DTIC AD0608056: PHASE AND ENVELOPE OF LINEAR FM PULSE-COMPRESSION SIGNALS FROM HIGH-VELOCITY TARGETS

Survival, Water, Medical Field Manuals

Military Manuals

Defense Technical Information Center

Document text

ESTI  FILE  COP 


ESTI  PROCESSED 


□  DDC  TAB  □  PRO J  OFFICER 

□  ACCESSION  MASTER  FILE 

ESD-TDR-  64-128  U - *  TM-03916 

DATE _ 

,  Ai  43563 

ESTI  CONTROL  NR_*? _ 

CY  NR  (  OF _ \__CYS 

PHASE  AND  ENVELOPE  OF  LINEAR  FM  PULSE- COMPRESSION 
SIGNALS  FROM  HIGH-VELOCITY  TARGETS 


TECHNICAL  DOCUMENTARY  REPORT  NO.  ESD-TDR-64-128 

NOVEMBER  1964 

ESD  RECORD  COPY 

RETURN  TO  M.  H.  Ueberschaer 

SCIENTIFIC  &  TECHNICAL  INFORMATION  DIVISION 
(ESTI),  BUILDING  1211 

COPY  NR. _  OF _ COPIES 

Prepared  for 

DIRECTORATE  OF  RADAR  AND  OPTICS 
ELECTRONIC  SYSTEMS  DIVISION 
AIR  FORCE  SYSTEMS  COMMAND 
UNITED  STATES  AIR  FORCE 
L.  G.  Hanscom  Field,  Bedford,  Massachusetts 


Project  750 
Prepared  by 

THE  MITRE  CORPORATION 
Bedford,  Massachusetts 
Contract  AF  19  (628)-2390 


Copies  available  at  Office  of  Technical  Services, 
Department  of  Commerce. 


Qualified  requesters  may  obtain  copies  from  DDC. 
Orders  will  be  expedited  if  placed  through  the  librarian 
or  other  person  designated  to  request  documents 
from  DDC. 


When  US  Government  drawings,  specifications,  or 
other  data  are  used  for  any  purpose  other  than  a 
definitely  related  government  procurement  operation, 
the  government  thereby  incurs  no  responsibility 
nor  any  obligation  whatsoever;  and  the  fact  that  the 
government  may  have  formulated,  furnished,  or  in 
any  way  supplied  the  said  drawings,  specifications, 
or  other  data  is  not  to  be  regarded  by  implication 
or  otherwise,  as  in  any  manner  licensing  the  holder 
or  any  other  person  or  corporation,  or  conveying 
any  rights  or  permission  to  manufacture,  use,  or  sell 
any  patented  invention  that  may  in  any  way  be  related 
thereto. 


Do  not  return  this  copy.  Retain  or  destroy. 


ESD-TDR-  64-128 


TM-03916 


PHASE  AND  ENVELOPE  OF  LINEAR  FM  PULSE- COMPRESSION 
SIGNALS  FROM  HIGH-VELOCITY  TARGETS 


TECHNICAL  DOCUMENTARY  REPORT  NO.  ESD-TDR-64-128 

NOVEMBER  1964 

M.  H.  Ueberschaer 


Prepared  for 

DIRECTORATE  OF  RADAR  AND  OPTICS 
ELECTRONIC  SYSTEMS  DIVISION 
AIR  FORCE  SYSTEMS  COMMAND 
UNITED  STATES  AIR  FORCE 
L.  G.  Hanscom  Field,  Bedford,  Massachusetts 


Project  750 
Prepared  by 

THE  MITRE  CORPORATION 
Bedford,  Massachusetts 
Contract  AF  19  (628)-2390 


FOREWORD 


The  author  wishes  to  acknowledge  the  help  of  Mr.  R.  D. 
Haggarty,  Dr.  R.  Manasse,  Dr.  J.  F.  A.  Ormsby,  and 
Mr.  R.  W.  Jacobus  who  also  suggested  the  investigation 
and  reviewed  the  manuscript. 


PHASE  AND  ENVELOPE  OF  LINEAR  FM  PULSE-COMPRESSION 


SIGNALS  FROM  HIGH-VELOCITY  TARGETS 


ABSTRACT 


Equations  for  the  phase  and  envelope  of  the  output  signal  from  a  linear  filter, 
matched  to  the  transmitted  signal,  are  derived.  The  transmitted  signal  is  assumed 
to  have  a  flat  band-limited  amplitude  spectrum  and  a  linear  group  delay.  The 
input  to  the  "matched  filter"  is  the  radar  echo  returned  from  a  moving  target 
whose  velocity  is  essentially  constant  during  the  illumination  time.  It  is  shown 
that  the  returned  signal  is  related  to  the  transmitted  signal  by  a  time  dilation. 

The  resulting  expressions  for  the  phase  and  envelope  are  functions  which  involve 
Fresnel  integrals.  Approximations  for  these  expressions  are  worked  out.  They 
are  shown  to  be  similar  in  form  to  those  which  are  obtained  when  the  returned 
signal  is  assumed  to  be  related  to  the  transmitted  signal  by  a  Doppler  shift. 


REVIEW  AND  APPROVAL 


This  technical  documentary  report  has  been  reviewed  and  is  approved. 


