a hi gh frequency si 9 nal. From Fi gure 5.2 (droppi ng the
low-pass and high-pass sealing blocks in the present example)

(E.7)

From Equations E.5 and E.7

E.1
 a[cos9]L + g[si n9]L ]
fL = 0 (E.8 )
-- [ a[sin9]L - g[cos9]L

Now since 9 has the form given by Equation E.4 it follows that

[sin9]L =0 (E.9)

[COS9]L = y (E.lO)

where y is a positive constant less than unity. Thus from Equations 5.30
and E.8 to E.lO

fL = y[a 0 _g]T (E.ll)

~SL = [0 9SL O]T (E.l2)

where 9SL is a constant. Al so it foll ows from Fi gure 5.2 that

cJi = [~A]hl (E.l3)

and from Equations E.6 and E.l3

•
<di = [0 [9]hl O]T
(E.l4)
•
= [0 e O]T

•
t he 1 atter form foll owi ng due to the hi gh frequency nature of 9. From
Equations E.l2 and E.l4 it can be seen that the simulator receives only
pitch Euler angle commands and thus

4>S = <iJS =0 (E.l5)

From Figure 5.2 and Equations 5.22, E.l4 and E.l5 it follows that

~ = [0 ë O]T (E.l6)

E.2
 • •
Beeause 9 is a high frequeney signal ~ passes unehanged through HP FILT #2
and thus from Figure 5.2

(E.l7)

and from Equations 5.15, E.12 and E.17

.ês = [0 9+9SL O]T (E.18)

Now from Figure ' 5.1 (dropping the high-pass sealing blad in the present
example) and Equations 5.11, 5.13, E.5, E.15 and E.18 it follows that

.!!. = !..AA + .9.s

= l aeos9 + gsin9 - gsin9s ]

asine - gc:se + gcoses

aeos9 + gsin9 - gsin9 eos9SL - geos9 sin9SL

= o (L19)

asin9 - geos9 + geos9 eos9SL - gsin9 sin9SL

Si nee 9SL and I a I are constants and 9 i s a hi gh frequeney si gnal then from
Figure 5.1

E.3
 a[cos9]hl + gs;n9 - gsin9 cos9SL - g[cose]hls;neSL

= o (E.20)

asin9 - g[COS9Jhl + g[cos9]hlcOS9SL - gsin9 sin9SL

(noting that [sine]hl ,.. sin9). Dropping the HP FILT #2 bloek in the present
example it follows from Figure 5.1 and Equation 2.2 that

.!SS = !!:!. - gS (L 21)

Combining Equations 5.13, L18, L20 and E.21

-sin9 coseSL - cose sin9SL

.!SS = 2.!i - 9 o
cose cos9SL - sine sineSL

a[cos9]hl + gsine + gsineSL(cos9 - [cose]h 'l )

= o (E.22)

asin9 - g[cos9]hl - gcoseSL(cos9 - [COS9]hl)

Taking note of the fact that

[cos 9h ,.. cos 9 - [cos 9Jh 1 ( E. 23)

it follows that

E.4
 a[cos9]h1 + gsin9 + gsin9SL[cOs9]L

!ss = o (E.24)

asin9 - g[cos9]h1 - gcos9SL[cOS9]L

Now how well Equation E.24 simulates the specific force experienced at the
origin of FAin the actual aircraft can be seen by examining Equation E.5.
The hi gh frequency part of !.AA is

a[cos9]h1 + gsin9

o (E.25)

asin9 - g[cos9]h1

because 9 is given by Equation E.4 and represents a high frequency signal.

The low frequency part of !.AA is

a[cos9]L

o (E.26)

-g[COS9]L

Comparing Equations E.25 and E.26 with Equation' E.24 it is seen that the
hi gh frequency pa rt of ~~ is simul ated exactly and the low frequency part
only approximately with a simulated by gsin9SL (or from Equations 5.30,
E.11 and E.15 a e [g/la 2 + g2] in the present case) and 9 simulated by
gcos9SL. Overall it appears that a reasonable simulation has been achieved,
especially if lal « g.

E.5
 APPENDIX F

PROGRAM DESCRIPTION FOR WASHOUT FILTER EVALUATION

F.l Introduction

In order to perfonn a prel imi nary eval uati on of vari ous washout
fi lters, program PLANE was written to cal cul ate the pl atfonn moti ons and
actuator extensions necessary to simulate various aircraft motions. This
evaluation process assisted in the choice of filter coefficients for the
classical filters, and penalty weights for the optimal and adaptive filters
which would provide the most realistic pilot vestibular sensations within
the constraint of allowable actuator travel. In the present application the
al gori thms were not desi gned to run in real time and thus there was no
attempt to minimize the running time by employing the best integration
schemes as indicated in Section 4. This real-time running problem will be
addressed in the second phase of the study. Provision was made to generate
motion time histories at the point in the aircraft corresponding to the
simulator centroid (the origin of Fe), and to perform the following
calculations:

(1) Eval ua te and compare the vesti bul ar responses of the ai rcraft and
simulator pilots.

(2) Apply the classical, optimal, or adaptive filters chosen by the user
to the aircraft motion to obtain the simulator motion.

(3) Generate the motion-base actuator time histories.

The program is interactive, and the following choices are available to

the user:

(1) Type of washout filter used -- classical, optimal, or adaptive.

(2) Duration of simulation -- maximum = 40s with a time step of 50 ms.

(3) Acceleration channel to be excited -- surge, sway, heave, roll, pitch

or yaw.

(4) Type of excitation step, pul se, double pul se, filtered random
noise, sinusoidal, or coordinated turn.

(5) Magnitude of input -- in m/s 2 for translational channels and r/s 2 for
rotational channels.

(6) Pulse duration (s), if applicable.

;
l The hierarchical organization of the program and its subroutines is
shown ·· in Figure F.l, · where the upper block of any two blocks joined by a
vertical line can call the lower block. The following is a description of
the structure and use of these programs.

F.l
F.2 Program Structure
PLANE is the control program which interacts with the user; performs
various calculations (see Figure F.2) to set up the problem according to the
options chosen; and then calls the appropriate subroutines. The procedure
followed is outlined in Section F.2.1 and Figure F.2, where the numbers in
each block correspond to the block numbers included in the program listings
in this Appendix.
In general , the WASHOUT , WASHOPT , WASHADPT and JACKDRVR routines are
similar to those which wil 1 be implemented in real time during simulator
operation, and are written using computationally fast algorithms. The
remainder of the software (PLANE, VESTIB, VESTIBI and RANDOM) are evaluated
using slower, more accurate algorithfs. Integrations in PLANE were
performed using the trapezoidal rule while those in VESTIB, VESTIBI and
RANDOM were solved by the IMSL DGEAR routine.
The variable names used in all programs were chosen to approximate the
variabl e names used in thi s report. Generally (but not rigorously), the
following rules were observed:
(1) Double letters were used to rep re sent capital letters in the variable
names, but not in subscripts (e.g., LLIS =~IS).
(2) C, A, P or S is appended to a variable name to represent the point at
which a motion is specified:
Crepresents the position in the aircraft (the origin of fame Fe)
corresponding to the simulator upper frame centroid (t~e
user-defined inputs are applied at this pOint).
Arepresents the origin of frame FA. This point is coincident with C
for the classical and adaptive filters and with P for the optimal
filters.
Prepresents the origin of frame Fp or Fp •
a s
S represents the origin of frame FS' and is coincident with the
simulator upper frame centroid for classical and adaptive filters;
and with the simulator pilotls head for optimal filters.
(3) A, S or I is then appended toa variable name to rep re sent components
in the aircraft (FA)' simulator (F S )' or inertial (F I ) reference
frames respecti vely (e.g., ASI =2..sI' the translational acceleration
of the origin of frame FS' given in lnertial frame components).
(4) D is appended to a variable name to represent lts derivative, or an

tIn the notation of Appendix A, the trapezoidal integration scheme is given

• •
by x = Xl + h(x + xl )/2.

F.2
 •
approximation to its derivative (e.g., BETAAD =~).
( 5) I is appended to a variable name to represent its integral (e.g.~
WHI = f~dt) •
(6) In the classical algorithm the filter parameters are given in Table
F.l and the filter fonnat in Table F.2.

F.2.l Program Operation

The following is a sequential description of the operation of program

PLANE and its associated subroutines, and follows the flow chart given in
Figure F.2 step by step. The first two pages of programming include all the
dimensioning, memory allocation and data initialization required, which are
perfonned at compile and link times. The flexibility to have many variables
avail abl e for pl otti ng comes at the expense of a 1 arge memory requi rement
(roughly 700 kb). The remainder of the program continues at execution time
as follows (where the (numbers) refer to those in the figures):

(1) The user chooses from various options of filter types and input
motions:

(a) Choose type of washout filter from classical, optimalor adaptive.

This fixes the location of the aircraft reference frame, FA' at
the centroid (the origin of Fe) for classical and adaptive
filters; or at the pilot's head (the origin of Fp ) for the
a
optimal fi 1 ters. As well, the location of FA and the
corresponding position of FS' dictate the values assigned to Bss,
..?I (0), and Ai S •
(b) Choose the simulation duration. This, along with the time
interval fi xes the number of steps in the simul ati on, whi ch must
not exceed 800.

(c) Choose the accel erati on channel to be excited from among surge,
sway, heave, roll, pitch or yaw. The test acceleration is applied
at the origin of FC' and all other accelerations are set to zero.

(d) Choose the type of test input from among step, pulse, double
pulse, filtered random, sine wave, or coordinated turn. If the
coordinated turn is chosen, input (c) is ignored, (e), (f) and (g)
are skipped, and the pilot's head is placed at the origin of FC'
If a filtered random input is chosen, subroutine RANDOM is called
to set up an array of pseudo-random nonnal deviates, using IMSL
subrouti ne GGNML. IMSL subrouti ne DGEAR is then call ed to
integrate the differential equations which describe the
first order low-pass filter through which the random noise is
passed. Note that for rotati onal channel s, the fi 1 tered random
noise input is applied as an angular displacement rather than an
acceleration, in order to conform to the optimal filter
fonnul ati on.

F.3
 (e) Choose the magnitude of the test input in m/s 2 for translational
inputs, and r/s 2 for rotational inputs.
(f) Choose the pul se durat i on, in seconds, in the case of single and
double pulse inputs.
(g) Choose the input frequency in the case of sine wave inputs.
•
(2) Once all the above choices have been made, the input arrays (~A' ~A)
are formed according to the user responses.
•
(3 ) The angular rates (~AA) are calculated by integrating ~A using the
trapezoidal rule. In the case of angular filtered random noise
inputs, !:!.AA is taken as an intermediate variable straight from the
IMSL DGEAR subroutine.
(4 ) Subroutine VMULT i.s called to transform the angular rates (~A) into
Euler angle rates (~) using T.
(5 ) The Euler angle rates are integrated using the trapezoidal rule to
yield the Euler angles (1A).
(6 ) Matrix.! is evaluated with~. At this point, the program backt~acks
t 0 obta in the est i mated angul ar rates and accel erat i ons (.!!!:..AA' ~A)
according to the procedure outlined in Section F.3.
(7) The translational accelerations at the origin of FC (aCA) are
transformed to pilot's head accelerations (~A) using Equatlons 5.46
to 5.48.
(8 ) The motion of the reference point (origin of FA) in the aircraft is
evaluated. Essentialy, ~AA = ~CA for the classical and adaptive
filters; while ~AA = ~PA for the optimal filters.
(9) and (10)

The vector..9.A is subtracted from ~A and ~PA to obtain!..AA and .!pA.

At this point, !!..!.PA and fE2pA' the estimated specific forces in the
aircraft are evaluated according to Section F.3.
(11) The aircraft pilot's vestibular response is generated from~A and
fE1tA for the classical filters; wEA~ and lYA for the optimal 1 ters;
an WAA and fE2 pA for the adaptive filters. In each case, two
vestibular modëTSare used: subroutines VESTIB1 and VESTIB, with and
without threshold blocks in the vestibular transfer function
respectively. In either case, the procedure used is outlined in
Figure F.3.
(12) Initial values are then set, and the appropriate washout routine is
cal led. The washout subroutines follow their flow charts given in
Figures F.5 to F.7, and are written using computationally fast
algorithrns. Trigonometrie functions are evaluated using truncated

F.4
 Taylor series expansions (see Appendix G), and integrations are
perfonned by Eul er or ABC methods. The outputs of the washout fi 1ters
•
are ~SI' ~I' !s' ~S, and ~S·
(13) Subroutine JACKDRVR is called with inputs ~I and .1!s to find the
corresponding actuator lengths and accelerations (~, ~est).

(14 ) Matrix .hSl is fonned by taking the transpose of .!:.IS.

(15) The translational acceleration of the origin of FS in body axes (~S)
is obtained from .!:.SI and ~I.
(16) The corresponding acceleration at the simulator pilotls head (~PS) is
evaluated using Equations 5.46 to 5.48. Note that if the optimal
filters were chosen, the origin of FS is at the pilotls head, and
~PS = ~S·
(17) The vector .[S is subtracted from ~PS to obtain .!PS.

( 18) The simulator pilotls vestibular response is generated from f pS and

wss using the vestibular routines VESTIB and VESTIB1 mentioned 1n step
11. It shoul d be noted here that wss and ~PS may not be exact val ues
s; nce the al gorithm used to generate them [dl fferent for each type of
•
washout fil ter) affects the exactness of wss and ~SS, as shown in
Section F.3. The aircraft pilotls vestibuTar responses are also
calculated us;ng the same approx;mat;ons in order to allow a valid
comparison of the two.

(19) Subroutine OUTP is called to plot the variables required by the user.
The vest i bul ar erro ris cal cul ated in that routine as the d ifference
between the aircraft and the simulator pilotls vestibular sensations.

The Perkin-Elmer FORTRAN VII compiler used for all the programs listed
in Section F.4 contains some extensions to the ANSI X3.9-1978 standard. Of
these, the foll owi ng have been used extensi vely and are critical to proper
understanding and running of the programs:
r

(i) Values of variables calculated in a subroutine are saved for a

I subsequent call to the subroutine without including those variables in
a SAVE or C<l1MON statement. Thi s feature is often used in the
integration and filtering processes in the washout filters.

(ii) A semicolon in a program line denotes that whatever follows on that

line is a non-executable comment.

F.2.2 Required Inputs

The inputs required to run these evaluation programs are:

(1) The user-variable inputs chosen at run time, described in Section

F.2.1.

F.S
(2) Other parameters whi eh requi re the mod ifi cati on of data statements
within the program, such as:
( a) Otol ith model parameters: a , 'tL , 'tS' dTH x ' dTHy ' dTHz·
K, 't

( b) Semi c i reul ar canal model parameters: TLp' TLq' TLr' Ts' Ta' öTH p '
öth q , öTHr·
( c) The iteration time step, OT.
(d) The washout fi lter coeffic i ents.

F.2.3 Available Outputs

The output variables available for plotting are as follows:
A A Ath Ath
(1) Aireraft variables: ~CA' !PA' .!!:AA' h, !PA' ~A' .!PA' .!!!pA·
A A Ath Ath
(2) Simulator variables: ~S I' !PS' is I' ~I' .ês, ~S, .!PS' ~S, .!PS' .!!!pS·
(3) Vestibular error: efA, e.hl.-' efAth, e.~h.
--P -~ --P --t'

(4) Actuator lengths: " .

9, .

(5) Adaptive filter parameters: PX ' PX ' PX ' PYl' PY2' PY3' PZ' Pijl.
1 2 3
Note in the above that sensations with a superscript ( )th represent results
with a threshold in the vestibular model and those without have no
threshold.

F.3 Calculation of Vestibular Sensations

The vestibular sensations consist of three translational otolith
sensations and three rotational semicircular canal sensations. The input to
the vestibular models are the simulator or aircraft motions at the pilot's
head, and the outputs of the model s are compared for the two cases.
However, since the simulator pilot's head motions are calculate differently
(i .e., with various degrees of approximation) for each of the washout
filters, the aireraft pilot's headmotions used in the vestibular models
must be determined in a similar manner in order to allow valid comparisons.
The procedure used to obtain the simulator pilot's sensations is outlined in
Figure F.8, where an 'estimated' variable implies one which has been
calculated with the following transfer function:

ut= W2 s x W
s + W2

F.6
 •
where w = integral of w
• •
<Af = est imate of w
w2 = 12.5 rls

Note that in Figure F.B, f pS is the exact specific force at the

pilotls head in FS components;.fEl pS is the same specific force, but has
been calculted using wE SS and wE SS (see also Secti~n 5.4); and fE2 pS has
been calculated using the exact ~S and the estimated <AfSS. The three will,
in general, be slightly different. In order to compensate for these
differences, the aircraft pilotls sensations must be calculated using
correspondingly estimated motions as follows:
(1) Classical washout uses:
(a) estimated angular rates (~A)

(b) fE1 pA calculated u.sing .!CA and estimated angular rates and
accelerations (~A' ~).
(2) Opt imal washout uses:
(a) estimated angular rates (~~A)

(b) exact 1pA

(3) Adaptive washout uses:
(a) exact angular rates ~A)

(b) fE2 pA calculated using !cA' exact.angular rates ~A) and

estimated angular accelerations (<Af AA)

F.7
 Table F.1

Filter Parameter Notation for the Classical Algorithm

Parameter - X Y Z Cl ~

Symbol Value Meaning

X T translational degree-of-freedom
R rotational degree-of-freedom
y L low-pass fi lter
H high-pass filter
z S filtering in FS
I filtering in FI
Cl 1,2, ••• order of the filter
1,2, ••• power of s in the filter transfer function
denominator which is multiplied by the
parameter

F.8
 Table F.2

High-Pass Filters Employed in Classical Washout

(M = N)
(see Sect ion 5.5)
Al gorithm
Case N Output y Xo xl x2
wl + u.tI 2 Xo WI* WH WHI WHII
(wl )** ( u:H)

•
~ -+- ~SH I xl BETAHDOT BETAHD BETASH
•
( ~H) ( ~SH)

al -+- aH 2 Xo Al AH AHI AHII

(~) (~)

~ + ~I 1 Jx ldt A2 SSIDD SSID

•
(~) (~IJ

f2 -+- fL 2 X F2 FLDD FLD FL

(with 1iJiti ng (fl.) (.!!:)
of xl)

1 LI J)DAt LIDD
( :i; )
est
*symbol used in computer code of Appendix F
**symbol used in Section 5 of report
I 11
tLIDOA = (~.1 - 2~·1 + ~.)/
1
llt 2
where Xl (t) = x(t - llt)
xll(t) = x(t - 211t)

F.9
F.4 Program Listings
Routine Name Function
PLANE main control program F.11
RANDOM generates filtered white noise F.17
VESTIB, FCNl, lNTERP non-threshold vestibul ar model routines F.19
VESlBl, FCN2, lNTERP threshold vestibul ar model routines F.21
WASHOUT cl ass ical was hout fi lters F.24
WASHOPT optimal washout fi lters F.26
WASHADPT adaptive washout filters F.29
JACKDRVR actuator length generator F.32
LIBS eva1 ua tes .hl S F.32
VMULT performs matrix mul ti pl ication F.33
VTRANSP performs matrix transpose F.33

F.I0
 1 C PROGRAM PLAIIE. FTN I
2 C 2
3 C PROGRAM TO GENERATE TEST Ale INPUTS. WASH OUT A/C MOTION 3
4 C INTO SIMULATOR MOTION. eOMPARE VESTIBULAR RESPONSES OF Ale 4
~ C AND SIMULATOR PILOTS. GENERATE THE REGUIRED ~ACK LENGTHS. 5
6 C AND PLOT ' RESUL TS AS REGUIRED. 6
7 C NOTE THAT TIE I""UT Ace &< ANG . VELS ARE AsSUMED TO BE AT THE 7
8 C POINT IN THE A/C CORRESPONDING TO TIE SIMULATOR CENTROID. 8
9 C 9
10 C PARAMETERS TO BE SET BY THE USER AT RUN TIME : 10
11 C IWASH= TYPE OF WASHOUT FILTER (I=CLA5sICAL. 2 z 0PTIMAI.. 11
12 C 3=A DAP TI VE ) 12
13 C NSEC= TOTAL DURATION OF SIMULATION (SEC) 13
14 C NEXC= CHANNEL TO BE EXCITED (I TO 6) 14
15 C NTYPE= TYPE OF TEST INPUT (l=sTEP. 2..pULSE. 3=DBLE PULSE 15
16 C 4=FILTERED RANDOM. 5zSINE) 16
17 C RMAG= MAGNITUDE OF TEST INPUT 17
18 C TDUR= DURATtON OF TEST INPUT (SEC) 18
19 C 19
20 C PARAMETERS WI-tICH CAN BE CHANGED FROM WITHIN THE PROGRAM ARE : 20
21 C DT.INVDT= TIME STEP (SEC). AND lTS INVERSE 21
22 C NPOINTs= TOTAL NO . OF STEPS 22
23 C ssI.ssIIN= INITIAL POsITION OF SIM . REF . POINT 23
24 C 24
25 0000001 REAL*4 FAA (3). WAA( 3) . ssI(3). 55 IIN( 3). VSI(3). AsI (3). LLsI (3. 3) 25
26 0000061 REAL*4 APs (3). AsS(3). RRS(3). TT (3.3). RR (3.3). AAA(3) 26
27 0000061 REAL*4 BETAA(3).BETAAD(3). BETAAI(3).BETAADE(3) 27
28 0000061 REAL*4 WAAI(3).WAAE(3).WAADE(3) 28
29 0000061 REAL*4 LI(6).LID(6) . LIDD(6) 29
30 0000061 REAL*4 AAIs(3.6) . BBII ( 3 . 6).LLIs(3.3).FUNC(6).DFUNC(6.6).LENGTH(6) 30
31 0000061 REAL*4 LENGTH_NEUTRAL 31
32 0000061 REAL*4 XN(6).XNPI(6) 32
33 0000061 REAL*4 RRsC(3).RRsP(3) .ssIINC(3).ssIINP(3).AAIsC(3.6). AAIsP(3. 6 ) 33
34 C 34
35 0000061 DIMENs ION PXIPL(801). P X2PL (801 ). PX3PL< 801). 35
36 *PYIPL(8011.PY2PL(801).PY3PL(8011.PZPL(801).PPsIPL(801) 36
37 00000 61 DIMENSION RMAG(6).PINK(BOI).PINKINT(BOI) 37
38 0000061 DIMENSION AXCA(BOll.AYCA(8011. AZCA(BOI) 38
39 0000061 DIMENSION AXPA(BOI).AYPA (80 1 1. AZPA(BOIl. 39
40 *sPPA( 801). sGPA ( BOl l. sRPA(8011. sFXPA ( BOI). sFYPA (801 I. 40
41 *SF ZPAI BOl ) 41
42 0000061 DIMENS ION AXAA (801). AYAA< 801 I. AZAA (BOl I. 42
43 *PHIA(8011. THETAA (8011. PsIA(801 I 43
44 0000061 DIMENs ION PHIAD ( BOll . THETAAD(8011. PsIAD(BOI) 44
45 0000061 DIMENsION FXAA ( BOI 1.FYAA(SOI). FZAA(BOII 45
46 000006 r DIMENSION FXPA(BOI1.FYPA(801).FZPA(BOII 46
47 0000061 DIMENsION FXPAEI(BOll . FYPAEH8011 . FZPAE1(801) 47
48 0000061 DI MENS ION FXPAE2 (801). FVPAE2(801). FZPAE2(801) 4B
49 0000061 DIMENSION PAA(801 ).GAA (801).RA A ( BOI ) 49
50 0000061 DIMENSION PAAD(801) .GA AD (8011. RAAD (BOI) ~O
51 000006 I DIMENS ION PAADE(801). (lAADE (80 1 I. RAADE< 801 ) 51
52 0000061 DIMENs ION VXsI (BOl). XS I (80 I). VysI (8011. YSI (eOI ). ~2
53 *VZSJ<BOll. ZsI(80l) 53
54 0000061 DIMENS ION SFXPs(6011 . sFYPS (8011. SF ZPS( BOll ~4
55 000006 I DIMEN s ION SPPS(BOll. SGPS( 801 I • sRPS(801 ) 55
56 000006 I DIMENSION R~(801 .6) 56
57 000006 1 DI MENS ION AXSI (801 ). AYSI (801). AZSI (801 ) 57
58 0000061 DIMENSION sFXPAUBOII. sFYPAH801) . SFZPA1<801l. sPPAl(801) . 58
59 *sGPAI f 801) • SRPAI ( 8011. SFXPsl< 80ll . sFYPSI (80l). sFZPsI (80ll. 59
60 *sPPsl ( 801 ) • SQPs1<80ll . sRPsl ( 80t) 60
61 0000061 DIMENSION PHIS (BOl). THETAs (BOl ). PS IS(801) 61
62 0000061 DIMENSION WSS ( 3 ).W SsD (3) 62
63 63
64 COMMON &< E(lUIVALENCE STATEMENTS . 64
65 0000061 COMMON I MATRIX !IAAIS. BBII. LLIS. FUNC . DFUNC. LENG TH. XN. XNPI . 65
66 1 LENGTH_NEUTRAL 66
67 000006 I COMMON I sINCOs l SINPSI.COsPSI. sINTH.COSTH. SINPHI.COSPHI 67
68 0000061 COMMON I FILTER I ASI.VSI 6B
69 69
70 0000061 COMMON ISIZEI NPOINTS 70
71 000006 I COMMIJN I CONTRDLI ICQNT 71
72 0000061 COMMON ITIME/ TIME(801 ) 72
73 0000061 COMMON IAMOTlONI FXPAP (80t ). FYPAP (8011. FZPAP(801) . 73
74 *PAAE( 801). (lAAE <BOl). RAAE( 801) 74
75 0000061 COMMON I SMOTlONI FXPS( 801) • FYPs (80 I). FZPs( 801). 75
76 *PSS(801). (lSS(801l. RsS(801 ) 7b
77 0000061 COMMON IVSTBLR I sCCTL I. SCCTL2. SCCTL3. SCCTS. sCCTA. OTOK. OTOTA . 77
'78 *OTOTL. OTOTS.OTOW2 . THP .TH<l. THR . THX . THY.THZ 78
79 0000061 COMMON I VSTWAsHI GSP . VI. V2 79
80 0000061 COMMON IANGDERI WSS.WsSD 80
81 0000061 COMMON IPDOT I WAAD (3) BI
82 0000061 COMMON I PXOUTI PXI . PX2 .PX3.PY I.PY2 . PY3.PZ.PPSI B2
83 C B3
84 C DATA STATEMENTS : 84
85 C CENTRO 10 AS SIMULATOR REFERENCE FRAME 85
86 0000061 DATA SSIINC/O . • 0 . • -1 . 8363 1 86
87 0000061 DATA RRSCI-0. 02.-0 . 465 . -I . 78301 87
88 0000061 DATA AAISC/-I.46647.-0.09525.0 . • -1 . 46647.0 . 09525.0 . • 0 . 65075. BB
89 I I. 3176. O. O. B9
90 1 0 . 81572. 1. 22237.0 .• 0 . 81572.-1.22237 .0 . • 0.65075.-1 . 3176.0 . 90
91 PILOT ' S HEAO AS SIMULATOR REFERENCE FRAME 91
92 0000061 DATA SSIINP/-O . 02.-0. 465.-3 . 61931 92
93 0000061 DATA RRSP/O .. O.• O. I 93
94 0000061 DATA AAISP/-I . 4464 7 •. 36975. 1. 783.-1.44647 •. 56025.1.783. 94
95 1 0 . 67075. I. 78260. 1. 783. O. 83572 . I. 68737. 1. 783. 95
96 1 0 . 83572. -0. 75737.1. 7 83. O. 67075. -0. 85260 . 1. 7831 96
97 0000061 DATA LENGTH_NEUTRALI 2 .34 I 97
98 0000061 DATA BBIl/-1. 21717. -1. 524 . O . • -1. 21717. 1.524. O.• -0 . 71124.1. 8161. 98
99 1 O. • 1. 9284. O. 2921 . 0 . 0.1. 9284. -0 .2921. O. O. -0 . 71124. -1. 8161. O. Ol 99
100 0000061 DATA FAA.WAA/O . • 0 . • -9 . 81.3*0. 01 100
lOl lOl
102 0000061 DATA SCCTLl . SCCTL2.SCCTL3 . SCCTS.SCCTA.OTOK.OTOTA.OTDTL.OTOTS. 102
103 *OTOW2. THP.TH<l.THR.THX.THY . THZ I 103
104 *6. 1 . 5 . 3 . 10 . 2 • . 1.30 .•. 4 .13. 2.5 . 33 •. 66. 104
105 *12 . 5 • . 0524 •. 0628 • . 0454 • . 17 • . 17 •. 28 1 lOS
106 *12 . 5,6*0. I lOl.
107 0000061 DATA TT 11. .3*0 .• 1. .3*0 .• 1. I 107
108 0000061 DATA RR/I. .3*0 .• 1. .3*0 .• 1. I. W2/12 . 51 108
109 0000061 DATA BETAA/3*0 . I.BETAAI/3*0 . I. WAAI 13*0 . I 109
IlO 0000061 DATA OT. INVDT. GI . 05.20.9 8t l 110
111 0000061 DATA RMAG/6*0 . I. TDUR/lOOO. I lil
112 112
113 0000061 113
114 044FC8I 114

F.11
115 044FOOI GSP SGRT «SCCTLl+SCCTS)**2 . + (1 . -SCCTLl*SCCTS)**2. 115
116 04503A 1 VI 1. /(SCCTLl*SCCTS) 116
117 04'0541 V2 (SCCTLl+SCCTS) I (SCCTLl*SCCTS) 117
118 C 118
119 C*** ******************* BLoeK NO. 1 ****************************** 119
120 C START PROGRAM--USER CHOOSES WASHOUT FILTER . NOTE THAT 120
121 C CLASSICAL ~ ADAPTIVE WASH OUT AT CENTROlO. WHILE OPTIMAL 121
122 C WAS~S OUT AT PILOT 'S HEAO . 122
123 C 123
124 0450741 161 WRITE (lW. 160) 124
125 04'0901 160 FORMAT <l1I12OX. ' WASHOUT FILTER EVALUATION PROGRAM ' I I I 125
126 *2X, 'ENTER DESIRED WASHOUT FILTER <I 1) ' / 2X . 126
127 *'(I=CLASSICAL. 2=OPTIMAL. 3=ADAPTIVE) : ') 127
128 0451141 RE AD (lW. 165) IWASH 128
129 0451341 165 FORMAT ( 11) 129
130 04513EI IF <IWASH. LT. !. OR . IWASH. GT . 31 GO TO 161 130
131 0451621 IREF=1 131
132 04516AI IF <IWASH. EG. 2) IREF=2 132
133 0451821 IF (IREF. EG. I) THEN 133
134 0451921 00 5 1-1 . 3 134
135 04519AI RRS (I )aRRSC (I I 135
136 0451861 SSIIN(I)=SSIINC(II 136
137 0451021 SSI(I)=SSIINC(I) 137
138 0451EEI 00 5 J~I,6 138
139 04'IF6 I AAIS( I. J)=AAISC (I. J) 139
140 04522AI 5 CONTINUE 140
141 045256 [ ELSE 141
142 04525C 1 00 6 1=1.3 142
143 0452641 RRS ( I) =RRSP ( 11 143
144 0452801 SSIIN(I)=SSIINP(I) 144
145 04529CI SSI(II=SSIINP(I) 145
146 0452881 00 6 J=1.6 146
147 0452C0 I AAIS(I.JI=AAISP(I.J) 147
148 0452F41 6 CONTINUE 148
149 0453201 ENDIF 149
150 C 150
151 C NOW CHOOSE THE SIMULATION DURATION : 151
152 C 152
153 0 45320 I 111 WR ITE (lW. 1101 153
154 04533C 1 110 FORMAT(2X. ' ENTER DURATION OF SIMULATION (MAX= 40 SEC) (12) : ') 154
155 0453781 READ (I W. 115) NSEC 155
156 0453981 I I 5 FORMAT (12) 156
157 0453A21 IF (NSEC . GT . 401 GO TO 111 157
158 0453B61 NPOINTS=NSEC*INVDT+l 158
159 C 159
160 e C HOOSE THE TEST INPUTS--FIRST CHOOSE THE CHANN EL TO BE EXCITED 160
161 C 161
162 0453C81 121 WRlTE (lW. 1201 162
163 0453E41 120 FORMAT (2X. ' ENTER CHANNEL TO BE EX C ITED (11)' 1 2X . 163
164 * ' (I=X-ACC. 2=Y-ACC. 3=Z-ACC. 4=ROLL-ACC. 5=PITCH-ACC , 164
165 * ' 6=YAW-ACC I : ' I 165
16b 0454581 REAO ( lW, 1251 NE XC 166
167 0454781 125 FORMAT (! 1) 167
168 0454821 IF ( NEXC . LT 1. OR . NEXC . GT. 61 GO TO 121 168
169 169
170 C NOW CHOOSE THE TYPE OF INPUT ( STEP.PULSE.RANDOM.SINE) 170
171 C 171
172 0454A6 1 131 WRlTE (IW.t301 172
173 0454CO[ 130 FORMA1(2X . ' ENTER TYPE OF INPUT ( l i l ·/2X. '( I=STEP, 2~PULSE . 173
174 *' 3~DBLE PULSE . 4~FILTERED RANDOM . 5=SINE WAVE . 6=COORD . T . I : 'I 174
175 045538 [ READ ( l W, 135) NTYPE 175
176 0455581 135 FORMAT (11 I 176
177 0455621 IF ( NTYPE . LT. 1 . OR . NTYPE . GT . 61 GO TO 131 177
HB 0455861 IF ( NTYPE . EG . 6 . AND . IWASH. EG . 2 ) THEN 178
179 0455A61 WRITE( IW. 1331 179
180 045%01 GO TO 161 180
181 045%41 133 FoRMAT(2X. ' CANNOT RUN OPTIMAL FILTERS ON COoRDINATED '1 181
182 2X. 'TURN; START AGAIN . 'I 182
183 045610 [ ENDIF 183
184 0456161 IF (NTYPE . EG. 61 GO To 132 184
185 C 185
186 C NOW CHoOSE THE MAGNITUDE OF THE INPUT 186
187 C 187
188 04562C 1 WR ITE (iW. 1401 188
189 0456481 140 FORMAT (2X. ' ENTER PEAK MAGNITUDE OF INPUT ' I 189
190 *2X . '( MI SI S FoR LINEAR. RAD /S/ S FOR ANG,ULARI : ') 190
191 04569EI READ (IW.*IRMAG(NEXCI 191
192 C 192
193 C CHOoSE THE INPUT DURAT ION FDR PULSE INPUTS 193
194 194
195 0456CCI IF (N TYPE . EG. 2 . DR . NTYPE. EG . 31 THEN 195
196 0456EC 1 WAITE (IW.150) 196
197 0457081 150 FORMAT(2X. ' ENTER TOTAL DURATION OF PULSE(SI') 197
198 0457341 *
READ <I W. )TOUR 198
199 0457501 ENDIF 199
200 c 200
201 C CHoOSE THE FREGUENCY FoR SINE WAVE INPUTS 201
202 C 202
203 045756 l IF ,NTYPE . EG. 51 TH EN 203
204 0457661 WA ITE (lW. 170 1 204
205 045780 l 170 FORMAT(2X. 'ENTER FREGUENCY OF SINE WAVE (RIS) : ' ) 205
206 0457BO 1 READ (lW •• ) OMEGA 206
207 0457CC 1 ENOIF 207
208 c 208
209 C********************** BLoeK NO. 2 ****************************** 209
210 C NOW SET UP THE INPUT ARRAYS--FIRST FOR STEP AND PULSE INPUTS 210
211 C 211
212 0457021 132 IF ( NTYPE . NE. 4 . AND . NTYPE. NE . 5 . AND . NTYPE. NE . 6) THEN 212
213 0458021 DO 20 1=I.NPOINTS 213
214 0458161 TIME(I)=FLOAT(I-ll*DT 214
21:i 04584E I ISIGN=O 215
216 045856 I IF (NTY PE. EG . 3) THEN 216
217 045866 I IF (TIME(II . GT. O. O. AND . TlME( IJ. LE . TDUR/2 . I ISIGN=1 217
218 0458A8l IF (TIME (I) . GT. TDUR /2. O. AND . TIME (I I. LE. TOUR I ISIGN=-1 218
219 045BE81 ELSE . 219
220 04~8EEI IF (TIME (I I . GT. o. AND . TIME( Il . LE . TOUR I IS IGNal 220
221 0459281 ENDIF 221
222 0459281 AXCA(I)=RMAG(II*ISIGN 222
223 0459441 AYCA(I)=RMAG(21*ISIGN 223
224 0459601 AZCA(II=RMAG(31*ISIGN 224
225 04597C 1 PAAD(I)=RMAG(4)*ISIGN 225
226 0459981 GAAD( I)=RMAG (5)*ISIGN 226
22 ;7 0 45984 I RAAD(II=R MAG (6)* ISIGN 227
228 0459001 20 CONTINUE 228

F.12
229 0459E8 I ENDIF 229
230 C 230
231 C NOW FOR RANDOM INPUTS : 231
232 C 232
233 0459EEI IF (NTYPE. EG. 4) THEN 233
234 0459FEI TCIN\l:0 . 2 234
23~ 045AOAI 1F (NEXC . EG . 4 . 0R . NEXC . EG. 5) TCINV=0.2 235
236 045A36I CALL RANDOM(NPOINTS.DT. TCINV.NEXC.PINK.PINKINT) 236
237 045A601 00 25 1=1. NPOINTS . 237
238 045A74I TIMEn) =FLOAT< 1-1l*DT 238
239 C NO TE THAT HE DIV1SION BY . 425 IS TO OBTAIN A MAGNITUDE OF ONE 239
240 045AAA1 AXCA(I)=RMAG(I)*PINK(I)I . 425 240
241 045AD2 I AYCA( I) =RMAG(2)*PINK( I) I . 425 241
242 045AFAI AZCA(I)=RMAG(3)*PINK(I)I . 425 242
243 045B221 PAAO(I)=RMAG(4)*PINK(I)I . 425 243
244 045B4A1 GAAD(I)=RMAG(5)*PINK(I) I. 425 244
245 045B721 RAAD( I) =RMAG(6)*PINK( I ) I . 425 245
246 045B9A1 25 CONTINUE 246
247 045BB2 I ENDIF 247
248 C 248
249 C FOR SI NE WAVE INPUTS : 249
250 C 250
251 045B88 I IF (NTYPE . EG. 5) TH EN 251
252 045BC81 DO 27 1=I.NPOINTS 252
253 045BDC 1 TIME( I) =FLOAT< I-I )*DT 253
:254 045C121 SINOM-SIN(OMEGA*TIME(I) ) 254
255 045C4A [ AXCA( I) =RMAG( 1 )*SINOM 255
256 045C64 [ AYCA(I)=RMAG(2)*SINOM 256
257 045C7E 1 AZCA(I)=RMAG(3)*SINOM 257
258 045C981 PAAD(I)=RMAG(4).SINOM 258
259 045CB21 GAAD(I)=RMAG(5)*SINOM ' 259
260 045CCC I RAAD(I)=RMAG(6)*SINOM 260
0,61 045CE61 27 CONTINUE 261
262 045CFEI ENDIF 262
263 263
264 AND LASTLY. FOR SIVAN'S TEST CASE 264
265 265
266 0450041 IF ( NT YPE . EG. 6) THEN 266
267 045D141 RAMP2=-1. 267
268 045D201 PO=0 . 2 268
269 045D2C I TMAX=2 . 5 269
270 045D381 AX C A(1 )=0. 270
271 045D441 AYC A(l)=O . 271
2 72 045D501 AZCA<1 )=0. 272
273 045D5C 1 DO 321 1=1.3 273
274 045D641 RRSC (I) =0 . 274
275 045D78I 321 RRS( I) =0 . 275
276 045DA21 DO 323 1=1. NPOINTS 276
277 045DB6 1 TIME (I )=FLOAT< I-I )*DT 277
2 78 045DEEI PAAD (I )=0 . 278
279 045E02I GAAD (I )=0 . 279
280 045E16I RAAD (I )=0. 280
281 045E2AI 1F (TIME(I ), GT 0 O. AND. TIME(I),LE. PO) PAAD(I)=1. 281
282 045E72I IF<TIME(I) . GT . ITMAX-PO ). AND . TIME(I), LE . TMAX) 282
283 PAAO( I )=1. *RAMP2 283
284 C T1ME ( I)=FLOAT ( !-ll*DT- 00001 284
285 C IF (I. EG. I) TIME(!)=O. O 285
286 C IF (TIME(I) . GT O. O. AND . TIME(J) . LE. (PO+ . OOOI) PAAD(J)=l. 286
287 ç 1F ' TIME (I) . GT. !TMAX-PO- . 0001>. AND . TIME ( I), LE . (TMAX- . 0001» 287
288 C PAAD ( I )=1 . *RAMP2 288
289 045EC81 32 3 CO NTINUE 289
29 0 045EEO I ENDIF 290
291 C 291
292 C NOW I NTEGRATE THE ANGULAR RATES. AND OBTAIN EULER ANGLES 292
293 C 293
294 045EE6 1 PHIA ( 1 )~O . 294
295 045EF2 J THETAA ( 1 ) =0. 295
296 045EFE I PSlA" )=0 . 296
297 045FOA r PHIAD ( I ) =0 . 297
298 045F16 I THETAA D (1 ) =0. 298
299 045F22 1 PSIAD( 1 )=0 . 299
3 00 045F2E [ PAA ( l ) =0. 300
301 045F3A I GAA (1 ) =0. 301
302 045F46 1 RAAl l ) =0. 302
303 045F52 I AXPA ( 1 ) =AXCA( I) 303
304 045F5E I AYPA ( 1 ) =AYCA( 1 ) 304
30 5 045F6A I AZPA ( 1 )=AZCA( 1) 305
306 306
307 C*********-!:t************ BLoeK NO . 3 ****************************** 307
308 THE ANGULAR ACCELERATIONS ARE 1NTEGRATED TO YIELD ANGULAR 308
309 RATES USING THE TRAPEZOIDAL RULE : 309
310 045F7 6 I DO 10 1=2. NPOINTS 310
311 045F8AI IF ( NTYPE. EG. 4 . AND . NEXC . GT . 3 ) THEN 311
312 045FAA I PAA ( I)=PINKINT ( I i *RMAG ( 4 ) 312
313 045FCC I GAA (I )=PINK INT< Il*RMAG ( 5 ) 313
314 045FEE I RAA(I)=PINKINT < I)*RMAG(6 ) 314
315 0460101 EL5E 315
316 046016 I PAA ( I ) =PAA ( 1-1 ) +DHO. 5* ( PAAD(! ) +PAAD( I-I) ) 316
317 04605E1 a4A(I)=GAA( I-l)+DT*O S*(GAAD(I)+GAAD(!-I» 317
318 0460A6 I RAA (I )=RAA( 1-1 ) +DHO . 5* IRAAD( I) +RAAD( 1-1» 318
319 0460EE I ENDIF 319
320 C 320
321 0460EEI WAA ( 1) =PAA (I) 321
322 046102 I WAA (2 ) =GAA ( 1> 322
323 0461161 WAA ( 3) =RAA (I) 323
324 C********************** BLoeK NO . 4 ******************************* 324
325 OBTAIN EULER ANGLE RATES FROM ANGULAR RATES : 325
326 04612A 1 CALL VMULT(TT.WAA . BETAAD. 3 . 3. 1) 326
327 C 327
328 C*********** *********** BLoeK NO. 5 ****************************** 328
329 C INTEGRATE EULER ANGLE RATE TO GET EULER ANGLE (TRAPEZOIDAL RULE) 329
330 04616C 1 PH1AD( I )=BETAAD<1) 330
331 0461801 THETAAD(I)=BETAAD(2) 331
332 0461941 PSIAD(I)=BETAAD(3 ) 332
333 0461A81 PHIA( I ) =PH IA( 1-1 )+DHO . 5*( PH1AO( I ) +PHIAD( I-I» 333
334 0461F01 THETAA ( I) =THETAA( I-!l+DT*O . 5+ <THETAAD ( I )+THETAAD( 1-1) ) 334
335 046238 I PS1A( I) =PS IA( I -t)+DHO . 5*( PSIAD( I ) +PSIAD( 1-1) ) 335
336 C UPDATE BETAA TO BE USED IN NEXT LOOP 336
337 0462801 BETAA ( 1 )=PHIA( I) 337
338 046294 1 BETAA ( 2)=THETAA(I) 338
339 0462A8 I BETAA ( 3)=PSIA ( I ) 339
340 c **~* *** *************., ~BLOCY, NO . 6 ****************************** 340
341 C EVALUATE THE T-MA TR IX ' 341
342 C UP DATE TT W!TH THE NE~ ANGLES 342