Acting  Chief,  Radar  Division 
Directorate  of  Radar  and  Optics 


iii 


PHASE  AND  ENVELOPE  OF  LINEAR  FM  PULSE-COMPRESSION  SIGNALS 
FROM  HIGH-VELOCITY  TARGETS 


SECTION  1 
INTRODUCTION 

This  document  is  a  first  step  in  exploring  the  possibilities  of  using  an 
"all  pulse-compression"  (linear  FM)  radar  system  to  obtain  accurate  estimates 
of  target  range,  radial  velocity,  and  radial  acceleration.  The  targets  of  pri¬ 
mary  interest  here  are  high-velocity  targets,  such  as  artificial  satellites. 

It  is  well  known  that  linear  FM  signals  have  an  inherent  coupling  between 
range  and  velocity.  However,  it  is  also  known  that  for  targets  whose  range  is 
varying  slowly,  and  for  small  pulse-compression  ratios,  this  can  be  overcome 
by  transmitting  alternately  FM  up  and  FM  down,  and  taking  the  sum  and  differ¬ 
ence  of  the  resulting  time-delays  to  obtain  unambiguous  estimates  of  target 
range  and  velocity,  respectively.  When  the  target  range  changes  rapidly  and 
the  pulse-compression  ratio  is  high,  certain  simplifying  assumptions  are  no 
longer  applicable,  and  the  problem  may  become  considerably  more  difficult. 

This  document  investigates  the  problems  from  a  fundamental  point  of 
view.  The  accuracies  required  in  the  ultimate  estimation  of  target  parameters 
dictate  the  necessity  of  measuring  the  phase  of  the  target  echo.  We  shall  derive 
several  expressions  (with  different  degrees  of  exactness)  for  the  phase  and 
envelope  of  the  output  signal  from  a  pulse-compression  system.  These  expres¬ 
sions  will  be  carefully  compared  and  interpreted  in  a  subsequent  document. 


1 


We  shall  first  review  briefly  the  concepts  of  autocorrelation,  matched 
filters,  and  correlation  functions.  Given  an  aperiodic  time  function  f(t),  such 
that 


f  (t)  dt  <  00 


(1) 


we  define  the  autocorrelation  function  of  f(t)  to  be 


0  (  t) 


f  (t)  f  (t  +  r)  dt 


(2) 


By  a  matched  filter  we  mean  a  linear  filter  whose  impulse  response  h(t)  is  a 
reflection  of  the  time  function  to  which  it  is  matched,  as  illustrated  in  Fig.  1. 


Input 


h(t)  =  f  (T-t) 


y(t) 


Output 


Fig.  1.  Matched  Filter 


The  quantity,  T,  is  a  constant  which  makes  the  filter  realizable  [i.  e.  ,  we 
require  that  h(t)  =  0  for  t  <  0] .  The  output  is  the  convolution  between  the  input 
and  the  impulse  response,  i.  e. , 


=  1 


y(t)  =  \  f  (x)  h  (t-x)  dx  . 


Thus, 


y(t)  =  \  f  (X)  f  (x+T-t)  dx 

*^—00 

Letting  r  =  T-t,  we  find  that 

y(t)  =  \  f(x)  f  (x+t)  dx  , 

—OO 

which  is  equivalent  to  0  (r)  defined  by  Eq.  (1). 


(3) 


(4) 


(5) 


2 


Thus,  it  is  seen  that  matched  filter  reception  and  autocorrelation  detection 
are  identical  processes,  provided,  of  course,  that  the  filter  is  truly  "matched" 
to  the  input  waveform. 

In  radar  applications,  the  filter  is  often  matched  to  the  transmitted  wave¬ 
form.  The  target  echo  may  be  quite  different  from  the  transmitted  signal.  If 
this  is  the  case,  the  output  of  the  "matched  filter"  is  no  longer  equal  to  the  auto¬ 
correlation  function  but  is  now  a  cross-correlation  function  between  the  trans¬ 
mitted  and  received  signals.  It  is  often  referred  to  loosely  as  an  autocorrelation 
function.  Regardless  of  its  name,  it  is  this  function  which  we  are  interested  in 
examining.  We  shall  denote  it  by  y(t)  throughout  the  rest  of  this  report. 


3 


SECTION  2 


PHYSICAL  MODEL 


Consider  the  model  shown  in  Fig.  2  t 


h(t) 


H(o>) 

Fig.  2.  Pulse-Compression  Model 


With  the  receiver  matched  to  the  transmitted  signal,  we  have, 

h(t)  =  s(T-t)  ,  H(o>)  =  S*(co)  exp  [-jcuT]  ,  (6) 

where  T  makes  the  filter  realizable,  and  the  star  denotes  the  complex  conjugate. 
A  linear-FM  rectangular  pulse,  lasting  from  time  t  =  0  to  t  =  T,  has  the 


form 


cos 


f(t)  = 


2*  fi +  K 


;  0  <  t  <  T 


(7) 


0  elsewhere 

The  Fourier  transform  of  f(t)  is  a  rather  complicated  function.  How¬ 
ever,  it  has  been  found*  that  the  amplitude  spectrum  of  f(t)  becomes  nearly 
2 

rectangular  as  KT  (the  time-bandwidth  product)  becomes  large  (>100).  The 
present  D-82  experimental  radar  facility  employs  a  pulse-compression  system 
with  a  time-bandwidth  product  of  1, 000.  A  future  system  is  proposed  which 


*cf.  Klauder,  et  al. ,  "The  Theory  and  Design  of  Chirp  Radars,  "  B.  S.  T.  J.  , 
July  1960. 


5 


will  have  a  time-bandwidth  product  of  10,  000.  Instead  of  synthesizing  a  simple 
function  of  time,  the  procedure  here  is  to  synthesize  a  simple  function  of 
frequency,  denoted  by  S(u>). 

The  function  S(u>)  has  approximately  a  rectangular  amplitude  spectrum 
A(oj),  and  a  linear  group  delay*  T(u>).  For  large  time-bandwidth  products,  this 
corresponds  approximately  to  a  linear  FM  pulse. 

For  our  analysis,  we  shall  use  the  model  shown  in  Fig.  3. 


Fig.  3.  Model  of  A(co)  and  T(u>) 


*  Group  delay  is  defined  as 

T(u>) 


d0(g?) 

do; 


where  0(u>)  is  the  phase  of  the  spectrum. 


6 


The  respective  equations  are: 


„  W  -  ,  W 

H  ;“o-  T  s  M  £“o  +  — 


A(u))  = 


(8) 


t) 


elsewhere 


TM  -  -  57  I 


2  I  ,1  t  /  w\ 

W  '  ^  W  yo  ~  2  )  ’ 

(-.*?)■ 


U).  —  —  — 


w 

2 


T  ,  ,  T 

w  +  W 


co  -  “  — 
0 


W 

2 


i  i  W 

|  co  |  +  ~2  for  FM  up 

i  i  W 

|  co  I  ^oj  +  —  for  FM  down 

U  L* 


elsewhere  (for  both  FM  up  and  FM  down) 

(9) 


The  spectrum  S(u:)  can  be  written  as 


S(o;)  =  A(oj)e 


j0(<^) 


(10) 


where  0(co)  =  -  C  T(u>)  do.-  plus  a  constant  of  integration.  Without  loss  of 
generality,  we  can  let  H  =  1  in  Fig.  3  and  Eq.  (8).  Doing  this,  we  obtain 


J0(w) 


S(w)  = 


w  ,  ,  w 

co  —  —  —  Cl)  —  co  +  — 

0  2  11  0  2 


elsewhere 


(11) 


It  is  clear  that  S(u>)  is  an  even  function,  since  A(u:)  and  T(a;)  are  both 
even  functions.  For  simplicity,  we  shall  work  only  with  the  positive  portion 
of  the  spectrum  throughout  the  rest  of  this  paper.  It  must  be  remembered  that 
the  actual  spectra  contain  negative  frequencies. 


7 


Integrating  T(u>)  and  restricting  ourselves  to  ui  ^  0,  we  obtain 


S(w)  = 


where 


j[c0  +  u  +  c2  w2] 


W  ^  ^ 

;  CO  -  —  —  CO  —  CO  + 

’  0  2  0 


elsewhere 


c  = 


c  = 


!(-o- 1)  jtorFMup  l 

~ff  (“o  +  t)  ^forFMdowni 


w 

2 


C  =  =F 

2  2W 


\  -  for  FM  up 


1  +  for  FM  down 


(12) 


(13) 


(14) 


and  Cq  is  a  constant  of  integration  which  will  cancel  out  later.  Equation  (12) 
and  Fig.  2  shall  be  applicable  throughout  this  paper. 

The  following  two  pulse-compression  systems  are  of  particular  interest: 


Pulse-Compression  Systems 

T 

W 

2ir 

Cl) 

_0 

27T 

No.  1 

1  msec. 

1  me. 

1280  me. 

3 

(10  system) 

No.  2 

2  msec. 

5  me. 

1280  me. 

4 

(10  system) 

The  length  of  the  transmitted  pulse  is  approximately  equal  to  T. 

We  shall  regard  the  targets  under  discussion  here  as  being  essentially 

point  targets.  Let  the  maximum  radial  acceleration  of  our  targets  be  200 
2 

m/sec  .  Suppose  a  pulse  of  duration  T  is  emitted  at  time  t  =  0.  Let  a  given 
target  be  at  range  R  when  the  leading  edge  of  the  pulse  strikes  it.  Assuming 


8 


free-space  propagation,  this  occurs  at  the  time  t  =  —  ,  c  being  the  velocity 

of  light.  Let  the  radial  velocity  and  radial  acceleration  of  the  target  at  that 
time  be  ef^ual  to  V  and  A,  respectively. 

During  the  time  interval  ( «T)  that  the  target  is  being  illuminated,  the 
change  in  radial  velocity  is 

AV  =  AT  ^  0.  2  m/sec  for  system  No.  1, 

^0.4  m/sec  for  system  No.  2, 

2 

(since  A  ^200  m/sec  ).  The  corresponding  change  in  range  due  to  the  accelera¬ 
tion  term  alone  is 

1  2 

AR  =  —  AT  ^  0.  0001  m  for  system  No.  1, 

0.0004  m  for  system  No.  2, 

This  is  only  a  few  degrees  of  phase  shift  for  the  1280-megacycle  radar,  which 

is  very  small  indeed.  We  are  primarily  interested  in  large  radial  velocities 
3  4 

(of  the  order  10  and  10  m/sec,  producing  a  AR  =  VT  of  the  order  of  1  to  20 
meters). 

Throughout  this  paper  we  shall  regard  the  target  velocity  as  being  constant 
during  the  illumination  time. 


9 


SECTION  3 


DOPPLER-SHIFTED  SIGNAL 


One  assumption  which  is  often  made  is  that  the  radar  signal  bounced  off 
a  moving  target  is  a  delayed  and  attenuated  replica  of  the  transmitted  signal, 
except  for  a  Doppler  shift.  *  This  is  only  an  approximation.  It  is  a  fairly  good 
one  for  small  velocities.  In  Appendix  I  it  is  shown  that  the  correlation  function 
(under  this  assumption)  is  equal  to 


The  envelope  has  the  familiar 


sin  x 
x 


form. 