F .13
343 04b2BCI SINBETl=SIN(BETAA( 1» 343
344 04b2DAI SINBET2=SIN(BETAA(2» 344
345 04b2Fbl SINBET3=SIN(BETAA(3» 345
3410 04103121 COSBETl=COS(DETAA(I» 346
347 04b32EI COSBET2=COS(BETAA(2» 347
348 04b34AI · COSBET3=COS(BETAA(3» 348
349 041036101 TANBET2=SINBET2/COSBET2 349
350 0463781 TT(I,2)mSINBETl*TANBET2 350
351 04638AI TT( 1,3 )=COSBET1*TANBET·2 351
352 04639C I TT(2, 2 ) =COSBET 1 352
353 0463A8 I TT (2, 3 )=-5 INDETl 353
354 0463BAI TT(3,2)-SINBETI/COSBET2 354
355 0463CCI TT(3,3)=COSBETI/COSBET2 355
356 C 356
357 C NOW FORM E5TIMATES OF P,G,R & POOT, GOOT,RDOT TO USE 357
358 C AS VESTIBULAR MODEL INPUTS 358
359 C 359
360 04630EI DO 11 J=I,3 360
361 0463E6 I DETAAOE(J) - W2*(BETAA(J)-DETAAl(J» 3101
362 0464141 11 DETAAI(J)=BETAAl(J)+OT*BETAAOE(J) 362
363 0464581 RR(I,3l=-SINBET2 363
364 04646AI RR (2, 2) -COSBETl 364
365 0464761 RR(2,3)=SINBETI*COSBET2 3105
366 0464881 RR (3, 2 )--SINDETl 31010
367 04649AI RR(3,3)=COSBETl*COSBET2 367
368 04b4ACI CALL VMULT(RR,BETAADE,WAAE,3,3, 1) 368
369 0464FOI DO 13 J=I,3 369
370 0464F8 I WAADE(J)=W2*(WAAE ( J)-WAAI(J» 370
371 0465261 13 WAAI (J )=WAAI(J)+OT*WAAOE(J) 371
372 04656AI PAAE( I )=WAAE( 1) 372
373 04b57EI GAAE( I )=WAAE(2) 373
374 0465921 RAAE(I )=WAAE(3) 374
375 04b5A61 PAADE ( I)=WAAOE(l l 375
376 0465BA I GAAOE( 1)=WAAOE(2) 376
377 0465CEI RAADE< 1)=WAADE(3 ) 377
378 C 378
379 04b5E2 I IF <NTYPE . EG. 10 ) THEN 379
380 0465F21 AXCA(I)=O. 380
381 04101006 I AYCA<I)=G.SINIPHIA(!» 381
382 04101042 I AZCA(I)=-G.SIN(PHIA(I».TAN(PHIA(I) ) 382
383 0466BC I ENOIF 383
384 C 384
385 C********************** BLoeK NO. 7 ****************************** 385
386 C CALCULATE THE PILOT 'S HEAO ACCELERATlONS : 386
387 C 387
388 0466C2 I AXPA <I)=AXCA(ll-RRSC (I).IGAA( l l.GAA ( l l +RAA(ll*RAA ( I » 388
389 +RRSC I2l*( PAA lll*G AA(ll-RAAD(l l) 389
390 +RRSC(3l. ( PAA ( !l·RAA(ll+GAAO(I» 390
391 C 391
392 0467841 AYPA I I )=AYC A( I l +RRSC ( 1 )*l PAA I I) *GAA( I) +RAAD< I l ) 392
393 -RRSC (2) * ( PAA< 1 )*P AA( I l +RAA <I> *RAA ( I l) 393
394 +RRSC<3l* <GAA(I)*RAA<I)-PAAO<ll) 394
395 C 395
396 0468441 AZPA ( ! I =AZCA( ! I+RRSC (I 1* <PAA( I I*RAA< I) -GAAO <I) I 3910
397 +RRSC(21*(GAA(!I*RAA<!I+PAAD(II I 397
398 398
399 0469041 10 CONTINUE 399
400 c 400
401 C*** ******************* BLoeK NO 8 ****************************** 4CI
402 C CALCULATE THE MOTION OF THE REFERENCE FRAME OR!GIN 402
403 C 403
404 04691C I 00 12 1=1, NPOINTS 404
405 0469301 IF <IREF. EG . 1 I THEN 405
406 0469401 AXAA< J) =AXCA< I I 406
407 04695C I AYAA ( I I =AYCA< I ) 4C7
408 0469781 AZAA(II=AZCA<1 1 408
409 0469941 ELSE 409
410 04699AI AXAA(I I= AXPA I! ) 410
411 0469B61 AYAA( I I =AYPA< I) 411
412 0469021 AZAA(I)=AZPAIll 412
413 0469EEI ENOIF 413
414 0469EEI 12 CONTINUE 414
415 C 415
416 C*** ******************* BLOCKS NO. 9 & 10 ************************ 416
417 C NOW OBTAIN THE SPECIFIC FORCES FROM THE ACCELERATIONS : 417
418 C 418
419 046A061 00 15 !=I,NPOINTS 419
420 046AIA! FXAA(I)=AXAA(I)+G*SIN<THETAA(I)I 420
4~1 046A641 FYAACI)=AYAA(I)-G*SINCPHIACI»*CQS<THETAA<I» 421
422 046AEBI FZAA <I) =AZAA(I)-G*COS<PHIAII»*COS<THETAA<I» 422
423 046B6C! FXPA <!) =AXPA(!)+G*SIN<THETAA(!» 423
424 04bBBB I FYPA(! I=AYPA<I)-G*SIN( PHIA<I »*COS< THETAA < I» 424
425 046C3C! FZPA <I)=AZPA<I)-G*COSIPH!A<I».COS<THETAA<I» 425
4210 C 426
427 C CALCULATE THE SPECIFIC FORCE AT THE PILOT ' S HEAO WITH 427
428 C APPROX!MATE ANGULAR RATES ANO ACCELERATIONS (FPAEll 428
429 c 429
430 04bCCOI FXPAEI(II=AXCA<I)-RRSC<I)* <GAAE<I)*GAAE<I)+RAAE<II*RAAE(I» 430
431 +RRSC ( 2)*<PAAE<!)*GAAE<I)-RAADE<!» 431
432 +RRSC(3 )*( PAAE<I)*RAAE(I)+GAAOE<I» 432
433 • +G*SIN(THETAA(II) 433
434 434
435 0460B4! FYPAEI(1)=AYCA<!)+RRSC(l)*(PAAE(!)*GAAE(I)+RAAOE(1I) 435
436 -RRSC(2)*<PAAE(I)*PAAE<I)+RAAE(I)*RAAE(I» 436
437 * +RRSC(3).<GAAE<!).RAAE<I)-PAAOE<I» 437
438 * -G*SIN(PHIA<!) )·'COS< THETAA< I» 438
439 C 439
440 046EOAI FZPAE1(1)~AZCA(I)+RRSC<1)* < PAAE(I)*RAAE<I)-GAAOE(I» 440
441 +RRSC(2).<GAAE<I)*RAAE<I)+PAAOE(I» 441
442 -RRSC ( 3 ) * I PAAE(I).PAAE<ll+GAAE(I)*GAAE<!I) 442
443 -G*COS(PH IA< I ) ).COSITHETAA( I) ) 443
444 C 444
445 C CALCULATE THE SP . FORCE AT THE PILOT ' S HEAD W/ EXACT 445
446 C ANGULAR RATES BUT APPROXIMATE ANGULAR ACCELERATIONS (FPAE2) 446
447 C 447
448 0470021 FXPAE2 (I I=AXCA< I )-RRSC (1)* (GAA( I ).GAA< I ) +RAA< I )*RAAI! I) 448
449 * +RRSC(2)*(PAA<I)*GAA<I)-RAADE(I» 449
450 • +RRSC(3).(PAA<I).RAA(!)+GAADE(I» 450
451 • +G*SIN<THETAA(I» 451
452 C 452
453 0 47OF4! FYPAE2(1)=AYCA<II+RRSC(I)*(PAA(I)*GAA(!)+RAAOE(I» 453
454 -RRSC (2). (PAA·<I ).PAA< I )+RAA< I ).RAA( I) ) 454
45~ * +RRSC(3)*<GAA<I)*RAA(!)-PAADE(!» 455
456 -G*SIN<PH IA< I) ).COSITHETAA( I) ) 456

F.14
457 457
458 04721AI FZPAE2 (ll-AZCA (I )+RRSC (1)* (PAA (I )*RAA( I )-(lAAOE (I» 458
4~9 * +RRSC (2)* (GAA (I )*RAA( ll+PAAOE (I» 459
460 -RRSC (3)* (PAA (I )*PAA( I )+GAA( I )*<lAA(l) ) 460
461 -G*COS(PHIA(I»*COSITHETAAII» 461
462 0473421 15 CONTINUE 462
463 C 463
464 C********************** BLOCK NO , 11 ***************************** 464
465 C GENERATE A/C PILOT'S VESTIBULAR RESPONSÉ. SET 465
466 C ICONTsl FOR PASSING TO INTERP THROUGH COMMON BLOCK CON TROL 466
467 C 467
468 04735AI ICONT m l 468
469 0473621 IF (IWASH. EG. 1) THEN 469
470 0473721 DO 16 I=I,NPOINTS 470
471 0473861 FXPAP(I)=FXPAEI(I) 471
472 0473A2 I FVPAP(I)=FVPAE1(I) 472
47 3 0473BEI 16 FZPAP( 1 )=FZPAE1(I) 473
474 0473F2 I CALL VESTIBINPOINTS.TIME.FXPAE1. FVPAE1. FZPAE1,PAAE. OAAE. 474
475 RAAE.SFXPA. SFVPA.SFZPA.SPPA.SQPA, SRPA) 475
476 04743C I CALL VESTIB1(NPOINTS.TIME.FXPAE1,FVPAE1.FZPAE1.PAAE,GAAE. 476
477 * RAAE. SFXPAI, SFVPAI. SFZPAI. SPPAI . SGPAI. SRPAI ) 477
478 0474881 ENDIF 478
479 04748EI IF ( IWASH. EG. 2) TH EN 479
480 04749E I DO 17 I=I.NPOINTS 480
481 0474B21 FXPAP(I )=FXPA(I) 481
482 0474CEI FVPAP( I )=FVPA( I) 482
483 0474EA 1 17 FZPAP(I)=FZPA(ll 483
484 04751EI CALL VESTIB I NPOINTS,TIME . FXPA,FVPA.FZPA.PAAE.GAAE.RAAE, 484
485 SFXPA,SFVPA,SFZPA, SPPA , SGPA,SRPA) 485
486 0475681 CALL VESTIBIINPOINTS.TIME. FXPA.FVPA.FZPA.PAAE,OAAE.RAAE. 486
487 * SFXPAI. SFVPAI. SFZPAI. SPPAI. SOFA I. SRPAI) 487
488 0475B41 ENDIF 488
489 0475BAI IF I IWASH . EO. 3 ) TH EN 489
490 0475CA 1 DO 18 I-I.NPOINTS 490
491 04 750E 1 FXPAPII)=FXPAE211) 491
492 0475FA I FYPAP<I)=FYPAE2<1) 492
493 047616 I 18 FZPAPII )=FZPAE211) 493
494 04764A I CALL VESTIB INPOINTS . TIME . FXPAE2. FVPAE2. FZPAE2. PAA. GAA. 494
495 * RAA. SFXPA, SFVPA . SFZPA. SPPA. SOFA. SRPA) 495
496 0476941 CALL VESTIB1(NPOINTS.TIME,FXPAE2.FVPAE2,FZPAE2.PAA.OAA. 496
497 RAA , SFXPAI . SFVPAI , SFZPAI , SPPAI. SOPAI . SRPAI) 497
498 0476E0 I ENDIF 498
499 C 499
500 C WASH OUT A /C Man ONS Ta GET SIMULATOR MOTIONS USING 500
501 C WOLF ' S CLASSICAL WASHOUT " JACK LENGTH ALGORITHMS . 501
502 C 502
503 C 503
504 C SET INITIAL VALUES 504
505 C 505
506 0476E61 XSI (1 ) =0. 506
507 0476F21 YS1(1 )sO. 507
508 0476FEI ZSIO )=0 . 508
509 04770A 1 VXSl( 1) =0 . 509
510 0477161 VY8l( 1) =0 . 510
51 1 0477221 VZSl( I) =0 . 511
512 04772E 1 AXSI (I ),,0 . 512
513 04773AI AYSI< I ) =0 . 513
514 0477461 AZS U 1) =0 . 514
515 047752 1 FXPSI I l =0 . 515
5 16 04775E I FYPS I 1) =0. 516
517 04776A I FZPS I I )=-G 517
518 04777A I PHIS I I )=0 . 518
519 0477861 THETAS 1 I ) =0. 519
520 0477921 PSIS I I)=O 520
521 04779C I PSS (I) =0. 521
522 0477A81 OS5 ( 1 ) =0. 522
523 04 77B4 r RSS II)=O 523
524 c 524
525 e START THE WASHOUT PROCESS 525
526 C 526
527 0477CO I DO 30 1=2. NPOINTS 527
5 28 c 528
529 0477041 FAA I I ) = FXAA 1 I) 529
530 0477E8 I FAA(2)=FVAAII) 530
531 0477FC I FAA (3) =FZAA (I) 531
532 532
533 047810 I WAAO ) =PAA I I) 533
534 047824 I WAA (2) =QAA I I) 534
535 0478381 WAA (3) =RAA I I) 535
536 c 536
537 04784C I WAAD ( I ) =PAAO< I) 537
538 047860 I WAADI 2 )=OAAO( I ) 538
539 0478741 WAAO l3 ) =RAAO 1 I ) 539
540 C 540
541 0478881 BETAA( I ) =PHIAI I) 541
542 04789C I BETAA (2) -THETAA ( I ) 542
5 43 0478BO I BETAA ( 3 ) =PSIAI I ) 543
544 C 544
54 5 04 78C4 I AAM I ) =AXAA I I ) 545
546 0478081 AAA (2)=AYAAII l 546
547 0478EC 1 AAA (3)=AZAA (I) 547
5 48 c 548
549 c THE APPROPRIATE WASHOUT ROUTINE IS NCW CALLED. 549
550 e USE WASHOUTI FOR FILTERING IN THE INERTIAL FRAME. 550
551 c AND WASHOUT2 FOR FILTERING IN THE CAB FRAME. 551
552 c AND WASHOUT FOR CAB ANO INERTlAL FILTERING 552
553 C 553
554 C CALL WASHOUTI (OT. FAA. WAA . SSI) 554
555 C CALL WASHOUT210T. FAA.WAA. SSI) 555
556 C********************** BLoeK NO . 12 ***************************** 556
557 C CALL THE APPROPRIATE WASHOUT FILTER 557
558 047900 I IF I IWASH. EG. I) CALL WASHOUTIOT. FAA.WAA.SSI) 558
559 0479341 IF (IWASH. EG. 2) CALL WASHOPTIOT,AAA.BETAA.SSI) 559
560 0479681 IF ( IWASH. EO. 3) CALL WASHAOPT(OT.FAA.WAA.SSI.SSIIN) 560
561 C**.******************* BLOCK NO . 13 ***************************** 561
562 OBTAIN THE ACTUATOR EXTENS IONS FROM SIMULATOR POSITIONS : 562
563 0479AOI CALL JACKORVR(SSI.LI . LID.LIOO. OT) 563
564 0479C8 1 XSIII)=SSIII)-SSIINII) 564
565 0479E2 I YSIII)=SSI(2)-SSIIN(2) 565
566 0479FC 1 ZSI(I)=SSI13)-SSIIN(3) 566
567 04 7AI6 I VXS I< I )=VSI (1) 567
568 047A2A I VYSI< 1 )=VSI (2) 568
569 047A3EI VZS III )=VSI13 ) 569
570 047A521 AXS I ( I ) =AS 1(1 ) 570

F .15
:171 047Ab61 AYSl< I )..,<\SII2) 571
:172
:173
574
047A7AI
c...................... BLoeK
C
AZSl< 1 )..,<\SI<3)

OBTAIN LSI BY TRANSPOSINQ LIS

NO. 14 .............................. 572
:173
:174
575 047A8EJ CALL VTRANSPILLIS.LLSI.3.3) :17:1
576 c........................ BLoeK NO . 1~ ............................... 576
577 C TRANSFOR~ ASI TO BODY AXES: 577
578 047ACOI CALL VMULTILLSI.ASI.ASS.3.3.1) 578
579
580
581
c...................... BLoeK
C

C
NO. 16 .............................
OBTAIN APS. THELPILOT'S HEAD ACCELERATIONS. FROH ASS
579
580
581
582 C 582
583 0478041 APSII)-ASSII)-RRSII).IWSSI2).WSSI2)+WSSI3).WSSI3)) 583
584 • +RRSIO!).IWSS(1).WSSI2)-WSSDI3)) 584
585 • +RRS(3).IWSSll).WSSI3)+WSSDI2)) 585
586 C 586
587 047B661 APS(2)-ASS(0!)+RRS(I).IWSS(I).WSSI2)+WSSDI3)) 587
588 -RRS(2).IWSSII).WS8(1)+WSSI3).weSI3)) 588
589 +RRS(3).(WSS(2).WSS(3)-W8SD(I)) 589
590 c 590
591 047Be61 APS(3)-ASSI3)+RRS(I).IWSS(I).WSS(3)-WSSD(2)) 591
592 +RRS(2).IWSS(2).WSS(3)+WSSDII)) 592
593 -RRS(3).(WSS(I).WSSII)+WSSI2).WSS(2)) 593
594 c 594
595 c . . . . . . . **** ••••••••••• BLoeK NO . 17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59:1
596 C OBTAIN SPECIFIC FORCES AT THE PILOT'S HEAD : 596
597 047e26I FXPSII)=APSll)+G.SINTH 597
598 047C44I FYPS(I)=APS(2)-G.SINPHI.cOSTH 598
599 047C6 .. 1 FZPS(I)=APS(3)-G.COSPHI.cOSTH 599
"600 C bOO
bOl 047C901 "PH IS( I )=XN (4) bOl
602 047CA41 " THETAS (!)=XN( 5) 602
603 047CB81 PSIS(I)=XN(6) b03
604 c 604
b05 047ecAI PSS ( I ) -wes ( \) b05
606 047CDEI GSS (I) -WSS (2) 606
b07 047CF21 RSS (I) =WSS (3) 607
608 608
609 0470061 DO 40 -J=1.6 609
610 04700EI 40 R-JL< r. -J)=Ll (-J) 610
611 ecccccccceccccccCCccccccccccccccccc ceccccccccccececeecceeeee 611
612 047D4C I PX IPL< I )=PXI 612
.,13 047D601 PX2PL ( I ) =PX2 613
614 047D741 PX3PL< I)=P X3 614
615 0470881 PY1PL < [)=PYl 615
616 04709C I PY:2PL ( [)=PY2 616
617 047DOOI PV3PL< I ) =pv3 617
618 047OC41 PZPL< l)=PZ "618
619 0470081 PPSIPL ( I)=PPSI 619
620 ccccccceccceccccecceeccececcccceecceccccccecccccececcceccec 620
621 047DECI 30 CONTINUE 621
622 c**~***.*****.********* BLOC~ NO 18 ***.**.******* ••••• ** •••••• ** 622
623 C GENERA TE SIMULATOR PILOT 'S VESTIBULAR RESPONSE " SET 623
624 C ICONT=2 TO PASS ra INTERP THROUGH eot1~ON BLOCK CONTROL. 624
62~ 625
626 047E041 ICONT=2 626
627 047F.:OC 1 eALL VESTIB(NPOINTG. fIME . FXPS . FYPS.FZPS.PSS.GSS,RSS.SFXPS . 627
628 *SFyP5. SFZPS, SPPS. SGPS, SRPS) 628
.,O!q 629
630 CALL VESTlBt(NPOINTS, TIME, FXPS, FYPS , FZPS, PSS, GSS, RSS , 630
.,31 .SF XPSI , SFVPSI,SFZPSI ,S PPS1 , SGPS1 , SRPSI) 631
<>32 c 632
633 C CALL OUTP TO PLOT O.. T....5 NECESSARV 633
.,34 C 634
63~ cccccccccecceccccccccccceccccccccccccccccccccceccccecccccceccccccc 635
.,36 047E.. 41 PXIPL ( I)=PXIPL(2 ) 636
037 047EBO [ PX2PL< I ):PX2PL<2) 637
638 047EBC 1 PX3PL< I ) =PX3PL (2) 638
639 047EC81 PVtPL l l ) =PVIPLI2 ) 639
640 047E041 PV2PL < 1)=PV2PLI2) 640
641 047EEOI PY3PL I I ):PV3f'L (2) 641
642 047EECI PZPL< I )=PZPLl2) 642
643 047EF81 PPSIPL(I)=PPSIPLI2 ) 643
644 cccccccccececcccccccccccccccecccccccccccccccceccceccccccccccccccccc 644
645 C********************** BLOCK NO. 19 *********.******************* 645
646 C C"LL SUBROUTINE OUTP TO PLOT THE DESIRED RESUL TS : 646
647 C 647
648 047F041 CALL OUTP(IR,IW,NPOINTS,TJME . .. XCA . AYCA,AZCA, 648
649 .FXP .. , FYPA, FZP .. ,PHIA,THETAA,PSIA,PAA,O .. A,R ..... 649
6S0 .SFXP" , SFYP",SFZPA, SPPA,SGP",SRPA. 650
651 .SF XP" t. SFVPI\l, SFZP"I , SPP"I, SGPAI, SRP"I, 651
652 *AXSI.AVSI, AZSI . FXPS , FVPS. FZPS. 652
653 ..vXSI , VVSI , VZSI , ~SI, VSJ, ZSI , PHIS, THETAS, PSIS ; 653
654 *PSS.OSS . RSS,SFXPS,SFYPS .SFZPS, SPPS.SGPS.SRPS. 654
655 .SFXPSI , SFVPSI,SFZPSI , SPPSI,SGPSI,SRPSI,R-JL , 655
656 *PXIPL , PX2PL,PX3PL,PVIPL,PV2PL, PV3PL. PZPL.PPSIPL) 656
657 c 657
658 0480241 STOP 65B
659 04802C I END 659

NO ERRORS : F70 R O~ -OO " 00 M.. INPROG " ~IN 12 / 12/84 14" 35 : 39 TABLE SPACE: 15 KB
STAT~ENT BUFFER : 20 LlNES/1321 BYTES ST .. C~ SPACE : 185 WORDS
SINGLE PREe 1610N FLOATING PT SUPPORT REGUIRED FOR EXECUTION

F .16
 1 0000001 SUBROUTINE RANDOM(NPOINTS.DT.TCINV.NEXC.PINK.PINKINT) 1
2 C 2
3 C THIS SUBROUTINE WILL QENERATE A RANDOM SIQNAL AND 3
4 C FILTER IT TO YIELD A TEST INPUT TO BE USED IN 4
3 C THE WASHOUT FILTER TEST PROClRAM. IHSL ROUTINE QQNML 5
6 C QENERATES RANDOH-NORMAL (0 . 1) SAMPLES WHICH WE 6
7 C ASSlA1E WERE GATHERED AT 20 SPS . 7
8 C 8
9 0000041 DIMENSION PINK(*).PINKINT(*) 9
10 0000041 DIMENSION YFIL(2). IWK(2).WK(26).YPRIME(2) 10
11 0000041 REAL*4 ..... 2 ( 1001 ) 11
12 0000041 EXTERNAL FCN3.FCNJ 12
13 0000041 COMMON /PINKIN/ AA.BB.CC.DD.WN2 13
14 000004 I CALL GGNML(31859.DO.NPOINTS+200.WN2) 14
15 C 15
16 C DGEAR IS NOW USED TO SOLVE THE DIFFERENTlAL EGUATION 16
17 C REPRESENTING THE WHITE NOISE FILTER. AND VALUES OF 17
18 C THE SIGNAL ARE STORED AT A RATE OF 20 SPS 18
19 C 19
20 C INITIALIZE VAR lABLES 20
21 C 21
22 OOOODOI CC=12 . 5 22
23 OOOODC I DD=TCINV 23
24 0000E81 IF (NE XC . EG. 1 . OR.NEXC . EG . 2 . ~ . NEXC . EG. 3) THEN 24
25 0001181 M=O . 25
26 0001241 BB-CC*DD 26
27 0001361 ELSE 27
28 00013C I AA-CC*DD 28
29 00014EI BB-O. 29
30 00015A I ENDIF 30
31 00015A I NFE<l-2 31
32 0001621 T=O . 32
33 00016EI YFIL< 1)=0. 33
34 0001781 YFIL(2)=O. 34
35 0001821 HD . 0001 35
36 00018EI TOL-. OOOOI 36
37 00019AI METH=2 37
38 0001A21 MITER-2 38
39 0001AAI INDEX= 1 39
40 0001B21 DO 12 I-I.NPOINTS 40
41 0001C61 TEND-FLOAT( Il*DT 41
42 0001E81 CALL DGEAR (NFEG. FCN3. FCNJ. T.H. YFIL.TEND.TOL. 42
43 *METH.MITER. INDEX. IWK.WK.IER) 43
44 0002341 IF (IER . GT . 100) STOP 44
45 00024C I IF (NEX C. EG. 1. OR. NEXC . EG. 2 . OR . NEXC . EG. 3) THEN 45
46 00027C I PINK(I ) =YFIL< I ) 46
47 0002941 PINKINT(I)-O. 47
48 0002AEI ELSE 48
49 0002B4 I CALL FCN3(NFEG.TEND.YFIL.YPRIME) 49
50 0002D81 PINK( I) =YPR lME( I) 50
51 0002FOI PINKINT(I)-YFIL(I) SI
52 0003081 ENDIF 32
S3 0003081 12 CONTINUE 53
54 0003201 RETURN 54
55 0003261 END 55
NO ERRORS : F7D ROS-OO. OO SUBROUTINE RANDOM 11/12/84 13: 25 : 05 TABLE SPACE : 2 KB
1 0000001 SUBROUTINE FCN3(N.T.Y.YPRIME ) 56
2 0000041 DIMENSION Y(N) . YPRIME(N) 57
3 OOOOOAI REAL*4 ..... 2 ( 1001 ) 58
4 OOOOOAI COMMON /PINKIN/ AA. BB. CC. OD.WN2 59
5 c 60
6 C FIRST INTERPOLATE FOR WHITE NOISE VALUES AT TIME T . 61
7 C ASSlA1E THAT THE STOREO VALUES WERE SAMPLED AT 20 SPS. 62
8 C 63
9 OOOOOA I ITO=INT(T*20. ) 64
10 00005AI TO=FLOAT< I TO) /20 . 65
11 00007C I WN2L=0 . 66
12 0000881 IF (ITO. GT . O) WN2L= ..... 2(ITO) 67
13 OOOOAC 1 WN2U=WN2( ITO+l) 68
14 000OC21 WN2T- WN2L+(T-TO)*(WN2U-WN2L)/0.05 69
15 C 70
16 C NOW SOLVE FOR Y (1) 71
17 c 72
18 OOOOEEI YPRIME(I)--(CC+DD)*Y(I)+Y(2)+AA*WN2T 73
19 0001321 YPRIME(2)= -CC*DD *Y(I) +BB*WN2T 74
20 00016C I RETURN 75
21 0001721 END 76

NO ERRORS : F7D ROS-OO. OO SUBROUTINE FCN3 11 /12/84 13: 25 : 06 TABLE SPACE : 2 KB

STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 130 WORDS
SINGLE PRECISION FLOATING PT SUPPORT REGUIRED FOR EXECUTION

F.17
FORTRAN-VIID ROS-DO . OO
11/12/84 13:2S : 02 PAGE 11
FORTRAN VIID : LICENSED RESTRICTED RIGHTS AS STATED IN LICENSE CL-D013
***. SEE DOCUMENTATION PACKAGE. 04-101M99.
1 0000001 SUBROUTINE FCN~(N.X.Y . PDl
2 C DUMMY SUBROUTINE REOUIRED BY DGEAR 77
3 0000041 DIMENSION Y(Nl.PD(N.Nl 78
4 OOOOOAI RETURN 79
5 0000321 END 80
81
NO ERRORS:F7D R05-00. 00 SUBROUTINE FCN~ 11/12/84 13: 25 : 06 TABLE SPACE: 1 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE: 48 WORDS

F. 18
 1 0000001 SUBROUTINE VESTIB(LINE.TIME.FXP.FYP.FZP.P.G.R.SFX. I
2 *SFY . SFZ.SP.SG.SR) 2
C 3
3
C ROUTINE TO RETURN VEST , RESPONSE GIVEN DRIVER'S HEAD MOTIONS 4
4
C 5
5
6 0000041 EXTERNAL ' FCNI.FCN~~ 6
7 0000041 DI MENS ION TIME (*). FXP( *). FYP(*). FZP (*). P(*). G( *). R (*). 7
*SFX (*). SFY (*). SFZ< *). SP (*). SG( *). SR (*). IWK (15). WK( 1(5). Y( 15) 8
8
COMMDN IVS TBLR I SCCTL1.SCCTL2. SCCTL3.SCCTS.SCC TA.OTOK.OTOTA. 9
9 0000041
10 *DTOTL.OTOTS.OTOW2.THP.THG.THR.THX.THY.THZ 10
11 0000041 COMMDN/COEFF/SCCRAO.SCCRAI.SCCRA2.SCCB2.SCCPAO.SCCPA1.SCCPA2. 11
*SCCYAO.SCCYAI.SCCYA2.0TOAO.OTOAI.OTOBO.OTOBI 12
12
13 C 13
14 C********************** BLoeK NO . Al ****************************** 14
15 C GENERATE THE VESTIBULAR TRANSFER FUNCTION COEFFICIENTS 15
16 C 16
17 0000041 SCCRAO=I , /(SCCTLI*SCCTS*SCCTA) 17
18 00036C I SCCRAI=I , I (SCCTLI*SCCTS)+I , I(SCCTA*SCCTLI)+1. I (SCCTA*SCCTS) 18
19 0003B21 SCCRA2=1. ISCCTLl +1. ISCCTS +1 , ISCCTA 19
20 0003EOI SCCB2~1. ISCCTS 20
21 C 21
22 0003F2 I SCCPAO=1. I (SCCTL2*SCCTS*SCCTA) 22
23 0004121 SCCPAI=I , I (SCCTL2*SCCTS)+I , I(SCCTA*SCCTL2)+I , /(SCCTA*SCCTS) 23
24 0004581 SCCPA2=1 , ISCCTL2 +I , /SCCTS +1 , ISCCTA 24
25 c 25
26 0004861 SCCYAO-I , /(SCCTL3* SCCTS*SCCTA ) 26
27 0004A6 1 SCCYAI-I , I(SCCTL3*SCCTS)+I , /(SCCTA*SCCTL3)+1./(SCCTA*SCCTS) 27
28 0004EC 1 SCCYA2-1. ISCCTL3 + 1. ISCCTS +1. ISCCTA 28
29 29
30 00051AI OTOAO=I , /(OTOTL*OTOTS) 30
31 0005341 OTOAI=I , / OTOTL +1 , l OTOTS 31
32 0005541 OTOBO-OTOK/(OTOTL*OTOTS) 32
33 00056E 1 OTOB1-0TOK*OTOTA/(OTOTL*OTOTS) 33
34 C 34
C*** ******************* BLoeK NO . A2 ***************************** 35
3' C SET UP FOR THE CALL TO DGEAR WHICH WILL EVALUATE THE 36
36
37 C 15 VESTIBULAR MODEL DIFFERENTlAL EGUATIONS : 37
3B C 38
39 00058EI NEG=15 39
40 0005961 T=TIME ( I) 40
41 0005A6 I DO 10 1=1. NEG 41
42 0005BAI 10Y(I>~O , 42
43 0005E4 I Y(14)=OTOK*FZP(I) 43
44 0005FBI Y(15 )- OTOAI*Y(14)-OTOBI*FZP(I) 44
45 0006181 SP ( U-Y(l) 45
46 0006261 SG (I )-Y(4) 46
47 0006341 SR (U=Y(?> 47
48 0006421 SFX( ll=Y( 10) 48
49 0006501 SF Y<1 ) -Y( 12) 49
50 00065EI SFZ(1 ) =Y(14) '0
51 00066CI H= , OOOI 51
52 0006781 TOL=, 00001 52
53 OOQ684t METH=2 53
54 00068C 1 MITER=3 54
55 0006941 INDEX=I 55
56 00069C 1 DO 15 1=2. LlNE 56
57 0006BOI TEND=T lME( I) 57
CALL DGEAR (NEG. FCNI. FCN~~ . T. H. Y. TEND. TOL. METH. 51l
58 0006CAI
59 *MI TER . INDEX. IWK . WK. IER) 59
60 0007141 IF ( IER , GT , 100) THEN 60
61 0007241 WRITE(2.99) I . NEG.T . H. TEND. TOL. METH. MITER. INDEX 61
62 0007641 STOP 62
63 00076C 1 ENDIF 63
64 000772 [ 99 FORMAT<2X • • 1= '. 13. 3X • • NEG- ' . 12. 3X. ' T= ' . F8, 3 . 3X. 'H=·. F8 , 5. 64
65 *3X . ' TEND= ' . FB, 3 / 2X . ' TOL= · . F8, ,. 3X . ' METH=', 12. 3X. 'M ITER = ' . 65
66 *12. ' INDEX='. 12) 66
67 000708 I INDEX=2 67
6B C 68
69 C CALCULATE THE SENSED SPECIFIC FORCES AND YAW RATE 69
70 C 70
71 0007EO I SP (Il = Y( 1) 71
72 0007F81 SG(!)=Y(4) 72
73 0008101 SR (!)-V(?> 73
74 0008281 SFX ( I )-V(IO) 74
75 0008401 SFV(I >=Y(12 ) 75
76 0008581 l ' SF Z ( I ) =Y ( 14) 76
77 0008881 RETURN 77
78 00088EI END 78

NO ERRORS : F70 RO S-CO , OO SUBROUTINE VESTIB 12 / 12 / 84 14: 36 : 02 TABLE SPACE : 4 KB

STATEMENT BUFFER : 20 LlNES / 1321 BYTES STACK SPACE : 136 WORDS
SINGLE PRECISION FLOATING PT SUPPORT REGUIRED FOR EXECUTION

F .19
 1 0000001 SUBROUTINE FCNl(NEG. T.Y.YPRIME) 79
2 C
3
80
C THIS ROUTINE OEFINES THE VESTIBULAR MODEL DIFFERENTlAL EGNS 81
4 C
~ 0000041 82
DIMENSION Y(NEG).YPRIMECNEG) 83
6 OOOooAI COMMON/COEFF/SCCRAO.SCCRAl.SCCRA2.SCCB2.SCCPAO.SCCPAI.SCCPA2. 84
7 *SCCYAO.SCCYAl.SCCYA2.0TOAO.OTOAI.OTOBO.OTOBI
8 OOOooAI 85
COMMeN ISIZE/ LINE 86
9 C
10 87
C********************** BLoeK NO . A3 ***************************** 88
11 C CALL INTERP TO INTERPOLATE FOR HEAD MOTIONS AT ANY TIME
12 89
C
13 OOOooAI 90
CALL INTERPCLINE.T.FXPT.FYPT.FZPT.PT.GT.RT) 91
14 C
1~
92
C THE FIRST 3 EGUATIONS ARE THE SCC MODEL. ROLL CHANNEL : 93
16 C 94
17 0000601 YPRIMECI)=-SCCRA2*YCI)+YC2)+SCCB2*PT 95
18 00009EI YPRIME(2)=-SCCRAI*YCI)+YC3)
19 0000021 96
YPRIME(3)=-SCCRAO*YCI) 97
20 C
98
21 C THE NEXT 3 EGUATIONS ARE THE SCC MODEL. PITCH CHANNEL :
22 99
C
23 000OF81 100
YPRIME(4)=-SCCPA2*YC4)+YC~)+SCCB2*GT
24 00013EI lOl
YPRIMEC~)=-SCCPAI*YC4)+YC6)
25 0001761 102
YPRIME(6)=-SCCPAO*YC4) 103
26 C
27 104
C THE NEXT 3 EGUATIONS ARE THE SCC MODEL. YAW CHANNEL : 105
28 C
29 0001AOI 106
YPRIME(7)=-SCCYA2*YC7)+YC8)+SCC B2*RT 107
30 000lE61 YPRIME(8)=-SCCYAl*YC7)+YC9)
31 00021EI 108
YPRIME (9)=-SCCYAO*YC7) 109
32 C
33 110
C THE NEXT TWO EGUATIONS ARE THE OTOLITH. SURGE CHANNEL : lil
34 C
35 0002481 112
YPRIMECIO)=-OTOAI*YCIO)+Y CII)+OTOB1*FXPT 113
36 00028EI YPRIME(11)=-QTOAO*YCI0) +OTOBO*FXPT
37 114
C
38 115
C THE NEXT TWO ARE FOR THE OTOLITH. SWAY CHANNEL : 116
39 C
40 0002C61 117
YPRIMECI2)=-OTOA1*Y(12)+YCI3)+OTOB1*FYPT 118
41 00030C I YPR IME C13) =-QTOAO*YC 12 ) +OTOBO*FYPT
42 119
C
43 120
C THE LAST TWO ARE FOR THE o TOL! TH. HEAVE CHANNEL : 121
44 C
45 0003441 122
YPRIMECI4)=-OTOAI*YCI4)+YCI5)+OTOB1*FZPT 123
46 00038AI YPR IME C15) =-QTOAO*YC 14 ) +OTOBO*FZPT
47 124
C
48 0003C2 I 125
RETURN 126
49 0003C8 I END 127
NO ERRORS : F7D R0 5 -00 . 00 SUBROUTINE FCNI 12/12/84 14: 36: 08 TABLE SPACE : 2 KB
STATEMENT BUFFER : 20 LlNES / 1321 BYTES STACK SPACE : 123 WORDS
SINGLE PREC lSION FLOATING PT SUPPORT REGUIREO FOR EXECUTION

I 0000001 SUBROUTINE FCNJJCN. X,Y,PD)

2 128
DUMMY SUBROUTINE REGUIRED BY DGEAR 129
3 0000041 DIMENSION YCN),POCN,N)
4 OOOooAI RETURN 130
5 000032 I END 131
132
NO ERRORS : F7D R0 5 -00. 00 SUBROUTINE FCNJJ 12 /12/ 84 14: 36: 08 TABLE SPACE . 1 KB
STATEMENT BUFFER : 20 LlNES / 1321 BYTES STACK SPACE : 48 WORDS

F.20
 1 0000001 SUBROUTINE VESTIBI(LINE.TIME. FXP . FYP.FZP.P . O.R.SFX. 1
2 *SFY.SFZ . SP.SO.SR) 2
3 C 3
4 C ROUTINE TO RETURN VEST . RESPONSE GIVEN DRIVER'S HEAO MOTIONS 4
5 C THIS MODEL INCLUDES THE INTERMEDIATE THRESHOLO BLOCK. :;
6 C 6
7 0000041 EXTERNAL FCN2.FCN~~~ 7
8 0000041 DI MENS ION TIME (*). FXP( *) . FYP(*). FZP (*). P(*). 0(*). R (*). 8
9 *SFX(*) . SFV (*). SFZ(*). SP(*). SO(*). SR(*) ; IWK( 18) . 101<.( 198). Y( 18) 9
10 0000041 eOMMON I VSTBLR I SCCTLI.SCCTL2.SCCTL3.SCCTS . SCCTA.OTOK.OTOTA. 10
11 *DTOTL . OTOTS. OTOW2. THP . THO. THR.THX . THY.THZ 11
12 0000041 COMMON / COEFFI/SCCRAO. SCCRA1 . SCCBI.SCCPAO. SCCPA1 . 12
13 *SCCYAO.SCCYAI.OTOAO.OTOAI.OTOBO 13
14 0000041 COMMON I SENSEI SRLIM. SPLIM.SYLIM.SFXLIM.SFYLIM.SFZLIM 14
15 C COMMON ITESTI FXPT . FX1.FXLIM.SSFXLIM 15
16 C 16
17 C********************** BLoeK NO . Al *******.*.******************* 17
18 0000041 SCCRAO=I ./ (SCCTL1*SCCTS) 18
19 000402 1 SCCRAI=l . I SCCTL1+1 . I SCCTS 19
20 0004221 SCCBI= 1. I seCTS 20
21 e 21
22 0004341 seCPAO=I ./ (SCCTL2*SCCTS) 22
23 00044EI seCPAI=I . I sceTL2+1 . I SCCTS 23
24 C 24
2~ 00046EI SeCYAO =I./ (SCCTL3*seCTS) 25
26 0004881 SCe YA1 =1. ISCCTL3+1. ISCCTS 26
27 C 27
28 0004A8 I OTOAO=l . / (OTOTL*OTOTS ) 28
29 0004C2 I OTOAI=I . /OTOTL +1 . l OTOTS 29
30 0004E2 I OTOBO=OTOK*OTOTA/( OTOTL*DTOTS ) 30
31 C 31
32 C*** ******************* BLoeK NO. A2 ***************************** 32
33 C SET UP FOR THE CALL TO DGEAR WHICH WILL EVALUATE THE 33
34 C VESTIBULAR MODEL DIFFERENTlAL EOUATIONS : 34
35 e 35
36 0005021 NEO=18 36
37 00050C I T=TI ME ( I) 37
38 00051C I DO 10 1=1. NEO 38
39 0005301 10 Y( 1l=0 . 39
40 00055AI Y(16)=OTOK*OTOTA*FZP (I) 40
41 00057 41 Y(17)=OTOAI*V(16) 41
42 000582 I V( 18)=-OTOW2*(V(16)+THZ ) 42
43 00059EI SP (1l =0 . 43
44 OOOSAEI SQ (1l =0 . 44
45 0005BE I SR(ll=O. 45
46 0005CEI SFX (I) =0. 46
47 0005DEI S F YO) =0. 47
4B OOOSEE I SFZ(I)=OTOK*FZP(I ) +THZ / OTOT A 48
49 0006161 H= . OOOI 49
50 0006221 TOL=. OOool 50
51 00062E 1 ME TH=2 SI
52 000636 1 MITER=3 52
53 00063EI INDEX=I 53
54 000646 1 DO 15 1=2. LINE 54
55 00065AI TEND=TIME ( I) 55
56 0006741 CALL DGEAR(NEQ.FCN2 . FCN~~~ , T .H, V,TEND,TOL.METH, 56
57 *MI TER , INDEX , IWK. WK. IER ) 57
5a 0006C01 INDEX=2 58
:;9 C 59
60 G CALCULATE THE SENSED SPECIFI C FORCES AND YAW RATE 60
61 C 61
62 C IF (TI ME (I) . GT . 8 . 0 . AND . TIME(!) . LT . 18.) THEN 62
63 C WRITE(6, 100) TIME(I ),FXPT. FXI, FXLIM. SSFXLIM 63
64 C 100 FORMAT< 2X, 'TIME=·,F8. 3 , 3X, 'MODEL INa·,FIO. 4,3X. 'LIMITER IN- ', 64
65 C F10. 4 , 3 X, ' LIMITER OUTo' , FIO. 4 , 3X, 'MODEL OUT= ' ,FIO . 4) 65
66 C ENDIF 66
67 0006C8 1 SP (I) =SRLIM 67
68 0006E21 SQ( I )=SPLIM 68
69 OOObFC l SR (I )=SVL IM 69
70 00071 6 1 SFX ( I ) =SFX LIM 70
71 0007301 SF Y ( I ) =SFVLIM 71
72 00074A I 15 SFZ ( I ) =SF ZLIM 72
73 00077C I RETURN 73
74 0007821 END 74