In  Eq.  (15) 

k 

W' 


t’ 


P 


some  constant  which  depends  on  the  radar  cross  section  of 
the  target, 


T 

/2V  \ 

-  for  FM  up 

W 

{-  V 

+  for  FM  down 

*By  this  we  mean  that  the  spectrum  of  the  returned  signal  X(o j)  is  related  to  the 
spectrum  of  the  transmitted  signal  S(oj)  by 

2R 

”  W  c  /  2V  \ 

X(w)  =  ke  c  S  fw  +  a>0)  , 

2V 

where  —  cj„  is  the  "Doppler  shift." 
c  0 


11 


2V 


Gl), 


6 

R 

V 

The  above 
it  with  the  more 


W 

2 


+  for  FM  up 
-  for  FM  down 


=  the  target  range  when  the  pulse  strikes  it,  and 


=  the  target  radial  velocity  when  the  pulse  strikes  it. 


expression  was  derived  primarily  for  the  purpose  of  comparing 
exact  expressions  to  be  developed  in  the  next  section. 


12 


SECTION  4 


TIME-DILATED  SIGNAL 


In  Appendix  II  it  is  shown  that  (under  the  assumption  of  essentially  constant 
velocity  during  the  illumination  period)  the  relationship  between  the  transmitted 
signal  s(t)  and  the  returned  signal  x(t)  is  given  by 

x(t)  =  k  s 


This  corresponds  to  a  "time  dilation. "  Here 


c+V  .  2R 

a  =  — —  ;  b  =  - 

c-V  c 


(18) 


and  k,  R,  V,  and  c  have  the  same  meanings  as  before.  In  Appendix  II  we  also 
obtain  the  Fourier  transform  of  x(t): 


X(w)  =  k  a  e"jal)  S  (aw)  . 

The  output  of  the  receiver  is  then 

Y(w)  =  X(w)  S*(w)  e"jcJr  =  k  a  e_jc0<b+T)  S(aw)  S*  (w) 


From  Eq.  (10)  it  is  clear  that 

V  * 

S*(cu)  = 

Similarly,  we  have 


10  (aw) 


w 

CO  -  -7T  —CO  —  CO  + 

0  2  o 


elsewhere 


W  _ 

w  „  -  —  —  aw  ^  w  + 

0  2  0 


W 

2 


W 

2 


S(aw)  = 


(19) 


(20) 


(21) 


(22) 


elsewhere 


13 


Since  a  >  0,  this  can  be  written  as 


W  W 

;  U)  -  —  U.'  +  — 

j0(aw)  ’0  2  0  2 

e  '  ;  -  - 


S(aw)  = 


0  elsewhere 

Combining  (21)  and  (23),  we  obtain 


i[0(ao.’)  -  0(a’)] 


S(aw)  S*(u>)  = 


Cl)  + 

w  o 

uj  -  —  ^  cc  —  - 

0  2  a 


W 


0  elsewhere 

Thus,  over  the  non-zero  interval,  we  have 

0(au>)  -  0(cj)  =  cQ  +  c^^  au>  +  c2  (aa)2  -  [cQ  +  o>  +  c2  a-2] 

2  2 

=  c^  u>  (a-1)  +  co  a;  (a  -  1) 


Let 


a  -  1  =  y  = 


2V 

c-V 


(23) 


(24) 


(25) 


a  -  1  =  (a-1)  (a+1)  =  y  \ 


y  = 


2c 

c-V 


Substituting  (24),  (25)  and  (26)  into  (20),  we  obtain 

2 


(26) 


W 


Y(w)  = 


Cl)  +  — 

\V  ^  0  2 

Cl)  -  ^  Cl)  —  - 

0  2  a 


elsewhere 


(27) 


14 


Let  go"  be  the  center  of  the  non-zero  region  in  the  go  domain,  and  let  W"  be  the 
bandwidth  of  Y(oj);  i.  e.  ,  let 


cj"  + 
0 


W” 

2 

W" 


=  GO  - 
0 


w 

2 


(28) 


(29) 


Combining  (28)  and  (29)  with  (18),  and  solving  alternately  for  go"  and  then  for 

W"  W 

in  terms  of  and  —  ,  we  obtain 


"0 

r 

2 


c 

"o  \cTv 
w 


w 

T 


c+v 


-or 


V 

c+V 

V 


0  Vc+V 


(30) 


(31) 


These  expressions  are  valid  for  a  receding  target.  For  an  approaching 
target,  we  would  obtain 


0 

W" 

2 


0  \c-V 

W  /  c 
2  (c-V 


W 


V 

c-V 


- 


V 

0  Vc-V 


The  output  spectrum  can  now  be  written  as 


2 

1j[u(-b-T+Cj  y]  +  a; 
ka  e 

0 

We  must  remember  that  Y(co)  in  Eq.  (32)  is  only  the  positive  half  of  the 
actual  output  spectrum.  But  since  we  have  an  even-frequency  function,  the 
output-time  function  y(t)  can  be  obtained  by  taking  twice  the  real  part  of  the 
Fourier  transform  of  the  positive  half  of  the  spectrum;  i.  e. , 


W" 

2 


GO  SCO, 


W" 

2 


(32) 


elsewhere 


15 


y(t)  =  2  Re 


ka  P 
2t r  J, 


a):'  + 


W' 


2/  jco[t  -  b  -  T  +  C1  y]  jcu  y  X 


dco  .  (33) 


W" 
2 


Here  Re  f(t)  means  "the  real  part  of  f(t).  "  Let 

t’  =  t  -  b  -  T 


=  t  -  —  -  T  , 
c 


(34) 


where  t'  =  0  corresponds  to  the  peak  of  the  autocorrelation  function  for  the 
case  of  zero  velocity.  We  then  have 


y(t)  =  2  Re 


ka  f 
2t r  J 


“o'+  ? 