NO ERRORS : F7D R05 -00 .00 SUBROUTINE VESTIB I 12/12/84 14: 36 : 34 TABLE SPACE : 4 KB
STATEMENT BUFFER 20 LINES/1321 BYTES STACK SPACE : 150 WORDS
SINGLE PRECI SION FLOATING P T SUPPORT REOUIRED FOR EXECUTION

F .21
 1 0000001 SUBROUTINE FCN2CNEG.T.V.VPRIME) 75
2 C 76
3 C THIS ROUTINE DEFINES THE VESTIBULAR MODEL DIFFERENTlAL EGNS 77
4 C THIS VERSION INCLUDES TfE INTERMEDIATE THRESliOLD BLOCK 78
~ C 79
6 0000041 ' DIMENSION VCNEG).VPRIMECNEG) 80
7 OOOooAI COMMON IVSTBLRI SCCTLl.SCCTL2.SCCTL3.SCCTS.SCCTA.OTOK.OTOTA. 81
8 +OTOTL.OTOTS.OTOW2.THP.THO.THR.THX.THV.THZ 82
9 OOOooAI COMMON/COEFFt/SCCRAO. S'CCRAI. SCCBl. SCCPAO. SCCPAI. 83
10 *SCCVAO.SCCVAl.0TOAO.OTOAl.0TOBO 84
11 OOOooAI COMMON IBENSEI SRLIM.SPLIM.SVLIM.SFXLIM.SFVLIM.SFZLIM 85
12 OOOooAI COMMON ISIZE/ LINE 86
13 C COMMON /TEST/ FXPT. FXl. FXLIM. SSFXLIM 87
14 C 88
15 C********************** BLoeK NO. A3 ***************************** 89
16 C CALL INTERP TO INTERPOLATE FOR fEAD MOTIONS AT ANV TIME 90
17 C 91
18 OOOooAI CALL INTERPCLINE.T.FXPT.FYPT.FZPT.PT.GT.RT) 92
19 C 93
20 C THE FIRST 3 EGUATIONS ARE TfE SCC MODEL. ROLL CHANNEL: 94
21 C 95
22 0000601 YPRIME(1)=-SCCRA1*VCl)+Y(2)+SCCBl*PT 96
23 00009EI YPRIME(2)=-SCCRAO*V(I) 97
24 C 98
25 000OC4 I IF (Y( 1) . QT. O. ) THEN 99
26 OOOOOAI RLIM~AMAXl(O . • Y(I)-THP) 100
27 00011EI EL SE 101
28 0001241 RLIM=AMINICO . • YCl)+THP) 102
29 000166 I ENDIF 103
30 C 104
31 0001661 VPRIME(3)=-1./SCCTA*(Y(3)+RLIM) 105
32 00019EI SRLIM=Y(3)+RLIM 106
33 C 107
34 C THE NEXT 3 EGUATIONS ARE THE SCC MODEL. PITCH CHANNEL : 108
35 C 109
36 0001B81 YPRIME(4)=-SCCPAl*Y(4)+Y(5)+SCCB1*GT 110
37 0001FEI YPRIME(5)=-SCCPAO*Y(4) 111
38 C 112
39 0002281 IF CY(4).Q~0 . ) TH EN 113
40 0002421 PLIM=AMAXICO . • Y(4)-THQ) 114
41 00028A1 ELSE 115
42 0002901 PLIM=AMINl(O . • Y(4)+THG) 116
43 0002061 ENOIF 117
44 C 118
45 0002061 YPRIME(6)=-I . / SCCTA* CY(6)+PLIM) 119
46 00030EI SPLIM=Y ( 6)+PLIM 120
47 C 121
48 C THE NEXT 3 EQUATIONS ARE THE SCC MODEL, YAW CHANNEL : 122
49 C 123
50 0003281 YPRIME(7)=-SCCYA1*Y(7)+Y(8)+SCCB1*RT 124
SI 00036E 1 YPRIME(8)=-SCCYAO*Y(7) 125
52 C 126
53 0003981 IF (V(7) . GT . O. ) THEN 127
54 0003821 YLIM=AMAX1(O . ,V(7)-THR) 128
55 0003FAI ELSE 129
56 0004001 VLIM=AMINl(O . ,V(7)+THR) 130
57 0004461 ENDIF 131
58 C 132
59 0004461 YPRIME(9)=-I . /SCCTA*(Y(9)+VLIM) 133
60 00047El SYLlM=Y(9)+YLIM 134
61 C 135
62 C THE NEXT 3 EQUATIONS ARE THE OTOLITH. SURGE CHANNEL : 136
63 C 137
64 0004981 YPRIME (10) =-QTDA1*Y (10) +v C11) 138
65 0004001 VPRIME(11)=-OTOAO*VCI0)+OTOBO*FXPT 139
66 C FX1=Y(10) 140
67 C 141
68 0005081 IF (YC lOl. GT. O. ) THEN 142
69 000~221 FXLIM=AMAXI (0 .• Y(10)-THX) 143
70 000~6AI ELSE 144
71 0005701 FXLIM=AMINI (0. , Y(10)+THX) 145
72 000~B61 ENDIF 146
73 C 147
74 0005B6 I YPRIME(12)=-OTDW2*(Y(12)+OTDW2*FXLIM) 148
75 OOO~C 1 SFXLIM=V( 12 )+( OTDW2+1 . / DTOTA)*FXLlM 149
76 C SSFXLIM=SFXLIM 150
77 C 151
78 C THE NEXT 3 ARE FDR THE DTDLlTH. SWAY CHANNEL : 152
79 C 153
80 0006181 VPRIME(13)=-OTDA1*Y(13)+Y(14) 154
81 0006~01 YPRIMECI4)=-OTDAO*YCI3)+DTDBO*FYPT 1~5
82 C 156
83 0006881 IF ( V( 13) . GT. O. ) THEN 157
84 0006A21 FYLIM=AMAXl ( 0 . • YCI3)-THY ) 158
85 0006EAI ELSE 159
86 OOObFOI FYLIM=AMINICO . • YCI3)+THY) 160
87 0007361 ENDIF 161
88 C 162
89 0007361 YPRIME(15)=-QTDW2*(Y(15)+OTDW2*FYLIM) 163
90 00076C 1 SFYLlM=Y( 15)+C DTDW2+1. /DTDTA)*FYLI M 164
91 C 165
92 C THE LAST 3 ARE FDR THE DTOLITH . HE AVE CHANNEL : 166
93 C 167
94 0007981 YPRIME(16)=-OTDAl*Y(16)+Y(17) 168
95 0007021 YPRIME(17)--OTDAO*Y(16)+OTOBO*FZPT 169
96 C 170
97 00080C 1 IF (Y(16) . GT. 0 . ) THEN 171
98 0008261 FZLIM=AMAXI (0 . • V(16)-THZ) 172
99 00086EI ELSE 173
100 0008741 FZLIM=AMINI (0 . • V(16)+THZ) 174
101 0008BAI ENOIF 175
102 C 176
103 0008BAI YPRIME(18)=-OTOW2*CYCI8)+OTDW2*FZLIM) 177
104 000BF41 SFZLIM=Y( 18)+COTOW2+1. /OTOTA)*FZLlM 178
105 C 179
106 0009221 RETURN 180
107 0009281 END 181

NO ERRORS : F7D RO~-oo . OO SUBROUTINE FCN2 12/12/84 14:36:42 TABLE SPACE : 3 KB

STATEMENT BUFFER : 20 LlNES/1321 BYTES STACK SPACE: 154 waROS
SINGLE PREC ISION FLDATlNG PT SUPPDRT REGUIRED FOR EXECUTlON

F .22
 1 0000001 SUBROUTINE FCN~JU(N,X,V,PD)
2 C DUMMV SUBROUTINE REOUIREO BV DGEAR IS2
3 0000041 DIMENSION V(N) , PD(N,N) 183
4 OOOooAI RETURN 184
:I 0000321 END 18~
186
NO ERRORS : F7D R05-oo . 00 SUBROUTINE FCN~~~ 12/12/84 14 : 36 : 43 TABLE SPACE : 1 KB
STATEMENT BUFFER : 20 LINES/1321 BVTES STACK SPACE : 48 WORDS

1 0000001 SUBROUTINE INTERP(LINE,T,FXPT,FVPT , FZPT,PT,OT,RT)

2 C 1
3 C ROUTINE TO INTERPOLATE FOR THE DRIVER'S HEAD MOTION AT 2
4 C TIME T 3
5 0000041 COMMeN leONTROLI ICONT 4
6 0000041 eOMMeN ITIMEI TIME(B01) ~
7 0000041 eOMMeN IAMOTIONI FXPA(801),FYPA(801),FZPA(801) , 6
8 *PAA(801),OAA(801),RAA(801) 7
9 0000041 eOMMeN ISMOTIONI FXPS(80U , FVPS(80U, FZPS(80U, 8
10 *PSS ( 801), OSS(80I), RSS(80U 9
11 C 10
12 C FIRST FIND THE LOCATION OF THE INTERVAL eONTAINING 11
13 T 12
C
14 0000041 DO 10 I-I,LINE 13
15 0000181 IF CT. LT. TIME(I) THEN 14
16 0000321 lU=I I~
17 00003EI IL= I-I 16
18 00004CI 00 TO 15 17
19 0000521 ENDIF 18
20 0000~8I 10 CONTINUE 19
21 C 20
22 C NOW INTERPOLATE FOR FXPT,FVPT,FZPT,PT,OT,RT 21
23 C 22
24 0000701 I~ IF (leONT. EO. 1) THEN 23
2~ 0000801 FXPT-FXPA(IL)+(T-TIME(IL»+(FXPA(IU)-FXPA(IL»I 24
26 * CTIME( lU)-TlME (lU) 25
27 0000E4I FYPT-FVPA( IU+(T-TIME( IU)* (FVPA(lU)-FVPA( IU)I 26
28 + (TIME( IU)-TlME(IU) 27
29 0001481 FZPTaFZPA( IU+( T-TlME( IL»* (FZPA( IU )-FZPA( IU) 1 28
30 + (TIME( IU)-TlME(IL» 29
31 OOOIAC I PT-PAA( IU+(T-TIP1E( IU )*(PAA(IU)-PAA(IU)I 30
32 * CTIME( IU)-TIME(IU) 31
33 0002101 OT~OAA( IU+(T-TIME( IU) *(OAA( IU )-OAA(IL»I 32
34 * (TIP1E( IU)-TlME(IU) 33
35 0002741 RT-RAA( IU+ (T-TIP1E( IU ) *(RAA( IU )-RAA( IU ) 1 34
36 * (TIME( IU)-TIME( IU) 35
37 0002081 ELSE 36
38 00020EI FXPT-FXPS( IU+(T-TIME( IU)* (FXPS(IU)-FXPS( IU)I 37
39 * (TIME( IU)-TlME(IU) 38
40 0003421 FYPT-FVPS(IL)+(T-TIME(IL»*(FVPS(IU)-FVPS(IL»I 39
41 * (TIME( IU)-TlME( IU) 40
42 0003A6 I FZPTaFZPS( IU+( T-TlME( IU)* (FZPS( IU )-FZPS( IL» 1 41
43 * (TIME(IU)-TlME(IU) 42
44 00040AI PT-PSS( IU+ (T-TIME( IU ) *(PSS( IU )-f'SS( IU)I 43
4~
* (TIME(IU)-TIME(IL» 44
46 00046EI 4~
OT-OSS ( IU+(T-TIME( IU )*(OSS(IU )-QSS(IU)I
47 * (TIME( IU)-TlME( IU) 46
48 0004021 RT-RSS( IU+(T-TIME( IU )*(RSS(IU)-RSS(IU)I 47
49 * (TIME(IU)-TIME(IL» 48
50 0005361 ENDIF 49
51 000~36 I RETURN ~O
:12 00053C I END 51
52
NO ERRORS : F7D R05-00.00 SUBROUTINE INTERP 11/12/84 13: 22:22 TABLE SPAeE: 2 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPAeE: 170 WORDS
SINGLE PRECISION FLOATING PT SUPPORT REOUIRED FOR EXEeUTION

F.23
 1 C------- - CLASSICAL WASHOUT ALQORITHM WITH LINEARIZ
2 C---- ATIONS 353
QIVEN FM AND WM. COMPUTE SSI AND EULER ANGLES.
3 C 354
4 C FILTERS 2ND ORDER IN TRANSLATION ON FM 355
5 C FIRST ORDER +DOUBLE INTEGR. IN INERTlAL FRAME IN 356
6 C TRANSLATION 357
2ND ORDER IN ROTATION ON WAA
7 C FIRST ORDER + INTEGRATION IN INERTIAL FRAME 358
8 C 359
9 0000001 SUBROUTINE WASHOUT(DT. PM.WM. SSI) 360
10 c 361
11 0000041 REAL*4 FAA(3). A1<3l.F2 (3). F1(3) 362
12 0000041 REAL*4 WAA(3).W l(3) 363
13 0000041 REAL*4 FIMAX(3) .WIMAX(3 ) 364
14 0000041 REAL*4 KW1.KF1.KF2 365
15 000004 I REAL*4 AH(3).AH I<3) . AHll(3) 366
16 0000041 REAL*4 ~I(3).WHII(3).WH(3) 367
17 0000041 REAL*4 FLD(3).F L(3).FLM AX(3) 368
18 0000041 REAL*4 SSID(3) 369
19 0000041 REAL*4 A2(3). ASI(3) 370
20 0000041 REAL*4 BETASL(3) . BETASH( 3).BETAS (3).BETAH DOT(3) 371
21 0000041 REAL*4 TT(3.3).V SI(3) . SSI(3) 372
22 0000041 REAL*4 MIS(3.6) .BBII(3.6 ).LLIS<3 .3).FUNC (6).DFUN 373
23 0000041 C(6. 6).LENGT H(6) 374
REAL*4 LENGTH_NEUTRAL
24 0000041 REAL*4 XN(6).XN P1(6) 375
25 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 376
26 0000041 CCCCCCCCCCCCCCCCCCCCC 377
REAL*4 BETAS1( 3).BETAS D(3).RR(3 .3)
27 0000041 REAL*4 WSSI(3).W SSO(3).W SS(3) 378
28 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 379
CCCCCCCC
29 0000041 COMMeN I MATRIXI/AAIS. BBII. LLIS.FUNC.DFUNC. LENG TH. CCCCCCCCCCCCC 380
30 1 XN.XNPI. 381
LENGTH_NEUTRAL
31 0000041 COMMeN ISINCOS/ SINPSIS.C OSPSIS.S INTHETA S. COSTHET 382
32 0000041 COMMeN IFILTER/ ASI.VSI AS.SINPH IS.COSPH IS 383
33 0000041 COMMeN IANGDERI WSS.WSSD 384
34 0000041 EQUIVALENCE (BETAS. XN(4» 385
35 C 386
36 0000041 EQUIVALENCE (PHISL. BETASL( 1». (THETASL. BETASL(2» 387
37 C 388
38 0000041 DATA TT/I . • 8*0 . Ol. RLS20.RL S21 / 9 . 6 1.8. 681 389
39 0000041 DATA THS20. THS21/9 . 61. 8.681 i LP FILTER CONSTS 390
40 0000041 i 2ND ORDER HP TRANSL. 391
DATA RHS20.RH S21/1 . 0404.2. 041
41 i 2ND ORDER HP ROT ' L
000004 I DATA . RHIlOIO. 21 392
i lRST ORDER HP ROT
42 000004 I DATA THII0/0. 21 393
43 0000041 DATA AHI.AHI I/3*0 . 0.3*0.01 394
44 0000041 DATA WHI/3*0. Ol 395
45 0000041 DATA WHII/3*0 . Ol 396
46 0000041 DATA FLD. FL / 3*0. 0.0 . • 0 . • -9 . 81 1 397
47 0000041 DATA BETASL / 3*0. Ol. BETASH / 3*0. Ol 398
48 0000041 DATA SSID/O . • O. • 0 . I i INITIAL VELOCITV IN INERTlAL
399
49 0000041 DATA KW1.KF1 .KF2/1..1 . .1 . I FRAME 400
50 0000041 DATA F1MAX.WIMAX/3*100 . • 8 . 29.8 . 29. 6 . 541 401
51 000004 I DATA FLMAX/3*1 . I TPD 5640 402
52 0000041 DATA VSI/3*0. 0 /. G/9 . 811 403
53 0000041 DATA SINPHIS.S INTHETA S .SINPSIS /O .• O.• 0 . I 404
54 0000041 DATA COSPHIS. COSTHETAS . COSPSIS /!. • 1. . 1 . I 405
55 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 406
56 0000041 CCCCCCCCCCCCCCCCCCC 407
DATA BETASI /3*0. I
57 0000041 DATA WSS1I3*0 . I 408
58 0000041 DATA RR /I . • 8*0 . I.W2/12. 5 / . TIME/O. I 409
59 CCCCCCCCCCCCCCCCCCCCCCCCCCCC CCCCCCCCCCCCCCCCCCCCCCCC 410
60 C
CCCCCCCCCCCCCCC 411
61 C******* ******** ******* BLDCK NO . B 1 ******** ******** 412
62 C---- SCALE AND LIMIT SPECIFIC FORCE FOR HP SEC TI ON
******** ***** 413
63 C 414
64 0000041 DO 10 1:1 .3 415
65 0001B41 Fl(l) a AMAX1(- FIMAX(!l .AMIN1(K F1*FAA(I ).FIMAX( 416
66 0002441 10 I») 417
CONTINUE
67 C 418
68 C******* ******** ******* BLOCK NO . 82 419
69 ******** ******** ******** ***** 420
C---- SC ALE LOW PASS SPECIFIC FORCE
70 C 421
71 00025AI F2 (1) FAA(I)*K F2 422
72 00026C 1 F2 (2) = FAA(2)*K F2 423
73 00027EI F2 (3) = FAA(3) 424
74 C 425
75 C******* ******** ******* BLOCK NO. 83 426
76 ******** ******** ******** ***** 427
C---- LP FILTER FOR F2 -> FL
77 C 428
78 0002BC 1 DO 50 -1:1.2 429
79 0002941 FLDD RLS20*(F 2(-I) - FL(-I» - RLS21*FL D(-I) 430
80 0002C81 FLD-I FLD(,,) + DT*FLDD 431
81 0002E6 1 FLD-I AMAX1(-FLMAX(-I) . AMIN1(FLMAX(-I).FLD-I» 432
82 C IF(-I. EQ . 2) FLD""=FL D" 433
83 0003521 FLD (,,) = FLD" 434
84 0003641 FL(-I) FL(-I) + DT*FLD-I 435
85 0003881 50 CONTINUE 436
86 00039EI FL(3) = F2 (3) 437
87 C 438
88 C******* ******** ******* BLoeK NO . 84 439
89 ******** ******** ******** *****
C---- HP SCALE AND LIMIT WAA 440
90 C 441
91 0003A6 1 DO 12 1=1 . 3 442
92 0003AEI Wl(l) = AMAX1(-W IMAX(I).A MIN1(KW I*WAA(I) .W1MAX( 443
93 0004401 12 I») 444
CONTINUE
94 C 445
95 c******* ******** ******* BLoeK NO . 85 ******** ******** 446
96 C---- HP FILTER 2ND ORDER FOR Wl -> WH ******** ***** 447
97 C 448
98 0004561 DO 40 1=1.3 449
99 00045EI WH(I) = Wl(l) - RHS21*W HI(I) - RHS20*W HII(I) 450
100 00049AI WHI(!) = WH!(!) + DT*WH( I) 451
101 0004C4 1 WHI!(I) - WHII(I) + DT*WHI( I) 452
102 0004EEI 40 CONTINUE 453
103 C 454
104 C******* ******** ******* BLOCK NO. 86 455
105 C---- COMPUTE TILT VARIABLES ******** ******** ******** ***** 456
106 C 457
107 0005041 THETASL = FL< 1) IG 458
108 0005101 PHISL =-FL(2)IG 459
i BETASL< 1)
109 C 460
110 C---- COMPUTE EULER ANGLES BASED ON ANGLES AT TIME N-l ) 461
111 C 462
112 0005241 TANT • SINTHETAS/COSTHETAS 463
113 C 464
114 C******* ******** ******* BLQeK NO. 87 465
******** ******** ******** ***** 466

F.24
115 0005521 TT(I.2) = SINPHIS*TANT 115
116 0005621 TT ( 1. 3) = COSPHIS* TANT 116
117 0005721 TT<2 ; ;1) - eOSPHIS 117
118 00057C I TT (2.3) =-SINPHIS 118
119 00058CI TT(3.2) = SINPHIS/COSTHETAS 119
120 00059C I TT(3. 3) ~ COSPHIS/COSTHETAS 120
121 C 121
122 C*** tt******_**** __ *_*** BLoeK NO . B8 ***************************** 122
123 0005AC I CALL VMUL T (TT. WH. BETAHOOT. 3.3. 1) 123
124 124
125 C********************** BLoeK NO. B9 ***************************** 125
126 C--- HP FILTER ANO INTEGRATE BETAHOOT 126
127 C 127
128 0005FOI DO 90 1=1.3 128
129 000~F81 BETAHO = BETAHOOT( J) - RHi1D*BETASH( I) 129
130 00061C I BETASH( I) = BETASH( I) + OT*BETAHO 130
131 0006401 BETAS (1) = BETASH( 1) + BETASL ( J) ; <- CURRENT EULER ANG 131
132 C********************** BLoeK NO . F 1 *tttt********_** ___ *"*_"**_** 132
133 0006661 BETASO( I ) =W2*(BETAS( 1 )-BETAS 1 (I) ) 133
134 0006901 BETASI ( I ) =BETASI (I )+OT*BETASO<I) 134
135 0006BAI 90 CONTINUE 135
136 C 136
137 C PREMUL TIPLY BETASD BY RR TO GET P . Cl. R 137
138 C 138
139 000600 I RR (1 .3 l=-SINTHETAS 139
140 0006EOI RR (2. 2) =COSPHI S 140
141 0006EA I RR(2 . 3)-SINPHIS*COSTHETAS 141
142 OOObFAI RR ( 3 . 21--SINPHIS 142
143 00070AI RR ( 3 . 3 )=COSPHIS~OSTHETAS 143
144 C 144
145 00071AI CALL VMULT(RR. BETASD. WSS. 3. 3 . 1) 145
146 146
147 C NOW USE THE OIFFERENTIATOR + L . P. FILTER TO GET WSSO 147
148 148
149 00075C I DO 55 1-1.8 149
150 0007641 WSSO( I )QW2*(WSS(I l-WSSI <I) ) 150
151 0007901 WSSl ( I )=WSSl (I )+DT*WSSD ( I) 151
152 0007BC I 55 CONTlNUE 152
153 CCC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCccccccceeeccecccc 153
154 c 154
155 .: ... ******************* BLOC .... NO . BlO **tt******"***_****_******** 155
156 ç --- COMPUTE LLIS 156
157 157
158 0007021 CALL L IBS 158
159 159
160 C*** ******************* BLoeK NO Bl1 ***** ..*-**-**-****** . . ***** 160
161 c---- FORM TRANSLATIONAL ACCELERATION VECTOR TO BE HIGH PASS FIL TEREO 161
162 C 162
163 00070AI AI (I) = Fl (1 I - G*SINTHETAS 163
164 0007EEI Al (2 1 = Fl(2) + G*COSTHETAS*SINPHIS 164
165 0008061 AI (31 = Fl (3) + G*COSTHETAS*COSPHIS 165
166 166
167 C********************** BLoeK NO 812 ***********-**************** 167
168 c---- HP FILTER 2NO ORDER FOR AI -; AH 168
169 169
170 00081EI DO 30 1:1.3 170
171 0008261 AH ( I I = Al (O - THS 2 1*AHI(II - THS20*AHII(II 171
172 0008621 f, HI< I ) AH I< 1 ) • OT*AH (I ) 172
173 00088C I AHl!(! 1 = AHII<l l + OT*AHI(l) 173
174 0008B61 30 CONTINUE 174
175 175
176 ( ...... ******************* BLOCK NO B 13 **************************** 176
177 C---- TRANSFORM HIGH PASS ACCELERATION TO INERTIAL FRAME . 177
178 178
179 ooosec I CALL VMUL T (LLIS. AH. A2. 3 .3. 1) 179
180 180
181 181
182 C********************** BLOCK NO . B 14 **************************** 182
183 FIRST ORDER HIGH PASS FILTER A2. 183
184 AND INTEGRATE FOR POSITION IN ONE STEP 184
185 185
186 0009101 DO 70 1 = 1.3 186
187 0009181 SS100 A2 ( I ) - THI 10*SSIO( I ) 187
188 00093C I SSIO(I)= SSIO ( I) + OT*SSIOO 188
189 0009601 SSI(I) SSI(II + OT*SSIO(II 189
190 00099AI VST(I) = 5510(1) 190
191 0009B41 ASl ( I) 55IOD 191
192 0009C81 70 CONTINUE 192
193 193
194 00090E1 RETURN 194
195 0009E4 1 END 195

NO ERRORS : F7D ROS -DO . OO SUBROUTINE WASHOUT 15/03/85 10: 43 : 18 TABLE SPACE: 5 KB
STATEMENT BUFFER 20 LINES / 1321 BYTES STACK SPACE . 200 WOROS
SINGLE PRECISION FLOATING PT SUPPORT REClUIREO FOR EXECUTION

F.25
 1 C**** SUBROUTINE WASHOPT . FTN
2
**** 1
C
3 C OPTIMAL WASHOUT ALGORITH1. 2
GIVEN AAA AI-a> BETAA.
4 C COMPUTE SSI AND SIMULATOR EULER ANGLES. 3
:5 C 4
6 C FILTERS CALL IN INERTlAL 5
FRAME):
7 C 6TH ORDER PHIA TO PHIS 6
8 C 6TH ORDER A2(2) TO PHIS 7
9 C 6TH ORDER PHIA ' 8
TO AVSI
10 C 6TH ORDER A2(2) TO AVSI 9
11 C 6TH ORDER THETAA 10
TO THETAS
12 C 6TH ORDER A2CI) TO THETAS 11
13 C 6TH ORDER THETAA TO AXSI 12
14 C 6TH ORDER A2C I) TO AXSI 13
15 C 4TH ORDER PSIA TO PSIS 14
16 C 4TH ORDER A2(3) TO AZSI 15
17 C 16
18 0000001 SUBROUTINE WASHOPTCDT.AAA.BETAA.SSI) 17
19 0000041 REAL*4 AAA(3). A2(3).BET AAC3). BETAAOLD(3). 18
20 *SSI(3).V SIC3).AS I(3).A2O LDC3) 19
21 0000041 REAL*4 XIC 10). XIOC 10). X2CIO). X20ClO). X3CIO). X30CIO). 20
22 *X4 C10). X40 (10). X5(8). X50(8) . X6 (8). X6Q(8) 21
23 0000041 REAL*4 BETAS(3 ).ASIOLD (3).VSIOL D(3) 22
24 0000041 REAL*4 AAISC3. 6). BBIl (3.6) • LLIS(3. 3). FUNC (6). DFUNC 23
25 *LENGTH (6) . LENGTH_NEUTRAL C6. 6). 24
26 0000041 REAL*4 XN(6),XN P1<6) 25
27 0000041 REAL*4 LLIAC3.3 ) 26
28 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 27
29 0000041 CCCCCCCCCCCCCCCCC 28
REAL*4 BETASI (3).BETASD C3).RRC3 . 3)
30 0000041 REAL*4 WSSI (3). WSSD(3). WSS (3) 29
31 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 30
32 CCCCCCCCCCCCCCCCC 31
C
33 0000041 COMMeN IMATRIXI I AAIS.BBI I.LLIS.FU NC.DFUN C.LENGT 32
34 H. XN.XNPI. 33
*LENGTH_NEUTRAL
35 0000041 COMMeN ISINCOS I SINPSIS.C OSPSIS.S INTHETA S. COSTHETA 34
36 S. 35
*SINPHIS . COSPHIS
37 0000041 COMMeN IFILTERI ASI.VSI 36
38 0000041 COMMeN I ANGDER I WSS.WSSD 37
39 0000041 EClUIVALENCE (BE TAS . XN(4)) 38
40 0000041 DATA LLIS/I ' .3_~. I .• 3*~. I . 39
41 0000041 DATA. VSII3*0. I.ASIOLD /3+O. I.VSIOLD /3*0 . 40
42 0000041 DATA BETAAOLD/3*0. I.A20LD / 3+O. I 41
43 0000041 DATA XID/I0*0 . I. X:10 / 10*0. I. X30/10*0 .1 . X40/I0*0 42
44 . 1. X50/8*0 . I . 43
*X60/S*O . I
45 0000041 DATA XN/6*0 .1 44
46 0000041 DATA R1NO.RIN I.RIN2.RI N3.RIN4. RIN5.RIN 6.R2NO.R :1NI.R2N 45
47 *R2N4. R2N5.R2N 6 . R3NO . R3Nl.R3N 2.R3N3 . R3N4.R3N 5.R3N6.R4 2.R2N3. 46
48 *R4N2. R4N3 . R4N4 . R4N5.R4N 6. RlDO .RI D1 . RID2.RID 3.RID4.R NO.R4NI. 47
49 * . 0039664 • . 09012571 • . 61391316 . 1. 68823851 .2 . 57059725 ID51 48
50 *2. 29123901 •. 99992324 . . 49
SI *-. 3397934E -03. - . 00813054 . - . 05687075 . - . 15559325 50
52 . - . 22012446 . 51
*-. 12396183 . - . 01035447 .
53 *0 . • O.• 0 . • - . 86898712 E-05 • . 22447328 E-03 • . 00684395 52
54 • . 03142471 . 53
*0 . • 0 .• 0 . • . 00831274 • . 17023042 • . 86387154 • . 8220245
55 . 54
* . 00373134 • . 08614651 . . 60061158 .1. 66604897 . 2 . 54961668
56 0000041 . 2 . 28167431 55
DATA PlNO. PINI . PIN2.PIN 3.PIN4.PI N5.PIN6. P2NO.P2
57 NI . P2N2.P2N 3. 56
*P2N4.P2 N5 . P2N6 . P3NO.P3N I . P3N2. P3N3.P3N 4.P3N5.P3
58 *P4N2.P4 N3.P4N4. P4N5.P4N 6.P1DO.P ID1 . PID2 . PID3.PID N6.P4NO .P4NI . 57
59 * . 00509459 • . 11301084 •. 7377372. 1.948118 98.2 . 8406302. 4.PID5 I 58
60 *2 . 41375911 . . 9999167 . 59
61 * . 47993E-0 3 • . 0109345 • . 07131254 . . 18585781 • . 25083196 60
62 * . 13723155 • . 01140347 . . 61
63 *0 . • 0 . • 0 . • . 32057E-0 4 • . 128476E -03.-. 00626852 .-. 62
64 03176104 . 63
+0 . • O.• O. • . 00986958 •. 193587 • . 90212164 • . 82633128
65 . 64
* . 00507335 • . 11263886 •. 73665436 .1 . 94664651 .2 . 83968359
66 0000041 DATA V1NO.VIN 1 .V IN2.VIN3 .V IN4.VIDO .VIDI.YI D2.YID3 .2 . 41371711 65
67 _0 . • 0 . • . 04402633 . . 54537972 •. 98238384 . 1 66
68 * . 00334042 • . 05499563 •. 27633653 • . 74417721 67
69 0000041 DATA HlNO . HIN1 . HIN2.HIN 3. HIN4.HID O.HIDI.H ID2.HID3 68
70 *0 .• 0 . • 0 . . . 05908784 • . 31829264 . 1 69
71 * . 03696848 • . 32958488 •. 9373682. 1 . 385481 70
72 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CCCCCCCCCCCCCCCC 71
73 0000041 CCCCCCCCCCCCCCCC 72
DATA BETASI/3 *0 . I
74 0000041 DATA WSSI/3*0 . I 73
75 0000041 DATA RR/I. .8*0 . I. W2/12 . 5 1 74
76 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 75
77 C
CCCCCCCCCCCCCCCCCCC 76
78 C******* ******** ******* BLoeK NO. Cl ******** 77
79 C FIRST . FORM LLIA AND PREMULTIPLV AAA TO OBTAIN ******** ******** ***** 78
80 C
A2 79
81 000004 1 CALL LIBACBET AA.LLIA) 80
82 0004081 CALL VMULTC LLIA.AAA .A2.3.3.1) 81
83 C 82
84 C******* ******** ******* BLoeK NO. ca 83
85 ******** ******** ******** *****
C FILTER FROM PHIA TO PHIS : 84
86 C 85
87 00045C 1 ANGAV z
O. 5*CBETAA (1)+BETAA OLDC1)) 86
88 C ANGAV=0. 5*CBETAAOLD(1)+BETAAOLDC1)) 87
89 0004761 XI(1)=XI0 CI)+DT*C -RID5*XI 0CI)+X20 CI)+CRIN 5-RID5*R 88
90 0004AEI XIAV=O. 5*CXICI) +XIOC1)) IN6)*ANG AV) 89
91 C XIAV=O . 5*C XIOC 1 )+XI OC1)) 90
92 0004C21 X2CI)=X2 0CI)+DT* C-RID4*X IAV +X3Q( 1 )+(RIN4-R ID4*RIN6 91
93 0004FCI X3CI)=X3 QCI)+DT* C-RID3*X IAV +X40(1)+C RIN3-RID )*ANGAV ) 92
94 0005361 3*RIN6)*A NGAV) 93
X4C U=X4Q( I )+DT*C-R ID2*XIAV +X50CI)+ (RIN2-RID
2*RIN6)*A
95 0005701 X5CI)=XS O(I)+DT* (-RIDI*X IAV +X6QCI)+ CRINI-RID I*RIN6)*ANGAV) 94
96 OOOSAAI X6CI)=X6 QCI)+DT* C-RIDO*X IAV NGAV) 95
97 +CRINO-R IDO*RIN6 )*ANGAV )
C 96
98 C******* ******** ******* BLoeK NO . C3 ******** 97
99 C FILTER FROM A2(2) TO PHIS : ******** ******** ***** 98
100 C 99
lOl OOOSEOI A2AV=0 . 5+(A2C2) +A20LD(2 )) 100
102 C A2AV=0 .5+C A20LD(2)+ A20LD C2)) lOl
103 0005F4 I XI(2)=XIO C2)+DT*C -RID5*XI O(2)+X20 (2)+CR2N 5-RID5*R 102
104 00062CI XIAV=O. 5*( XlC2 )+XIO(2 ) ) 2N6)*A2A V) 103
105 C XIAV=O . 5*CXIO( 2)+XI OC2) ) 104
106 0006401 X2(2)-X2 0(2)+DT* (-RID4*X IAV +X3Q(2)+ CR2N4-RI D4*R2N6) lOS
107 00067AI X3(2)=X3 QC2)+DT* C-RID3*X IAV +X40(2)+C R2N3-RID *A2AV) 106
3*R2N6)*
108 0006841 X4(2)mX4 QC2)+DT* (-RID2*X IAV +XSO(2)+(R2N~-RID~*R2N6) A2AV) 107
*AC!AV)
109 OOObEEI X5(2)=XS OC2)+DT *(-RID1*X IAV +X6Q(2)+ (R2NI-RID 1*R2Nb)* 108
110 0007281 X6(2)-X6Q C2)+DT*( -RIDO*XI AV A2AV) 109
111 +(R2N0-R IDO*R2N6 )*A2AV)
110
112 C******* ******** ******* 1i1
113 C
BLoeK NO . C4
FILTER FROM PHIA TO AYS!:
__ ***.*••• ******** ***.**** *** ' 112
114 C 113
114

F.26
115 00075EI XI (3)=Xl0(3)+DT*(-RID5*Xl0(3)+X20(3)+(R3N5-R1D5*R3N6)*ANGAV) liS
116 0007961 XIAV=O . 5*( XI (3)+XID(3» 116
117 C XIAV=0 . 5*(XIO(3)+XIO(3» 117
118 0007AAI X2(3)=X20(3)+DT*(-RID4*X1AV +X30(3)+(R3N4-RID4*R3N6)*ANGAV) 118
119 0007E4 I X3(3)=X30(3)+DT*(-RID3*XIAV +X40(3)+(R3N3-RID3*R3N6)*ANGAV) 119
120 00081EI X4(3)=X40(3)+DT*(-RID2*XIAV +X50(3)+(R3N2-RID2*R3N6)*ANGAV) 120
121 0008581 X5(3)=X50(3)+DT*(-RIDI*XIAV +X6Q(3)+(R3NI-RID1*R3N6)*ANGAV) 121
122 0008921 X6(3)=X6Q(3)+DT*(-RIDO*XIAV +(R3N0-R1DO*R3N6)*ANGAV) 122
123 C 123
124 C*** ******************* BLoeK NO. C5 *****************************. 124
125 C FILTER FROM A2(2) TO AYSI : 125
126 C 126
127 000ec8I XI (4)=XI0(4)+DT*(-RID5*XI0(4)+X20(4)+(R4NS-RIDS*R4N6)*A2AV) 127
128 0009001 XIAV=0 . 5*(X1<4)+XIO(4» 128
129 C XIAV=0 . 5*(XlO(4 )+XIO(4» 129
130 0009141 X2(4)=X20(4)+DT*(-RID4*XIAV +X30(4)+(R4N4-RID4*R4N6)*A2AV) 130
131 00094EI X3(4)=X30(4)+DT*(-RID3*XIAV +X40(4)+(R4N3-RID3*R4N6)*A2AV) 131
132 0009881 X4(4)=X40(4)+DT*(-RID2*XIAV +X50(4)+(R4N2-RID2*R4N6)*A2AV) 132
133 0009C2I X5(4)=X50(4)+DT*(-RIDI*XIAV +X6Q(4)+(R4NI-RIDI*R4N6)*A2AV) 133
134 0009FC I X6(4)=X6Q(4)+DT*(-RIDO*XIAV +(R4NO-R1DO*R4N6)*A2AV) 134
135 C 135
136 C *~******************.. BLoeK NO. eb ***************************** 136
137 C FILTER FROM THETAA TO THETAS: 137
138 C 138
139 000A321 ANGAV=O . 5*(BETAA(2)+BETAAOLD(2» 139
140 C ANGAV=O . 5*(BETAAOLD(2)+BETAAOLD(2» 140
141 000A4C I Xl (5 ) =XI0(5)+DT*(-PID5*XIO(S)+X20(S)+(PIN5-PID5*PIN6)*ANGAV) 141
142 000A841 XIAV=0 . 5*(Xl(5)+XI0(5» 142
143 C XIAV=O . 5*( XI0( 5)+XI0(5» 143
144 000A98I X2 ( 5)=X20(5)+DT*(-PID4*XIAV +X30(5)+(PIN4-P1D4*PIN6)*ANGAV) 144
145 000AD2I X3 ( 5)=X30(5)+DT*(-PID3*XIAV +X40(S)+(PIN3-P1D3*P1N6)*ANGAV) 145
146 OOOBOC I X4 (5)=X40( 5)+D1*( -P ID2*XIAV +X50(S)+(P1N2-PID2*PIN6)*ANGAV) 146
147 000B461 X5(5 )- X50(S'+DT*(-PIDl*XIAV +X60(5)+(PINI-PIDI*PIN6)*ANGAV) 147
148 000B801 X6 (5)=X6Q( 5)+DT*(-PIDO*XIAV +(PINO-PIDO*PIN6)*ANGAV) 148
149 149
150 C********************** BLoeK NO . C7 ***************************** ISO
151 C FILTER FROM A2( 11 TO THETAS : 151
152 C 152
153 000BB6 1 A2AV=0 . 5*< A2( 1 ) +A20LD ( 1» 153
154 C A2AV=0 . 5*( A20LD( 1l+A20LD( 1» 154
155 OOOBCAI Xl ( 6)=XI0(6)+DT*(-P1D5*XI0(6)+X20(6)+(P2N5-PID5*P2N6)*A2AV) 155
156 000C021 XIAV=O . 5*( X1<6 )+XI0(6 » 156
157 C XIAV=O . 5*( XI0(6)+X 10(6» 157
158 000CI61 X2(6)=X20(6)+DT*(-P1D4*XIAV +X30(6)+(P2N4-PID4*P2N6)*A2AV) 158
159 000C501 X3(6)=X30(6)+DT*(-PID3*XIAV +X40(6)+(P2N3-PID3*P2N6)*A2AV) 159
160 000C8AI X4(6)=X40(6)+DT*(-P1D2*XIAV +X50(6 )+(P2N2-PID2*P2N6)*A2AV) 160
161 000CC41 X5(6)=X50(6)+DT*(-PIDl*XIAV +X6Q(6)+(P2NI-P1Dl*P2N6)*A2AV) 161
162 OOOCFEI X6 (6)=X6Q( 6 )+DT*(-P1DO*X1AV +(P2N0-P1DO*P2N6)*A2AV) 162
163 C 163
164 C*** ******************* BLoeK NO . CB ***************************** 164
165 C FILTER FROM THETAA TO AXSI : 165
166 C 166
167 000D34I Xl ( 7)=XI0(7 )+DT* ( -PID5*XI0(7)+X20(7)+(P3N5-PID5*P3N6)*ANGAV) 167
168 000D6C 1 XIAV=O. 5*< XI (7 ) +XI0(7 » 168
169 XI AV=O . 5*(XI0(71+Xl0(7» 169
170 000D80 1 X2(7) - X20(7)+DT*(-PID4*X1AV +X30(7)+(P3N4-P1D4*P3N6)*ANGAV) 170
171 OOODBAI X3 ( 7)=X30(7)+DT*(-PID3*XIAV +X40(7)+(P3N3-PID3*P3N6)*ANGAV, 171
172 000llF41 X4 ( 7 ) =X40(7 )+DT*(-PID2* X1AV +X50(7)+(P3N2-PID2*P3N6)*ANGAV) 172
173 000E2EI X5(7)=XSO(7)+DT*(-PIDI*XIAV +X6Q(7)+(P3NI-PIDI*P3N6)*ANGAV) 173
174 000E68I X6 (7 )=X6Q(7)+DT* ( -PIDO*XIAV +(PJNO-PIDO*P3N6)*ANGAV) 174
175 C 175
176 e. ·It. ***_**** ******** *__ BLQeK NO. C9 ***************************** 176
177 FILTER FROM A2(1I TO AXSI : 177
178 178
179 000E9E I XI ( 8 ) =XIO(8 )+DT*(-P1D5*XIO ( 8)+X20(8)+(P4N5-PID5*P4N6)*A2AV) 179
180 000ED61 XIA V=O. 5*( X1(8)+XIO(8» 180
181 X1AV=0 . 5*(XIO(8l+Xl0 ( 8» 181
182 OOOEEAI X2 ( 8)=X20(8)+DT*(-PID4*XIAV +X30(8)+(P4N4-PID4*P4N6)*A2AV) 182
183 000F24I X3(8)=X30(8)+DT*(-PID3*XIAV +X40(B)+(P4N3-PID3*P4N6)*A2AV) 183
184 000F5EI X4(8 ) =X40(8)+DT*(-PID:2*XIAV +X50(8)+(P4N2-P1D2*P4N6)*A2AV) 184
185 000F98I X5(8)=X50(8)+DT*(-PIDl*XIAV +X6Q(8'+(P4NI-PIDI*P4N6)*A2AV) 185
186 000FD21 X6 ( 8)=X6Q(8)+DT*(-P1DO*XIAV +(P4NO-PIDO*P4N6)*A2AV) 186
187 187
IBB C*** ******************* BLoeK NO . eto **************************** 188
189 C FILTER FROM PSIA TO PSIS : 189
190 C 190
191 0010081 ANGAV=O . 5*(BETAA(3)+BETAAOLD(3» 191
192 C ANGAV=O . 5 * (BETAAOLD(3) +BETAAOLD(3) ) 192
193 0010221 Xl (9)=XIO(9)+DT*(-YID3*XIO(9)+X20(9)+(YIN3-Y1D3*Y1N4)*ANGAV) 193
194 00105AI XIAV=O . 5*( X1(9)+XIO(9» 194
195 XIAV=O . 5* (XI0(9)+Xl0(9» 195
196 OOIObEI X2 ( 9)=X20(9 )+DT*(-YID2 *X1AV +X30(9)+(YIN2-Y1D2*YIN4)*ANGAV) 196
197 0010A81 X3(9)=X30(9)+DT*(-YIDl*XIAV +X40(9)+(YINI-Y1DI*YIN4)*ANGAV) 197
198 0010E21 X4(9)=X40(9)+DT*(-Y1DO*XIAV +(YINO-YIDO*YIN4)*ANGAV) 198
199 C 199
200 C********************** BLoeK NO. Cll **************************** 200
201 FILTER FROM A2(3) TO AZS1 : 201
202 202
203 0011181 A2AV=0 . 5* (A2( 3) +A20LD ( 3) ) 203
204 A2AV=0 . 5*(A20LD(3)+A20LD(3» 204
205 OOII:2C I XI (10)=XIO(10)+DT*(-HID3*XI0(10)+X20(10)+(HIN3-H1D3*HIN4)*A2AV) 20S
206 0011641 XIAV=O . 5*( XI (10)+Xl0( 10» 206
207 XIAV=O . 5*< XIO( 10)+XI0 ( 10» 207
208 0011 7 81 X2(10)=X20(!0)+DT*(-HID2*XIAV +X3D(10)+(H1N2-H1D2*HIN4)*A2AV) 208
209 0011B21 X3 ( 10)=X30(10)+DT*(-HIDI*X1AV +X40(10)+(HINI-HIDI*HIN4)*A2AV) 209
210 0011ECI X4 ( 10 ) =X40(10)+DT*(-HIDO*XIAV +(HINO-H1DO+HIN4)*A2AV) 210
211 211
212 C********************** BLoeK NO. C12 **************************** 212
213 NOW ADD THE APPROPRIATE COMPONENTS TO OBTAIN THE SIM. MOTION 213
214 214
215 0012221 BETAS(I)-XI(I)+R1N6*BETAA(I)+Xl(2)+R2N6*A2(2) 215
216 0012481 BETAS(2)=XI(5)+P1N6*BETAA(2)+Xl(6)+P2N6*A2(1) 216
217 00126EI BETAS(3)=Xl(9)+Y1N4*BETAA(3) 217
218 218
219 0012861 ASI(I)=XI(7)+P3N6*BETAA(2)+Xl(8)+P4N6*A2(1) 219
220 0012AC I AS1 (2) =Xl (3)+R3N6*BETAA( I) +Xl (4)+R4N6*A2(2) 220
221 0012D21 ASI(3)= Xl(10)+H1N4*A2(3) 221
222 222
223 UPDATE ALL THE DUMMY VARIABLES : 223
224 224
225 0012E41 DO 5 1=1 . 10 22S
226 0012ECI XI O(I ) =Xl( I ) 226
227 0013041 X2 0 ( I' - X2 ( I ) 227
228 00131C I X30 (I ) =X3 ( 1 ) 228

F.27
 229 0013341 5 X40(1)=X4( l )
230 0013621 DO IS 1=1.8 424
231 00136AI X50(!) =X5( I) 425
232 0013821 15 X60( I) =X6( I) 426
233 0013BOI DO 20 1=1. 3 427
234 0013B81 BETAAOLD(I)=BETAA(!) 428
235 0013081 20 A20LD(I)=A2(1) 429
236 c 430
237 C EVALUATE THE TRANSFORMATION MATRIX LIBS 431
238 C 432
239 0014061 CALL LIBS 433
240 C 434
241 435
C********************** BLoeK NO . C13 **************************** 436
242 C NOW INTEGRATE ASI TO GET VSI AND SSI
243 C 437
244 00140EI DO 30 1=1.3 438
245 0014161 VS I (!) =VSI (U+DHO . 5*< ASI ( I )+ASIOLD( I) ) 439
246 00145AI SSI(I)=SSI(I)+DT*O . 5*(VSI( 1)+VSIOLD(l» 440
247 0014AAI ASIOLD (I )=ASI (I) 441
248 0014C41 30 VSIOLD(I)=VSJ(I) 442
249 443
250
C********************** BLoeK NO . F1 ***************************** 444
C NOW USE DIFFERENTIATOR + L . P. FILTER TO OBTAIN BETASDOT :
251 c 445
252 0014F41 DO 90 1=1. 3 446
253 OOI4FC I BETASD(I )=W2*(BETAS( 1)-BETAS1(l» 447
254 0015261 BETASI(I)=BETASI(I)+DT*BETASD(I) 448
255 0015501 90 CONTINUE 449
256 C 450
257 C PREMULTIPLY BETASO BY RR TO GET P.G.R 451
258 C 452
259 0015661 RR (1.3) =-S INTHETAS 453
260 0015761 RR(2.2)=COSPHIS 454
261 001580r RR'2.3)=SINPHIS*COSTHETAS 455
262 00"901 RR,3.2)=-SINPHIS 456
263 0015AOI RR(3.3)=COSPHIS*COSTHETAS 457
2 64 C 458
265 OOISBOl CALL VMULT(RR.BETASD. WSS , 3 . 3 . I) 459
2 66 C 460
267 461
C NOW USE THE DIFFERENTIATOR + L . P FILTER TO GET WSSD
268 C 462
269 0015F41 00 55 1=1. 3 463
270 OOI5FC I WSSO i l)=W2*(WSS ( I ) -WSSI ( I » 464
2 71 0016281 WSSI ' I ) =WSSI'I)+DT*WSSD(I ) 465
272 001654 I 55 CONTINUE 466
2 73 CC CC CCCCCCC C CCCCCCCCCC CCCC CC CCCC CCC C CCCCCCCCCCCCCCCCCCCCCCCCC C
46 '1
2 74 OOI66AI RETURN 468
27 5 0016701 END 469
470
NO ERRORS ' F7D I'L15 -00 . 00 SUBROUTINE WASHOPT 15 / 03 / 85 10: 59 : 29 TABLE SPACE : 7 K.B
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK. SPACE : 175 WORDS
SINGLE PRECISION FLOATING PT SUPPORT REGUIREO FOR EXECUTION

I 0000001 SUBROUTINE LIBA(BETAA.LIA)

2 C 276
3 C COMPUTE THE TRANSFORMATION MATRIX LIA 277
4 C 278
5 0000041 REAL*4 BETAA(3).LIA(3. 3) 279
6 0000041 280
COMMeN /SINCOS / SINPSIS.COSPSIS.SINTHETAS.COSTHETAS.
7 *SINPHIS.COSPHIS 281
8 C 282
9 0000041 PHI2 BETAA(I)*BETAA(I) 283
10 00002EI THETA2 BETAA(2)*BETAA(2) 284
11 C PSI2 BETAA(3)*BETAA(3) 285
12 C 286
13 0000481 SINPHI BETAA(U*(l. + PHI2 287
14 0000761 SINTH
*(-0 . 1666665 + O. 00833332*PHI2
BETAA(2)*(1. + THETA2* (-0. 1666665 + O.
» 288
15 00833332*THETA2» 289
C SINPSI BETAA (3)* (1. + PSI2 *<-0 . 1666665 + O. 00833332*PSI2
Ilo » 290
C SINPSI SINPSIS
17 0000A41 SINPSI SIN(BETAA(3) ) 291
18 0000021 COSPHI 1. + PHI2 292
* (-0 . 5003465 + O. 04166666 *PHI2 293
19 000OF61 COSTH 1. + THETA2 * (-0 . 5003465 + O. 04166666 *THETA2)
20 C COSPSI 1. + PSl2 294
* (-0 . 5003465 + O. 04166666 *PSI2 )
295
21 C COSPSI COSPSIS
22 00011A 1 COSPSI = COS(BETAA (3» 296
23 C 297
24 00014AI LIA<1.1) 298
COSPSI *COSTH
25 0001601 LIA(2. 1> SINPSI *COSTH 299
26 0001761 LIA(3. I) -SINTH 300
27 00018C I LIA(I.2) 301
COSPSI*SINTH*SINPHI - SI"f'SI*COSPHI
28 000lB61 LIA(2.2) 302
SINPSI*SINTH*SINPHI + COSPSI*COSPHI
29 OOOIEOI LIA(3. 2) 303
COSTH*S INPH I
30 000lF61 LIA<1.3) 304
COSPSI*SINTH*COSPHI + SI"f'SI*SINPHI
31 0002201 LIA(2.3) 305
SINPSI*SINTH*COSPHI - COSPSI*SINPHI
32 00024AI LIA(3.3) = COSTH*COSPHI 306
33 0002601 RETURN 307
34 0002661 END 308
309
NO ERRORS:F7D R05-00 . 00 SUBROUTINE LIBA 11/12/84 13:24:13 TABLE SPACE : 2 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 145 WORDS
SINGLE PREC ISION FLOATlNG PT SUPPORT REGUIRED FOR EXECUTlON

F.28
 1 C**** SUBROUTINE WASHADPT . FTN **** WASHADP3 VERSION ********
2 C 2
3 C ADAPTIVE WASHOUT ALGORITHM . GIVEN FAA AND WAA. 3
4 C COMPUTE SS I AND SI MULATOR EULER ANGLES . 4
5 C FILTERS ARE ALL IN INERTIAL FRAME 5
6 C 6
7 0000001 SUBROUTINE WASHADPTCDT.FAA . WAA.SSI .SS IIN) 7
B 8
9 000004 I REAL*4 FAA (3). A2(3). A20(3). BETACDC 3). BETACDO(3). FN (2). FNO(2). 9
10 *SS I (3) • VSI C3). ASI C3). BETAS C3) . SSIIN(3). WAA C3). TTC3. 3) 10
11 0000041 REAL*4 AAI SC3. 6). BBII C3.6). LLI S C3. 3). FUNC C6). DFUNC C6. ó). 11
12 *LENGTH C6). LENGTH_NEUTRAL 12
13 0000041 REAL*4 XNC ó). XNPI CÓ) 13
14 000004 [ REAL*4 KXI.KX2 . KYI.KY2.KZ1.KZ2.KZ3.KPSII.KPSI2 14
15 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 15
16 000004 I REAL*4 BETASD(3) . RRC3 . 3) 16
17 000004 I REAL*4 WSSI(3).WSSDC3 ). WSSC3) 17
IB CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 18
19 C 19
20 000004 I COMMON / MATRIXI I AAIS . B8II . LLIS.FUNC.DFUNC.LENGTH. XN.XNPI. 20
21 *LENGTH NEUTRAL 21
22 0000041 COMMON-/SINCOS I SINPHIS.COSPHIS. SINTHETAS . COSTHETAS. 22
23 *SINPSrS.Cospsrs 23
24 0000041 COMMON I FILTER I ASI. VSI 24
25 0000041 COMMON /A NGDER I WSS. WSSD 25
26 0000041 COMMON / PXOUTI PXI.PX2.PX3.PYI . PY2 . PY3.PZ.PPSI 26
27 0000041 EClUIVALENCE CBETAS. XN(4» 27
28 28
29 000004 I DArA X IO. X20. XIPX10.XIPX20.XIPX30. X2PXIO.X2PX20.X2PX30. 29
30 *THESO. DTHSDPX20 . DTHSDPX30.PX10 . PX20 . PX30.PXI0.PX20.PX301 30
31 * i l *o .• 1. O. O. 16404. O. 5 . I O. 0 . 16404. O. 51 31
32 32
33 0000041 DATA YIO.Y20.YIPYIO.YIPY20.Y1PY30.Y2PYIO.Y2PY20.Y2PY30. 33
34 *PHI SO . DPHSDPY20. DPHSDPV30.PVIO.PV20.PV30.PV10.PY20.PV301 34
35 *11*0 . 1.0.-0. ló404.0 5.1.0.-016404 .0. 51 35
36 C 36
37 00 0004 [ DA rA liD. Z20.Z30.Z IP ZO.Z2PZO. Z3PZO.PZO. PZO 1 6*0 . • 2*1. I 37
3B 000004 1 DATA PSIlO. peleo. pelIPPSIO . PSIPPSIO . PPSIO. PPSIO /4*0 .• 2*1. 38
39 C 39
40 000004 1 DATA BETACDO/3*0. I. A20 / 3*0. I. FNO/-9. 81. O. / • HNVO / O. 40
41 0000041 DATA LLIS/I. . 3*0 • I ' . 3*0 .• 1. I TTll. . 3*0 . • 1. .3*0 . • 1.1 41
42 000004 1 DATA XN/ó*O . / • G/ 9 . 811 42
43 C 43
44 C SE T CONTSTANTS IN ADAPTlVE FILTERS : 44
4!i C 45
46 000004 1 DATA WX3. WX2. WX 5. WX4. KX1 . KX2. WXI / 46
47
4B 000004 1
•DATA 5. O, Q. 2 . Q . O. 1 0, 1 . O. 2 . 828. O. 00929/
GX 1. GX2 . GX3. WX6 . WX7. WX81
47
48
49
* o 5 1668 . 0 . 005793. 0.0 1 07"4 . 0 . 19354. 17. 262 . 46 . 451 1 49
50 50
51 0000041 DA TA WY3 . WY2. WY5 . WV 4. KVI . KY2 . WVII 51
52 5 . O. 0 . 2 . 0 O. 1. O. I . O. 2 . 828 . 0 . 00929/ 52
~3 0000041
* DATA GY l. GY2. GY3. WY6 . WV7. WYBI 53
54 * 0 5166 8 . 0 . 005793. 0 . 010764. O. 19354. 17. 262. 46 . 4511 54
55 C 55
'6 0000041 DA fA WZ2. WZl . KZ 2. KZ3. KZ1. GZ . \.IZ3 / 56
'7
58
O. 5. 5 . O. 3 . O. 4 . 5. O. 5 . 0 . 3. O. 333331 57
58
59 000004 [ DATA WPSI2 . WPSII . KPSI2. KPS l ! . GP5I. WPSI31 59
60 1. O. O. O. 0. 4242. 0 . 09 . 1. O. 1. O l 60
61 61
62 000004 1 DArA TILTOMAX I. 10191 62
63 000004 r DA TA S INPHIS. SINTHETAS . SINPSIS 13*0 I 63
ó4 000004 [ DATA C OSPHIS.C05THETAS. C05PSI S 13*1.1 64
65 CCCC CCCCC CCCCCCCCCCCCCCCCCCCCCCCCCC C CCCCC CCCCCCCCCCCCCCCCCCCCCCCCCC 65
6" 000004 [ DA T A WSSI /3*0. I 66
6 ·' 0000041 DA T A RR/I. . 3*0 .• 1. .3*0 .• 1. I. W2/ 12 5 1 67
68 CCCC CCCCC CCC CCCCCCCCCCCCCCC CCCCCCC CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 68
69 C 69
70 C********************** BLoeK NO. El ***************************** 70
71 C FIRST. PREMULTIPLV FAA BY LLIS AND ADD G TO OBTAIN A2 71
72 C 72
73 000004 1 CAI_L VMUL TC LLIS . FAA. A2. 3. 3. I) 73
74 00027C I A2 C 3'~A2(3)+G 74
75 75
76 NOW TRANSFORM FAA TO OBTAIN FN 76
77 77
7B 00028BI FN ( 1 ) -FAAC 1) *SI NTHETAS*COSPSIS 78
79 +FAAC2 l *SINPHIS*COSTHETAS*COSPSIS 79
BO +FAA(3)*COSPHIS*COSTHETAS*COSPSIS 80
BI 0 002E6 I FN (2) - FAAC2 )*CCOSPH IS*C OSTHETAS *SINPSIS-SINPHIS*COSPSIS) 81
B2 -F AA C3 )*C S INPH IS*S INTHETAS*SINPSIS+COSPHIS*COSPSIS) 82
B3 0003401 HN r CWAA(2 ) *COSPHIS - WAA(3)*SINPHIS) * SINTHETAS /COSTHETAS 83
84 84
B5 C********************** BLoe K NO . E2 ***************************** 85
86 C OBTAIN THE CDMMANDED ANGULAR RATES ' B6
87 C 87
BB 000374 I CALL V MULT CTT . WAA . BETACD.3.3. 1) 88
89 c 89
90 C********************** BLoeK NO . E3 ***************************** 90
91 START THE ADAPTI VE FILTER I NG PR!JCESS--FIRST DO SURGE-P ITCH 91
92 92
93 0003CBI XTERMl = ( 1 . -PXIO'*A20( 1 ) +KX 1*)' lO+KX2*X20 93
94 00031"0 [ XTERM6 =A20 ( 1 )-KXl*XlPX 10-KX2*X 2PXl 0 94
95 00040EI XTERM 7=PXIO*FNO CI ) *DTHSDPX2D-KXI*XIPX20-KX2*X2PX20 95
96 000434 I XTERM8 = PXIO*FNOCI) *DTHSDPX30- KXI*X IPX30-KX2*X2PX30 96
97 97
9B 00045AI Xl XtO + DT * X20 98
99 00046EI X2 X20 + DT * ( PX10*A20C I) - KX1*XIO - I<.X2*X20l 99
100 00049AI XIPX, XIPXIO + DT * X2 PXIO 100
101 0004AE I XIPX2 XIPX20 + DT * X2PX20 lOl
102 0004C21 XIP X3 XIPX30 + OT * X2PX30 102
103 0004D61 X2PXl X2PXI0 + OT * XTE RM6 103
104 0004tC I X2PX2 Xf:P X20 + OT * XTERM7 104
105 000502 [ X2P X3 X2P X30 + OT * XT ERM8 105
106 C*** ******************* BLoeK NO E4 ***************************** 106
107 00051131 THESDT = AMAX1 C-TILTDMAX.AMINICPX20*A20CI).TILTDMAX» 107
lOB 000576 [ THES : THESO + DT * lTHESDT + PX30*BETACDO(2» 108
109 0005941 FA CX= I. 109
110 0005AOI IF CABS CTHESDTl . EO . T1LTDMAX) FACX=O. 110
111 0005CEI XTERM2 = WX 1 *< (1 ·~ PX30) *BETACDOc 2 )-THESDT) lil
112 0005ECI XT EPM3:a <A20 (1) .. ·PX20*FNO ( 1 ) *DTHSOPX20) *FACX 112
113 0006081 XT ERM4=WX4*CTHESDr+PXJO*BETACDO( 2» 113
114 0006201 X TERM~=PX20*FN O(1)* DTHSDPX30*FACX+BETACDOC2) 114

F .29
 115 00063CI OTHSOPX2 = OTHSDPX2o + OT * XTERM3 115
116 0006521 OTHSDP X3 ~ DTHSDPX30 + DT * XTERM5 116
117 0006681 PXl - PXIo + DT * GX1* CXTERM1*XTERM6 - WX3*XIo*XIPXIo 117
118 * - WX2*X2o*X2PXlo 118
119 * - WX6*CPXIO-PXI0 11 119
120 0006B41 PX2 ~ PX20 + DT * GX2* CXTERM1* 120
121 CXTERM7-FNoC 1 1*DTHSDPX201 + XTERM2*XTERM3 121
122 * - WX3*XID*XIPX2o - WX2*X20*X2PX2o 122
123 * - WX5*THESo*DTHSDPX2o - XTERM4*XTERM3 123
124 * - WX7*CPXêO-PX20» 124
125 0007341 PX3 :ar PX30 + OT * GX3* (XTERMl* 125
126 * CXTERM8-FNoCl1*DTHSOPX301 + XTERM2*XTERM5 126
127 * - WX3*XID*XIPX30 - WX2*X20*X2PX3D 127
128 * - WX5*THESo*OTHSDPX3o - XTERM4*XTERM5 128
129 - WX8*CPX30-PX301) 129
130 C 130
131 C Now DO THE SWAY-RoLL 131
132 C 132
133 0007841 YT~RMl-(l . -PYIO)*A20(2)+KYl*YlO+KY2*Y20 133
134 0007DC 1 YTERM6-A2oC21-KYI*YIPYlo-KY2*Y2PYlo 134
135 0007FAI YTERM7-PYIo*FNoC21*DPHSDPY2D-KYI*YIPY2o-KY2*Y2PY2o 135
136 000B201 YTERMS-PYIo*FNoC21*DPHSDPY30-KYI*YIPY3o-KY2*Y2PY3o 136
:37 137
138 0008461 YI = Ylo + OT * Y20 13B
139 00085AI Y2 Y20 + OT * CPYlo*A2DC21 - KYI*Ylo - KY2*Y2D1 139
140 0008861 YIPYI = YIPYlo + OT * Y2PYtD 140
141 00089AI YIPY2 = YIPY20 + OT * Y2PY20 141
142 OOOBAEI YIPY3 = YIPY30 + OT * Y2 PY30 142
143 000BC21 Y2PYI ~ Y2PYtD + OT * YTERM6 143
144 000BD81 Y2P Y2 Y2P Y20 + OT * YT ERM7 144
145 OOOBEEI Y2PY3 ~ Y2PY3o + OT * YTERMB 14~
146 C********************** BLO eK NO . E4 ***************************** 146
14 7 000904 I PHISDT = AMAXIC-TILTDMAX.AMINICPY2o*A2DC21.TILTDMAXl I 147
148 0009621 PHIS = PHISo + OT * CPHISOT + PY30*BETACDoCll l 148
149 0009801 FACY=I . 149
150 00098C I lF CADSCPHISOT 1. EO . TIL TOMAX I FACY=O . 150
i51 0009BAI YTERM2=WYI*CCI . -PY301*BETACOOCIJ-PHISDTl 151
152 0009081 YTERM3=CA2oC21+PY2o*FNoC2J*DPHSDPY201*FACY 152
l!i3 + PY30*HNYO*DPHSDPY20 153
154 OOOA021 YTERM4=WY4*CPH1SDT+PY30*BETACDOC IJ I 154
15 5 OOOAIA I YTERM5=PY2o*FNo C21 *DPHSDPY30*FACY+BoTACOO Cll 155
156 * + PY30*HNyo*DPHSDP Y3o 156
157 000A44! DPHSDPY2 DPHSDPY~D + OT * YTERM3 157
15B 000A5Al DPHSDPY3 DPHSDPY3D ... OT * YTERM5 158
159 000A701 PYl PY10 + OT * GY1* (YTERM1*YTERM6 - WY3*Yl0*Y1PY10 159
1,,0 - WY2*Y20*Y2PY10 160
161 - WY6*(PYID-PYIO» 161
162 OOOADC I PY 2 PY20 + OT * GY2* ( YTERM1* 162 .
11>3 CYTERM7-FNO ( 21*DPHSDPY201 163
164 + YTERM2*<YTERM3-HNYO*OPHSOPY20) 164
16!i - WY 3 * Yl0*Y1P Y2 0 - WY2*Y2D*Y2PY20 165
166 - WY5*PHISO*OPHSOPY20 - YTERM4* YTERM3 166
167 - WY7* ( PY20-PY20» 167
t 68 0008461 PY3 PY30 + OT * Gv3* (Y TERM1* 16B
169 (YTERMB-FNO (2) *DPHSDPY30) 169
170 + YTER H2* <YTERM!j -HNYO*DP HSDPY30 } 170
\71 - WY3*Yl0*Y1PY30 - WY2*Y20*Y2PY30 171
172 - WY5*PHISO*DPHSDPY30 - YTERM4*YTERM5 172
173 - WY8*CPY30-PY3011 173
174 C
* 174
175 C Now DO HEAVE 175
176 C 176
177 000SE21 ZI - Z10 + OT * Z20 177
178 000SF61 Z2 c Z2D + OT * Z30 17B
179 OOOCOAI Z3 - Z30 + OT * CPZo*A2oC31-KZ2*Z2o-KZ3*Z30-KZI*Zlol 179
180 000C401 ZIPZ - ZIPZo + OT * Z2PZo lBO
181 000C54I Z2PZ • Z2PZO • OT * Z3PZO 181
182 000C68I Z3PZ • Z3PZo + OT * CA20C31-KZ2*Z2PZO-KZ3*Z3PZO-KZI*ZIPZOI 182
183 000C9AI PZ - PZO + OT * GZ* CCCI.-PZol*A2oC31 183
184 + KZ2*Z20 + KZ3*Z30 • KZ1*Z10) 184
185
* * CA2DC31 - KZ2*Z2PZo - KZ3*Z3PZO - KZI*ZIPZol IB5
18b - WI2*12D*Z2PZO - Wll*Z3D*13PZO lB6
187
* - WZ3*CPZo-PZOl 1 IB7
188 C
* 18B
189 C AND FINALLY. DO YAW IB9
190 C 190
191 OOOD2AI PSII + OT * PSISO
- PSIIO 191
192 000D3EI PSIS + OT * CPPSIo*BETACDo..c31 - KPSI2*PSISo
- PSISo 192
193 - KPSI'*PSIlol 193
194 OOOD6AI *PSIIPPSI - PSIIPPSID + OT * PSIPPSID 194
195 00007EI PSIPPSI = PSIPPSIO + OT * CBETACDoC31 - KPSI2*PSIPPSID 195
196 - KPSll*PSIIPPSIol 196
197 000DA6I
*PPSI = PPSIo + OT * GPSI* CCCI . -PPSI01*BETACDoC31 197
198 + KPSI2*PSISo + KPSI1*PSIlol 198
199
* * CBETACOOC31 - KPSI2*PSIPPSIo - KPSII*PSIIPPSIol 199
200
* - WPSI2*PSISo-PSIPPSIO - WPSII*CPPSIo-BETACOoC31 200
201
* - KPSI2*PSISo - KPSII*PSIIol 201
202 * CBETACOO(31 - KPSI2*PSIPPSIo - KPSll*PSIIPPSIol 202
203
* - WPSI3*CPPSIo-PPSIOll 203
204 C
* 204
205 C NOw UPDATE ALL THE OLD VARIABLE VALUES 205
206 C 206
207 000E50I DO 10 1=1.3 207
208 000E5BI A2oC!) -A2C !) 20B
209 000E70I 10 BETACDoCI1-BETACDCIl 209
210 000E9EI FNoCl l-FNC I I 210
211 000EA6I FNOC21-FNC21 211
212 OOOEAEI HNYO - HNY 212
213 C 213
214 OOOEBBI XIO XI 214
215 000EC2I X20 = X2 215
216 OOOECCI XIPXIo = XlPXI 216
217 000ED61 XIPX20 XIPX2 217
218 OOOEEOI XIPX30 = XIPX3 218
· 219 OOOEEAI X2PXIo - X2PXI 219
220 000EF4I X2PX20 = X2PX2 220
221 OOOEFEI X2PX30 X2PX3 221
222 OOOFOBI THESo = THES 222
223 000FI2I DTHSDPX:W = OTHSOPX2 223
224 000F1C I OTHSDPX3(] = OTHSOPX3 224
225 000F26I PX10 = PXI 225
226 000F30I PX20 = PX2 226
227 000F3Al PX30 ~ PX3 227
228 C 228

F .30
 229 000F44I YIO - Yl 229
230 000F4EI Y20 - Y2 230
231 000F5S1 YIPYI0 = YIPYl 231
232 000F62I YIPY20 - YIPY2 232
233 000F6CI YIPY30 - YIPY3 233
234 000F761 Y2PYI0 • Y2PYl 234
23~ OOOFSOI Y2PY20 · = Y2PY2 23~
236 OOOFSAI Y2PY30 - Y2PY3 236
237 000F94I PHISO • PHIS 237
238 000F9EI DPHSDPY20 • DPHSDPY2 238
239 000FA81 DPHSDPY30 • DPHSDPY3 239
240 000FB21 PYI0 - PYl 240
241 OOOFBC I PY20 - PY2 241
242 000FC61 PY30 - PY3 242
243
243
244 OOOFDOI Z10 - ZI 244
245 OOOFDAI Z20 - Z2 245
246 000FE41 Z30 - Z3 246
247 OOOFEEI ZIPZO - ZIPZ 247
248 000FF81 Z2PZO = Z2PZ 248
249 0010021 Z3PZO • Z3PZ 249
250 00100CI PZO • PZ 250
251 C
251
252 0010161 PSII0 = PSIl 2~2
253 0010201 PSISO = PSIS
253
2~4 00102AI PSIIPPSIO = PSIIPPSI
254
255 0010341 PSIPPSIO • PSIPPSI 255
256 00103EI PPSIO PPSI
256
257 C
257
258 C NOW THE REGUIRED RESULTS ARE FORMED FROM THE VARIABLES 25S
259 C
259
260 0010481 BETAS( 1) - PHIS
260
261 0010541 BETAS (2) • THES
261
262 0010601 BETAS( 3) - PSIS
262
263 C
263
264 00106CI BETASD(!)- PHISDT + PY3*BETACD(!)
264
265 0010801 BETASD(2)· THESDT + PX3*BETACD(2)
265
266 0010941 BETASD(3)· PPSI*BETACD(3) - KPSI2*PSIB - KPSI1*PSll 266
267 C
267
268 0010BAI ASI(I) = PX1*A2(!) - KX1*Xl - KX2*X2 268
269 0010E21 ASI(2) - PY1*A2(2) - KY1*Yl - KY2*Y2 269
270 00110A I ASI(3) • PZ *A2(3) - KZ2*Z2 - KZ3*Z3 - KZ1*ZI
271 270
C
271
272 00113EI VSI(1) • X2
273 00114AI VSI(2)
272
- Y2 273
274 0011561 VSI(3) - Z3
275 274
C
275
276 0011621 551 (1) - Xl + SSlIN( 1) 276
277 00117C I SSI (2) • Yl + SSIIN(2) 277
278 0011961 551 (3) ,.. Z2 + SSIIN(3)
279
278
C
280
279
C EVALUATE THE TRANSFORMATION MATRIX LIBS
281 280
C
282 0011801 281
CALL LIBS
282
283
283
284 EVALUATE THE T-MATRIX :
284
28~
286 0011B81 28~
TANTHETAS=SINTHETAS/COSTHETAS
287 0011CAI 2S6
TT (L 2) SINPHIS*TANTHETAS
287
288 0011DAI TT ( 1.3) COSPHIS*TANTHETAS
289 0011EA I 288
TT ( 2.2) COSPHIS
;290 00 IIF4 J 289
TT(2.3) -SINPHIS
291 0012041 290
TT (3.2) SINPHIS/COSTHETAS
292 001214 1 291
TT(3. 3) = COSPHIS/C05THETAS
292
~3 C
;293
294
295
C********************** BLoeK NO. ES ************************* .... 294
C PREMUL TIPLY BETASO BY RR TO GET P. G. R ;295
296 C
297 0012241 296
RR(I . 3)=-SINTHETAS
298 0012341 297
RR (2. 2) =COSPHI 5
299 OOI23EI 298
RR (2.3) =SINPHIS-cOSTHETAS
299
300 00124EI RR (3. 2)=-SINPHIS
301 00125E 1 300
RR (3.3 )=COSPHI5*COSTHETAS
302 301
303 00126EI 302
CALL VMULT(RR. BETASD. WSS. 3. 3. I)
304 303
C
305 304
C**** . . . ********.****** BLoeK NO. Eb ****************************. 305
306 C NOW USE THE DIFFERENTIATOR + L . P. FILTER TO GET WSSD
307 306
C
308 0012801 307
DO 55 I-I. 3
309 00128S1 308
WSSD (I )=W2*(WSS(I ) -WSSI ( I ))
310 0012E4I 309
WSSI ~ I ) =--WSSI (I )+DT*WSSD{ I)
311 0013101 310
55 CONTINUE
312 0013261 311
RETURN
313 00132C I 312
EN D
313
NO ERRORS : F 7D R05-00 . 00 SUBROUTINE WASHADPT 30/09/S~ 14: 13: 55 TABLE SPACE : 9 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : IS9 !.ORDS
SINGLE PREC ISlON FLOATING PT SUPPORT REGUIRED FOR EXECUTION

F.31
 1 0000001 SUBROUTINE JACKDRVR(SSI.LI.LID.LIDD. DTlHE) 1
2 C 2
3 C COHPUTE THE ~ACK EXTENSION COMMANOS BASED ON POSITION IN INERTlAL 3
,
4 C
C
FRAME

REAL*4 LI(6).LIDD(6).LID(6)
4
:5
6 0000041 6
7 0000041 REAL*4 LINl(6).LIN2(6) . LIDDA.LIDA 7
B 0000041 REAL*4 SSI (3) 8
9 0000041 REAL*4 DUMMY3IA(3. I) 9
10 0000041 REAL*4 LIl (3.6) 10
11 0000041 REAL*4 AAIS(3. 6). BBl I (3.6). LLIS(3. 3). FUNC (6). DFUNC (6. 6). LENQTH(6) 11
12 0000041 REAL*4 XN(6).XNPI(6) 12
13 0000041 REAL*4 LENQTH_NEUTRAL 13
14 0000041 REAL*4 LI.L2.L3.LENQTHTOT 14
l' 0000041 COMMaN /MATRIXI/AAIS.BBII.LLIS.FUNC.DFUNC.LENQTH. XN.XNPI. l'
16 I LENQTH lEUTRAL 16
17 0000041 DATA LINI.LIN2/6*0. 0.6*0. 0/ 17
lB C 18
19 C 19
20 C 20
21
22
23
0000041

OOOQABI
c.. *********............
DO 200 JACK - 1.6
BLoeK NO. 01
CALL VMULT(LLIS.AAIS(I.~ACK).DUMMY31A.3.3 . 1)
. . . . . . . . . . . . . . . . . . . . . . . .**...
21
22
23
24 00010BI DO lOB 1=1. 3 24
2' c . . . . . . . . . . . . . .****.... BLeeK NO. 02 . . . . . . . . . *** •••• ***** ...****. 25
26 0001101 lOB LII(I.~ACK) - DUMMY3IA(I.I) + SSI(I) - BBIl(I.JACK) 26
27 C 27
2B 00017BI LI = LII(I.~ACK) 28
29 0001901 L2 = LII(2.~ACK) 29
30 0001ABI L3 = LII(3.~ACK) 30
31 0001COI . LENQTHTOT = SOOT (L1*L1 + L2*L2 + L3*L3) 31
32 C*** ••*****-*********** BLoeK NO. 03 *********.******************* 32
33 00020AI LI(~ACK) - LENQTHTOT - LENQTH-NEUTRAL 33
34 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 34
35 C LID - ESTIMATE OF ~ACK VELOCITY FOR OFF-LINE CHECKINQ 35
36 C 36
37 00022AI LIDA = (LI(JACK) - LINI(~ACK»/DTIME 37
3B 0002561 LID(~ACK) - LID(~ACK) + O. 625*(LIDA - LID(~ACK» 38
39 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 39
40 C***************.** •• ** BLoeK NO . 04 *************.*.*****.** •• *** 40
41 C COMPUTE ~CK ACCELERATIONS BASED ON THE PAST ~ACK DISPLACEHENTS 41
42 C 42
43 00029EI LIDDA = (L1(~ACK) - 2. *LIN1(~ACK) + LIN2(~ACK»/(DTIME*DTIME) 43
44 0002E6 I LIN2(~ACK) = LIN1(~ACK) 44
45 0002FEI LINI(~ACK) • LI(~ACK) 45
46 00031EI LIDD(~ACK) = LIDD(~ACK) + 0 . 625*(LIDDA - LIDD(~ACK» 46
47 0003661 200 CONTINUE 47
4B C 48
49 00037C I RETURN 49
50 0003821 END 50
NO ERRORS : F7D R05-CO. 00 SUBROUTINE ~ACKDRVR 04/01/85 14: 31 : 07 TABLE SPACE : 2 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE: 175 IoLlRDS
SINQLE PREC ISION FLOATING PT SUPPORT REQUIRED FOR EXECUTION

1 000000 I SUBROUTINE LlBS 52

2 C 53
3 C COMPUTE THE TRANSFDRMATION MATRIX LIB 54
4 C 55
5 0000041 REAL*4 AAIS(3 . 6).BBII(3.6).LIBe3.3).FUNCe6).DFUNCe6.6).LENGTH(6) 56
6 0000041 REAL*4 LENGTH_NEUT RAL 57
7 0000041 REAL*4 XN(6). XNPI (6) 58
8 0000041 COMMON / MATRIXI /AA IS. BBII . LIB. FUNC.DFUNC.LENGTH .XN. XNPI . 59
9 1 LENGTH NEUTRAL 60
10 0000041 COMMON /SINCOSI SINPSI . COSPSI. SINTH .COSTH.SINPHI.COSPHI 61
11 0000041 EQUIVALENCE (XN(I>. X). (XN(2). Y). (XNe3). Z). (XN(4).PHI) . (XN(5). 62
12 1 THETA). eXN(6).PSI) 63
13 64
14 0000041 PHI 2 z: PHI*PHI 65
15 0000161 THETA2 THETA*THETA 66
16 000028 I PSI2 PSI*PSI 67
17 C 68
18 00003AI SINPHI PHI *<1. + PHI2 *e-o . 1666665 + 0 . 00833332*PHI2 » 69
19 0000641 SINTH THETA* ( 1. + THETA2*( - 0 . 1666665 + O. 00833332*THETA2» 70
20 OOOOBEI SI NPSI PSI *<1 , + PSI2 *(-0 . 1666665 + 0 . 00833332*PSI2 » 71
21 0000B81 COSPHI 1. + PHI2 * (-0 . 5003465 + 0 . 04166666 *PHI2 ) 72
22 OOOODC I COSTH 1. + THETA2 * (-0 . 5003465 + 0.04166666 *THETA2 ) 73
23 0001001 COSPSI 1. + PSI2 * e-o . 5003465 + 0 . 04166666 *PSI2 ) 74
24 C 75
25 0001241 LIBC! .I> COSPSI *COSTH 76
26 0001361 LIB e2. I> SINPSI *COSTH 77
27 0001481 LIB(3 . I> -SINTH 78
2B 00015AI LIBel.2) COSPSI*SINTH*SINPHI - SINPSI+COSPHI 79
29 0001801 LIB e2. 2) SINPSI*SINTH*SINPHI + COSPSI+COSPHI 80
30 0001A61 LIB(3. 2 ) COSTH*SINPHI BI
31 000lB81 LIBC! .3) COSPSI*SINTH*COSPHI + SINPSI*SINPHI 82
32 OOOlDEI LIB e2. 3) SINPSI*SINTH*COSPHI - COSPSI*SINPHI 83
33 0002041 LIB(3.3) COSTH*COSPH I 84
34 0002161 RETURN 85
35 00021C 1 END 86
NO ERRORS : F7D R05-00 . 00 SUBROUTINE LIBS 11 /12/ 84 13: 23:33 TABLE SPACE : 2 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 135 IoLlRDS
SINGLE PRECISION FLOATING PT SUPPORT REQUIRED FOR EXECUTION

F.32
 1 0000001 SUBROUTINE VMULT(A,B,C, K,L , HI
2 0000041 87
DIMENSION A(K,LI,B(L,M),C(K, MI
3 OOOOOAI 88
DO 20 KK- l , K
4 00005CI DO 20 Mf'to=l , H
89
~ 0000701 C(KK, MM) - O.
90
6 00009AI 91
DO 20 LL- l,L
7 OOOOAEI 92
20 C(KK, MMI - C(KK, MM) + A(KK,LL)*B(LL.HH)
8 0001861 RETURN
93
9 00018C 1 94
END
95
NO ERRORS : F7D R05-00 . 00 SUBROUTINE VI'tULT 11/1:!/84 13: 23 : 34 TABLE SPACE : 1 KB
STATEMENT BUFFER : 20 LINES/13:!1 BYTES STACK SPACE : 172 I.()RDS
SINGLE PRECISION FLOATING PT SUPPORT REGUIRED FOR EXECUTION

1 0 000001 SUBROUTINE VTRANSP (A, B, 1'1 , N)

2 0000041 DIMENSION A(H,N) , B(N, M) 96
3 OOOOOAI DO 10 1=1 , H 97
4 00004 C I DO 10 ')~I,N 98
5 0000601 10 B(,).I) - A (I,,) 99
6 0000D81 RETURN 100
7 OOOODEI END 101
102
NO ERRORS : F7D R05-00. 00 SUBROUTINE VTRA NSP 11/12/84 13: 23:35 TABLE SPACE : 1 KB
STATEMENT BUFFER : 20 LINES/1321 B YTES STACK SPACE : 150 I.()RDS
SINGLE PREC ISION FLOATING PT SUPPORT REGUIRED FOR EXECUTION

F.33
 PLANE
User interactive control program

I
WASHADPT WAS HOUT WASHOPT JACKDRVR
impl ements implements implements computes
adaptive classical optimal actuator
filters fi 1ters filters extensions

I - -_ _ _--Jf~

I I
LIBS
calculates
.hlS

I
VMULT
multiplies
matrices

FlGURE F.l PROGRAM STRUCTURE

 PLANE
User interactive control program

I I
VESTIB VESTIBl RANDOM VTRANSP
calculates calculates generates
vestibular vestibular colored transposes
coefficients coefficients noise matrices

I
OUTP
DGEAR plots all
IMSL results
integration
routine
I
DGEAR GGNML
IMSL IMSL
I integration random number
FCNJJJ routine generator
dUll111y
routine
PLOT, PLOTS, PAGE
SCALW, BOX, LABEL
SYMBOL, NUMBER
plotting routines

I I I
FCNl FCN2 FCN3 FCNJ
supplies supplies supplies
vest i bul ar vestibular filter dummy
equations equations equations routine
I I

INTERP
interpolates
for f and w
at any time

FIGURE F.l CONCLUDED

 ----------~-~- !SI

P H.P. 2nd ORDER lst ORDER 1/5

SCALE AND LIMIT al H.P. aH b~H a2 H.P. !SI (EULER)
R FIL TER FILTER B14
(ABC METHOD) B13 (ABC METHOD)
0
G
B1
B12 B14 lis I
1/5
~I
t..-
.2.s (EULER) I -
R
1---1 FORM ~I B14
A .[S

MILg_ f2
2nd ORDER
L. P. fL
TILT 1.--
FORM
lIS

P I
.
SCALE
B2 -- FIL TER
(ABC METHOD)
---,;; COORDINATION
~
B6 1-1- - - - - - - ,
BlO

A
B3 I iS
), + .. ~
N
.!!!.AA 1-------- w1
~
I
2nd ORDER
H.P.
I <J-I
T <J-I
• ~
H.P.
S

~~ H.P.
SCALE AND LIM IT FILTER FILTER ~ .5/ (5+<&)
(ABC METHOD) B8 AND 1/5 (ABC METHOD)
BI (ABC METHOD)
B5 F1
B9

.
T B.~
B7 F1

t .. uf SS
~.s/(s+uQ)
(ABC METHOD)
F1 ~ ~ss
FIGURE F.S CLASSICAL WASHOUT FILTER.
- i

Z+Z (\I~
4th ORDER 4th ORDER
.OPTIMAL (\IA OPTIMAL
FI LTER FIL TER
P (\IS
CU CIO lis
R !SI (TRAPEZOIDAL)
LI t"\ I'~

0 aZ
SI Cl3
x+9 9+8
G 6th ORDER 6th ORDER
OPTIMAL OPTIMAL 151
R ~ FILTER FILTER
a2 z
r<
A

M
!AA .!:IA
I--
a2 x
a2Y
C7 W
h C6

+€hs-
9S
1/5
(TRAPEZOIDAL)

C13
--'!
~I
Cl I - - x+x 9+x
6th ORDER 6th ORDER
P OPTIMAL OPTIMAL
!.., FILTER FILTER
L 9A
Cg CB + 1-1 ~ <& 05/(5+<&)
A

~l\-
(ABC METHOD)
IJ IV t"\
N a SI F1
Y+4> 4>+4>
E 6th ORDER 6th ORDER 0
OPTIMAL OPTIMAL I3E
FILTER FILTER -S
I
C~
C3 C2
b. + I~
Rx
I-J tv v
4>S F1
y+y 4>+Y
6th ORDER i 6th ORDER
OPTIMAL OPTIMAL uf SS
~
FILTER FILTER
4>A
C5 C4 w2,s/(s+w2)
+ (ABC METHOD) I- ~S
W~IY ~~
aY
F1
: All fil ters were impl emented usi ng the Improved Eul er' s Method. SI
l!s
FIGURE F.6 OPTIMAL FILTERS
 Jtl

r---
SURGE, SWAY .!SI
P
R
0
!AA
lIS x
El re a2
-
AND HEAVE
FIL TERS
(EULER)
E3
isI
~I
~SS
G
.~
R uQ·s/(s+u2)
LIM
A (ABC METHOD)
E4
M E6

P
•
L 1s
~ ROLL, PITCH -R x ~S
A AND YAW
Tx FILTERS
N \ (EULER) ES
!!!AA E2 ~~C 1s
E
I~
E3 - 1s
~

FIGURE F.7 ADAPTIVE FILTERS.