2/  jw[t'  +  c  -y]  j  co  c2  y\ 


dco  . 


(35) 


W 
0  2 


This  expression  is  exact  for  a  constant-velocity  target,  and  very  nearly 
exact  if  the  percentage  change  in  velocity  over  the  pulse  duration  is  small. 
However,  it  cannot  be  integrated  in  closed  form.  It  is  convenient  to  introduce 
the  variable 


S2  =  a;  -  cii^'  , 


(36) 


so  that  y(t)  becomes 


where 


y(t)  =  2  Re 


ka  jcr  P 
27r  6  J 


W" 

2 


j[c  Sl+c  if] 
e  dS2 


(37) 


W" 

2 


a  =  (t'  +  y  c1)  cu”  +  y\c2  (cd^')  , 

cg  =  t'  +yc1  +  2y\c2  to”  ,  and 


C4  =  YXC2  • 


(38) 


16 


Equation  (37)  can  be  manipulated  into  the  form 


where 


y(t)  =  2  Re 


u  .  a 

i  j»  — 

e 


=  2  Re  kx  [  Z  (ux)  -  Z  (u2)] 


dx 


(39) 


Z(x) 


■I 


X  j7T 

e 

0 


a 


dx 


(40) 


is  the  complex  Fresnel  integral.  The  resulting  expression  can  then  be  put  in 
terms  of  the  simple  Fresnel  integrals, 


C(x) 


-f 

J0 


cos  f  7T  ^-jdx  ;  S(x) 


■r 

Jo 


sin  ur  —  J  dx  ,  (40a) 


which  are  tabulated  functions.  This  was  done  (the  details  are  given  in  Appendix 
III),  yielding 

X"i  /. 

(41) 


y(t)  = 


0 


^[C(x3)  -  C(x4)]  cos  xi  +  [S(x3)  -  S(x4)]  sin  x^ 


for  FM  up 


x 

y(t)  =  —  ^c(x3)  -  c(x4)]  cos  Xx  ~  ts(x3)  "  s(x4)]  sin  x^ 
2  / 


(42) 


for  FM  down 


where 


ka 

•>y  27r  ■  ’ 

(t’)2  +  2y\Ci  t' 
4y\c2 


17 


X 


3 


x 


4 


+ 


W" 

~2 


,  and 


W" 

“ 


(43) 


The  other  symbols  are  the  same  as  before.  We  can  also  express  y(t)  in  the 


following  form: 


y(t) 


R  cos  (x^  -  0)  , 


(44) 


where 


R  =  -]/  [C(x3)  -  C(x4)]  2  +  [S(x3)  -  S(x4)] 


(45) 


for  both  FM  up  and  FM  down,  and 

-1 


+  tan 


S(xg)  -  S(x4) 


0  = 


-  tan 


-1 


C(X3)  "  C(X4) 


S(x3)  -  S(x4) 
C(x3)  -  C(x4) 


for  FM  up 


for  FM  down 


(46) 


These  expressions  are  exact  for  a  target  whose  velocity  is  constant  over  the 
pulse  duration.  Tables  of  the  Fresnel  integrals  are  available  so  that  numerical 
answers  of  the  correlation  function  y(t)  can  be  obtained  directly. 

It  would,  however,  be  desirable  to  obtain  approximate  "closed-form" 
expressions  for  both  the  envelope  R  and  the  carrier  term  cos  (x^  -  0).  We 
are  primarily  interested  in  the  representations  which  are  valid  in  the  vicinity 
of  the  peak  of  the  envelope. 


18 


In  Appendix  III  we  show  that  the  peak  occurs  when  c  =  0.  This  corresponds 

u 


to  the  time 


(Note  that  y  =  — —  =  0  when  V  =  0).  Thus,  we  let  t"  =  c  in  order  to  empha- 

C  —  V  o 

size  that  the  peak  occurs  at  t"  =  0;  i.  e. ,  let 


t"  =  t  -  —  -  T  +  vc  +  2  v X c  o>" 


(48) 


In  Appendix  III  it  is  also  shown  that  (for  the  two  pulse-compression  sys¬ 


tems  we  consider  here)  the  arguments  x^  and  x^  of  the  Fresnel  integrals  are 


close  to  unity  at  the  peak  of  the  envelope  (i.  e.  ,  at  t"  =  0)  for  large  velocities. 
Thus,  asymptotic  approximations  for  R  and  0  which  are  valid  in  the  immediate 
vicinity  of  the  peak  are  unsuitable  except  for  very  low  target  velocities.  How¬ 
ever,  the  analysis  of  a  simple  Doppler-shifted  signal  [with  Eq.  (15)  as  the 
resulting  ambiguity  function]  is  expected  to  be  adequate  for  low  velocities.  We 
shall,  therefore,  not  attempt  to  get  "closed-form"  approximations  for  (45)  and 
(46),  which  are  valid  in  this  region. 