 1r.A' ~A .!PA' .êA

CLASSICAL OPTIMAL ADAPTIVE

WASHOUT AT WASHOUT AT WASHOUT AT
CENTROID PILOT'S HEAO CENTROID

• •
.!ss ~SS wE SS !Ps wESS .!ss ~S wE SS

TRANSFORM TO TRANSFORM TO
PILOT'S HEAO PILOT'S HEAO

fElpS wE SS fE2 pS ~S

VESTIBULAR VESTIBULAR VESTIBULAR

MODEL MODEL MODEL

CLASSICAL OPTIMAL ADAPTIVE

FIGURE F.8 INPUTS TO THE VESTIBULAR MODEL FOR THE SIMULATOR PILOT.
 APPENDIX 6

SINE AND eOSINE APPROXIMATIONS

In order to speed up the real-time computation of sines and cosines

they are approximated in the moti on dri ve al gorithms by the fi rst three
terms in their infinite series expansions.

sinx" x - ~ + ~ (G. 1)
3! 5!

COSX" 1 - ~ + ~ (G.2)
2! 4!

with x in radians. The goodness of fit over the range ±45° is quite
acceptable as shown in Figures G.1 and G.2. Timing tests on the PE 3250
computer indicated that the above al gorithms took approximately 17 IJS to
implement while the computer library routine (without argument checking)
took approximately 49 ~s.

6.1
 0.7

0 .6
Sine
0.5 ·

0.4

0.3

0 .2
t> Approximation

- Exact
O. I

o ______.-____-.______.-____-.______ ____ ______,-____ ______

~~ ~ ~ ~ ~

o 5 10 15 20 25 30 35 40 45
Angle (deg)

FIGURE G.l APPROXIMATION TO SINE

 1.00-- ~ --
eosine
0 ..95

0.90

0.85 .

0.80
t> Approximation
- Exact
0.75

0.70
o 5 10 15 20 25 30 35 40 45
Angle (deg)

FIGURE G.2 APPROXIMATION TO eOSINE

 APPENDIX H

THE OPTIMAL SIMULATOR PROBLEM

The fOllowing development is based on material contained in Reference

6.3. Consider the system

•
x =Ax + Bu + Hn (H.I)

(H.2)

where n and hence H n is white noise. The problem is to select the control
u to mTn i i ze the cost funct i ona 1

(H.3)

where G ) 0 and R > 0 and both G and Rare symmmetric. In order to

restructure the problem into a- standard form make the following
substitutions noting that all of the above matrices are taken to be time
invariant:

_1
R = ___
CT G C (H.4)

!!.12 = CT§. Q (H.5)

R
-2 = -R + -OT -_
G0 (H.6)

-1 T
lil = !!.1 - !!.12 li2 !!.12 (H.7)

UI
-1 T
= .!! +!!.2 !!.12 ~ (H. 8)

AI -1 T
= A - !!!.2 !!.12 (H. 9)

H.I
 From Equations H.2 to H.6 it follows that

(H.la)

and from Equations H.l, H.B and H.la it follows that

tI I
C1 = E {f - T _1
[X R -X R U I ]dt }
+ _U I T _2_ (H.U)
to

where from Equations H.I, H.8 and H.9

•
X = AI X + B Ui + H n (H.12)

The above follows by the direct applietion of substitution and matrix

manipulation. In this devel0P-!'fnt we will require that Rl' B.I ' and R2 be
symmetrie, B.~ > 0 and that B.'2 exists. Sinee..§. and .R-are symmetrl'e it
follows immedlately from Equations H.4 and H.6 that B.l and B.2 are symmetrie.
Sinee ..§. ) a and B. > 0 i~l follows from Equation H.6 that B.2 > a. Sinee
.!i2 > a it follows that B.2 exists. Also from Equation H.l it is seen that
B.I ' is symmetrie sinee B.I and B.2 are symmetrie.
Thus sinee B.L' and B.2 are symmetrie and B.2 > a it follows that
Equations H.U and H.12 represent a form of the Stoehastie Linear Optimal
Regulator Problem whieh is solved in Referenee 6.3. In partieular following
Referenee 6.3 it can be shown that for thi s form of the probl em the optimal
control is

Ui = -!.(t)~ (H.13)

where
-1 T
[ ( t) = ~2 .!!. f. (t ) (H.14)

and t(t) is the solution of the matrix Rieeati equation

(H.15)

with
(H.16)

H.2
It is further stated in Reference 6.3 that the steady-state control law
represented by

-1 T_
=
F _R2 _
_ B P (H.I7)

where P is the sol ut ion of the algebraic Riccati equation

1
_ o = _1
R I P B R-2
_ ___ _BT _P + _AlT _P + __
P AI (H.18)

minimi zes both

tI
lim I E{f [x T R I x + uiT R ulJdt} (H.19)
tI +co t 1 - tOt 0 - _1 - - _2_

and
tlim
+CD E{x
_
T _1
R I _x + _uiT _2_
RUl} (H.20)

Furthermore if x(t) and u(t) [and hence ~I(t)] are ergodic, it follows that
Equation H.19 can be repTäced by

E{x
_ T_
R1 I _X + _uiT _2_
RUl} (H.2I)

In the above it is interesting to note (see Equation H.7) that it is

not true in general that !1 1 is positive semi-definite. Thus it is possible
that situations could arise in which some elemènts of x become infinite.
This should be checked whenever this controller is utilized.

From Equation H.8, H.13 and H.14 it follows that the optimal control
is

(H.22)

Using Equation H.17 it can be similarly shown that the steady-state control
1aw is

-1[ T T ] (H.23)
u = -B.2 ~ f. + B.12 ~

H.3
and from Equations H.3, H.lI and H.21 the corresponding steady-state co st
functional (for! and ~ ergodic) is

(H.24)

H.4
 APPENDIX I

THE UNCOUPLED SYSTEM EQUATIONS

1.1 Pitch/Surge Eguations - 6th Order

(I.1 )

(1.2)

(I.3 )

( 1.4)

(I. 5)

•
x=Ax+Bu (I.6 )

-T Iq 1 0 0 0 0

-T 3q 0 0 0 0 0

0 0 -Box 0 0 0
A= (I.7)
0 o. 0 0 1 0

0 0 0 0 0 1

0 0 0 0 0 0

1.1
 -T lq T:lq 0

-T:lq T~q 0

gG Ox (Aox - BOx) Gox{Aox - Bux)

B = ( 1.8)
0 0

0 0

0 1

s ]T
iis = [X16 S
x17
S
x18
(I. 9)
= [JJusrdt2 JUsrdt usr]T

(1.10)

o o o o
(1.11)
o 1 o o

o ] (1.12 )
Gox

•
xn = .:..:n
A xn
- + B n
=.n -
(1.13)

(I. 14)

(1.15)

1.2
 • IV IV

X = A X + B US + Hn (1.17)

A
-0

:J
IV

A = A (1.18 )
-0

-0 -0

o
IV

B = B (1.19)

y = CX + 0 uS (I. 20)

c
(I. 21)

(1.22)

(1.23)

1.3
 o
H = 0 (I. 24)

J = E f.rT §. 1. + pusT .R. -,!S}

= Et!T Q ~ + p(~ST .R. ~s + EST E..ct ES)} (I. 25)

(1.26)

G = [: (I.27)

The solution for ~(s) then follows directly frOOl Section 6.2.

1.2 Pitch/Surge Eguations - 7th Order

(I. 28)

(1.29)

( 1.30)

(I. 31)

(I. 32)

•
x=Ax+Bu (I. 33)
-- --

1.4
 G = [: p:J (I. 54)

The solution for ~(s) then follows directly from Section 6.2.

1.3 Roll/Sway Eguations - 6th Order

(I. 55)

(I. 56)

(I. 57)

fnPS ]T (I. 58)

1..a = [ PPA
" (1. 59)

•
x=Ax+Bu (1.60 )

-T lp 1 0 0 0 0

_T 3p 0 0 0 0 0

. 0 0 0 0 0
-Boy
A = (I. 61)
0 0 0 0 1 0

0 0 0 0 0 1

0 0 0 0 0 0

1.8
 0
,....

-B = B (I. 46)

-0

y = Cx + 0' US (I.47)

(I.48)

(I.49)

(1.50)

o
H = 0 (1.51)

J = E {r.T §. 1. + pusT B. ~S}

= E~T.Q.!. + p(.!!ST R..!!s + EST R.ct ES)} (I.52)

(I. 53)

1.7
 (1.37)

o o o o o
(I. 38)
o 1 o o o

(1.39 )

(1.40 )

(1.41 )

(1.42 )

(I. 43)

.X = A X~ ....
+ B uS + H n (I. 44)

A o B
....
A = o A o (I. 45)

o 0 bJ

1.6
 -T Iq 1 0 0 0 0 0

-T 3q 0 0 0 0 0 0

0 0 -Box 0 0 0 0

A = 0 0 0 0 0 0 0 (1. 34)

0 0 0 0 0 1 0

0 0 0 0 0 0 1

0 0 0 0 0 0 0

-T Iq T2q 0

-T 2q T3q 0

gGOX(AOX - BOX) Gox(Aox - Box)

B = .1 0 (I. 35)

0 0

0 0

0 1

(1.36)

1.5
 -T lp T 2p 0

-T 2p T 3p 0

-gGoy{Aoy - Boy) Goy{Aoy - Boy)

B = (1.62)
0 0

0 0

0 1

S sST
liS = [xl9 x20 x 2l J
(I. 63)
= [JJ vSldt 2
JVSl dt vS1J T

(1.64)

o o o o
(I. 65)
o 1 o o

o = [ T 2p o ] (I. 66)
-gG oy Goy

•
xn =~
A -xn + B n
.::.n - (1.67)

(I.68)

(1.69)

1.9
 · ,...
x = A X + B US
,...
+ Hn (I.71)

,...
-A -0 -B
-A = -0 A
-0 (1.72)

0
-0 ~

,...
-0
-B = B (I. 73)

-0

y = ë X + 0 uS (I. 74 )

,. . [-CO
C=
C
(1.75)

(1.76 )

DO[:] (1.77)

1.10
 o
H= o (I. 78)

J = EfrT §..!. + pu s T E. ~s }
= E~T Q ~ + p(.!!ST E..!!s + EsT ~ ES)} (1.79 )

(I. 80)

G = [: p:J (I. 81)

The solution for ~(s) then follows directly from Section 6.2.

1.4 Yaw Equations - 4th Order

(1.82 )

.!!s = <PS (1.83)

(I. 84)

S A
1.. = rpS (I. 85)

(I. 86)

•
x =Ax + Bu (I. 87)

1.11
 -T Ir 1 0 0

-T 3r 0 0 0
A = (1.88 )
0 0 0 1

0 0 0 0

-T 1rT2r

-T 2r T3r
-B = 0
(1.89)

vndS = [x S14 xS]T

.t-=.. 15

= [f f<Ps dt 2 (I. 90)

(1.91 )

C = [1 0 0 0] (I. 92)

o = T2 r
- (I. 93)

•
xn = ~-
-
A xn +. =-n-
B n (I. 94)

(I. 95)

(I. 96)

S
x15
]T (I. 97)

I.12
 x = Ax + BUS + Hn (I. 98)

-A -0 B
IV

A = -0 -A -0 (I. 99)

-0 0 &

-0
'"
-B = -B (1.100)

-0

y = CX + DuS (1.101)

c
(1.102 )

(I.1 03)

(1.104 )

H = -0 (1.105)

.ê.n

1.13
 J = E{yT §.!. + pusT ~.!!s}
= E ~T !L! + p(.!!ST ~.!!s + 1!!,ST R.d 1!!,S)} (1.106)

(1.107)

p:J (1.108)

The solution for ~(s) then follows directly from Section 6.2.

1.5 Heave Eguations - 4th Order

(1.109)

(1.110)

(1.111)

(see Equation 6.81) (1.112)

(see Equation 6.82) (1.113)

•
x=Ax+Bu (I.114)

1.14
 J = E{rT .§.! + p.!!sT ~.!!S}

= E {eT ~ ~ + p(.!!sT ~.!!s + EsT ~ ES)} (1.133)

(1.134)

G = [: (1.135)

The solution for ~(s) then follows directly from Section 6.2.

1.17
 .... ...
x = A X+ B US + Hn (I .125)

-A 0 B
""
A= -0 A
-0 (I.126)

0 0 ~

-0
...
B = B (I. 127)

-0

y =ëX+ 0 US (1.128)

C
"" [-CO
C= (I .129)

(1.130 )

n.[:] (1.131)

o
H = 0 (1.132 )

1.16
 -B oz o o o
0 o 1 o
A = (I.115)
0 o o 1

0 o o o

o
B= (1.116)
o
1

(1.117)

(1.118)

C = [1 o o 0] (1.119)

-o = Go z (1.120)

-; = .:..:n-
A xn + B n
~-
(I.121 )

(1.122)

x = ~aT xsT !.nT]I (I.123)

(1.124)

1.15
 APPENDIX J

PROGRAM DESCRIPTION FOR THE DESIGN OF THE OPTIMAL WASHOUT FILTERS

Unlike the classical washout filters which are designed mostly by

trial and error, the final forms of the optimal filters are obtained
methodically. This process involves the expansion of Equation 6.37 which
yields the washout filter polynomials W(s). The following describes the
programs developed to obtain this solutfön, the inputs they require and the
outputs they yield.

The simulator motion equations, when developed for the linearized

case, decouple from the original six degrees-of-freedom equations into the
following four separate sets:

(1) pitchjsurge equations

(2) rolljswayequations

(3) yaw equations

(4) he ave equations

As we 11 , the pitchjsurge and ro11jsway equations are found to be

identical (with ~ corresponding to e; and v to u) except for an opposite
g-vector polarity (see Appendix I). A single program was therefore
developed to handl e sets 1 and 2 above.

The full desi gn software was developed for three independent cases
(pitchjsurge or rolljsway; yaw; and heave). Certain programs are identical
in function but differ in the degrees-of-freedom they are meant to handle.

The design sequence for the optimal filters consists of three steps:

(1) Given vestibular model parameters and cost functional penalty weights
find the solution to the algebraic Riccati equation.

(2) Using the same parameters and weights, and the Riccati solution, solve
the washout system equations to obtain the W(s) filter transfer
function numerator and denominator polynomials. -

(3) Obtain the roots of the polynomials found in (2), cancel half the
po 1es and zeros and expand the rema i ni ng ones to obta in the fi na 1
washout filter transfer function polynomials.

Once the final filter transfer functions are obtained, they can be
solved using the evaluation program PLANE described in Appendix F. The
three design steps are discussed bel ow, the computer codes used in the
solution of the pitchjsurge and ro11jsway cases are given and their use
described.

J.1
J.l Solution of the Algebraic Riccati Equation
Purpose:
To solve for the P matrix in the Riccati equation (from Equations
6.26, 6.27 and 6.29):

(J.l )

Three programs were developed to solve this equation. They are

identical in function but differ in the order of the matrices used.
(1) WASHSPSR for the pitch/surge or roll/sway cases.
(2) WASHYAW for the yaw case.
(3) WASHEAVE for the he ave case.
Method:
The Jnp~t pa~amEl.ters (see Table J.1) are used to form the required
matrices A, B, B, C, D, G, pRo These matrices are manipulated, as shown in
- - - - - - - -1
Figure J.1 to obtain ~1' ~12 and ~2 which appear in the Riccati equation.
The following routines are used:
(1) MATML to perform matrix multiplication.
(2) MATAD to perform matrix addition and subtraction.
(3) MATTR to trans pose matrices.
(4) LINV2F an IMSL routine to invert matrices.
The Riccati equation is formed according to Equation J.1 and solved
using the Schur method outlined in Reference J.l. The following routines
are used:
(1) Five routinest included in the program listings: RICAT, HQR4, EXCHNG,
SPLIT, QRSTEP.
(2) Nine IMSL library routines: LEQT2F, LUDATN, LUELMN, LUREFN, UERTST,
UGETIO, VXADD, VXMUL, VXSTO.

tDeveloped by Mr. K. A. Carroll of UTIAS.

J.2
 (3) Four EISPACK library routines: BALANC, ORTHES, ORTRAN, BALBAK.
Finally, the solution to the Riccati equation (matrix P in Equation J.!) is
prirted and used in the next step of the design-process. A sample
pitch/surge solution is shown in Table 6.3.

J.2 Solving For Washout Filter Polynomials

Purpose :

To solve for the washout filter polynomials W{s}, as separate

numerator and denominator terms, in the system Equation 6~37:

Again, three programs were developed to manipulate this equation.

They are identical in function but differ in the order of the matrices
used:

(l) WASHl2 for pitch/surge or roll/sway cases.

(2) WASH8Y for the yaw case.
{3} WASH8H for the heave case.
Method:

Since certain terms in Equation J.2 are made up of symbol ic rather

than numeri cal variables, the equation was solved using the FORMAC compiler
{FORmula MAnipulating Compiler} which is based on the PLl language, and
available on the University of Toronto IBM 3033. Other symbolic
manipulators such as MACSYMA, SCRATCHPAD or SMP (see Reference J.2) could be
used instead of FORMAC if these were more readily available. Attempts to
dispense with computer symbol ic manipulation in favour of manual
manipulation would doubtlessly prove excessively cumbersome.

The programs developed use no subroutines and follow the flow chart
shown in Figure J.2. Matrices~,!,.f,.Q,..Q, pR,.§, .Q and B.2 are fonned
usi ng the same input from the user as suppl ied in Secti on J.l and shown in
Table J.l. As well, the Riccati solution matrices i2I' in and i23 found in
Secti on J.l are input by the user and used to fonn i.I' i 2 and i3. Equati on
J.2 is then solved. The required matrix inversions are perfonned by a
permutati on al gorithm developed accord i ng to Reference J.3. The fi nal
result, ~(s), is printed as a numerator and adenominator where:

J.3
 (J.3)

and
Denominator = Dl • O2 (J.4)

where
(J.5)

02 = determinant(s.!. - A) (J.6)

J.3 Reducing the Order of the Washout Filter Polynomials

Purpose :
To obta in the pol es and zeros and frequency response curves of the
washout filter polynomials found in Section J.2; cancel half the zeros and
pol es from the nl.nerators and denomi nators respecti vely; and thereby obta in
lower order fil ters.
prog rams ANAL YZE and POL YEX were developed to handle all cases with
appropriate control parameters.

Method:
Given the coefficients of the numerator and denominator nth order
polynomials of the W(s) elements, ANALYZE uses the IMSL routine ZPOLR to
find the n roots. As wel 1 , the s-operator in the polynomials is replaced by
jw, and w is varied from .01 rls to 10 rls to obtain the gain and phase lag
of the filters at various frequencies.
The n roots found by ANALYZE are visually inspected, and n/2 identical
roots are cancelled from the numerators and the denominator. Note that n =
12 for pitch/surge and roll/sway cases, n = 8 for yaw and heave cases.
The remaining n/2 roots are used to expand the new lower order (n/2)
polynomial using program POLYEX written in FORMAC.
The following describes the inputs requi red and outputs given by
ANALYZE and POLYEX.

J.4
ANALYZE
Inputs NORD = 4 for pi tch/surge and roll/sway cases
= 1 for yaw and heave cases
NOEG = 12 for pi tch/surge and roll/sway cases
= 8 for . yaw and heave cases
COEF(I,J) where J =1 for the WIl numerator
=2 for the Wl2 ntJllerator
=3 for the W21 numerator
=4 for the W22 numerator
=5 for the denominator
and I = 1 to 13 for the 13 coefficients of the 12th
order polynomi al. Si nce n = 8 for the yaw and
heave cases, zero coefficients are used for
unused terms (i.e., I = 10 to 13).
Outputs (1) The n roots of the nth order polynomials in the numerator and
denominator of ~(s), as shown in Tables 6.4 to 6.7.
(2) Tables and plots of the frequency response of the filters, as
shown in Figures 6.7 to 6.11 for various cases.
POLYEX

Inputs (1) NORD = 4 for pitch/surge and roll/sway cases

=1 for yaw and heave cases
(2) The factored polynomials obtained after hand-cancelling n/2 of
the n roots found by ANAL YZE
Outputs The expanded (n/2)th order polynomials which make up the numerators
and denominator of ~(s).
References
J.1 Laub, A. J., "A Schur r1ethod for Solving Algebraic Riccati Equations",
IEEE Transactions on Automatic Control , Vol. AC-24, No. 6, December
1979, pp. 913-921.
J.2 Pavelle, R., Rothstein, M., and Fitch, J., 11 Computer Algebra",
Scientific American, Vol. 245, No. 6, December 1981, pp. 136-152.
J.3 Anton, H., "Elementary Linear Algebra", John Wiley & Sons, New Vork,
1977 •

J.5
 TableJ.1

Inputs From User For Riccati Solution (Section J.1)

and Polynomial Solution (Section J.2)

Parameter Program
WASHSPSR WASHYAW WASHEAVE
(for pitch/surge or
roll/sway cases) (for yaw case) (for heave case)
Vestibular
Surge or Sway Heave Otol ith
Otol ith Channel Channel

Pitch or Roll Yaw

Semicircul ar Canal Semicircular Canal
Channel Channel
9 (mIs 2) +9.81 in surge/pitch +9.81 +9.81
-9.81 in sway/roll
Random No; se
Filter Time
Constants

Surge or sway Heave accel era-

accel erati on tion
Pitch or roll angle Yaw angl e
pena lty We i ghts

Ru Rollor pitch angle Yaw angl e

R22 Sway or surge accel- Heave accel era-
eration tion
Rd II Integral of surge or Integral of
sway positi on heave position
R Surge or sway Integral of yaw Heave position
d22
positi on angle
Rd 33 Surge or sway Doubl e i ntegral Heave velocity
velocity yaw angle
P on RII to Rd on Ru' Rd ,Rd
33 22 33
Semicircular canal Semicircul ar canal
.' error error
Oto 1 ith erro r Otol ith error

J.6
J04 Program Listings

Routine Name Function Page

WASHSPSR control program for Riccati solution J08
MATML performs matrix multiplication Jolt
MATAD performs matrix addition or subtraction Jolt
MATTR performs matrix transpose J 012
RICAT Riccati solver Jo12
HQR4 Jo14
EXCHNG Riccati sol ver subroutines to rearrange the Jo17
SPLIT eigenvalues of the solution Jo18
QRSTEP J020
*WASH12 solves for washout filter transfer functions,
given Riccati solution J021
ANALYZE finds roots and poles of the transfer
functions, and plots frequency response J028
*POLYEX expands the lower order polynomials J030
*Written in PL/l and FORMAC

J07
 1 C*** * PROGRAM WASHSPSR . FTN * .... 1
2 C 2
3 C THIS PROGRAM SOLVES FOR P IN THE STEAOY STATE RICCATI 3
4 C EQUATION: RIPRIHE - P.T*P + P*APRIME + APRIHETRANS*P K 0 4
:; C THIS VERSION SHOULO BE USED FOR PITCH-SURQE I< SWAY-ROLL CASES. 5
6 C THIS FORMULATION INCLUDES THE INTEQRAL OF SURQE OR SWAY 6
7 C DISPLACEMENT AND USES A 6TH ORDER SYSTEM MATRIX . 7
8 C INCQRPORATES CHANQES IN SEQUENCE OF EQUATIONS 8
9 C 9
10 0000001 IMPLICIT REAL.8 (A-H.O-Z) 10
11 0000061 DIMENSION AABAR(14. 14) . BB(6.2). BBBAR(14.2). 11
12 *DDBM(5.2). QQ(5. S). CCBAR(5.14). 12
13 * RHORR(2 . 2) . TEMPI(2.2) . BBBART(2.14) . 13
14 *CC BAR T <14. 5). DDBART<2.5). CBTQ(14.5) . RRI (14.14). 14
15 *RRI2(14 . 2). DBTQ(2. Sj. RRI2T(2 . 14) . RR2(2 . 2). IS
16 <tWKAREA(10) . RR2INV(2.2). TEHP2(2.14). TEHP3(14 . 14). Ilo
17 *RIPRIM (14.14). APR IHE( 14.14) . PP( 14.14). 17
18 *S(28). 0(28). ER(28). EI(28). Z(28 . 2B) . U(28. 28) . 1T(24) 18
19 C
19
20 C• • * * ENTER DIMENSIONS AND VARIABLES .*.* 20
21 C
21
22 0000061 11=2 22
23 00678CI -1-1=5 23
24 0067941 KK=6 24
25 00679C 1 LL=6 25
26 0067A41 MM=14 26
27 0067ACI AO=0. 076 27
28 0067881 BO=O. 19 28
29 0067C41 QO=5. 86 29
30 C
30
31 C PITCH SCC PARAMETERS : 31
32 C Tl=(S. 3+0. 1)/(5. 3.0. 1) 32
33 C T2=86 . 31/(5. 3.0. I) 33
34 C T3=1. 0/(5 . 3*0. 1) 34
35 C G= 9 . 81 35
36 C
36
37 C ROLL SCC PARAMETERS : 37
38 0067001 Tl= ( 6 . 1+0. 1)/(6. 1.0. 1)
38
39 0067DCI T2=118 . 55 / (6. 1*0. 1) 39
40 0067E8 1 T3= 1-. 0 / (6 . 1*0. 1) 40
41 0067F4 1 G= -9. 8 1
41
42 C
42
43 C OLD VESTIBULAR PARAMETERS ( SIVAN ' S) : 43
44 C GO=2 . 16 44
45 C TI = 5. 903/ (5 . 9* . 003)
45
46 C T 2 =233 . /( 5 . 9* . 003 ) 46
47 C T3=1 . 0 /( 5 . 9 • . 003) 47
48 C 48
49 0068001 AN11 - -0 . 2
49
50 006BOC 1 AN22=-1. 0 50
51 006B181 Qll=O . 4472 SI
52 0068241 Q2 2 =0. 8944 52
53 0068301 RHO=1.
53
54 00683C I Rll= . 9 54
SS 0068481 R22 =. 1581 SS
56 0068541 RD 11=. 1581 56
57 0068601 RD22= . 1581 57
58 00686C I R033= . 1581 58
59 0068781 WR ITE (2. 400> 11. -I.J . KK. LL . MM 59
60 006BA81 WRlTE (2. 410) AD. BO. GO
60
lol 0068001 WRITE (2. 415) Tl . T2 . T3.c:
61
62 006BFCI WRITE ( 2 . 420) -I. / ANIt.-1. / AN22
62
63 0069541 WRITE (2. 430) QII . Q22 . RHO . RII . R22 . RDII.RD22.R033 63
64 C 64
65 C*** * END OF VAR IABLE ENTRY ; NOW SET UP THE REQUIRED MATRICES **** 65
66 C 66
67 C FIRST THE A-BAR MATRIX : 67
68 0069901 DO 5 1=1 . MM 68
69 0069A41 DO 5 .J=I . MM
69
70 0069B8 I 5 AADAR(I • .J)=O. O
70
71 OObAOBI AADAR( I . I ) c-Tl
71
72 006AIAI AABAR( I . 2)=1. 0
72
73 006A261 AABAR(I.13)=-TI*T2
73
74 006A401 AADAR(2.1)=-T3
74
75 006A52 1 AADAR(2 . 13)=-T3.T2
75
76 006MCI AABAR (3. 3) =-BO
76
77 006A7EI AABAR(3.13)=QD.Q. (AO-BO) 77
78 006A9EI AABAR(3 . 14)=Q0.(AO-BO)
7B
79 006AB61 AADAR ( 4 . 5)=1. 0
79
80 006AC21 AABAR ( 5 . 6)=1. 0
80
81 006ACEI AABAR (6.14 )=1. 0
81
82 006ADAI AABAR(7.7 ) =-Tl
82
83 006AECI AABAR(7.8)=1 . 0
B3
84 006AFBI AA8AR (8.7) =-T3
84
85 006BOAI AABAR (9. 9) =-DD
85
86 006BICI AABAR(IO. II)=I . O
86
87 006B2BI AABM(II . 12) - 1 . 0
B7
88 006B341 AABAR( 13. 13)=ANll
88
89 006B401 AABAR(14.14)=AN22
89
90 006B4CI WR ITE (2.570)
90
91 006B6BI WRITE (2.575) «AABAR(I • .J) • .Ja>I.HH),J=I.I1I1) 91
92 C 92
93 C NOW MATRIX D: 93
94 006C24I DO 7 1-1. KK
94
95 006C38I DO 7 .J=I . I I 95
96 006C4CI 7 BB (I • .J)=O. 0
96
97 006C9CI BB (I. I )=-TUT2
97
98 006CB61 BB (2. I )=-T3H2
98
99 006COOI BB(3.1)=GO*Q*(AO-BO ) 99
100 006CFOI BB(3. 2)-G0*(AO-BO)
100
lOl 0060081 BB (6.2)=1. 0
lOl
102 0060141 WRITE (2.500) «BB(I • .J) • .J=I.!I) . I=I.KK) 102
103 C 103
104 C NOW MATRIX B-BAR : 104
105 006DCCI DO 10 1=1. MM
105
106 006DEOI DO 10 .J=1. ! I
106
107 006DF41 10 BBBAR(I.-I)=O. O
107
108 006E44I DO 11 1=1. LL
108
109 006E5BI DO 11 .J=I.11
109
110 006E6CI 11 BBBAR(I+KK • .J)=DB(I • .J)
110
111 006EDBI WR!TE (2.510)
111
112 006EF41 WR!TE (2.515) «BBBAR(I • .J).-J=I,JI),J=I . I1I1)
112
113 C
113
114 C NOW MATRIX C-BAR :
114
J. 8
115 006FACI DO 12 I.I • .J.J 115
116 006FCOI DO 12 .J~I.MM 116
117 006FD4I 12 CCDARII • .J) ~O. O 117
118 0070241 ceBM ( 1. 1) --1. 0 118
119 0070301 CCDAR I I. 7) ~1. 0 119
120 00703CI CCDARI I. 13)--T2 120
121 00704EI CCDAR I 2.3) .-1. 0 121
122 00705AI CCDARI2. 9)-1. 0 122
123 0070661 CCDARI2.13)--GO.o 123
124 0070801 CCBAR (2. 14) --GO 124
125 0070921 CCBAR(3, 10)-1. 0 125
126 00709EI CCBAR(4. 11 )-1. 0 126
127 0070...AI CCBARI 5.12)-1 . 0 127
128 0070861 WRITE 12.580) «CCBARII • .J) ...... 1.t1M).I-I • .J.!) 128
129 C 129
130 C NOW MATRIX D-DAR : 130
131 00716C I 00 1:5 101 • ..J.J 131
132 0071801 DO 1:5 .)=1.11 132
133 0071941 1:5 DOOM (1 • .J) cO. 0 133
134 0071E41 DOOAR( 1.1 ,-T2 134
135 0071FOI DOOM (2. 1 ) -OO*Q 135
136 007202 I DDBM (2.2) =-00 136
137 00720EI WRlTE (2. 540) (IOODARII • .J).~I. 11). I-I • .J.!) 137
138 c 138
139 C NOW MATRIX G: 139
140 0072C4 I DO 20 I-I • .J.J 140
141 0072081 DO 20 .J-I • .J.J 141
142 0072ECI 20 GGII • .J)~ . 0 142
143 00733CI GGll.1)..o11 143
144 0073481 GG(2. :;!)~22 144
145 0073541 GG I 3. 3 ) -RH01tRO 11 145
146 0073661 GG 14.4 )-RH01tR022· 146
147 0073781 GG(S,5)-RHO*RD33 147
148 00738AI WRlTE (2. :5:50) «00(1,"),.1-1,..),).1-1,")') 148
149 C 149
ISO C NOW MATRIX RHO*R : 150
151 0074401 RHORRC I. 1 )-RHO-RII 151
152 0074521 RHORR I 1.2)-0. 0 152
153 00745EI RHORR(2.1>-O.O 153
154 00746AI RHORRI2.2)·RHO*R22 154
155 00 747CI WRlTE 12.560) I IRHORRI I • .J) . ..... I. 11). 1=1. 1I) 155
156 C 156
157 C**** END OF MATRIX SETUP **** 157
158
159
160 0075341
C
C**** CALCULATE TRANSPOSES OF B-BAR. C-BAR.
CALL MATTRII1I1. 11. DDDAR . DBDART)
D-BAR .*.. 158
159
160
161 0075581 CALL MATTRI.J.!.MM. CCDAR.CCDART) 161
162 00757CI CALL MATTR I.J.!. II. OODAR. OODART! 162
163
164
165 0075"01
C
c •••• CALCULATE Rl=C-BAR-T*Q*C-BAR : .*.*
CALL MATML 11111 • .J.J • .J.J. CCDART • GO. CeTG)
163
164
165
166 0075CC I CALL MATML 11111 • .J.J.MM. CDTG. CCDAR. RRI) 166
167 007SF81 WR ITE 12.590) 167
168 0076141 WRITE 12.575) I IRRIII • .J) • .J-I. MM). 1=1. MM) 168
169 C 169
170 C** •• CALCULATE R12-C-BAR-T*O*O-BAR. AND R12-TRANSP 170
171 007óCCI CALL MATMLII1I1 • .J.J. II.CDTG.00DAR.RRI2) 171
172 007bF81 WRITE 12.600) 172
173 0077141 WRlTE (2.515) I IRRI21l • .J) • .J-I. I1). I-I. MM) 173
174 0077CCI CALL MATTR 11111. II. RRI2. RRI2Tl 174
175 C 175
176 C**** CALCULATE R2=RHO*R+O-8AR-T*O*O-BAR ** •• 176
177 0077FOI CALL MATMLI II • .J.J • .J.J. OODART. GO. ODTG) 177
178 00781C I CALL MATMLI II • .J.J. II. ODTO. OODAR. TEMP!) 178
179 0078481 CALL MATAO I II. II. RHDRR. TEMPI. RR2. 1) 179
180 00787C I WRlTE 12.610) IIRR211 • .J) • .J=I. l I l . I-I. I1) 180
181 0079341 CALL LINV2FIRR2. II. 1I. RR2INV. O. WKAREA. IER) 181
182 0079bCI IF liER . GT . 128) STOP 182
183 0079841 WRlTE 12.620) I IRR2INVII..J) • .J-I. 11). I-I.Il). IER 183
184 C 184
185 C***. NOW CALCULATE THE MATR ICES REOUIREO FOR USE IN RlCAT **** 185
186 C 186
187 C FIRST MUST FINO RI-PRIME-RI-R12*R2-INV*RI2-TRANSP 187
188 007A481 CALL MATMLI Il. II. MM. RR2INV. RRI2T. TEI"P2) 188
189 007A74I CALL MATMLIMM. II.MM.RRI2.TEI"P2.TEMP3) 189
190 OQ7AAOI CALL MATAD(MM,MM.RRl.TEHP3.RlPRlM, -1) 190
191 007...041 WRlTE 12.630) 191
192 007AFOI WRlTE 12. ~7~) IIRIPRIMII • .J),.J-I.MM).!=I.MM) 192
193 C 193
194 C NOW FIND A-PRIME-A-BAR-B-BAR*R2-INVitRl:2-TRANSP 194
195 0078A81 CALL MATMLIMM. II.MM.DDDAR . TEMP2.TEMP3) 195
196 007804 I CALL MATAOIMM.MM.AADAR.TEMP3.APRIME.-I) 196
197 007C081 WR lTE 12. 640) 197
198 007C24I WRlTE lOl. ~7~) I (APRIMEII • .J) • .J-I.l1M). I=I.MM) 198
199 199
200 NOW FINO TmB-BAR.R2-INV*D-DAR-TRANSP 200
201 007COCI CALL MATMLI II. 11 . MM. RR2INV. DBDART. TEMPOl) 201
202 0070081 CALL MATMLII1I1. II.MM.DDDAR.TEMP2.TEMP3) 202
203 0070341 WRlTE 12.650) 203
204 0070501 WRlTE 12.575) IITEMP3II • .J) • .J=I.I1I1).I=I.MM) 204
205 C 205
206 C•••• WE NCW HAVE THE RICCATI EOUATION IN THE FORM: ****.* 206
207 e RIPRIM - P*TEt1P3*P + P*APRHE + APR!MT*P .. 0 207
208 C**** SO ~ CALL SUBROUTIPE RICAT TO SOLVE FOR MATRIX P ****** 208
209 C 209
210 007E08I CALL R I CAT< APR II'E. TEMP 3. RI PRIM. Z. U. 1111. 2*11M. MM. 2*11M. 210
211 *S.O.ER.EI. IT.!. 00-12. IER) 211
212 007E84I WRlTE 12.660) IER 212
213 007EA4I IF IIER.NE . O) STOP 213
214 C 214
215 C PRINT THE P-MATRIX WHICH IS STOREO IN Tel' LEFT CORNER OF Z 215
216 007EDCI WRlTE (2.57:5) «Z(I,J),J=l,I1'1),I-l,l"t1) 216
217 C 217
218 C CREATE MATRIX PP OF PROPER DIMENSIONS TO USE IN MAT/t... 218
219 C ANO PRINT OUT THE SUDMATRICE9 OF INTEREST P21. P22. P23. 219
220 007F741 DO 35 1-1.1'111 220
221 007F88I DO 35 .)-l.Ht1 221
222 007F9CI 35 PPII • .J)·ZII • .J) 222
223 0080001 WRlTE (2.700) 223
224 0080lCI WRlTE 12. 70~) IIPPII • .J) • .J=1.KK).I-KK+I,)UHLLl 224
225 0080E0I WRlTE 12.710) 225
226 0080FC I WRITE (2.705) lIPPIl • .J) • .J=KK+ld\K+LLl. looKK+I. KK+LLl 226
227 0081001 WR lTE (2.715) I IPP I I • .J) • .J-KK+LL+I. KK+LL+Il). l-KK+I.I\K+LLl 227
228 228
J .9
 :/29
C.... ** NOW CHECK THAT THE RICCATI SOLVER HAS BEEN SUCCESSFLL****
230 C 229
231 C RR 1=f'*T 230
232 0082110 I CALL MATML(MM.MM. MM.PP . TEMP3.RRI) 231
233 C TEMP3 s P*T*P 23:2
234 0082DCI CALL MATML(MM . MM.MM . RRI.PP.TEMP3) 233
23~ C RR 1..p*A~R IME 234
236 0083081 CALL MATML(MM.MM. MM . PP.APRIME.RRI) 235
237 C AABAR=APRIMTRANS 236
238 0083341 CALL MATTR(MM.MM. APRIME . AABAR) 237
239 C APRIMEzAPRlMETRANS*P 23B
240 0083~81 :239
CALL MATML(MM.MM.MM . AABAR.PP.APRIME)
241 C TEMP3 s RIPRlME-P*T*P 240
242 0083841 CALL MATAD(MM.MM.RIPRIM.TEMP3.TEMP3._I) 241
243 C TEMP3=RIPRlME-P*T*P+P*APRIME 24:2
244 0083881 CALL MATAD(MM.MM.TEMP3.RRI.TEMP3. I) :243
243 C 244
TEMP3=RIPRIME - P*T*P + P*APRIME + APRlMETRANS*P
246 0083ECI CALL MATAD ( MM.MM.TEMP3.APRlME.TEMP3.1) 245
247 C 246
248 :247
C THE P-DOT MATRIX (WHICH SHOLLD BE =0) IS NOW PRINTED
249 0084201 WRlTE(2.670) 24B
250 00843C I WRlTE (2. ~75) «TEMP3(1 • .J).J=I""M).I=I.MM) :249
251 C 2~0
252 C. . . . FORMAT STATEMENTS **** :251
2~3 C 2~:2
254 0084F4 I :253
2~~
400 FORMAT ( • RICCATI EOUATION SOLVER- P-S OR R-S CASE'11113X.
*'MATRIX DIMENSIONS : ·llex. ·IIm·. 13. 5X. 254
2~6 255
*'.J.J='. 13. ~X. 'KK· ' . 13. 5X. 'LL. · . 13. 5X. 'MM. ' . 13)
257 C :256
2~8 008~741
410 FORMAT (/'- OTOLITH MODEL : ' I IBX •• AC-'. F9. 4. ~X. 2~7
2~9 25B
*'SO='. F9. 4. 5X. ' 00- ' . F9 . 4)
260 008~B21 259
415 FORMAT(/ ' - SCC MODEL : · IIBX. 'Tl.'. F9. 3. 5X.
261 *·T2='. F9. I. 5X. 'T3= ' . F9 . 4 . ~X . 'G=·. F9. 3) 260
262 0085F4 I 261
420 FORMAT(/'- NOISE FILTER TIME CONSTANTS: 'IIBX.
263 262
*'-I/ANII (ON ANG)='.FB . 3 . 5X. ' -lIAN22 (ON ACC).'.FB . 3)
264 00B6521 263
430 FORMAT (/'- WEIGHTING ELEMENTS : • I IBX. '011 (ON ESCC).'.
26~ 264
*Fe . 4 . T43. '022 (ON EOTO)· ' . FB. 411BX. 'RHO (ON Ril TO RD33).'.
266 263
*FB . 4. T43, 'Ril (ON ANQ)= '. FB. 4. T70. 'R22 <ON ASI ) s ' ,
267 266
*FB . 4 1 1BX. ' RDll : ON l)m '. FB . 4 . T43. ' RD22 (ON DISP).'.
26B *FB . 4. T70. ' RD33 <ON VELl =·. FB. 4) 267
269 C 26B
270 00B726I 500 FORMAT (' l MATRIX B: ' l I l 269
271 270
*2X , 2(IPGI5 . 5) 112 X. 2(IPGI5. 5) 1 12X.2(IPGI5. 3)112X.2(IPGI5. 5)11
27:2 *:2X . 2(IPGI5 . 5)/ /:2 X. 2<lPGI5. 5» 271
273 C 272
274 00B7BC I 510 FORMAT(/ II I '- MATRI X S- SAR : ' ) 273
275 00B7AC I 515 FORMAT(/I I :2X . 2(IPGI5. 5) // 2X.2(IPGI5. 5>11 274
276 275
277 *:2X. :2( I PGI5 . 5) I 12X . :2<lPGI5. 5) I 12X. 2 (lPG 15. 5) 112X.:2( IPGI5. 5) I I
*:2X. 2( IPG15 . ~) I/ :2X . 2( IPGI5 . 5) 1 I :2X. 2 (IPGI5. 5>1 12X. 2( IPGI~. 5>1 I 276
:27B
*:2 X• 2 ( I PGI5 . ~>I 1 2X . 2( IPGI5 . 5) I I:2X. 2 (IPGI5. 5) 112X.:2( lPGIS . 5) ) 277
279 C 27B
2BO OOBB6C I 279
2BI 540 FORMAT (/11 ' - MATRIX D-BAR : ' 1112X. 2<lPGI5. 5)112X.2(IPGI5. 5>11
*2X . 2(IPGI5 . 5>1 I:2 X. 2(IPGI5 . 5>112X.2(IPGI5. 5» 280
2B2 OOBecC I 281
2B3 550 FORMAT (/1 1 ' - MATRIX G: '1 112X. 5(IPGI5. 5)112X. 5(IPGI5. 5)/1
*2X . 5(IPGI5 . :l>l 1 2 X. 5(IPGI5. 5>112X. 5(IPGI5. 5» 282
284 00B92BI 283
560 FORMAT< II I ' - MATRIX RHa_R : '1 //2X . 2(IPGI5. 5>11
285 *2X. 2( IPGI5 . 5» 284
286 00B962I 570 FORMAT ( ' 1 MATRI X A- BAR : ' ) 285
287 00B97EI 575 FORMAT (/ I I 286
28B 287
*2X . 7 ( IPGI5 . 5)/ :2 X.7 (IPGI5. 5>112X . 7( IPGI5. 5>12X. 7(IPGI5. 5)11
289 28B
*2X. 7( IPGI5 . 5 >1 :2X. 7 (IPGI5. 5) I I:2X. 7( IPGI5. 5>12X. 7( IPGI5. 5)/ I
290
*:2X. 7( I PGIS . ~>l2X . 7 (IPGI~. 5 ) 11 2X. 7( lPGIS. 5) I:2X. 7( lPGIS. 5)/1 289
291 290
*:2X . 7 ( I PGI5 . 5>12X, 7 ( lPG 15. 5) I I :2X . 7( lPGI5. 3) 12X. 7( lPGI5. 5)11
:292
*2X . 7( IPGI~ . 5 ) 12X. 7 (IPGI~. ~) // :2X . 7( lPGIS. S>I:2X. 7( lPGIS. 5)/ I 291
293
*2X . 7( lPGIS . ~>l2X . 7 ( lPG I~ . 5 >1 12X . 7( IPGI~ . 5) 12X. 7( IPGI~ . 5)1 I 29:2
:294 :293
*:2X . 7( IPGI~ . 5>1:2X . 7 (lPG 1:5. :5 >I 12X. 7( IPGI:5. "/2X. 7( IPGI:5. 5»
295 C 294
296 00BAE61 :5BO FORMAT< ' I MATRI X C- SAR : '1 1 1 295
297 296
*2X . 7( IPG15 . ~) I:2X . 7 (lPG I~. 5) I 12X. 7( IPGI~. ~) I:2X. 7( IPGI~ . ~)I I
:29B 297
*2X. 7( I PGI5 . 5) I:2X . 7 (IPGI~ . ~ >I 1 2X . 7( lPGI:5. "/2X. 7( lPGI~ . 5)1 I
299 *2X.7(IPGI5. 5)/2X.7(lPGI5. ~» :29B
300 C 299
301 OOBSBOI 590 FORMAT ( '1 MATRIX RI : ' ) 300
30:2 00BS9BI 600 FORMAT ( 'I MATRIX RI:2 : ' ) 301
303 00BSS:21 302
610 FORMAT(/II'- MATRIX R:2 : ' / / / :2X.:2(IPGI5. 5)II:2X.:2(IPGI5 . 5»
304 OOBSEBI 620 FORMAT(/II'- MATRIX R2- INV : ' 1/1:2X . 2(lPGI5.:5)11 303
30~ 304
*:2X. 2(IPGIS . 5) I I:2X . 'IER=· . 13)
306 00BC2EI 630 FORMAT('I MATRIX RI-PRIME : ') 305
307 00BC4CI 640 FORMAT< ' I MATRIX A-PRIME : ') 306
30B 00BC6AI 650 FORMAT ( 'I MATRIX T: ' ) 307
309 00BCB21 30B
310 660 FORMAT C' I THE SOLUT I ON OF THE RICCATI EON (MATRIX P) IS : '11
*:2X. 'IER= · . 15) 309
311 OOBCCAI 310
670 FORMAT( ' I THIS SHOULD BE A ZERO MATRIX (MATRIX P-ooT) : ')
312 00B004I 700 FORMAT ( ' I MATRIX P-21 : ' l i l 311
313 00BD201 70:5 FORMAT (:2X. 6(IPGI~ . :5)II:2X. 6(IPGI5. 5)II:2X.6(IPGI5. ~>l1 31:2
314 313
*:2X, 6( IPGI5 . :5) 1 12X . 6<lPGI5. 5)II:2X. Ó (IPGI5. 5) )
315 00BD74I 710 FORMAT('- MATRIX P-22: ' 11) 314
316 00B0901 315
715 FORMAT ( '- MATRIX P-23 : '11:2X. 2(IPGI5. 5)II:2X.2(lPGI:5. 5)11
317 316
*2X. 2( 1 POlS . :5) I 12X. :2( IPGI5. 5)1 12X. 2 (lPGI3. 3) 112X. 2( IPGI5. 5) )
31B C 317
319 OOBDFBI STOP 31B
320 OOBEOOI END 319
320
NO ERRORS : F7D R05-00. 00 MAINPROG . MAIN 21/12/B4 09:00:07 TABLE SPACE: 9 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 150 WORDS
DOUBLE PRECISION FLOATING PT SUPPORT REOUIRED FOR EXECUTION