As  we  move  away  from  the  peak,  the  absolute  values  of  x  and  x  increase 

u  4 

quite  rapidly,  allowing  us  to  use  asymptotic  approximations  of  C(x)  and  S(x),  for 
large  x,  yielding  approximate  expressions  for  R  and  0  which  might  provide 
some  insight.  This  was  done  in  Appendix  III  with  the  result 


(49) 


where 


(50) 


19 


(51) 


In  Eq.  (44)  we  expressed  the  correlation  function  by 


y(t)  = 


o 


R  cos  (x^  -  0)  . 


This  can  be  written  as 


y(t)  =  E  cos  6  .  (52) 

In  Appendix  III  the  following  approximate  expressions  for  the  envelope  E  and  the 
phase  6  were  obtained: 


e  «  oo”  t"  ± 


/  2V\  /  2c  \  /  T 
\c-V/  \c-V/\2W 


sin 


W" 

~ 


(■"  ?) 


for  FM  up 
for  FM  down 


(53) 


(54) 


(55) 


These  approximations  are  good  for 

.-I  >“•- (sxa,)  (»)-“(?)’■ 

It  is  interesting  to  note  that  the  (t")^  term  in  the  expression  for  0  cancelled 
with  (t1 )  term  in  the  expression  for  x^  [see  Eq.  (43)],  so  that  6  =  x^  -  0  has 
only  linear  time  dependence. 

This  is  a  direct  result  of  the  approximations  we  have  made.  The  conclu- 

2 

sion  to  be  drawn  is  that  the  (tM)  term  is  negligibly  small  in  the  region  indicated 
by  Eq.  (55).  This  may  or  may  not  be  the  case  in  the  immediate  vicinity  of  the 
peak. 


20 


Let  us  consider  a  different  method  for  approximating  the  output  function. 
Going  back  to  (37)  we  have 


Y(t)  =  2  Re 


ka  j  a  C 
27 r  6  J 


W" 

~2~ 


j [c  ft  +  c,  ft  ] 

e  3  4  dn  . 


W" 

~2~ 


It  would  be  nice  if  we  could  simply  ignore  the  ft  term.  Suppose  we  say  that  the 
answer  we  get  by  doing  so  is  a  good  approximation  if 


|c3  S2|  >  100  |c4  ft  I  . 

Letting  t"  =  c  and  substituting  for  c  ,  we  find  that  (56)  becomes 

O  rr 


(56) 


|tM  n|  >  ioo 


Clearly,  for  any  particular  value  of  t"  this  inequality  is  most  difficult  to  satisfy 

ii  W" 

when  |  S2|  is  largest;  i.e.  ,  when  ft  =  ±  . 


Thus,  (56)  is  satisfied  if 


:*•!>*"  (5)  (5)  (#?  -"(?)*■ 


(57) 


When  the  ft  term  is  ignored,  (37)  is  integrated  readily  to  yield,  after  taking 
the  real  part, 


y(t)  w  a 


sin  | 

wff 

L‘  — J 

W"- 

~2~ 

cos  a 


(58) 


We  note  that  the  envelope  above  is  identical  with  Eq.  (53).  Let  us  compare  the 
phase  with  Eq.  (54).  We  have,  from  (38), 

.2 


a  =  (t'  +  yc1)  wj'  +  y  Xc2  (w”) 


(38) 


21 


When  we<  express  t'  in  terms  of  t",  with  the  aid  of  (48),  we  have 


co’’  V  =  co’’  t"  -ycx  co’’  -  27\c2  (co’’) 


Substituting  (59)  mto  (38),  we  obtain 


co^'  t”  -yXc2  (co^') 


(59) 


(60) 


This  becomes 


(y  -  (jj 


o'  *  £r)  (£)  © 


,  2  +  for  FM  up 

0  -  for  FM  down 


(61) 


Comparing  this  expression  with  the  phase  9  in  Eq.  (54)  we  find  that  the  two 


expressions  are  the  same  except  fbr  the  last  factor  which,  in  (54) 

2 


■  W 


,  is  j(co”) 


instead  of  just  (oo^' )  .  However,  since  co^'  «  co^  (very  closely)  and 


W"  <  W,  and  since  co^  is  about  three  orders  of  magnitude  greater  than  W,  we 
see  that  cr  and  0  are  identical  for  all  practical  purposes. 

Note  that  the  region  over  which  (58)  is  valid  is,  from  (57), 


i"i 

In  (55)  we  required  |t"|  to  be  twice  as  large  as  that,  which  was  a  little  more 
conservative. 

It  is  quite  remarkable  that  the  simple  way  of  approximating  y(t)  shown 
above  yields  virtually  the  same  result  as  the  rather  involved  procedure  of 
approximating  the  Fresnel  integrals. 


22 


SECTION  5 


SUMMARY 


We  have  derived  three  different  expressions*  for  the  output  function  y(t). 
Each  expression  is  of  the  form 

y(t)  =  E  cos  0  . 


(1)  Simple  Doppler  Shift:  X(c o)  =  k  e 


2R 

-jcu  - 


where 


2k  W'  Sm 

(t*  +  P ) 

W’l 

2~J 

7 r  2 

W'  1 

(  +  P  )  2 

=  <t'  +/3)«J 

+  <5 

[» f  *  ?  4 


W' 


w 


2V 


2 

2 

—  CO 

c  0 

> 

t' 

=  t  - 

2R 

c 

-  T  , 

P 

=  T 

2V  T 

“  ^0  w 

j  -  for 
|  +  for 

6 

2V 

c 

[•5 

T 

±  — 
w 

<x>  ' 

0 

=  CO 

0 

?)■ 

(62) 


*Note:  The  expressions  listed  here  may  differ  slightly  in  appearance  from  the 
corresponding  ones  in  the  test.  They  are  identical,  however. 


23 


c  =  the  speed  of  light, 

R  =  the  target's  range  when  the  pulse  strikes  it,  and 

V  =  the  target's  radial  velocity  when  the  pulse  strikes  it. 


-jw 

(2)  Time  Dilation:  X(u>)  =  k  e 


E  = 


The  exact  expressions  (within  the  limits  of  our  assumptions)  are: 

(t)  ‘  •(»";)  -i'vl*!  ■ 

©  -d 

# 


t"  ±  y  X 

±  L - 


±  tan 


S(x3)  -  S(x4) 


C(x3)  -  C(x4) 


+  for  FM  up 
-  for  FM  down 


(63) 


The  terms  C(x)  and  S(x)  are  the  Fresnel  integrals,  defined  by  Eq.  (40a),  and 
c+V  2V 


oo 

0 

W" 

2 


c-V 


y  = 


c-V 


X  = 


2c 

c-V 


(£r)  -  ?  (ii)  ■ 

(c~7v)  ■  “o  (jTv)  ’ 

(w)[“o-  ¥  - 


w 

2 


2R 

=  t  -  -  -  T  ±  y 

c 


X  oj" 
0 


?  ©  ’RJ) 

1  w  L-  (£) 


+ 


W^n  j-  for  FM  up  I 
2  I  |+  for  FM  down  j 


+  for  FM  up  | 

-  for  FM  down  )  ’ 


W 

~2 


] 

?] 


and 


|  -  for  FM  up  | 
|+  for  FM  down  ^ 


24 


(3)  Approximations  for  E  and  6  in  Eq.  (63). 


fa\ 

WT,1 

r  tJ 

i) 

l  ft- 

W"1  J 

9  «  t"  ±  y  X 


tf)  [<-;•)• -('f 


|+  for  FM  up  | 
|  -  for  FM  down  ^ 

These  approximations  are  expected  to  be  quite  good  for 


(64) 


|t"|  >50W"y\  ^L\  M  100  T  . 

They  may  be  adequate  for  considerably  smaller  values  of  |t"|  . 

The  "exact"  expressions  in  (63)  are  not  easily  interpreted.  This  will 
have  to  be  done  numerically.  We  did,  however,  determine  that  the  peak  of  the 
envelope  occurs  at  the  time  when  t"  =  0.  The  peak  of  the  approximate  expres¬ 
sion  for  the  envelope  in  (64)  occurs  at  exactly  the  same  time.  It  is  interesting, 
also,  to  note  that  the  expressions  in  (64),  for  both  the  phase  and  the  envelope, 
are  virtually  identical  to  those  in  (62)  when  the  target  velocity  is  small. 

Due  to  the  length  of  this  document,  we  shall  reserve  the  detailed  interpre¬ 
tation,  comparison,  and  application  of  the  above  expressions  for  a  subsequent 
document,  ESD-TDR-64-129. 


%4  Ktfa 


L  ¥  (1*1  < 


M.H.  Ueberschaer 


25 


APPENDIX  I 


ANALYSIS  OF  A  DOPPLER-SHIFTED  SIGNAL 

We  assume  that  the  relationship  between  the  transmitted  and  received 
Fourier  spectra  is 

2R 

-j  —  c o 

X(co)  =  k  e  C  S(co  —  co  )  ,  (I—  1) 

d 


where 


k  is  the  attenuation  constant, 

R  is  the  range  of  the  target  when  the  pulse  strikes  it,  and 
co,  is  the  "Doppler  shift",  given  by 


2V 

“d  -  -  -  "» •  ‘!-2> 

where  V  is  the  radial  velocity  of  the  target  when  the  pulse  strikes  it.  (We  are 
here  using  the  convention  that  V  is  positive  when  the  target  is  receding  from 
the  radar. )  Consider  the  case  of  a  receding  target,  and  let 


2V 

a  =  —  co.  =  —  or  . 

d  c  0 

If  Eq.  (11)  is  substituted  for  the  transmitted  spectrum,  (1-1)  becomes: 

2R 


-jco 


k  e 


c  jp(co  +  a) 
e 


X(u>)  = 


W  W 

CO  -  —  —  CO  +  Q!  —  CO  +  — 

0  2  0  2 


elsewhere 


(1-3) 


(1-4) 


27 


Again,  we  must  remember  that  this  is  only  the  positive  half  of  the  frequency 
spectrum.  Equation  (1-1)  can  be  written  as 


2R 


k  e 


^  c  j  0  {oo  +  a) 


W 


Wj 


X(w)  = 


We  have 


Y(u>)  =  X{oo)  S*  (uj)  e 
By  use  of  Eq.  (1-5),  this  becomes 

[k  (' c“  +  T)  ej[0(w+a) -0(w)] 

Y(W)  = 


;  a,0-a-  T  +  T 

elsewhere 
-jcoT 


(1-5) 


(1-6) 


W 


W 


;  oo  -  —  —  U)  —  oo  -  a  +  —  J 
’0  2  0  2  1 


elsewhere 


Now, 


0  (oj)  =  Cq  +  oo  +  c^  oo  , 


with  c  and  c  given  by  (13)  and  (14).  Thus, 

X  Li 


(1-7) 


(1-8) 


{oo  +  Q! )  —  <f>{oo)  =  C  +  C  (ii)  +  a)  +  C  (ft)^  +  2  U  Q  +  a^)  -  [  c  +  c  w  +c.w^] 

V  1  Li  U  1  Li 

=  oo  ^2 c^  of^  +  (c^  a  +  c  a2) 


(1-9) 


Let 


2  c  a  =  p ,  c  a  +  c  a  =  <5  . 