J . lO
 I
1 0000001 SUBROUTINE t1ATHL<L. H. N. A. B. Cl
C. . . . SUBROUTINE TO PERFORH HATRIX HULTIPLICATION •••• 2
2 3
3
4
5
0000041
0000041
OOOOOAI
I1'1PLlCIT REAL.8 lA-H.O-Il
DlHENSION AlL. Hl. Bll'1. Nl. CIL. NI
DO I I-I. L
,
4

6
6 00005CI DO 1 .)- 1. N 7
7 0000701 C ( l , ,)) -0. 0 8
8 00009AI DO I K- I.H 9
9 OOOOAEI C (1 ; '-') -CC I • .)+A( I. K)*B (K, ..J)
10
10 0001861 RETURN 11
11 00018CI END
NO ERRORS : F7D R05-00 . 00 SUBROUTINE HATI'1L 21/12/84 09:2':13 TABLE SPACE: I KB
STATEMENT BUFFER : 20 LlNES/1321 BYTES STACK SPACE : 172 WORDS
DOUBLE PRECISION FLOATINQ PT SUPPORT REQUIRED FOR EXECUTION

SUBROUTINE t1ATADlL . I'1 . A. B. C. Kl 12

I 0000001
2 c**** SUBROUTINE TO PERFORH HATRIX ADDITION Sc SUBSTRACTION **** 13
14
3
4
C**** SET K- I FOR ADDIT ION
C K- -I FOR SUBSTRACTION
II'1PLIC IT REAL.8 lA-H . O-Zl
.... 15
16
5 0000041 17
6 0000041 DII'1ENSION AlL. Hl. BeL. tU, CIL. Hl
18
7 OOOOOAI DO I I-I. L 19
8 0000'01 DO I .i- I.H 20
9 0000641 C( I . J) .. A( I • .)+K*B( 1 • ..J)
21
' 10 0001081 RETURN
22
11 OOOIOEI END
NO ERRORS: F7D RO'-OO . OO SUBROUTINE MATAD 21/12/84 09: 25: 14 TABLE SPACE : I KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 158 WORDS
DOUBLE PRECISION FLOATINQ PT SUPPORT REQUIRED FOR EXECUTION

J.U
 1 0000001 SUBROUTINE MATTR(L.M.A.B)
2 23
C**** SUBROUTINE TC PERFORM MATRIX TRANSPOSITION *** 24
3 0000041 IMPLICIT REAL*8 (A-H.O-Z)
4 0000041 DIMENSION A(L.M). B(M.L) 25
5 OOOOOAI · DO I I=I.L 26
6 00004CI DO I ,)-1. M 27
7 0000601 B(,).I)-A(I.,) 2B
8 0000D81 RETl*IN 29
9 OOOODEI END 30
31
NO ERRORS:F7D R05 - 00.00 SUBROUTINE MATTR 21/12/84 09:25: 15 TABLE SPACE: I KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 150 WORDS
DOUBLE PRECISION FLOATING PT SUPPORT REOUIRED FOR EXECUTION

I 0000001 SUBROUTINE RICAT(A.G.H.Z. U. NA.NZ. NNA.NNZ.S.O.ER.EI.T.EPS.IER)

2 0000041 32
REAL*8 A(NNA.NA).G(NNA.NA).H(NNA.NA).Z(NNZ.NZ).U(NNZ.NZ)
3 OOOooAI REAL*B S(NZ). 0 (NZ). ER (NZ). EI (NZ). EPS 33
4 OOOooAI INTEGER NA.NZ.NNA.NNZ. IER.T(NZ).I.,) 34
5 OOOooAI INTEGER LOW. IGH 35
6 C 36
7 C 37
8 C THIS ROUTINE SOLVES THE ALGEBRAIC RICCATI EOUATION : 3B
9 C 39
10 C 40
11 C 41
12 C THE METHOD USED IS DECOMPOSITION OF THE MATRIX 42
13 C 43
14 C = [A • -G ] 44
IS C [-H.TRANSPOSE(-A)] 45
16 C 46
17 C INTO S=TRANSPOSE(U)*Z*U 47
18 C 4B
19 C 49
USING ORTHOGONAL TRANSFORMATIONS. THIS METHOD IS DESCRIBED
20 C BY A.,). LAUB. IN UA SCHUR METHOD FOR SOLVING ALGEBRAIC 50
21 C RICCATI EOUATIONS u • FROM VOL. AC-24. NO. 6. OF THE IEEE 51
22 C TRANSACTIONS ON AUTOMATIC CONTROL. P. 319. 52
23 C 53
24 C THE SUBROUTINES BALANC 54
25 C ORTHES 55
26 C ORTRAN 56
27 C AND BALBAK ARE REOUIRED FROM E ISPACK. 57
28 C SB
29 C THE DOUBLE-PRECISION SUBROUTINES LEOT2F 59
30 C 60
LUDATN 61
31 C LUELMN
32 C 62
LUREFN 63
33 C UERTST
34 C 64
UGETIO
35 C 65
VXADD 66
36 C VXI'f.JL
37 C AND VXSTO ARE REOUIRED FROM IMSL .
67
38 C 6B
39 C IN ADDITION THE SUBROUTINES HOR4 69
40 C 70
EXCHNG
41 C 71
SPLIT
42 C 72
AND ORSTEP ARE REOUIRED. THE LAST THREE 73
43 C OF THESE ARE AVAILABLE AS PART OF ALGORITHM 506: HOR3 AND EXCHNG.
44 C FROM THE COLLECTED ALGORITHI'1S FROM THE ASSOCIATION FOR 74
45 C 75
COMPUTING MACHINERY. THE FIRST ONE. HOR4. IS A MODIFICATION
46 C OF HOR3. DESIGNED TD ORDER THE EIGENlIALUES OF THE UPPER 76
47 C HESSENBERG MATRIX IN REVERSE ORDER. BY THE REAL PART OF 77
48 C THE EIGENVALUES. 7B
49 C 79
50 C THE PARAMETERS IN THE CALLING SEOUENCE ARE (STARRED BO
51 C PARAMETERS ARE AL TERED BY THE SUBROUTINE) : BI
52 C B2
53 C A \ B3
54 C B4
55
G > MATRICES DEFINING THE RICCATI EOUATION.
B5
C H /
56 C B6
A DOUBLE-PRECISION MATRI X. OF DIMENSION NNZ BY NZ .
57 C B7
ON OUTPUT. THE UPPER LEFT NA BY NA PORTION OF Z
BB
J.12
 CONTAINB nE SOLUTION OF TIE RICCATI EOUATION. X. B9
5B C
C tIU A DOUBLE-PRECISION. NNZ DY NZ WORI'SPACE VARIABLE. 90
59
NA nE ORDER OF nE RICCATI EOUATION I1I\TRICES. 91
100 C
NZ AN I NTEOER THAT MUST BE SET EOUAL TO 2*NA. 92
lol C
NNA AN INTEOER VARIADLE GREATER THAN OR EOUAL TO NA. 93
102 C
6ET EOUAL TO THE ROW DIMENSION OF A. 0 AND H 94
103 C
A6 DEFINED IN THE CALLINO PROORAM. 95
104 C 96
b~ C 'NNZ AN INTEOER. OREATER THAN OR EOUAL TO Nl. SET
EOUAL TO THE ROW DIMENSION OF U I\ND Z. AS 97
blo C 98
107 C DEFINED IN THE CALLINO PROORAM.
A DOUBLE-PRECISION WORK6PACE VECTOR. OF LENOTH 99
loB C
AT LEAST NZ. 100
109 C 101
70 C A DOUBLE-PRECISION WORKSPACE VECTOR. OF LENOTH
AT LEAST Nl. 102
71 C 103
72 C \ ON OUTPlIT. THE DOUBLE-PRECI610N VECTORS ER AND
/ EI. OF LENOTH AT LEAST NZ. CON,,,IN THE REAL AND 104
73 C
IMAOINARY PARTS OF THE EIOENVALUES OF (A-OXI 105
74 C
106
7~ C (IE. THE CLOSED-LOOP SPECTRUM I IN THE FIRST
NA POSITlONS. 107
76 C
108
77 C AN INTEOER VECTOR WORKSPACE. OF LENG TH AT LEAST NZ.
THE DESIRED ACCURACY TO WHICH THE EIGENVALUES 109
7B C
OF (A-QX I. AND THE VALUE OF EACH ELEMENT OF X. 110
79 C lil
BO C ARE TO BE CALCULATED.
AN INTEGER ERROR-COMPLET ION CODE. 112
BI C *IER
IER-O INDICATES A NORMAL COMPLETlON. 113
B2 C 114
B3 C x-I INDICATES TH AT NA. Nl. NNA OR NNZ IS
DEFINED INCORRECTLY. 115
B4 C
B~ C =1 WHERE [0 < I < NZ+I] INDICATES THAT _ 4 116
FAILED WHEN WORKINO ON EIOENVALUE .1. 117
Bb C
-I WHERE 1--34. -129 OR -131 INDICATES THAT I1B
B7 C 119
BB C LEOT2F FAILED <Uil WAS UNINVERTABLEI.
RETURNINO A WARNINO OF -I (6EE THE IMSL 120
B9 C
121
90 C DOCUMENTATIONI .
122
91 C
IF(NA. OT. NNAI GaTO 10 123
92 oooooAI
IF(NZ. OT. NNZI GOTO 10 124
93 OOOOBBI
IF(NI . NE. 2*NAI WTO 10 125
94 OOOQAOI
126
9~ OOOOBCI GOTO 20
IER--I 127
910 000OC2[ 10
RETURN 12B
97 OOOOCAI
DO 30 I-I. NA 129
9B OOOODOI 20
DO 30 .)ol. NA 130
99 0000E41
Z([,,)l-A(I,,}) 131
100 000OF81
Z([+NA. ')1--HO.')1 132
101 0001401
Z([. ')+NAl:-O (I.,) I 133
102 0001941
Z(I+NA.'-'+NA).-AC,J. I) 134
103 OOOIEBI
CONT[NUE 135
104 0002421 30
CALL BALANC(NNZ.NZ.Z.LOW. IOH.Sl 136
10~ 0002721
CALL ORTHES(NNZ.NI.LOW. IOH.Z.Ol 137
lOb 0002C0[
CALL ORTRAN(NNZ.NI.LOW.IGH.Z.O.Ul 13B
107 0003101
CALL HOR4(NNZ.NZ.EPS.T.Z.ER.EI.U. IERI 139
lOB 000374 [
IF(IER . NE. Ol RETURN 140
109 0004041
CALL BALBAK(NNZ.NZ.LOW. IOH.S.NZ.Ul 141
110 00041AI
DO sa 1-1. NA 142
111 oo04bCI
DO 50 -.lxl. NA 143
112 0004901
Z(I. ,)l-G( 1.,)1 144
113 0004941
114 0004DCI z([. ,)+NAl-H( 1.,)1 145
G([. ,)l-U(,). I I 146
115 00052AI 147
1110 000~721 H([. ,)1=U (-.l+NA. I I
CONTINUE 14B
117 000~01 ~O
149
118 000:lF01 IDGT-DLOOI0(EPSI
CALL LEOT2F(O.NA. NA.NNA.H. 10GT.U. IERI 150
119 00010101
IF(IER . NE. Ol IER--IER 151
120 00010781 152
121 0006961 DO 60 l-l,NA
DO 60 ,)Kl. NA 153
122 OOOMAI 154
123 0006BEI eHl, ".I)-Z(I • ..)
ZU. ,J)-(H<I, ,J)+H(,J. I) )/200 155
124 0007061
CONTINUE 156
125 0007781 60
157
126 0007A81 DO 70 Ixl.NA
DO 70 -.1=1. NA 15B
127 0007BCI 159
12B 0007DOI Hn. ,J).Z (t • .J+NA)
CONTINUE 160
129 OOOBIEI 70
161
130 OOOB4EI RETURN
162
131 000B54I END

NO ERROR6: F7D ROS-CO . OO SUBROllTINE RICAT 21/12/B4 09:2~ : IB TABLE SPACE : 3 KB

STATEMENT BUFFER : 20 LINES/1321 BYTES 6TACK SPACE : IBI WORD6
DOUBLE PRECISION FLOATINO PT SUPPORT REOUIRED FOR EXECUTION

J . 13
 1 0000001 8UBROUTINE HOR4(NA . N.EPS.TY PE.A.ER . EI.V.IER R)
2 0000041 INTEQER N. NA. TYPE(N). IERR 163
3 ooOOOAI REAL*8 A(NA.N) .EI(N).ER (N).EPS.V (NA. N) 164
4 C 165
5 C 166
6 C 167
7 C Th. followin g prog r .m i s • modifica tion of th. below- 168
8 C mentione d algortthm : 169
9 C 170
10 C 171
11 C COLLECTED ALGORITHMS FROM ACM (THE ASSOCIATION FOR 172
12 C COMPUTING MACHlNERY) 173
13 C 174
14 C ALQORITHM 506 17~
15 C 176
16 C HGR3 AND EXCHNG : FORTRAN SUBROUTINES FOR CALCULATINQ 177
17 C AND ORDERING THE ElQENVA!..UES OF A REAL UPPER HESSENBE 178
18 C MATRIX CF2] RG 179
19 C 180
20 C Q. W. STEWART. UNIVERSITY OF MARYLAND 181
21 C 182
22 C PUBLISHED IN THE ACM TRANSACTIONS ON MATHEMAT 183
23 ICA!.. SOFTWARE. 184
C VOL. 2. NO. 3. SEPTEMBER 1976.PAQ ES 275-280.
24 C 185
25 C SUBROUTINE HGR3 186
26 C 187
27 C 188
28 C 189
29 C HGR3 .REDUCES THE UPPER-HESSENBERG MATRIX A TC 190
30 GUASI- 191
C TR IANGULAR FORM BY UNITARY SIMlLAR ITY TRANSFOR
31 C THE EIGENVALUES OF A. WHICH ARE CONTAINED IN
MATIONS. 192
THE
32 C AND 2*2 DIAGONAL BLOCKS OF THE REDUCED MATRIX. ARE1*1 193
33 C ORDERED IN DESCENDING ORDER OF MAGNITUDE ALONG 194
34 THE 195
C DIAGONAL . THE TRNSFORMATIONS ARE ACCUMULA
35 TED IN THE 196
C ARRAY V. HGR3 REGUIRES THE SUBROUTINES
36 EXCHNG. 197
C GRSTEP AND SPLIT.
37 C 198
38 C HOR4, on the other hand. orders th. eigenval 199
39 u.s a f A 200
C accordin g to thltil' ,.ea1 portions . Th. eigenval
40 ue with the
C most negative real part app •• rs fir.t i that with 201
41 the most 202
C positive 1' • • 1 part appears last .
42 C Th. purpose foT' th15 modifica tion Is to 203
43 order th.
C etgenval u •• 50 th . t thev ca" be u •• d to
c.leut.t . the
204
44 C solutio" ta the algebrat t Rfceat! ltC[uatton , 205
45 accordin g
C to a method develope d by A . ~ . Laub. 1n UA SchuT' 206
46 M_thod fo~ 207
C Solving AlgebrAi c Riccatl Eq.uation .", 'rom Vol. AC-24,
47 C No . 6 . of th. IEEE T~.nsActions on AutofYti 208
48 c Cont~ol. p . 913.
C 209
49 C THE PARAMETERS IN THE CA!..L ING 210
50 C SEGUENCE ARE (STARRED PARAMETERS ARE ALTERED BY THE 211
51 C SUBROUTINE) : 212
52 C *A AN ARRAY THAT INITIALLY CONTAINS THE N * N 213
53 c UPPER HESSENBERG MATRIX TO BE REDUCED. ON 214
54 C RETURN A CONTAINS THE REDUCED. GUASI- 215
55 C TRIANGULAR MATR IX . 216
56 C AN ARRAY THAT CONTAINS A MATRIX INTO WHICH 217
57 C THE REDUC ING TRANSFORMAT IONS ARE TO BE 218
58 C MUL TIPLIED. 219
59 C N THE ORDER OF THE MATRICES A AND V. 220
60 C EPS A CONVERgENtE CR I TER ION. 221
61 c *ER AN ARRAY THAT . ON RETURN. CONTAINS THE REAL 222
62 C PARTS OF THE ElgENVALUES . 223
63 C EI AN ARRAY THAT . ON RETURN. CONTAINS THE 224
64 C lMAGlNARY PARTS OF THE ElQENVA!..UES. 22~
65 C *TYPE AN INTEGER ARRAY WHQSE I-TH ENTRY IS : 226
66 C o IF THE I-TH EIQENVALUE IS REAL 227
67 C 1 IF THE I-TH ElGENVALUE IS COMPLEX 228
68 c WITH POSITIVE IMAGINARY PART. 229
69 C 2 IF THE I-TH ElQENVALUE IS COMPLEX 230
70 C WITH NEGATIVE IMAGINARY PART. 231
71 C -1 IF THE I-TH EIQENVALUE WAS NOT 232
72 C CALCULATED SUCCESSFULLY. 233
73 C NA THE FIRST DIMENSION OF THE ARRAYS A ANC V. 234
74 C THE CONVERGENCE CR lTER ION EPS IS
USED TO DETERMINE 235
75 C WHEN A SUBDIAGONAL ELEMENT IS NEGL IGIBLE. 236
76 C SPECIFICALLY. A(I+l . I) IS REGARDED 237
77
AS NEgLIGIBLE IF
C DABS( A 0+1. I) ) . LE. EPS*<DABS(A< 1. I) +DABS(A( 1+1. 238
78 C TH IS MEANS THAT THE FINAL MATR IX RETURNED
1+1) ) . 239
79 C PROGRAM WILL BE EXACTLY SIMILAR TO
BY THE 240
80 A + E. WHERE E IS
C OF ORDER EPS*NORM(A) . FOR ANY REASONABLY
BALANCED NORM. 241
81 C SUCH AS THE ROW-SUM NORM. 242
82 C ON A NORMAL RETURN. IERR- O. IF FOR SOME ElGENVALUE 243
83 C FAILED TO CONVERQE AFTER 30 ITERATIONS . IERR
THE 244
84 IS SET TO THE 245
C INDEX OF THAT EIQENVALUE. ER AND EI
85 WILL THEN CONTAIN THE
C EIQENVALUES FDR INDICES IERR+l UP
TO N. THEY WILL BE ORDER ED. 246
86 C VALID EIQENVALUE9. BUT MAY NOT REPRE9ENT THE 247
87 ElQENVALUES OF 248
C THE SYSTEM WITH THE GREATEST REA!.. PART .
88 C 249
89 C INTERNAL VARIABLES : 250
90 OOOOOAI INTEGER I . IT.L.MU.NL.NU.MU2 . FLAQ.NV.NLOW.NUP 251
91 OOOOOAI REAL*8 El . E2. P.G.R.S. T . W.X.Y. Z. EI2 . E22 252
92 OOOOOAI LOQICAL FA IL 253
93 OOOOOAI NLOWool 254
94 00004CI NUP-N 25~
95 0000581 NV-NA 256
96 C INITIALI ZE 257
97 0000641 DO 10 I-NLOW.NUP 258
98 00007CI TYPE( 1)*-1 259
99 0000921 10 CONTINUE 260
100 ooOOAAI T-ODO 261
101 C MAIN LOOP; FIND AND ORDER EIQENVALUES. 262
102 0000B61 NU-NUP 263
103 000OC21 20 IF(NU. LT. NLOW) GO TO 240 264
104 OOooDAI ITaO 265
105 C GR LOOP . FIND NEGLlgIBLE ELEMENTS AND PERFORM GR 266
106 0000E21 30 CONTINUE STEPS . 267
107 C SEARCH BACK FOR NEQLIGIBLE ELEMENTS . 268
108 0000E21 L=NU 269
109 OOOOEEI 40 CONTINUE 270
110 OOOOEEI IF(L. EG. NLOW) GO TO 50 271
111 0001061 IF(DABS( A(L. L-l» . LE. EPS-(DA BS(A(L-l .L-l»+DA 8S(A(L.L 272
112 1 TO 50 l ») gO 273
113 0002001 L=L-l 274
114 00020EI GO TO 40 275
276
J.14
115 0002121 50 CONTINUE 277
116 C TEST TO SEE IF AN ElgENVALUE OR A 2*2 BLOCK HA8 BEEN FQUoID . 278
117 0002121 X- A<NV. NJ) 279
118 oo023CI IF(L. EO. NU) gO TO 160 280
119 0002'41 V - A (NU-I, NU-I) 281
120 000;ze21 W-A(NU. NJ-I )*A(NJ-I. NU) 282
121 000:m41 IF (L. EO. NU-I) CIO TO 100 283
122 C TE8T ITER"'TION CooNT. IF IT 18 30. OOIT. IF IT 18 284
123 C 10 OR 20. SET UP AN "'O-I-«JC SHIFT. 28'
124 00021'01 IF (IT . EO. 30) gO TO 240 286
IF (IT. NE. 10 . ANO. IT. NE. 20) go 'ro 70 287
12' 0003061
126 C ",O-I-«JC SHIFT 288
127 OOO:l2C 1 T - T+X 289
128 oo033EI 00 60 I-NLOW. NU 290
129 0003'61 ... (I.I)-I\(I.I) - X 291
130 00031\41 60 CONTINUE 292
131 0003BCI S_OABS( ... (NU.NU-l»+DABS( ... (NU-l.NJ-2» 293
132 0004'81 X- O. 7:5*9 294
133 oo046AI v- x 29'
296
134 0004761 W_ _ O. 437:5.S*5
13' 0004961 70 CONTINUE 297
136 0004961 IT- IT+l 298
137 C LOOK FOR TWO CONSECUTIVE SI1ALL. SUBOI ... OONAL ELEIENT8. 299
138 OOO4A4 1 NL - NU-2 300
139 0004B21 80 CONTINUE 301
Z- "'(NL . NL) 302
140 oo04B21
R-X-Z 303
141 oo04DCI
142 0004EEI S-V-Z 304
P-(R*S-W)/ACNL+l,NL)+ACNL.NL+l) 30~
143 000~1
144 000~41 Q- ACNL+l. NL+1) -Z-R-9 306
R- ACNL+2. NL+!) 307
14' 000~41
S-OABS (P )+OABS (O)+OABS (R) 308
146 000'021
147 0006341 p-ptS 309
148 0006461 0.- 0/9 310
R- RIS 311
149 0006'81
IF (NL. EO. Ll 00 TO 90 312
1'0 00066AI
IF ( DABS(A(NL. NL-U )*(O ...B8(O)+0 ... BS(R) l. LE. EPS*O ... BS(P)*(OABS (jIo( 313
1'1 0006821
1 NL-I.NL-I»+OABS(l)+O ...B8( ... (NL+I.NL+I»» CIO TO 90 314
152
NL - NL-l 31~
153 00080CI
00 TO 80 316
1'4 oo081AI
155 00081EI 90 CONTINUE 317
C PERFORI1 ... OR 8TEP BETWEEN NL "'NO NU. 318
1'6 319
1'7 00081EI C"'LL ORSTEP(·.... V. P. O. R. NL. NU. N. N.... NV)
1'8 oo087CI 00 TO 30 320
C 2 *2 BLOCK FOUNO. 321
1'9 322
160 0008801 100 IF(NU. NE.NLOW+I) "'(NU-l . NU-2)-ODO
161 0008C4 I A(NU, NU) - ACNl). NU)+T 323
162 0009121 ... (NU-I. NU-I ).A(NJ-I. NU-l )+T 324
163 0009681 TVPE(NU)-O 32'
164 00097EI TVPE( NU-I) - O 326
16' 0009961 I1U - NU 327
166 C LOOP TO POSITION 2*2 BLOCK . 328
167 0009A21 IlO CONTINUE 329
168 0009A21 NL - /1U-I 330
169 C "'TTEMPT TO SPLIT THE BLOCK INTO TWO RE"'L ElQENVALUES. 331
170 0009BOI CALL SPLITCA,V,N, NL, El.E2,NA, NV) 332
C IF THE SPLIT W"'S SUCCESSFUL. gO ANO OROER 333
171
C THE REAL EIClENVLAUES. 334
172
IF ( A(/1U.r1U-I) . EO. 0) CIO TO 170 335
173 000A081
C TEST TO SEE IF THE BLOCK 1 S PROPERLY POSIT IONEO. 336
174
C I F NOT . EXCHANCIE IT. 337
175
176 OOOMOI IF(/1U. EO. NUP) GO TO 230 338
177 000A581 IF(/1U. EO. NUP-l) GO TO 130 339
178 000A74 I IF ( A(I1U+2 . I1U+ll.EO . 0) GO TO 130 340
179 C THE NEXT BLOCK IS 2*2 341
180 OOOMEI I1U2-MU+I 342
181 ooOABCI CALL SPLIT(A. V.N. I1U2.EI2. E22. N.... NV) 343
182 000B141 IF (El. LE. E12) GOTO 230 344
183 000B2CI CALL EXCHNO(A.V.N.NL.2.2.EPS. FAIL.N.... NV) 34'
184 OOOB9CI IF ( . NOT. FAILl CIO TO 120 346
18' OOOBAEI TYPECNLJII:I-l 347
186 000BC41 TYPE(NL+l )·-1 348
187 OOOBOCI TYPECNL+2)--1 349
TYPECNL+3)=-1 3~0
188 000BF41
00 TO 240 3~1
189 OOOCOCI
190 00OC121 120 CONTINUE 3'2
191 00OCl21 HU- HU+;! 353
192 000C201 GO TO 150 3'4
193 000C26I 130 CONTINUE 3'5
194 C THE NEXT BLOCK IS 1*1 356
IF(EL LE. A(/1U+I.I1U+l» C:OTO 230 3~7
195 00OC26 I
196 OOOC601 CALL EXCHNOCA. V,N,NL,2 . 1,EPS, FAIL,NA,NV) 3'8
197 OOOCOO I IF L NOT. FAILl 00 TO 140 3'9
198 ooOCE21 TYPE(NU--I 360
199 000CF81 TYPE(NL+l )·-1 361
200 0000101 TYPE(NL+2)--1 362
201 0000281 00 TC 240 363
202 OOOO2EI 140 CONTINUE 364
I1U - /1U+I 36~
203 OOOO2EI
204 OOOOOC I 150 CONTINUE 366
20' OOOOOCI gO TO 110 367
206 C SI NelLE EIgENVALUE FOUNO 368
207 0000401 160 NL - O 369
208 0000481 AC NU, NU) - ACNU. NU)+T 370
209 0000961 IF(NU. NE. NLOW) A(NU.NU-U-ODO 371
210 0000041 TYPE(NU) - O 372
211 ooODEAI I1U - NJ 373
212 C LOOP TO P08ITlON ONE OR TWO REAL ElQENVALUES. 374
213 000DF61 170 CONTINUE 37'
214 C POSITION THE EIClENVALUE LOC"'TEO AT ... (NU. NU l. 376
21' 000DF61 180 CONTINUE 377
216 000DF61 IF(/1U. EO. NUP) GO TO 220 378
217 ooOEOEI IF(/1U. EO. NUP-l) gO TO 200 379
,Zl8 oo0E2AI IF(A(I1U+2. /1U+Il. EO. 000) 00 TO 200 380
219 C THE NEXT BLOCK IS 2*2. 381
220 000E641 I1U2-MU+I 382
221 000E721 CALL SPLIT(A. V.N.I1U2.EI2.E22.NA.NV) 383
222 00 OEC a I IF(A(I1U. I1Ul.LE . EI2) OOTO 220 384
223 OOOEFEI CALL EXCHNQ( .... V.N.I1U. I. 2 . EPS. F",IL. N.... NV) 38~
224 00OF6CI IFL NOT. FAIL) gO TO 190 386
225 OOQF7EI TYPE(I'IU)·-I 387
226 000F94 I TYPE(I'IU+l )--1 388
227 OOOFACI TYPE(I1U+2)--1 389
228 000FC41 GO TO 240 390

J.15
 229 OOOFCAI 190 CONTINUE
230 OOOFCAI MU=!'tJ+2 391
231 000FD81 GO TO 210 392
232 OOOFDEI 200 CONTINUE 393
233 C THE NEXT BLOCK IS 1*1. 394
234 OOOFDEI IF(A(MU.MUl . LE.A(MU+I.MU+111 GOTU 220 395
235 0010361 396
CALL EXCHNG(A.V.N.MU. I. I.EPS.FAIL.NA.NVI
236 0010A41 MU=MU+l 397
237 0010B21 210 CONTINUE 398
238 0010B21 GO TO lBO 399
239 0010B61 220 CONTINUE 400
240 0010B61 MU=NL 401
241 001OC21 NL-O 402
242 0010CAI IF(MU. NE. Ol GO TO 170 403
243 404
C GO BACK ANO GET THE SECONO EIGENVALUE OF THE PAIR.
244 00100EI 230 CONTINUE 405
24~ OOIOOEI NU-L-l 406
246 OOIOECI GO TO 20 407
247 408
C ALL OF THE EIGENVALUES HAVE NOW BEEN FOUNO ANO ORDEREO .
248 C COMPUTE THEIR VALUES ANO TYPES . 409
249 001OFOI 240 IF(NU. LT.NLOWI GO TO 260 410
250 0011081 00250 IaNLOW.NU 411
251 0011201 A(I. Il=A(I. Il+T 412
252 00116EI 250 CONTINUE 413
253 0011861 260 CONTINUE 414
254 0011861 NU=NUP 415
25~ 0011921 FLAG>-O 416
256 00119AI IERR=NUP 417
257 0011A61 270 cONTINUE 418
258 0011A61 IF(TYPE(NUl.NE.-11 GO TO 280 419
2~9 0011CAI FLAG=I 420
260 0011021 NU=NU-I 421
261 0011EOI GO TO 310 422
262 0011E61 280 CONTINUE 423
263 0011E61 IF(NU. EG.NLOWI GO TO 290 424
264 0011FEI IF(A(NU.NU-II.EG.ODOI GO TO 290 425
265 C 2*2 BLOCK: 426
266 0012361 CALL SPLIT(A.V.N.NU-I.EI.E2.NA.NVI 427
267 OOI29CI IF(A(NU.NU-Il. EG. 0001 GO TO 290 428
268 0012041 ER (NU)=El 429
269 OOI2EEI EI (NU-I I=E2 430
270 OOI30AI ER (NU-l )=ER(NU) 431
271 0013341 EI (NUl =-EI (NU-I) 432
272 0013641 TYPE(NU-ll=1 433
273 OOI37CI TYPE(NU)=2 434
274 0013921 NU=NU-2 435
27~ OOI3AOI GO TO 300 436
276 0013A61 290 CONTINUE 437
277 C SI NGLE ROOT: 438
278 0013A61 ER (NU) =A(NU. NU) 439
279 00130EI EI (NU)=OOO 440
280 0013F81 NU=NU-l 441
281 0014061 300 CONTINUE 442
282 0014061 310 CONTINUE 443
283 0014061 IF (FLAG. EG . Ol IERR=NU 444
284 0014221 IF(NU. GE.NLOWI GO TO 270 445
285 0014381 RETURN 446
286 00143EI END 447
448
NO ERRORS : F7D RO~-oo . 00 SUBROUTINE HGR4 21/12/84 09 : 2~ : 26 TABLE SPACE: 4 KB
STATEMENT BUFFER : 20 LINES/1321 BYTES STACK SPACE : 182 WORDS
DOUBLE PRECISION FLOATING PT SUPPORT REOUIRED FOR EXECUTION

J .16
 1 0000001 SUBROUTINE EXCHNO(A.V. N.L. BI. B2. EPS. FAIL. NA.NV) 449
2 0000041 INTEQER BI.B2.L. NA.NV 450
3 0000041 REAL*8 A(NA. N). EPS. V(NV. N) 451
4 OOOooAI LoQICAL FA IL 452
5 C 453
6 C 454
7 C 455
8 C CoLLECTED ALGoRITHMS FRDH ACM (THE ASSOCIATION FoR COHPUTINQ 456
9 C MACHINERVl 457
10 C 4~8

11 C ALGOR ITHM 506 459

12 C 460
13 C HGR3 AND EXCHNQ : FORTRAN SUBROUTINES FOR CALCULATINQ 461
14 C ANO ORDERING THE EIGENVALUES OF A REAL UPPER HESSENBERQ 462
I~ C MATRIX CF2] 463
16 C 464
17 C G. W. STEWART. UNIVERSITY OF MARYLAND 46~
18 C 466
19 C PUBLISHED IN THE ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE. 467
20 C VOL. 2. NO . 3. SEPTEMBER 1976. PAGES 275-280. 468
21 C 469
22 C SUBROUTINE EXCHNO 470
23 C 471
24 C 472
25 C 473
26 C GIVEN TI-E UPPER HESSENBERG MATRIX A. WITH CtNSECUTIVE 474
27 C Bl*BI ANO B2*B2 DIAOONAL BLDCKS (Bl.B2 . LE . 2) 475
28 C STARTING AT A(L.L). EXCHNG PRO DUCES A UNITARY 476
29 C SIMILARITY TRANSFORMATION THAT EXCHANGES THE BLDCKS 477
30 C ALoNG WITH THEIR ElQENVALUES. nE TRANSFORMATION 478
31 C IS ACCUMULATEO IN THE MATRIX V . EXCHNG REGUIRES THE SUBROUTINE 479
32 C GRSTEP . THE PARAMETERS IN nE CALLING SEGUENCE ARE 480
33 C STARRED PARAMETERS ARE AL TERED BY THE SUBROUTINE) : 481
34 C 482
3~ C THE MATRIX WHOSE BLOCKS ARE TC BE INTERCHANGED. 483
36 C THE ARRAY INTO WHICH THE TRANBFORMATIONS 484
37 C ARE TO BE ACCUMULATED . 485
38 C N THE ORDER OF THE MATR IX A. 486
39 C L THE POSI TION OF nE FIRST BLOCK. ALONG THE 487
40 C o IAGONAL OF A. 488
41 C BI AN INTEGER CONTAININQ THE SIZE OF THE FIRST BLOCK. 489
42 C B2 AN INTEGER CONTAINING THE SIZE OF THE SECONO BLOCK. 490
43 C EPS A CONVERGENCE CR I TER ION (CF . HGR3) . 491
44 C *FAIL A LOQICAL VARIABLE SET To . FALSE. FOR A NORM AL 492
4~ C RETURN. IF THIRTY ITERAT I ONS IERE PERFoRMED. 493
46 C WITHOUT CONVERQENCE. FAIL IS SET TC . TRUE.. ANO 494
47 C TI-E ELEMENT A(L+B2. L+B2-1) CANNOT BE ASBUMED 495
48 C To BE ZERO. 496
49 C NA THE FIRST DIMENSION OF THE ARRAY A. 497
50 C NV THE FIRST DIMENSION OF THE ARRAY V. 498
51 C 499
~2 C INTERNAL VARIABLES : 500
53 OOOooAI INTEGER I. IT. J. Ll. M 501
~4 OOOooAI REAL*8 P, G. R, S. W, x. V, Z 502
~~ OOOooAI FA IL- . FALSE . 503
56 00004CI IF (BI. EG. 2) GO TO 70 504
~7 0000621 IF(B2 . EG. 2) GO TO 40 505
58 C INTERCHANGE 1*1 AND 1*1 BLOCKS 506
59 0000781 Ll - L+I 507
60 0000861 Q:zACL+ 1. L+l )-A CL, L) 508
61 0000081 P- A(L. L+I) 509
62 0001041 R- DMAX 1 <DABS(P). OABS(G) ) SlO
63 00015EI IF (R. EG . 000) RETURN 511
64 0001761 P- P/R 512
65 0001881 G=O/R 513
66 00019AI R-DSORTep**2+0*.2 ) 514
67 000lE61 P-P/R SIS
68 0001F81 G- O/R ~16
69 00020AI DO 10 J=L. N 517
70 0002221 S - P*ACL. J)+O*A(L+l. J) 518
71 0002801 ACL+l • .,J)=P*ACL+l. J)-O*ACL. J) 519
72 0002FEI ACL. ,J)=9 520
73 0003281 10 CONTINUE 521
74 0003401 DO 20 Iml. L1 522
75 0003541 S=P*ACI,L)+O*ACI, L+l> 523
76 0003B2 I A (t. L+l )=P_AC I. L+l )-O_A( I. L) 524
77 0004301 A <I. Ll=S 525
78 00045AI 20 CONTINUE 526
79 0004721 DO 30 r-l. N 527
80 0004861 S-P*VCI, L)+()*VU , L+l) 528
81 0004E4 I ve J, L+l )=p*ve I, L+! )-G*V( I , L) ~9

82 0005621 V <1. UOOS 530

83 0005BCI 30 CONTINUE 531
84 0005A4 I A(L+l,L)·ODO 532
85 0005001 RETU'lN 533
86 0005061 40 CONTINUE ~4
87 C INTERCHANQE 1*1 ANO 2*2 BLoCKS 535
88 0005061 X-A(L. Ll ~36
89 0006001 P-IDO 537
90 00060CI 0 .... 100 538
91 0006181 R- lDO 539
92 0006241 CALL ORSTEPeA,V,p,Q,R.L,L+2, N,NA,NV) 540
93 0006901 IT " O 541
94 0006981 :50 IT-IT+! 542
9~ 0006A6 I IF (IT. LE. 30) GO TO 60 543
96 0006BCI FAIL- . TRUE . ~44
97 0006C4 I RETURN 545
98 0006CAI 60 CONTINUE 546
99 0006CAI P-ACL, L)-X 547
100 0006FAI a - A(L+1. L) 548
lOl 0007261 R-ODO 549
102 0007321 CALL OR6TEP(A.V.P. G.R.L.L+2.N. NA. NV) ~50
103 oo07AOI IF(DABS(ACL+2.L+l».GT.EPS*(DABSCACL+l,L+!»+DABS(A(L+2. 551
104 I L+2»» GO TO sa 5~2
lOS 00089EI A(L+2 , L+l ) .. 000 553
106 ooOBCCI RETU'lN 554
107 0008021 70 CONTINUE 555
108 INTERCHANGE 2*2 AND 2*2 BLoCKS 556
109 0008021 M-L+2 557
110 OOOBEOI IF (B2. EO. 2) M-M+I 558
111 OOOBFEI X-A(L+1.L+l) 559
112 00092C I YmA(L . Ll 560
113 0009561 W-ACL+l,L)*A(L,L+!) 561
114 0009A81 P-IDO 562

J.l7
 115 0009B41 (l=I DO
116 0009C01 R=IDO 563
117 0009CCI 564
CALL ORSTEP(A.V . P. O. R. L. M. N. NA . NV) 565
118 OOOA2CI !T- O
119 000A341 80 !T=IT+I 566
120 000M21 IF(IT. LE. 30 ) GO TO 90 567
121 000A581 56B
FAIL= . TRUE. 569
122 000A601 RETURN
123 000A661 90 CONTINUE 570
124 000A661 Z=A(L. L> 571
125 000A901 R=X-Z 572
126 000M21 S-Y-Z 573
127 000AB41 P- CR*S-W)/A(L+l , L)+A(L.L+l) 574
128 000BI81 G- ACL+l,L+ l ) - Z- R- S 575
129 000B58I R- ACL+2, L+l ) 576
130 000B86I S- DABS(P)+DABS ( O)+DABS(R) 577
131 000BE81 P DP/S 578
132 OOOBFAI O- O/S 579
133 OOOCOC I R- RIS 5BO
134 OOOCIEI 581
CALL ORSTEP( A.V. P . (l. R. L. M. N. NA.NV) 582
135 000C7CI IF(DABS(A(M- I . M-2» . GT . EPS *( DABS(A(M-I . M-I»+DABS(A(M_2.
136 1 M-2»» GO TO 80 583
137 000D7AI A(M-l.M-2) - ODO 584
138 000DA81 RETURN 585
139 OOODAEI END 586
587
NO ERRORS : F7D R05-00. 00 SUBROUTINE EXCHNG 21/12/84 09: 25:33 TABLE SPACE: 2 KB
STATEMENT BUFFER : 20 L1NES/1321 BYTES STACK SPACE : 182 WORDS
OOUBLE PRECISION FLOATING PT SUPPORT REOUIRED FOR EXECUTION

1 0000001 SUBROUTINE SPLI T( A. V. N. L. EI . E2. NA.NV )

2 0000041 INTEGER L. N. NA . NV 588
3 0000041 589
REAL*B A(NA.N) .V( NV. N). EI. E2 590
4 C
5 C
591
6 C 592
7 593
C COLLEC TED ALGORI THMS F ROM ACM ( THE ASSOCIATION FOR 594
8 C COMPUTING MACHI NERY)
9 C 595
10 C ALGORITHM 506 596
11 C
597
12 59B
C H(lR3 AND E XCHNG : FORTRAN SUBROUTINES FOR CALCULATING 599
13 C AND ORDERING THE EIGENVALUES OF A REAL UPPER HESSENBERG
14 C MATRI X [F2J 600
15 C 601
16 CG. W. STEWART. UNIVERSITY OF MARYLAND 602
17 C 603
18 604
C PUBLISHED IN THE ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE . 605
19 C VOL. 2. NO. 3. SEPTEMBER 1976. PAGES 275-280 .
20 C 606
21 C SUBROUTINE SPLIT 607
22 C 60B
23 C 609
24 C 610
25 C GIVEN THE UPPER HESSENBER G MATRIX A. WITH A 2*2 BLOCK 611
26 C STARTING AT A( L. L) . SPLIT DETERMINES IF THE 612
27 C 613
CORRESPONDING EIGENVLAUES ARE REAL OR COMPLEX . IF THEY 614
28 C ARE REAL. A ROTATION IS DETERMINED THAT REDUCES THE
29 C BLOCK TO UPPER TRIANGUILAR FORM WITH THE EIGENVALUE 615
30 C OF LARGEST ABSOLUTE VA LUE APPEARING FIRST. THE 616
31 C 617
ROTATION IS ACCUMULATED IN V. THE EIGENVALUES (REAL 618
32 C OR COMPLEX) ARE RETURNED I N El AND E2. THE PARAMETERS
33 C IN THE CALLING SEOUENCE ARE (STARRED PARAMETERS ARE ALTERED 619
34 C BY THE SUBROUTINE) : 620
35 C 621
*A THE UPPER HESSENBERG MATRIX WHOSE 2*2 622
36 C BLOCK IS TO BE SPL IT.
37 C 623
THE ARRAY IN WHICH THE SPLITTING TRANS- 624
38 C FORMATION IS TO BE ACCUMULATED .
39 C 625
THE ORDER OF THE MATRIX . A. 626
40 C THE POSITION OF THE 2*2 BLOCK ON A ' A DIAGONAL .
41 C 627
ON RETURN. IF THE EIGENVALUES ARE COMPLEX. 628
42 C EI CONTAINS THEIR COMMON REAL PART. ANO
43 C 629
E2 CONTAINS THE POSITIVE IHAGINARY PART. 630
44 C IF THE EIGENVALUES ARE REAL. EI CONTAINS
45 C 631
THE ONE LARGEST IN ABSOLUTE VALUE. ANO E2 632
46 C CONTA I NS THE SMALLER ONE.
47 C 633
NA THE FIRST OlMENSION OF THE ARRAY A. 634
48 C NV THE FIRST OlMENSION OF THE ARRAY V.
49 C INTERNAL VARIABLES : 635
50 OOOOOAI INTEGER I . J . L1 636
51 OOOOOAI REAL*B P.O . R.T.U.W.x.Y . z 637
52 OOOOOAI XIO;tACL+L L+l) 638
53 0000721 Y=A(L. L> 639
54 00009CI W=A(L , L+l >*A(L+l,L) 640
55 OOOOEEI P= (Y-X )/200 641
56 0001061 G=P**2+W
642
57 0001201 IF«l. GE. ODO) GO TO 10 643
644
J . 18
 C COMPLEX EIOENYALUE
645
5B 646
59 0001391 EI-P+X
E2-DSORTC -0)
647
60 oo014AI 64B
61 OO017AI RETlAlN
649
62 0001601 10 CONTINUE
C TWO REAL EIOENVALUE9. BET UP THE TRANBFCJR"ATION. 6'0
63
Z-DBClRHO) 6'1
64 ooOIBOI
IF(P.LT. ODO) 00 TO 20 6'2
6' 00019EI 6'3
66 0001B61 Z-P+Z
00 TC 30 6'4
67 ooOICBI
6B ooOICEI 20 CONTINUE 6"
Z-P-Z 6'6
69 ooOICEI 6'7
70 0001EOI 30 CONTINUE
IF(Z. EO. ODO) 00 TO 40 6'6
71 0001EOI
6'9
72 ooOIFBI R - -~."'Z 660
73 0002121 00 TO ~
661
74 0002161 40 CONTINUE
662
75 0002161 R-ODD
663
76 0002241 ~ CONTINUE
IF(DABS(X+ZI. OE. DABS(X+R)) Z-R 664
77 0002241
66'
7B 0002961 V=V-X-Z
666
79 ooO:/AEI X- -Z
667
60 0002<:01 T- ACL. L+l)
bbB
BI OOO2ECI U- ACL+l, L)
IF(DABS(V)+DABS(U) . LE . DAB9(T)+DABS(X)) 00 TO 60 669
62 0003161 670
B3 0003A6 I O-U
671
B4 0003821 P-V
672
B5 00038EI gO TC 70
673
Bb 0003C4 I 60 CONTINUE
674
67 0003C4 I O-X
paT 67'
BB 0003001 676
69 0003DCI 70 CONTINUE
677
90 0003DCI R-DSGRTCP**2+0*.2)
676
91 0004261 IF(R.gT. ODO) 00 TO BQ
679
92 00043EI EI-A(L.U
E2-ACL+l, L+1)
bSO
93 0004661 bSl
94 0004961 ACL+l.L)cODO
bS2
9' 0004C2 I RETURN
bS3
96 0004C6 I BO CONTINUE
bS4
97 0004C61 P- P/R
oo04DAI (I-OIR bS'
9B bS6
99 C PR EI'Ul. TIPL V
DO 90 .laL. N .
bS7
100 0004ECI bSB
101 000~41 Z-ACL,.)
A CL • ..J )..p*Z+Q*".CL+l,.J) bS9
102 OOO~EI
A(L+l. ~J)-P.A(L+1. ,»-O_Z 690
103 OOO~CI
691
104 OOO~CI 90 CONTINUE
692
10' POBTI1UL TIP LV
693
106 0006041 LI - L+l 694
107 0006121 DO 100 1- 1. L1
Z- ACI,L) 69'
106 00Qb261 696
109 0006'01 A (I, L ,..p*Z+O*A( I. L+l)
A (I . L+l )oP.AC I. L+l )-O*Z 697
110 OOObAEI 696
111 00070EI 100 CONTINUE
699
112 ACCUI1ULATE THE TRANSFORI1ATlON IN V
700
113 0007261 DO 110 1- 1. N
701
114 oo073AI Z- Vel , L)
VeI.L)..p_Z+Q*V(I , L+l) 702
11' 0007641 703
116 0007C21 V ( J. L+l )=P*VC Jo L+l ) - O*Z
704
117 0006221 IlO CONTINUE
705
116 ooOB3AI ACL+I, L)-ODO
706
119 000Bb61 El=A(L. U
707
120 0006901 E2 - ACL+l , L+l)
706
121 OOOSDEI RETURN
709
122 000BC41 END

NO ERRDRB : F7D RO'-oo. OO SUBROUTINE SPLIT 21/12/94 09: 2' : 37 TABLE SPACE : 2 KB
STATEI1ENT BUFFER : 20 L1NES/1321 BVTES STACK SPACE: 144 WORDS
DOUBLE PRECISION FLOATINO PT SUPPDRT REGUIRED FOR EXECUTION

J.19
 I 0000001 SUBROUTINE QRSTEP(A.V.P.O.R.NL.NU.N.NA.NV) 710
2 0000041 INTEGER N.NA.NL.NU.NV 711
3 0000041 REAL*8 A(NA.N).P.O.R.V(NV.N) 712
4 C
713
5 C
714
6 C
7 715
C COLLECTED ALGORITHMS FROM ACH (THE ASSOCIATION FOR 716
8 C COMPUTING MACHlNERY)
717
'I C
10 718
C ALGORITH'I 50b
11 719
C
720
12 C HOR3 AND EXCHNG ; FORTEAN SUBROUTINES FOR CALCULATING
13 721
(- C AND ORDERING THE EIGENVALES FOR A REAL UPPER HESSENBERG
14 722
C MATRIX [F21
15 723
C
16 7024
C G. W. STEWART. UNIVERSITY OF MARYLAND
17 725
C
18 726
C PUBLISHED IN THE ACH TRANSACTIONS ON MATHEHATICAL SOFTWARE.
1'1 727
C VOL . 2. NO. 3. SEPTEMBER 1976. PAQES 275-280.
20 728
C
21 729
C SUBROUTINE QRSTEP
22 730
C
23 731
C OR STEP PERFORMS ONE IMPLICIT OR STEP ON THE
24 732
C UPPER HESSENBERG MATRIX A. THE SHIFT IS DETERMlNED
25 733
C DY THE NUMDERS P. 0 AND R. AND THE STEP IS APPLIED TO
26 734
C ROWS AND COLUMNS NL THROUGH NU . THE TRANSFORHATIONS ARE
735
27 C ACCUMULATED IN THE ARRAY V. THE PARAMETERS IN THE
28 736
C CALLING SEOUENCE ARE (STARRED PARAMETERS ARE ALTERED
2'1 737
C DY TH!, SUBROUTINE) ;
30 738
C
31 739
C THE UPPER HESSENDERG MATRIX ON WHICH THE
32 740
C OR STEP 15 TO BE PERFORMED.
33 741
C THE ARRAY IN WHICH THE TRANSFORMATIONS
34 C
742
ARE TO BE ACCUMULATED. 743
35 C PARAMETERS THAT .
36 744
C *0 ... DETERHINE .
37 745
C *R ... THE SHIFT .
38 746
C NL THE LOWER LIMIT OF THE STEP .
3'1 C 747
NU THE UPPER LIHIT OF THE STEP .
40 748
C N THE ORDER OF THE MATRIX A.
41 C 749
NA THE FIRST DIMENSICJ'I OF THE ARRAY A.
42 750
C NV THE FIRST DIMENSICJ'I OF THE ARRAY V.
43 751
C
44 752
C INTERNAL VARIABLES ;
45 OOOooAI 753
INTEGER 1.~.K.NL2.NL3.NUMI
46 OOOooAI 754
REAL*8 5. X. Y. Z
47 OOOooAI 755
LOGICAL LAST
48 OOOooAI 756
NL2 c NL+2
4'1 0000521 757
DO 10 I-NL2.NU
50 OOOOóAI 758
A( 1.1-2)-000
51 0000'161 759
10 CONTINUE
52 OOOOAEI 760
IF(NL2 . EO . NU) GO TO 30
53 000OC61 761
NL3=NL+3
54 0000D41 762
DO 20 I=NL3. NU
55 OOOOEC I 763
A(I .I-3)=000
56 0001181 764
20 CONTINUE
57 0001301 765
30 CONTINUE
~e
766
0001301 NUMl=NU-l
5'1 00013EI 767
DO 130 K=NL.NUMI
60
768
COETERMINE THE TRANSFORMATION
61 0001561 769
LAST=K . EO . NUMI
62 0001701 770
IF(K . EO.NL) GO TO 40 771
63 0001881 P=A(K. K-I)
64 0001B41 772
O=A(K+l. K-l)
65 000lE21 773
R=ODO
66 OOOIEEI 774
IFLNOT. LAST) R=A(K+2.K-l) 775
67 0002281 X-DABS(P)+DABS(O)+DABS(R)
68 00028C I 776
IF(X. EO. ODO ) GO TO 130 777
6'1 0002A4 I P=P/X
70 0002661 778
O=G/X
71 0002C8 I 779
R=R/X
72 0002DAI 780
40 CONTINUE
73 0002DAI 781
S=DSGRT(P**2+Q**2+R**2) 782
74 00033EI IF(P . LT.ODO) 5=-5 783
75 0003621 IF(K . EO.NL) GO TO 50
76 00037AI 784
A (K. K-l)=-S*X
77 0003B41 785
GO TO 60
78 0003BAI 786
50 CONTINUE
79 0003BAI 787
IF(NL . NE. 1) A(K. K-l)--A(K. K-Il 788
80 00041C I 60 CONTINUE
81 00041CI 789
P=P+S
82 00042EI 790
X-PIS
83 0004401 791
V-O/S
84 0004521 792
Z=RIS 793
85 0004641 Cl -ClIP
86 0004761 794
R=R/P
87 795
C PREMUL TIPL V
88 0004881 796
DO 80 ~K.N
8'1 0004AOI 797
P-A(K.~)+Cl*A(K+I.~)
'10 0004F6 I 798
IF(LAST) GO TO 70 799
'11 0005081 P..p+R*A(K+2.~)
'12 0005401 800
A(K+2,~)aA(K+2,~)-P*Z
'13 00059AI 801
70 CCJ'ITINUE
'14 0005'1AI 802
A(K+l.J)aA(K+l ,J )-P*Y 803
'15 0005F4 I A(K . ~)=A(K.~)-P*X
'16 OOOMAI 804
80 CCJ'ITINUE
'17 805
C POSTMULTIPLV
'18 0006621 806
~=MINO(K+3. NU)
99 0006921 807
DO 100 I=I.~
100 000bA61 808
P-X*A(I,K)+V*A(I,K+l' 809
101 0007041 IF(LAST) GO TO 90
102 0007161 810
P=P+Z*A( I. K+2)
103 00074EI 811
A(I,K+2)-A(I,K+2)-P*R 812
104 0007A81 90 CCJ'ITINUE
105 0007A81 813
A<I . K+l )=A( I. K+l )-f'*0 814
106 0008021 A(I. K)=A(I. K)-f'
107 0008501 815
100 CCJ'IT I NUE
108 816
C ACCUMULATE THE TRANSFORMATICJ'I IN V
109 0008681 817
DO 120 1-1. N
110 00087CI 818
P=X*V(I . K)+V*V(I.K+l) 819
lil 0008DAI IF(LAST) GO TO 110
112 OOOBECI 820
P..P+Z*V( I. K+2)
113 0009241 821
V(I.K+2)=V(I.K+2)-f'*R 822
114 00097EI 110 CCJ'ITlNUE
823
J.20
115 00097EI V(I. K+U-V( I. K+l )-f'.O 1124
116 0009D81 VO. K)-V(I.K)-f' 1125
117 000A261 120 CIINTIMJE 1126
118 OOOA3EI 130 CONTI NI.E: 1127
119 OOOA56 I RETlJlN lI2B
120 OOOA5CI END 1129

NO ERRORB :F7D R05-DO. 00 SUBROUTINE ORSTEP 21/12/94 09:25: 41 TABLE SPACE: 2 KB

STATEMENT BUFFER : 20 LlNES/1321 BYTES STACK BPACE : 159 WORDB
DOUBLE PRECISION FLOATINQ PT BUPPORT REOUIRED FOR E~ECUTION

00120 WA SHI2: PROCEDURE OPTIONS(MAIN)I

00130
00140 PROGRAM FOR WA SHOUT FILTER DESIGN FOR 12TH ORDER CASE
00150 TO BE USED FOR SURGE/PITCH & SWAY/ROLL CASES
00160 *1
00170 1* DECLARE AND SET OPTIONS *1
00180 DECLARE <Z,J,K,L,M,N> BINARY FIXED (31,0);
00190 DECLARE (II,JJ,Il,Kl,Ml,Nl,SIGN,NINV) BINARY FIXED <31,0);
00200 DECLARE (V.W.X.Y.Z) BINARY FIXED (31.0)/
00210 FORHAC_OPTIONS 1
00220 CONVERT FIXED C!;JiKiLiMiN);
00230 CONVERT FIXED ClliJJilliKl;MliNliSIGNiNINV);
00240 CONVERl FIXED (V;W;X;YiZ);
00250 OPTSET (LINELENGTH=721 EXPND)/
00260 PUT FILE (SYSPRINT) PAGE LIST
00270 ('SOLUTION TO WASHOUT FILTER EONS - - P-S OR R~S CASE')I
00280 1*
00290 *****INPUT VARIABLES:****I
00300 OPTSET (PRINT);
00310 PUT FILE (SYSPRINT> SKIP(4) LIST
00320 ('VESTIBULAR MODELS:')I
00330 PUT FILE (SYSPRINT) SKIP(2)1
00340 LET (Tl =(6.1+0.1)/(6.1*0.1»1
00350 LET (T2=118.55/(6.1*0.1»1
00360 LET (T3=1.0/(6.1*0.1»1
00370 LET (AO=0.076) 1
00380 LET (80=0.19) 1
00390 LET (GO=5. 86);
00400 LET (G=-9. 81>1
00410 PUT FILE (SYSPRINT) SKIP(4) LIST
00420 ('WEIGHTING ELEMENTS:')I
00430 PUT FILE (SYSPRINT) SKIP(2)1
00440 LET (RHO = I) 1
00450 LET <011 =0.4472; Q22=0.8944);
00460 LET (Rl1 =O.9; R22 = O.1581; R33=O.1581);
00470 LET (R44=0.15811 R55=0.1581)1
00480 OPTSET (NOPRINT)/
00490
00500 NOW ENTER THE RICCATI SOLUT ION P21(A).P22(B).P23(C).
00510 NOTE THAT WE ASSUME THAT P22 IS SYMMETRIC AND THAT
0052 0 THE FIRST 3 COLUMNS OF P21 ARE THE
00530 NEGATIVE OF THOSE IN P22. *1
00540 LET (AI4 = 1.71730E-171 AI5 =-5.44946E-17/ AI6 =-1.26503E-16)/
00550 LET CA24 =1 . 60997E-IB; A25 =1.62490E-17; A26=7.99180E-17);
00560 LET CA34=Oi A35=-S.93386E-16; A36=-S.52014E-1S);
00570 LET (A44 =3.96887E-161 .A45=1.28971E-15/ A46=9.08523E-16)/
00580 LET (A54 =6.41125E-16/ A55=3.00326E - 151 A56 =3.58809E-15)1
00590 LET (A64= 7 .63244E-17/ A65=3.10014E - 151 A66=4.87612E-15)/
00600 1* *1
00610 LET (Bll =6 . 59145E-031 BI2=1.49452E- 03/ 813=2.17136E-02)1
0062 0 LET (BI4 =- 1.94450E-02/ 815=- 7.456ó8E- 02/ BI6=·-0.13326>1
00630 LET (B22=1.32641E-02/ 823=2.7526 4E-031 B24 =1.97001E-03)1
00640 LET (B25=6.92058E-031 826=7. 199BOE-03/ B33=7.38617E-02)1
00650 LET (B34= - 6.62170E-02; B3 5 =- 0.25374; B36=-0.45274);
00660 LET (844=0.57661/ B45=0.97478/ 846=0.90847)/
00670 LET (B55 =2 .6648; 856=2.9608; 866=4.2497);
00 6 80 1* *I
00690 LET (Cll=-5.77612E-04; CI2--7.81343E-02)1
* J.21
00700 LET (C21=-6.04130E-05; C22=2.24924E-03);
00710 LET (C31=5.47594E-02' C32=-0.26522)'
00720 LET (C41=-1.74365E-04' C42=0.35219)'
00730 LET (C51=-9.39297E-03. CS2=1.2708)'
00740 LET (C61=-7.83120E-02; C62=2.0393);
00750
00760 *****FoRM MATRICES*****
00770
00780 1* MATRIX A *1
00790 PUT FILE (SYSPRINT) SKIP(4) LIST
OOBOO ('NON-ZERO ELEMENTS . oF A:');
008 10 PUT FILE (SYSPRINT) SKIP(3);
00820 L=5; M=2; N=6;
00830 DO 1=1 To N'
00840 DO J =l To N'
00850 LET (AA("I","J")=O);
00860 END.
00870 END;
00880 oPTSET (PRINT>'
008 90 LET (AA(l,l) =-T l' AA(1,2)=1; AA(2,1)=- T3)'
00900 LET (AA(3,3)=- Bo. AA(4,5)=1' AA(S,6)=1).
00910 1* *1
00920 1* MATRIX B *1
00930 PUT FILE (SYSPRINT) SKIP(S) LIST
00940 ( ' MATRIX 13:');
009S0 PUT FILE (SYSPRINT) SKIP(3)'
00960 LET (8B(1,1)=-T1*T2; 88(1,2)=0; BB(2,1) =- T3*T2; 88(2,2)=0);
00970 LET (88(3,1)=GO*G*(AO-BO)' BB(3,2) =Go*(AO-BO»'
00980 LET (BB(4,1)=0; B8(4,2)=0; 88(5,1)=0);
00990 LET <88(5,2)=0; B8(6,1)=0; BB(6,2)=1);
01000 1* *1
01010 1* MATRIX C *1
01020 PUT FILE (SYSPRINT) PAGE LIST
01030 ('MATr.:IX Cr');
01040 PUT FILE (SYSPRINT) SKIP(3)'
01050 LET (CC(l,1)=!; CC(l,Z)=Oi CC(1,3)=O; CC(1,4)=O);
01060 LET (CC(l,S)=O; CC(1,6)=O; CC(2,1)=O; CC(2,Z)=O);
01070 LET (CC(Z,3)=1; CC(2,4)=O; CC(2,S)=O; CC(Z,6)=O);
01080 1* *1
01090 1* MATRIX D *1
01100 PUT FILE (SYSPRINT) SKIP(5) LIST
01110 ('MATRIX D:');
01120 PUT FILE (SYSPRINT) SKIP(3).
01130 LET (DD(1,1)-T2. DD(1,2)=0)'
01140 LET (DD(2,1)=GO*G~ DD(2,2)=GO);
01150 1* *1
01160 1* MATRIX Q *1
01170 PUT FILE (SYSPRINT) SKIP(S) LIST
01180 ('MATRIX Q:');
01190 PUT FILE (SYSPRINT) SKIP(3)'
01200 LET (QQ(1,1)=011; 00(1,2)=0; 00(2,1)=0; 00(2,2)=022 );
01210 1* *f
01220 1* MATRIX P21 *1
01230 PUT FILE (SYSPRINT) PAGE LIST
01240 ('MATRIX P211').
01250 PUT FILE (SYSPRINT) SKIP(3)'
01260 LET (PP21(1,1) =- Bll; PP21(1,2) =- BIZ; PP21(1,3)=-BI3);
01270 LET (PP21(1,4) =A14; PP21(1,S) =A1S; PP21(1,6)=A16);
01280 LET (PP21(2,l)= - B12; PP21(2,2) =- B22; PP21(2,3)=-B23);
01290 LET (PP21(2,4)=A24; PP21(2,S)=A2S; PP21(2,6)=A26);
01300 LET (PP21(3,1)=-B13; PP21(3,2) =- B23; PP21(3,3) =- B33);
01310 LET (PP21(3,4) =A34' PP21(3,S) =A35' PP21(3,6) =A36).
01320 LET (PP21(4,1)= - B14; PP 2 1(4,2)= - B24; PP21(4,3)=-B34);
01330 LET (PP21(4,4)=A44; PP21(4,5) =A45; PP21(4,6) =A46);
01340 LET (PP21(S,1) =-B I5; PP21(S,2)=-B25 ; PP21(S,3)=- B35);
01350 LET (PP21(S,4) =AS4; PP21(5,5) =A55; PP21(S,6)=A56);
01360 LET (PP21(6,1) =- B16. PP21(6,2) =-826' PP21(6,3)= - B36)'
01370 LET (PP21(6,4) =A6 4; PP21(6,S)=A6S' PP21(6,6) =A66)'
01380 1* *I
01390 1* MATRIX P22: *1
01400 PUT FILE (SYSPRINT) PAGE LIST
01410 ('MATRIX P2Zt');
01420 PUT FILE (SYSPRINT) SKIP(3)'
01430 LET (PP22(1,1)=Bl1; PP22(1,2)=B12; PP22(1,3)=B13);
01440 LET (PP22(1,4)=B14; PP22(1,S)=B1S; PP22(1,6)=B16);
01450 LET (PP22(2,1)=B12; PP22(2,Z)=B22; PP22(2,3)=B23);
01460 LET <PP2Z(Z,4)=B24; PP22(2,S)=B2S; PP22(2,6)=B26);
01470 LET (PP22(3,!)=B13; PP22(3,2)=B23; PP22(3,3)=B33);
01480 LET (PP22(3,4)=B34; PP22(3,S) =B35; PP22(3,6) =B36);
01490 LET (PP22(4,1)=B14; PP22(4,2)=B24; PP22(4,3)=B34);
01500 LET (PP22(4,4)=B44; PP22(4,S)=B45; PP22(4,6)=B46)i
01510 LET (PP2Z(S,1)=B1S; PP22(S,Z) =B2S; PP22(S,3)=B35);
01520 LET (PP22(S,4) =B4S; PP22(S,S) =B55; PPZ2(S,6)=BS6);
01530 LET (PP22(6,1)=B16; PP22<6,2)=B26; PP22(6,3)=B36);
01540 LET (PP22(6,4)=B46; PP22(6,S)=B56; PP22(6,6)=B66);
01550 1* *1
01560 1* MATRIX P23 *1
01570 PUT FILE (SYSPRINT) PAGE LIST
01580 ('MATRIX P23:')'
01590 PUT FILE (SYSPRINT) SKIP(3)'
01600 LEl (PP23(1,1)=C11' PP23(1,2)=C12' PP23(2,1)=C21)
01610 LET (PP23(2,2)=C22; PP23(3,1)=C31; PP23(3,2)=C32)
01620 LET (PP23(4,1)=C41; PP23(4,2)=C42; PP23(5d )=C51)
01630 LET (PP23(5,2)=C52' PP23(6,1) =C61' PP23(6,2)=C62)
01640 1* *1
01650 1* MATRIX RHo*R *1
01660 PUT FILE (SYSPRINT) SKIP(5) LIST
01670 ('MATRIX RHO*R:').
01680 PUT FILE (SYSPRINT) SKIP(3)'
01690 LET (RHORR(l,l)=RHo*Rll. RHoRR(1,2)=0)'
01700 LET (RHORR(2,1)=0; RHORR(2,2)=RHO*R22);
01710 1* *1
01720 1* MATRIX G *1
01730 PUT FILE (SYSPRINT) SKIP(5) LIST
01740 ('NON-ZERO ELEMENTS OF MATRIX G:' );
01750 PUT FILE (SYSPRINT) SKIP(3)'
01760 oPTSET (NOPRINT);
01770 DO 1=1 To U
01780 DO J=1 Ta U
01790 LET (GG(III,IJI)=O);
01800 END;
01810 END.
01820 oPTSET (PRINT>.
01830 LET (GG(l,l)=Qll; GG(2,2)=Q22' .GG(3,3)=RHO*R33);
01840 LET (GG(4,4)=RHo*R44. GG(5,5)=RHo*R55).
01850 1* *1
*
J.22
01860 / * MATRIX D-BAR */
01870 PUT FILE CSYSPRINT) PAGE LIST
01880 C'MA TRIX D-BAR; ')I
01890 PUT FILE CSYSPRINT) SKIP(3)I
01900 LET CDDBARC1.l)=T21 DDBARC1.2)=0)I
01910 LET CDDBARC2.1)=GO*GI DDBARC2.2)=GO)I
01920 LET <DDBAR(3,1)=O; DD8AR(3,2)=O);
01930 LET CDDBARC4, 1)=O; DDBAR (4,2)= O);
01940 LET CDDBARC5 ,1) =O; DDBARC5,2) =O);
01950 /* */
01960 / * OBTA1N THE REQUIRED TRANSPOSES */
01970 OPTSET CNOPRINT)I
01980 DO J=1 TO MI
01990 LET CJ= 'J');
02000 DO 1=1 TO NI
02010 LET CI= 'I');
02020 LET CBBTTC..J,I)=BBCI,J»;
02030 END;
02040 DO 1=1 TO MI
02050 LET C1=' I ' );
020 60 LET CDDTTCJ.I)=DDCI.J»I
02070 END;
02080 DO 1=1 TO U
02090 LET CI='I ');
02100 LET CDDBARTCJ.I)=DDBARCI.J»;
02110 END;
02120 END;
02130 / * */
02 140 / * CALCULATE THE MATRIX R2; FIRST DBART*G */
02150 DO 1=1 TO M;
02160 DO ..J = l TO L;
02170 LET CI=-I-; J='J");
02 180 LET CDBTGCI.J)=O)I
02190 DO K=l TO Li
02200 LET CK= ' K')I
02210 · LET (DBTGC I ,J)=DBTG( I ,J).+DDBARTC I ,f().GGCK,J»;
02220 END ;
02230 END;
02240 END;
0 2250 /* */
02260 /* NOW DBART*G*DBAR */
02270 DO 1=1 TO MI
02280 DO J=1 TO MI
02290 LET <1='1'; J =" ..J");
02300 LET CDBTGDBCI.J)=O)I
02310 DO K= 1 TO L;
02320 LET CK= 'K') I
02330 LET CDBTGOBCI.J)=DBTGDBCI.J)tDBTGCI.K)*ODBARCK. J» I
02340 END;
02350 END;
02360 ENOI
02370 /* */
02380 /* NOW FIND RHO*RtDBART*G*DBAR=R2 */
02390 DO 1=1 TO MI
02400 DO J =1 TO MI
02410 LET CI= "I'; J = 'J'>;
02420 LET CRR2CI,J)=RHORRCI.J )tDBTGDBCI.J»I
02430 END;
02440 END;
02450 /* */
02460 /* NOW INVERT R2 USING ADJOINT AND DET SCHEME */
02470 /* FOR A 2X2 MATRIX. THIS CAN BE DONE IN ONE STEP */
02 48 0 LET CDETRR2=RR2Cl.l)*RR2C2.2)-RR2Cl.2)*RR2C2.1»;
02490 OPTSET CPRINT);
02500 PUT FILE CSYSPRINT) SKIP(5) LIST
02510 C' THE INVERSE OF MATRIX R2; ' )I
02520 PUT FILE CSYSPRINT) SKIP(3)I
02530 LET <RR2INVC1,1)= RR2C2,2) / DETRR2) ;
02540 LET (RR2INV(1,2)=-RR2(1,2)/DETRR2);
02550 LET CRR2INVC2.1)=-RR2C2.1)/DETRR2),
02560 LET CRR2INVC 2.2)= RR2Cl.1)/DETRR2)1
02570 /* */
02580 /* NOW CALCULATE Ft: FIRST FINII OUQ */
02590 OPTSET CNOPRINT),
02600 DO 1=1 TO MI
02610 DO J =1 TO MI
02620 LET <1=-1 1
; J=IJI);
02630 LET CDDTTUQCI.J) =O)I
02640 110 K=1 ra MI
02650 LET CK= 'K');
026 60 LET CDDTTQQCI.J)=DDTTQQ CI.J)tIlDTTCI.K)*UQCK.J»1
02670 END;
02680 END;
02690 END;
02700 /* */
02710 /* NOW GET OT*Q*C */
02720 DO 1=1 TO M;
02730 DO J =1 TO NI
027 40 LET (IclIl; J=IJI);
02750 LET <DTOC ( I,J) =O);
0 2760 DO K= l TO M;
02770 LET CK= ' K') I
02780 LET (DTOC(I,J) =DTQC(I,J)+DDTTQQ(I,K)*CC(K,J»;
02790 END;
02800 END;
02810 END;
02820 / * */
02830 / * NOW GET BT*P21 */
02840 DO 1=1 TO MI
02850 DO J =1 TO NI
02860 LET (I = I .; J= I J I ) ;
I

02870 LET CBTP21CI .J) =0)I

02880 DO K=1 TO NI
02890 LET <KcIKI);
02900 LET CBTP21CI.J)=BTP21CI.J)tBBTTCI.K)*PP21CK. J »I
02910 END!
02920 END;
02930 END;
02940 / * */
02950 /* NOW CALCUL ATE CBT*P21) -CD T*Q*C) */
02960 DO 1=1 TO MI
02970 DO J =1 TO NI
02980 LET (1= -1 J=IJI);
1
;

02990 LET (BPDOC(!,J)=BTP2 1(I , J) - DTOC<I,J»;

03000 END;
03010 END;
*
J.23
 03020 1* *1
03030 1* NOW EVALUATE F1=R2INV*(BT*P21-DT*a*C) *1
03040 PUT FILE (SYSPRINT) PAGE LIST
03050 ('MATRIX F1:' >;
03060 PUT FILE (SYSPRINT) SKIP(3)I
03070 DO 1=1 Ta MI
03080 DO J=1 Ta NI
03090 LET (1='1'1 J='J')I
03100 LET (FF1(I,J)=0)I
03110 DO K=1 Ta MI
03120 LET (K='K');
03130 LET (FF1(I,J)=FF1(I,J)+RR2INV(I,K)*BPDaC(K,J»I
03140 ENDI
03150 PRINT_OUT (FF1(I,J»I
03160 ENDI
03170 ENDI
03180 1* *1
03190 1* NOW FIND F2. FIRST CALCULATE BT*P22 *1
03200 DO 1=1 TO MI
03210 DO J=1 Ta NI
03220 LET (1='1'1 J='J')I
03230 LET (BTP22(I,J)=O);
03240 DO K=1 Ta NI
03250 LET (K='I('>;
03260 LET (BTP22(I,J)=BTP22(I,J)+BBTT(I,K)*PP22(K,J»I
03270 END;
03280 ENDI
03290 ENDI
03300 1* *1
03310 1* NOW CALCULATE (BT*P22)+(DT*a*C) *1
03320 DO 1=1 Ta tll
03330 DO J=1 TU NI
03340 LET (I='Z '; J='J');
03350 LET (BP2DgC(I,J)=BTP22(I,J)+DTQC(I,J»I
03360 END;
03370 ENDI
03380 1* *1
03390
0 3 400
1* NOW EVALUATE F2=R2INV*(BT*P22+DT*g*C) *1
PUT FILE (SYSPRINT) SKIP(5) LIST
03410 ('MATRIX F2:' >;
03420 PUT FILE (sYsPRINT) sKIP(3)I
03430 DO 1=1 Ta MI
03440 DO J=1 Ta NI
0 3450 LET (1='1'; J:::'J');
03460 LET (FF2(I,J) =0)I
03470 DO K=1 Ta MI
03480 LET (1( = '1(')1
0 3490 LET (FF2(I,J) =FF2(I,J)+RR2INV(I,K)*BP2DQC(K,J»I
03500 ENDI
0351 0 PRINT_OUT (FF2(I,J»I
03520 END;
03530 ENDI
03540 1 >1< */
03550
03560
1* NOW CALCULATE F3: FIRsT FIND DT*a*D *1
DO 1=1 TO M;
03570 DO J=1 TD MI
03580 LET (1 = '1'; J='J');
03590 LET (DTaD(I,J)=O)I
DO 1(= 1 TO MI
03610 LET (t( = 'K');
03620 LET (DTaD(I,J) =DTaD(I,J)+DDTTaa(I,K)*DD(I( , J»;
03630 END;
03640 ENDI
03650 END;
03660 / * *1
0 36 70 1* NOW FIND BT>I<P23 * /
03680 DO 1 =1 Ta 1'1;
03690 DO J=1 Ta 1'1;
037 00 LET (1 = '1', J='J');
03710 LET (BTP23(I,J)=0);
03720 DO K= l TO N;
03730 LET (K='K');
03740 LET (BTP23(I,J)=BTP23(I,J)+BBTT(I,K)*PP23(I(,J»;
03750 ENDI
03760 ENDI
03770 ENDI
03780 / * */
03790
03800
/* NOW CALCULATE (BT*P23) - (DT*a*D) *1
DO 1=1 Ta 1'1;
03810 DO J=1 Ta 1'1;
03820 LET <1 = '1'; J = 'J');
03830 LET (BPDQD(I,J) =BTP23(I,J)-DTQD(I,J»I
03840 END;
03850 ENDI
03860 /* *1
03870 /* NOW EVALUATE F3=R2INV*(BT*P23-0T*a*D) *1
03880 PUT FILE (sysPRINT) PAGE LIST
03890 ( ' MATRIX F3:');
03900 PUT FILE (sysPRINT) sKIP(3)I
03910 DO 1 =1 TO 1'1;
03920 DO J=1 TO MI
03930 LET (I='Z'; J:::'J');
03940 LET (FF3(I,J)=0);
03950 DO K=1 Ta MI
03960 LET (K = 'K');
03970 LET (FF3(I,J) =FF3(I,J)+RR2INV(I,K)*BPDQD(K,J»;
03980 ENDI
03990 PRINT_OUT (FF3(I,J»;
04000 END;
04010 ENDI
04020
0 4030 ******WE NOW START THE CALCULATION OF W*****
04040
04050 /*FIRsT SET UP A NEGATIVE 2X2 IDENTITY MATRIX */
04060 LET (IIN(1,1)=-1; IIN(1,2)=0; IIN(2,1)=Ol IIN(2,2)=-1)I
04070 1* */
04080 /* NOW SET UP THE MATRIX Ta BE INVERTED: (S*I-A+B*F2)
04090 FIRsT SET UP 5*1: */
04100 DO 1 = 1 TO NI
04110 DO .J=1 TO NI
04120 LET (1='1'; J:::'J');
04130 LET (sII(I,J)=O);
04140 END;
041 50 LET (511(1,1)=5)1
041 6 0 END;
04 170 1* */

'" J.24
04180 /* NOW EVALUATE AND PRINT S*I-A */
04190 PUT FILE (SYSPR INT) PAGE LIST
04200 ( ' THE MATRIX S*I-A IS:');
04210 PUT FILE (SYSPRINT) SKIP(3);
04220 DO 1= 1 TO 1'1;
042 30 DO J= l TO Ni
04240 LET (1="1"; J="J");
04250 LET (SIHAC!,J)=S IICI,J) -AA(!,J»;
04260 PRINT_OUT (SIMA(I.J»;
04270 END;
04280 END;
04290 /* *I
04300 / * NOW GET B*F2 */
04310 DO 1 =1 TO 1'1;
04320 DO J = l TO 1'1;
04330 LET (1="1"; J="J");
04340 LET (BBFF2(I,J) =O);
04350 DO K=l TO M;
04360 LET <K="K");
04370 LET <BBFF2(I,J) =B BFF2(I,J)+BDCI,K>*FF2<K,J»;
04380 END;
04390 END;
04400 END;
04410 / * */
04420 /* EVALUATE ANII PRINT THE MATRIX TO BE
04430 INVERTED: «S*I-A)+(B*F2» */
04440 PUT FILE (SYSPRINT) PAGE LIST
0 4450 ( ' THE MATRIX S*I-A+B*F2 IS: ');
04460 PUT FILE (SYSPRINT> SKIP(3n
04 4 70 DO 1=1 TO 1'1;
04480 DO J=l TO 1'1 ;
04490 LET (1="1"; J="J");
04500 LET (SIABF2(I, J )=SIMA(I ,J)+BBFF2(I,J»;
04510 PRINT_OUT (SIABF2(I,J»;
04520 ENDi
04530 END;
04540 1* */
04550 / * WE NOW EVALUATE THE INVERSE OF MATRIX (S*I-A+ B*F2)
04560 USING THE ADJOINT AND DETERMINANT METHOD, THE
04570 RESULT IS GIVEN AS NUMERATOR AND DENOMINATOR * /
04580 /* */
0 4590 / * FIRST FIND THE MATRIX OF COFACTORS */
04600 DO Ml = l TO 1'1;
04610 DO 1'11 =1 TO 1'1;
04620 11 = 1; KI=!;
04630 DO IJ = 1 TO Ni
04640 DO JJ = l TO Ni
0 4650 IF II =Ml
04660 THEN GO TO NEXT1;
04670 ELSE IF JJ =Nl
04680 THEN GO TO NEXT 2 ;
0 4690 EL SE DO;
04700 LET (DC"Il","Kl")=SIABF2( -IP,-J J·»;
04710 Kl =K 1+lI
04720 END;
0 4 730 NEXT2! END;
0 4 7 40 Kl=1; 11 =11+1;
04750 NEXTl: END;
04 760 /* */
0 4 770 /* THE 5X5 MINOR OF Ml.Nl HAS BEEN CONSTRUCTED.
04780 lTS DETEBMINANT YIEL DS THE COFACTOR OF Ml.Nl * /
0·4 790 LET (DET=O);
04800 DO V= l TO 5;
04810 DO 101 =1 TO S;
04820 I F W=V
04830 THEN GO TO NEXTtH
04840 DO X=l TO 5;
04850 IF X=V : X= W
04860 TH EN GO TO NEXTX;
04870 DO Y= 1 TC S;
04880 IF y=v : Y= W : y=x
04890 THEN GO TO NEXTY;
04900 DO Z=l TO 5;
049 10 IF Z=V : Z= W : Z=X : Z=Y
04920 THEN GO TO NEXTZ;
04930 /* *I
04940 / * NOW FINO THE NUMBER OF INVERSIONS */
04950 NINV=O;
04960 IF Y>Z
04970 THEN NINV =NINV+l;
04980 IF X>Y
04990 THEN NINV =NINV+l;
05000 IF x>z
050 10 THEN NINV =NINV+1 ;
05020 IF W>X
05030 THEN NINV =NINV+l ;
05040 IF W>Y
05050 THEN NINV =NINV+l;
05060 IF W>Z
05070 THEN NINV =NINV+l;
05080 IF V>W
05090 THEN NINV =NINV+l;
05100 IF V>X
05110 THEN NINV =NINV+l;
05120 IF V>Y
05130 THEN NINV =NINV+l;
05 140 IF V>Z
05150 THEN NINV =NINV+l;
05160 IF NINV =O : NINV=2 NINV =4 NINV =6 NINV=8 NINV=10
05170 THEN SIGN=J;
05180 IF NINV = l : NINV =3 : NINV=5 NINV=7 NINV=9
05190 THEN SIGN=-l;
05200 LET (V=-V-; W= ·W· ; X=-X ·; y = .y-; Z=·Z-; SIGN= ·SIGN- );
05290 LET (DET =DET+SIGN*D(1.V)*0( 2 .W)*D(3.X)*D(4.Y)*D(S.Z»;
05220 NEXTZ: END;
05230 NEXrr: END;
05240 NEXTX: EN[I;
05250 NEXTW: END;
05260 END;
05270 LET (COF(·Ml -,·Nl-) =(-1>**( -M1-+-N1- >*DET);
05280 END;
05290 END;
05300 / * */
05310 / * NOW OBTAIN THE TRANSPOSE OF COF TO GET THE ADJOINT */
05320 DO 1=1 TO 1'1;
05330 DO J=l TO 1'1;
* J.25
05340 LET (1="1"1 J:"J")'
05350 LET (ADJ(J.I)=COF(I.J»I
05360 END;
05370 END;
05380 / * */ /* NOW OBTAIN THE DETERMINANT OF (S*I-AtB*F2)
05390
05400 (DENOMINATOR OF THE INVERSE) */
05410 /* WE RECALL THAT HATRIX COF(MM.NN) CONTAINS
05420 ALL THE 4X4 DETERHINANTS CORRESPONDING TO ELEMENTS
05430 . MM.NN. WE THEREFORE USE IT TO EXPAND BY lST ROW */
05440 LET (DET=OH
05450 DO J=l TO N'
05460 LET (J="J"H
05470 LET (DET=DET+SIABF2(1.J)*COF(1.J».
05480 END;
05490 /* */ /* THE INVERSE HAS NOW BEEN CALCULATED WHERE ADJ
05500
05510 IS THE NUMERATOR AND DET IS THE DENOMINATOR */
05520 /* CONTINUE BY FINDING F2*ADJ */
05530 DO 1=1 TO M'
05540 DO J=l TO N'
05550 LET (1="1"; J="J");
05560 LET (F2ADJ(I.J)=0).
05570 DO K=1 TO Ni
05580 LET (K="K").I
05590 LET (F2ADJ(I.J)=F2ADJ(I.J)tFF2(I.K)*ADJ(K.J»f
05600 END;
05610 END;
05620 END;
05630 / * */ /* NOW FIND (F2*ADJ)*B */
05640
05650 DO 1=1 TO H.
05660 DO J=l TO H.
05670 LET (1=·1·; J="J");
05680 L"ET (F2ABCI,J)=O);
05690 DO 1<=1 TO N;
05700 LET O( = "K");
05710 LET (F2ABC!,J)=F2ABCI,J)+f2ADJCI,K>*BB(K,J»;
05720 END;
05730 END;
05740 ENDi
05750 /* */
/* NOW MULTIPLY THE DENOMINATOR (DET) TIMES -I * /
05760
05770 DO 1= 1 TO M'
05780 DO J = l TO Mi
05790 LET "CI="I"; J="J");
05800 LET (DETIINCI,J)=DET*IINCI,J»;
05810 ENDi
05820 ENDi
05830 1* */
05840 /* FINALLY, ADD (DET*-I)+(F2*ADJ*B)
05850 (NUMERATOR OF FIRST TERM OF W) */
05860 DO 1=1 TO 'H
05870 DO J=l TO M;
05880 LET (1="1"; J="J");
05890 LET (NHR1(I,J)=F2ABCI,J)+DETIINCI,J»;
05900 END;
05910 END;
05920 /* *I
05930 /*T HE NUMERATOR AND DENOMINATOR OF THE FIRST TERM OF
05940 W HAVE NOW BEEN CALCULATED. WE USE THE SAME
05950 PROCEDURE FOR THE SECO ND TERM. */
05960 / * WE NOW EVALUATE THE INVERSE OF MATRIX (S*I-A)
05970 USING THE ADJOINT AND DETERMINANT HETHOD. THE
05980 RE SULT IS GIVEN AS NUMERATOR AND DENOMINATOR */
05990 /* FIRST FIND THE MATRIX OF COFACTORS * /
06000 DO Hl=1 TO N;
06010 DO Nl=l TO N'
06020 11=1; K1=U
06030 DO 11 = 1 TO N;
060 40 DO JJ=l TO N'
06050 IF II =Ml
06060 THEN GO TO NEXT11.
06070 ELSE IF JJ=Nl
06080 THEN GO TO NEXT22;
06090 EL SE DO;
06100 LET (D(·I1·,·K1·)=sIMA(·II·,·JJ·»;
0611.0 K1=K1+1 ;
06120 END;
06130 NEXT22: END;
06140 Kl=l; Il=Il+U
06150 NEXT11: END I
06160 /* *1
06170 /* THE 5XS MINOR OF Ml.Nl HAS BEEN CONSTRUCTED.
06180 lTS DETERMINANT YIELDS THE COFACTOR OF Hl.Nl */
06190 LET (DETSIA=OH
06200 DO V=1 TO 5;
06210 DO W=l TO S;
06220 IF W=V
06230 THEN GO TO NE XT WWI
06240 DO X= l TO 5;
06250 IF X=V : X=W
06260 THEN GO TO NEXTXX'
06270 DO Y=l TO SI
06280 IF Y=V : Y=W : Y=X
06290 THEN GO TO NEXTYY'
06300 DO Z=l TO 5.
063 10 IF Z=V : Z=W : Z=X : Z=Y
06320 THEN GO TO NEXTZZ'
06330 /* */ 1* NOW FIND THE NUHBER OF INVERSIONS */
06340
06350 NINV =O ;
06360 IF Y>Z
06370 THEN NINV=NINV+1'
06380 IF X>Y
06390 THEN NINV=NINV+1'
06400 IF X>Z
06410 THEN NINV=NINV+1'
06420 IF W>X
06 430 THEN NINV =NINV+1'
06 440 IF W> Y
06450 THEN NINV =NINV+1'
06460 IF W>Z
06470 THEN NINV=NINV+1'
06480 IF V>W
06490 THEN NINV=NINV+1'
* J . 26
 06500 IF V>X
06510 THEN NINV=NINV+l1
06520 IF V>Y
06530 THEN NINV=NINV+l;
06~40 IF V>Z
06550 THEN NINV=NINV+l1
06560 IF NINV=O : NINV=2 NINV=4 NINV=6 NINV=8 I NINV=10
06570 THEN SIGN=l;
06580 IF NINV=l : NINV=3 : NINV=5 NINV=7 NINV=9
06590 THEN SIGN=-l;
06600 LET (V='V'; W='W'; X='X'; V= 'Y'; l='Z'; SIGN~'SIGN');
06610 LET (DETSIA=DETSIA+SIGN*D(1,V)*D(2,W)*D(3,X)*D(4,Y)*D(5,Z»;
06620 NEXTZZ: END I
06630 NEXTYri END;
06640 NEXTXX: END;
06650 NEXTWW: END;
06660 ENDi
06670 LET (COFSIA('Ml','Nl') =(- l )**( 'Ml'+'Nl')*DETSIA)I
06680 ENDi
06690 END;
06700 1* * 1
0 6 710 1* NOW OBTAIN THE TRANSPOSE OF COFSIA TO GET THE ADJOINT *1
06720 DO 1=1 TO N;
06730 DO J=l TO Ni
067 40 LET (1='1'; J = 'J');
06750 LET (ADJSIA(J,I)=COFSIA(I,J»;
06760 END;
06770 ENDi
067 80 1* *1
06790 1* NOW OBTAIN THE DETERMINANT OF (S*I-A)
06800 (D ENOMINATOR OF THE INVERSt) *1
06810 1* WE RECALL THAT MATRIX COFSIA(MM,NN) CONTAINS ALL THE
06820 SIGNED 4X4 DETERMINANTS CORRESPONDING TO ELEMENTS
06830 MM,NN, WE THEREFORE USE IT TO EXPAND BY lST ROW *1
06840 LET CDETSI A=O);
06850 DO J=l TO N;
06660 LET (J='J');
06870 LET (DETSIA=DETSIA+SIMA(l,J)*COFSIA(l,J»;
061lBO ENDi
06890 1* */
06900 1* THE INVERSE HAS NOW BEEN CALCULATED WHERE ADJSIA
06910 IS THE NUMERATOR AND DETSIA IS THE DENOMINATOR *1
06920 1* CONTINUE BY FINDING Fl*ADJSIA *1
06930 DO 1=1 TO M;
06940 DO J=l TIJ Ni
06950 LET (1= ' 1'; J='J');
06960 LET (FIADJCI, J )=O);
06970 DO K= 1 TO N;
06980 LET (I( == 't< ' );
06990 LET (FtADJ ( I,J)=F1ADJ (I,J)+FF1(I, K) *ADJSIA(K,J»;
07000 EN1H
070 10 END;
07020 END i
07030 /* */
07040 1* NOW FIND (F1*ADJSIA> *B *1
07050 DO 1=1 Ta Mi
07060 DO J =l TO Mi
07070 LET ( I =' I '; J=' J' ) ;
07080 LET (FIA BCI,J )=O) ;
07090 00 K=1 TO Ni
07100 LET (1(= ' 1( ' );
07110 LET (F1 AB (I, J )=F1AB(I ,J)+F1ADJ(I,K)*BB( K,J »;
07120 ENDi
07130 ENDi
07 140 ENDi
07150 1* *I
07160 1* NOW MULTI PLY THE DENOMINATOR (DETSI A) TIMES F3 *1
07170 DO 1=1 Ta Mi
07180 DO J =l TO M;
07 190 LET (1='1'; J ='J I);
07200 LET (DETFF3(I,J)=DETSIA*FF3( I,J »;
072 10 END;
07220 END;
07230 /* */
07240 1* FINALLY, ADD (DET*F3)+(F l*ADJSIA*S)
07250 (NUMERATOR OF SECOND TERM OF W) *1
07260 DO 1=1 TO M;
07270 DO J=l TO M;
07280 LET (1= ' 1'; J = 'J');
072 90 LET (NMR2(I, J )=F1AB(I,J)+DETFF3(I,J»;
07300 ENDi
07310 END;
07320 1* *1
07330 1* NOW THAT WE HAVE THE TWO TERMS WHICH HAKE UP W,
07340 WE MULTIPLY THEM OUT TO OBTAIN AND PRINT
07350 DUR FINAL ANSWER *1
07360 PUT FILE (SYSPR INT) PAGE LIST
07370 ('THE NUHERATOR OF WIS:' ) ;
07380 PUT FILE (SYSPR INT) SKIP(3) ;
07390 DO 1=1 TO Mi
07400 DO J=l TO H;
07410 LET (1=-1-; J=.J e );
07420 LE T ( NMRTR(I,J) =O);
07430 DO K= l TO M;
07440 LET (K='K');
07450 LET (NMRTR(I,J)=NMRTR(I,J)+NMR1(I,K)*NMR2(K,J»'
07460 END;
07470 PRINT_OUT(NHRTR(I,J»;
07480 PUT FILE (SYSPRINT) SKIP(4);
07490 END;
07500 END;
07510 PUT FILE (SYSPRINT) PAGE LIST
07520 ('THE DENOHINATOR (DETERMINANT) IS:');
07530 PUT FILE (SYSPRINT ) SKIP(3);
07540 LET (TOTDET=DET*DETSIA)'
07550 PRINT_OUT(TOTDET);
07560 END WASH1 2 ;
07570 1*
07580 II

J.27
 1
2
c ..
C
*. PROGRAM ANALYZE.FTN 1
2
3 C PROGRAM TO CALCULATE. PR I NT ~ PLOT THE TRANSFER FUNCTION 3
4 C BODE PLOTS ~ERE THE TRANSFER FUCT I~S ARE GIVEN AS 4
5 C POLYNOMIALS OF ORDER 14. IF A N-TH ORDER POLYNOMIAL 5
I. C IS REGUIRED. SET ALL COEFF HIGHER THAN N TO ZERO. I.
7 C THE ROOTS OF THE TRANSFER FUNCTIONS AND NODES OF THE 7
8 C DETERMINANT ARE CALCULATED USING IMSL ' 5 ZPOLR . 8
'I C NOTE THAT COEFll.:I) TO COEFll:1.5) ARE THE DETERMINANT COEFFS. 9
10 C COEFI1. 1> TO COEFI 15. 1) ARE FOR WIl. 1). 10
11 C COEFI1. 2 ) TO COEFI15, 2 ) ARE FOR WI1.2). 11
12 C COEFll.3) TO COEFI15, 3) ARE FOR WI2.1). 12
13 C COEFll . 4) TO COEFI15.4) ARE FOR WI2.2). 13
14 C 14
1:1 C SET NOROaI FOR lXI W IHEAVE OR YAWl 15
Ilo C NORD-4 FOR 2X2 W IPITCH- SURQE OR SWAY-ROLL) Ilo
17 C SET NDEç.b OR 12 FOR PITCH-SURGE OR SWAY-ROLL 17
18 C NDEç.4 OR 8 FOR YAW OR HEAVE 18
1'1 C 1'1
20 C THIS PROORAM WAS NOT CHANGED WHEN THE SEGUENCE OF THE 20
21 C EGUATIONB WAS CHANOED. A OIVEN SYMBOL ON THE PLOTS NOW 21
22 C REPRESENTS THE DIAO~ALLY OPPOSITE ELEMENT . 22
23 C 23
24 0000001 DIMENSI~ OM(33) . WPHI31>. COEFI 15, 5). AR (31,4). 24
25 *DB (31. 4). PHASE 131.4). DETAMP 131 ). DETPHI 31> . 25
21. *PHSTRI (33). PHSTR2(33) , PH5TR3(33). PHSTR4(33). 21.
27 *DBSTRI(33).DBSTR2133) . DBSTR3133).DBSTR4133). 27
28 *AA I 15). BBl 15). CCI 15).00(15). EE(15) 28
2'1 000001.1 COMPLEX Z (14) 2'1
30 000001.1 INTEOER*4DATES(3) , 1114) 30
31 000001.1 DATA PHSTR 1(32) . PHSTR2 I 32). PHSTR31 32) , PH5TR4(32) 14*31.0 . 31
32 000001.1 DATA PHSTRI(33).PHSTR2133) . PHSTR3133) . PH5TR4133)/4*120 . 32
33 000001.1 DATA DBSTRI(32) . DBSTR2132).DBSTR3132).DBSTR4132)/4*180 . I 33
34 000001.1 DATA DBSTRI(33) . DBSTR2133) . DBSTR3133) , DBSTR4(33) 14*30. I 34
35 000001.1 DATA OM(32). OMI 33). Pil . Ol • . 3333.3. 1415'11 35
31. 000001.1 DATA NDEO.NORD / 8.11 36
37 C 37
38 C-** ENTER VARIABLES 38
3'1 C 39
40 000001.1 DATA· COEF I 40
41 *-. 27133E-35 •. 322b2E-3 4 . - . 12258E-Ib.-. 225'1IE-15 • . 04325473. 41
42 * . '18041778.1.. 511.65007. 10. 41.57425, 0 . '18238384.1.*0 . • 42
43 * . bb27'1E-20 • . 18084E-18 •. 2Ob8b4E-17. -. 10345E-03.-. 003b7b23. 43
44 *- . 04'151.'163 . - . 332'14203. -1 . 21428238. -2. 48468571. -2. 87759138. 44
45 *-1 . 54111'107 . - . 2312473 . - . 01035447. O .• O.• 45
41. *0 . • o. . 0 .• - . 46b7bE-18. -. Ibb47E- lb. - . 20457E-15. - . 25bOJE-Ob. 46
47 * . 3b787bE-05 • . 2b8'l4bE-03 • . 003451.31. • . 01733754 • . 03142471 . 47
48 *. 002'17252. O.• O.• 48
4'1 *-. 48b'lOE-35 . - . 255 1'1E-33.- . 53280E-32 • . b'l'104E-18 • . 17806E-Ib. 49
50 * . 1058IE-15 •. 00258914 •. 08270122 • . '1b2'13454,5 . 11130'15.12. 04'17117. 50
51 *'1 . 37505785 • . 8220245. O. • O. • 51
52 *0 . • O.•. 0032831.5 • . 0877'13'18 • . 8303515'1 . 3 . 5771018.8. 77433'101.. 52
53 *10 . 8424844.1. • 6*0 . I 53
54 C 54
55 C-** END OF VARIABLE ENTRY 55
:Ilo C CALCULATE THE AMPLITUDE RATIOS AND PHASE ANOLES : 51.
:17 C FI RST DO FOR THE DETERMINANT: 57
58 C 58
5'1 000001.1 DO 2 ,)=1 . 31 59
1.0 OOOEEC I OM I .)=IO. * *1 -2 . I+FLOATI.) I 10. ) 60
lol 000F34I DETI=COEFI2.5)*OMI.)-COEFI4.5)*OMI.)**3. +COEFI6.5)*OMI.)**5 . lol
1.2 * -COEFI8. S >*OM I .),.*7 . +COEFI 10. S ) *OMI.)**'1. 1.2
1.3 * -COEFI12.5)*OMI.) * *I1. + COEFI 14. 5)*OMI.),.*13 . 63
1.4 0010401 DETR~COEF(I.5 ) -COEF I 3 . 5 ) *OMI.)**2. +COEFI5. 5)*OMI.)**4 . 1.4
1.5 *-COEFI 7 . 5 ) *OMI .) * *1. . +COEF I '1. 5) *OMI .)**8 . -COEFI 11 . 5 ) *OM 1.)**10. 65
blo *+COEFI 13. 5)*OMI.)*1tl2 . - COEFI 15. 5H.OMI.)*1tl4 . 61.
1.7 00115EI DETAMP(~) - (DETR.*2 . +0ETI**2 . )**0. 5 1.7
1.8 001lA81 2DETPHI.)sATAN2IDETI.DETR ) 68
1.'1 C 6'1
70 C NOW CALCULATE THE TRANSFER FUNCTIONS: 70
71 C 71
72 OOllEbl DO 5 1=1. NORD 72
73 001lF81 DO :I ,)=1 . 31 73
74 0012001 WI - COEFI2, I )*OM ( .) -COEF I 4 . I )*OMI.) **3 . +COEFI6. I )*OMI.) **5. 74
75 * -COEFI8. I) *OMI .)**7. +COEFI 10. I ) *OM 1.)**'1. 75
71. -COEFI 12. I )*OMI.)**I1. + COEFI 14, I )*OMI.)**13. 76
77 0013bEI WR - COEFI I. I )-COEF I 3. I) *OMI .)**2 . +COEFI 5. I) *OMI .)**4 . 77
78 *-COEF( 7 . I) *0f1( ~)**6 . +COEF( 9. I) *OM( ~)**8 . -COEF( 11 , 1 ) *OM (~)**10 . 78
7'1 _COEF I 13. I ) *OM 1.)>**12. - COEFI 15. I )*<lMI.)**14 . 79
80 OOI4FCI AMP=(WR**2 . +WI * *2 . )**0 . 5 80
81 0015401 AR I.). I) ~AMP/DETAMPI') 81
82 OOI5bAI OBI.). 1)=20 . *ALOOIOIAMP/DETAMPI.)) 82
83 0015Bbi WPHI.)=ATAN2IWI . WR) 83
84 0015DC 1 5 PHASEI.). I)=IWPHI.)-DETPHI.))*57. 2'16 84
8:1 C 85
81. C SMOOTH THE PHASE DATA : 86
87 C 87
88 0011.401 DO 15 I - I.NORD 88
8'1 0011.521 IIIIl - 1 8'1
'10 OOlbbOI DO 15 ,)=1.31 '10
'11 00lbb8I IFI.). EG. 1>00 TO 15 '11
'12 00lb7EI IF I IPHASEI.). I) -PHASEI.)-I . I) l. OT . 450. ) '12
'13 *PHASEI.). Il=PHASEI.), 1)-720. '13
'14 00lbF41 IFI IPHASEI.). I)-PHASEI.)- l . I» . GT . '10 . ) '14
'15 *PHASEI.).I)~PHASEI') . I)-3bC. '15
'Ilo 0017bAI IF I IPHASEI.). I) -PHASEI.) - I . I ) ) . LT. -4:10. ) 'Ilo
'17 *PHASEI.).I) - PHASEI.). Il+72O. '17
'18 0017E01 IF I IPHASEI.). I) -PHASEI')-l, I) l. LT. -'10. ) '18
'1'1 *PHASEI.).I ) - PHASEI.). Il+3bC. '1'1
100 001851.1 IF IPHASEI.).!) . OT. 0 . ) IIII)-o 100
101 0018881 1:1 CONTINUE 101
102 001881.1 DO 25 l=l.NORD 102
103 0018C8I DO 25 ,)=1.31 103
104 0018001 25 PHASEI.). I) - PHASEI.). I ) +FLOATI II I I) >*31.0 . 104
105 C 105
lOb C CREATE A TABLE OF RESULTS AND OBTAIN ROOTS OF POLYNOMIALS 106
107 C FIRST RESET POLYNOMIAL COEFFICIENTS FOR USE BY ZPOLR : 107
108 C 108
10'1 00191.01 GO TO '1'1 10'1
110 001'11.1.1 DO 50 1=1. NDEG+l 110
111 00197AI AA(I)=COEFINDEG+2-I.l) 111
112 001'1'181 BB I 1 ) =COEFINDEG+2-1 , 2) 112
113 001981.1 CC 1!)=COEFINDEQ+2-!. 3) 113
114 001'1041 DO I I )=COEFINDEG+2-!. 4) 114
J. 28
115 00191'21 ~o EEtI)-COEFCNDEQ+2-1.5) 11'
11b C FIRST DO FOR WC 1. 1) : 116
117 001A29I CALL DATE CDATES) 117
11 9 001AAOI WRlTE C2.10') DATES(01).OATESC3).DATES(1) 11B
119 001Ab9I WRlTE (2. 110) 119
120 001AS4I DO 100 1=1.31 120
121 001ASCI 100 WRlTE (01.115) OM(Il.DBCI.ll.AR(I.I). PHASE(I.l) 121
122 OOIBlEI WRlTE (2.1010) 122
123 001B3BI CALL ZPOLR(AA.NDEQ.Z. IER) 123
124 001B'CI WRlTE (01.1015) CZ(I).I-l.ND EQ ) 124
12' 001BDBI IF (NORD. EG . l) go TO bO 12'
12b C NOW FOR WCl.01) : 126
127 001BECI WR lTE (2.130) 127
12B 001C09I WRlTE (2. 110) 128
129 001C24I DO "0 1=1.31 129
130 001C2CI 1:)0 WRlTE (2.115) OMCIl.DB(I.2).AR (I .01 ) . PHASE(I.01) 130
131 001CBEI WRlTE (2.1010) 131
132 001CDBI CALL ZPOLR(BB.NDEg.Z. IER) 1301
133 001CFCI WRlTE (2.101') (Z(I), I-l. NDEg) 133
134 NOW FOR W( 01.1) : 134
135 001D79I WR lTE (01. 135) 13'
13b 0010941 WR lTE (2.110) 136
137 001DBOI 00 lbO 1-1.31 137
139 001DB9I lbO WRlTE (2.115) OM(I).DB(I.3).ARCI.3).PHASE(I.3) 139
139 001E4AI WRlTE (2,120) 139
140 001Eb4I CALL ZPOLR(CC.NDEG.Z.IER) 140
141 001EB9I WRlTE C2.12') (ZCI).I-l.NDEG) 141
142 NOW FOR W(01.2) : 142
143 001F04I WR lTE (01. 140) 143
144 001F20I WR ITE C2. 110) 144
145 001F3CI DO 170 1-1,31 145
14b 001F44I 170 WRITE (01.115) OM(I).DH(I.4).AR(I.4).PHASE(I.4) 146
147 001FDbI WR ITE (2. 1010) 147
149 001FFOI CALL ZPOLRCDD.NDEg. Z. IER) 14B
149 0020141 WR ITE C2.1015) (ZC Il. I-I. NDEG) 149
150 C FIND THE POLES OF THE DETERMINANT: 150
"I 002090 I bO WRlTE (2.145) 151
152 0020ACI CALL IPOLRCEE.NDEG.Z. IER) 152
153 0020DOI WRlTE (2.125) CIC I ). I-l .NDEG) 153
154 C**** NOW PLOT THE RESUL TB: 154
155 C FIRST PLOT THE PHASE ANGLE--FREG CURVES 155
15b 00214C I 99 CALL PLOTS (0. O. 0) 15b
157 0021941 CALL PLOT(7 .•. 7.-3) 157
15B 0021C4I CALL PLOT(-3. • 0 . • +3) 158
159 0022041 CALL PLOT(-3.• 9 . • +01) 159
lbO 000101441 CALL AXIS(O .• O. • 11 HPHASE (DEG).+11. b . O.lBO . • -PHSTRl(32). 160
lb! ttPHSTR 1 (33) ) 161
lb2 C CREATE STRINGS TO USE IN PLOTTING (NEG SION FOR AXES ROTATlON) 162
lb3 00012091 DO 10 1- 1.31 Ib3
lb4 002:zE0 I PHSTRl(I)--PHASECI.1) 164
IbS 002O!FE I IF (NORD. EG . l) go TO 10 165
lbb 00013121 PHSTR2( I ).-PHASEC 1.2) Ibb
lb7 0023301 PHSTR3 CI ) - -PHASEC 1 . 3) 167
lb9 00234EI PHSTR4 CI ) - -PHASEC 1.4) 16B
lb9 00013bCI 10 CONTINUE 169
170 0023921 CALL LGAXIS(O . • 0 . • lOHFREG (RIS).-1 0 .9. 0 . 90 . • DI't(32) . Dl't(33» 170
171 0024041 CALL LQLINE(PHSTRl , QM,31. 1 , +1,0,+ 1) 171
172 00014'CI IF CNORD. EG . l) GO TO 70 172
173 0024701 CALL LGLINE(PHSTR2. OM. 31. 1. +1.1. +1) 173
174 00014CBI CALL LGLINEePH9TR3.OM,3t. 1.+1.2.+ 1) 174
175 0025201 CALL LOLINEePHSTR4 . OH . 31. 1. +1.5,+ 1) 175
17b 002'7BI 70 CALL PLOT ( O .0 .• -999) 176
177 C**** NOW DRAW THE AMPLITUDE RATID--FREG CURVES: 177
179 002'091 CALL PLOTC7 . • . 7. -3) 17B
179 0001!5F9I CALL PLOT( - b . • 0 . • +3) 179
lBO 00010391 CALL PLOT( - 6 .• 9 . . +2) 180
IBI 00 01b7BI CALL AXIS(O . • O. • BHA. R. (DB). +8. 6 . O. lBO .• -DBSTRI (32). 191
lB01 *DBSTRI (33» 192
lB3 C CREATE STRINGS TO USE IN PLOTTING CNEg SION FOR AXES ROTATlON) lB3
lB4 00017041 DO:ZO 1 - 1.31 lB4
lB5 00270CI DBSTRI (I ) - -DB( 1. 1) lB5
lBb 00017:zA I IF (NORD. EG . l) go TO 20 IB6
lB7 00273EI DBSTR2 (Il - -DB( I. 2) lB7
lBB 00275CI DBSTR3( I) - -DB( I . 3) lBB
lB9 00277AI DBSTR4 (Il - -DB( I . 4) 189
190 000179BI 20 CONTINUE 190
191 00017AEI CALL LGAXIS(O .• 0 . • 1OHFREO CRIS).-1 0.9. 0 . 90 . • OM(32),[»1(33» 191
192 00018301 CALL LGLINE(DBSTRl.0M.31. 1.+1.0.+1) 1901
193 0001BSBI IF (NORD.EO. 1) go TO 80 193
194 0001B9CI CAL L LGLINECDBSTR2.QM. 31.1.+1.1.+1) 194
19' 0001BF41 CALL LGLINECOBSTR3. OM. 31. 1. +1. 2. + 1) 195
196 00294CI CALL LGLINEeOBSTR4 . OM. 31. 1, ... 1. 5, + 1 ) 196
197 0029A4I 90 CALL PLOT(0 .• 0 .• +999) 197
19B C**** FORMAT STATEI'1ENTS: 198
199 00019E4 1 110 FORMAT (01X. ' FREOUENCY (RIS) '. ,X. 'OAIN (DB)'. 'X . 199
200 * ' AMPLITUDE RATIO'. 'X. 'PHASE ANGLE (DEO) '1 1) 200
201 0001AJbI 115 FORMAT<2X. IPG15 . ,. 3X, lPGIS . 5. O1X. lPGI'. ,. 6X. lPG". 5) 201
202 0001A5EI 120 FORMAT< ' - ROOTS : ' I / ) 202
:Z03 002A74I 125 FORMAT (2X. 2(IPQI!5. ')/12X. 2 (lPQI!5. ' l i l 203
204 *2X . 2C lPG15 . 5)112X. 2( lPO". 5)112X. 2( lPOl~. 5 )112X. 2( lPGI5 . , ) / I 204
205 *2X. 2C lPOlS . 5)112X. 2( lPOI~ . ')112X. 2C1PG15. ')//2X. 2( lPG" . ,) /1 20'
:ZOb *2X. 2( lPQl~ . ,) 112X. 2( lP015. ')112X. 2 (lPG 15. ,) 1/2X. 2( lPG" . ~) ) 206
207 0001B30I 105 FORMAT ( '1 ANALYSIS OF W( 1. 1) : '. 20X. 12. ' 1'.101. '1'. 121 /) 207
20B 0001Bb01I 130 FORMAT< ' I ANALYSISOFWC1.01):'11l 209
209 0001BBbI 13' FORMATC'l ANALYSIS OF W(01.!) : 'lil 209
210 002BAAI 140 FORMAT< ' I ANALYSIS OF W(2.2) : 'lil 210
211 0001BCEI 145 FORMAT ( ' I POLES OF THE DETERMINANT: '11l 211
212 002BF9I STOP 212
213 0001COOI END 213

NO ERRORS: F7D ROS-OO . OO MAINPROQ . MAIN 21/ 1011B4 09: 02 : 41 TABLE SPACE: 6 KB
STATEI'1ENT BUFFER : :zo LINES/13011 BYTES STACK SPACE: lBB WORDS
SINGLE PRECISION FLOATING PT SUPPORT REOUIRED FOR EXECUTION

J . 29
00100 POLYEX: PROCEDURE OPTIONS (HAIN);
00110 /* PROGRAH TO EXPAND AN ARBITRARY POLYNOHIAL */
00120 /* SET NORD=l FOR HEAVE OR YAWl NORD=4 FOR OTHER CASES */
00130 DECLARE (NORD) 8INARY FIXED(31.0);
00140 FORHAC_OPTION5 ;
08150 OPT5ET (LINELENGTH=72; EXPND);
00160 NORD=!;
00170 /* */
00 180 /* FIRST DO THE NUHERATOR OF W */
00190 /* WO.ll: */
00200 LET (E=5t.45712; F=S**2; G=St.0980395);
00210 LET (H=5**2tO,7 1578*StO.59568);
00220 LET ( W(1 . 1)=0.98238384*E*F*G);
00230 IF NORD=l
00240 TH EN GO TO DETC;
00250 /* */
00260 /* W(1,2): */
00270 LET (I=St9.9887; J=Stl.000l; K=St.16786; Kl=St.0704788);
00280 LET (K2=S**2tO.74468*StO.27767);
00290 LET (W(1.2)=-0.01035447*I*J*K*Kl*K2);
00300 /* *I
00310 /* W(2,1): */
00320 LET (L=St.16392; H=S-.0221822; N=St.0760511);
00330 LET (0=5**3);
00340 LET (W(2.1)=.03142471*L*H*N*0);
00350 /* */
00360 /* W(2,2): *1
00370 LET (P=5t.81091; Q=St.16392; R=St.0760773);
00380 LET (T",S**3);
00390 LET (W(2,2)=.8220245*P*g*R*T);
00400 /* *I
00410 DETC: /* NO W DO THE DETERHINANT */
00420 LET <A=S+.31402; 8=5+.0978172; C=5**2t.33234*5t.l0875);
00430 LET (D=5**2t.71586*5t.59552);
00440 LET (DET=A*B*C);
00450 f* */
,00460 /* PRINT OUT RE5ULT5: */
00470 PUT FILE (5YSPRINT) PAGE LIST
00480 ('THE EXPANDED VERSION OF THE POLYNOHIAL5:');
00490 PUT FILE (SY5PRINT) 5KIP(3);
00500 PRINT_OUT (W(l.l»;
00510 IF NORD=l
00520 THEN GO TO PRINT;
00530 PRINT_OUT (W(1,2); W(2,1); W(2,2»;
00540 PRINT: PRINT_OUT (DET);
00550 END POL.YEX;
*

J , 30
 i INPUTS FROM USER (SEE TABLE J.l) r

FORM REQUIRED MATRICES

,-w ,..",..",....,

OBTAIN MATRICES REQUIRED

FOR RICCATI SOLUTION

R
_1 =CTGC
__ _

SOLVE FOR P IN THE ALGEBRA IC MATRIX

RICCATT EQUATION (EQN. J .. l)
USING THE SCHUR METHOD

I PRINT SOLUTION !' f.21' f.22' E.23 I

FIGURE J.l SOLUTION OF MATRIX RICCATI EQUATION.

 INPUTS FROM USER AND RICCATI SOLUTION
f21' ~22' ~23 (FIGURE J.l)

FORM REQUIRED ~ATRICES:

A, ~, f, Q, Q, pR, §.

OBTAIN MATRICES REQUIRED IN WASHOUT EQUATION J.2

R
_2 =pR+DTGD
___ _

f1 = B.; 1 (~T f21 - OT Q f)

1
f2 = B.; (B T f22 + .Q.T Q f)
1
f3 = B.; (B T f23 _ .Q.T Q.Q.)

SOLVE FOR THE NUMERATOR AND DENOMINATOR POLYNOMIALS

OF W(s) ACCORDING TO EQNS. J.3 AND J.4
- AND PRINT THE RESULTS

FIGURE J.2 SOLVING FOR THE WASHOUT POLYNOMIALS.

UTIAS Report No. 296 UTiAS Report No. 296

..";,, " """". '""'Mo ,,,

~, Dufferin Street, Downsview, Ootario,
4925 ~"' Canada,.."
"ooI
M3H.. 5T6
'""'" ~ ~"m", " W''''.
4925 Dufferin '''''Mo
Street, Downsview, ,,, ~"'••" """" "','"
Ootario, Canada, M3H 5T6 ~
FLIGHT SIMULATION MOTION-BASE DRIVE ALGORITHMS: PART 1 - OEVELOPING AND TESTING THE EOUATIONS"", FLIGHT SIMULATION MOTION-BASE DRIVE ALGORITHMS: PART 1 - OEVELOPING AND TESTING THE EOUATIONS ' """

Reid, L.O., Nahon, M. A. Reid, L.O., Nahon, M. A.

1. Flight simulator 2. Motion system 3. Motion-Drive Algorithms 1. Flight simulator 2. Motion system 3. Motion-Drive Algorithms

I. Reid, L.O., Nahon, M. A. 11. UTIAS Report No. 296 I. Reid, L. 0., Nahon, H. A. 11. UTiAS Report No. 296

Thi s report covers the fi rst phase of a project aimed at developing and testing flight simulator This report covers the first phase of a project aimed at developing and testing flight simulator
motion-base algorithms suited to commercial jet transports. Full six-degrees-of-freedom motion of a motion-base algorithms suited to commercial jet transports. Full six-degrees-of-freedom motion of a
synergistic motion-base i s being studied. Three forms for these algorithms are be i ng considered : synergistic motion-base is being studied. Three forms for these algorithms are being considered:
(1) classical linear washout, (1) classical linear wa s hout,
(2) optimal control, (2) optimal control,
(3) coordinated adaptive washout. (3) coordinated adaptive washout •
It is feIt that the latter two techniques may provide some advantages over the classical, which is It is felt that the latter two techniques may provide some advantages over the classical, which is
currently employed in most commercial flight simulators. The aims of this project are to : currently employed in most commercial flight simulators. The aims of this project are to:
(1) develop the necessary equations, (1) develop the necessary equations,
(2) implement the necessary software in real-time (2) implement the necessary software in real-time
(3) evaluate the performance of the software with the help of piloted testing in a complete flight (3) evaluate the performance of the software with ' the help of piloted testing in a complete flight
simulation. simulation.
The present report describes the development of the equations and the software used for testing them. The present report describes the development of the equations and the software used for testing them.
At thi s stage the eval uat i on of the performance of the al gori thms is based on predi cted pil ot mot i on At this stage the eval uation of the performance of the algorithms is based on predicted pilot motion
sensations and mot1on-base travel. The motion sensat10ns are fonnulated from linear models of the human sensat10ns and motion-base travel. The motion sensations are fonnulated from l1near models of the hLl11an
vest 1bul ar system. At present the actual mot i on-base hardware i 5 not empl oyed. vest1bular system. At present the actual motion-base hardware is not employed.
The results of many test runs of the algor1ttms are presented to demonstrate their proper operation and The resul ts of many test runs of the al gor1 thns are presented to demons trate the1 r proper operat 1on and
their effect;veness. Several modifications to ex1sting formulations are sU9gested. their effectiveness. Several modifications to existing formulations are suggested.
Available copies of this report are limited. Return this card to UTIAS, if you reqUire a copy. Available copies of th is report are limited. Return this card to UTIAS, if you require a copy.
UTlAS Report No. 296 UTlAS Report No. 296

University of Toronto, Institute for Aerospace Studies (UTlAS)

""'~ '"' '00"'""' ~ .~ ""''' '"' '''''''"' ~

University of Toronto, Institute for Aerospace Studies (UTlAS)
4925 Dufferin Street, Downsview, Ootario, Canada, H3H 5T6 4925 Dufferi n St reet, Downsv iew, Ootari 0, Canada, H3H 5T6

FLIGHT SIMULATION MOTION-BASE DRIVE ALGORITHMS : PART 1 - OEVELOPING

M" FLIGHT SIMULATIOtI MOTION-BASE DRIVE ALGORITHMS :
Reid, L. 0., Nahon, M. A.
PART 1 - DEVELOPING

Reid, L. 0., Nahon, M. A.

1. Fl i9ht simulator 2. Mot i on system 3. Motion-Drive Algorithms 1. Flight simulator 2. Motion system 3. Motion-Drive Algorithms

I. Reid, L. 0., Nahon, H. A. 11. UTI AS Repo rt No. 296 I. Reid, L.O., Nahon, M. A. I I. UTlAS Report No. 296

This report covers the first phase of a project aimed at devel opi ng and test i ng fl ight s imul ator This report covers the first phase of a project aimed at developing and testing fllght simulator
mot ion-base algorithms suited to commercial jet transports. Full six-degrees-of-freedom motion of a motion-base algorithms suited to commercial jet transports. Full six-degrees- of-freedom motion of a
synergistic motion-base is being studied. Three forms for these algorithms are being considered : synergistic motion-base is being studied. Three forms for these algorithms are being considered:
(1) classical linear washout, (1) classical linear washout,
(2) optimal control, (2) optimal control,
(3) coordinated adaptive washout. (3) coordinated adaptive washout.
It is felt that the latter two techniques may provide some advantages over the classical, which is It is felt that the latter two techniques may provide some advantages over the classical, which is
currentl y employed i n most cammerc i a 1 fl i ght simulators. The aims of thi s proj ect are to : currently employed in most commercial flight simulators. The aims of this project are to:
(1) develop the necessary equations, (1) develop the necessary equations,
(2) implement the necessary software in real-time (2) implement the necessary software in real-time
(3) evaluate the performance of the software with the help of piloted testing in a complete flight (3) eval uate the performance of the softwa re wi th the hel p of pil oted test i ng i n a compl ete fl i ght
simulation. simulation.
The present report describes the development of the equations and the software used for testing them. The present report describes the development of the equations and the software used for testing them.
At this stage the evaluation of the performance of the algorithms is based on predicted pilot motion At this stage the eval uat i on of the performance of the al gorithms is based on predi cted pil ot mot i on
sensations and motion-base travel. The motion sensations are formulated from linear models of the human sensati ons and mot i on-base travel. The mot i on sensat i ons are formul ated from li near model s of the human
vestibular system. At present the actual motion-base hardware is not employed. vestibular system. At present the actual motion-base hardware is not employed.
The results of many test runs of the algorithms are presented to demonstrate thei r proper operation and The results of many test runs of the algorithms are presented to demonstrate their proper operation and
their effectiveness. Several modifications to existing formulations are suggested. their effectiveness. $overal modifications to existing formulations are suggested.

Available copies of this report are limited: Return this card to UTIAS, if you require a copy. Available copies of this report are limited. Return this card to UTIAS, if you require a copy.