Li  X  Li 


(1-10) 


28 


Further,  let  co^  be  the  center  of  the  non-zero  region  of  Y(co),  and  let  W1  be  the 
bandwidth  of  Y(co);  i.  e.  ,  let 


a  W'  W-a 

“o  u0  ~2  ’  T  ~2~ 


(1-11) 


Then  we  obtain 


Ik  ej6  e 


--PM., 


Y(co)  = 


jco/3  ,  W  _  W' 

e  ;  co  1  —  —  —  co  —  co  1  +  — 

’0  2  0  2 


(1-12) 


0  elsewhere 

Our  desired  time  function  y(t)  (i.e.  ,  the  correlation  function)  is  twice  the  real 
part  of  the  inverse  Fourier  transform  of  Y(co);  i.  e.  , 


yd)  =  2  Re  L  J 


Y(co)  dco 


Substituting  (1-12),  we  have 

±r° 


,  W’ 

+  2  id  ^ 

y(t)  =  2Re  ^  J  w,  k  e  e 

"o*  T 


t-  HE 

c 


-  T  +  P 


Let 


f  =  t  -  —  -  T 
c 


We  can  integrate  y(t)  directly,  obtaining 


k  id 
y(t)  =  -  Re  e 


is 

j  [t  +  p]  J 


CO 


W' 

o  "  T 


W' 

co'  -  — 

0  2 


(1-13) 


dco  (1-14) 


(1-15) 


W' 


W' 


k  id  j[t’  +  /3]w(;  ej[t'+/3]  2  _e-j[t' +  /?1  2 

—  Re  eJ  e  \ - rrr, - vi - 

t r  \  j[t'  +p] 


(1-16) 


29 


from  which  we  finally  obtain  the  desired  expression 


y(t) 


/  sin 

.(f  +  (3  )  2  

W'~ 
<t'  +0)  ^ 

|cos  [(t'  +  /3)  +  6] 


(1-17) 


30 


APPENDIX  H 


RELATIONSHIP  BETWEEN  TRANSMITTED  AND  RECEIVED  RADAR  SIGNALS 

For  simplicity,  we  shall  consider  a  point  target  with  constant  cross  sec¬ 
tion.  Suppose  the  transmitter  sends  an  impulse  in  the  direction  of  the  target 
at  time  t  =  0.  The  transmitted  signal  is  given  by 

(t)  =  5(t)  ,  (II-l) 

where  6(t)  is  the  delta  function.  Assume  free-space  propagation.  Let  the 
target  be  at  a  distance 

Rj  =  (II— 2) 

when  the  impulse  hits  it,  c  being  the  velocity  of  light.  The  returned  echo  will 
then  also  be  an  impulse  which  will  strike  the  antenna  at  time  2t^  ,  i.  e. , 

xi  (t)  =  k  8  (t  -  2^)  =  k  6  ^t  -  20  ,  (II— 3) 

where  the  constant  k  depends  on  the  radar  cross  section  of  the  target.  Let 
another  impulse  be  transmitted  a  short  time  t  later,  so  that 

s2  (t)  =  8  (t  -  r)  .  (II— 4) 

For  t^  >  t,  this  impulse  will  be  at  a  position  in  space  equal  to  R^  -  cr  when  the 
first  impulse  strikes  the  target.  Let  the  target  have  radial  velocity  V  and 
radial  acceleration  A  at  time  t  .  For  V  «  c,  this  second  impulse  will  strike 
the  target  at  approximately  an  interval  t  after  the  first  impulse  did.  We  have 
shown  in  the  text  that  for  the  targets  and  the  radar  parameters  of  interest  to  us, 
the  acceleration  has  negligible  effect  (over  the  illumination  time)  on  the  velocity 
and  range.  Here  we  assume  t  to  be  shorter  than  the  total  illumination  time. 


31 


Thus,  the  relative  velocity  between  the  second  impulse  and  the  target  at 


time  t 
AR  = 


is  AV  =  c  -  V.  The  distance  between  the  target  and  this  impulse  is 
cr.  The  time  required  for  this  impulse  to  "catch  up"  with  the  target  is 


At  = 


AR 

SV 


CT 

C-V 


(II-5) 


The  target  will  be  at  range 

R„  =  R,  +  VAt 
2  1 


R. 


+ 


Vc 

c-V 


T 


(II-6) 


when  this  impulse  strikes  it.  The  echo  returned  from  this  impulse  is  then 
given  by 

x2<*>  - k  5  (»-r-  ^r)=  -  0*  T)  •  <«-7> 

/  2V-\ 

Thus,  the  two  returned  pulses  are  a  time  t  1 1  +  — -^1  apart,  while  the  two 
transmitted  pulses  are  only  a  time  t  apart.  A  time-dilation  has  taken  place. 


Since  we  can  think  of  our  actual  transmitted  signal  as  consisting  of  a 
sequence  of  impulses,  we  see  that  the  target  may  be  regarded  as  a  time- 
varying  ideal  delay  line  in  cascade  with  a  time-invariant  attenuator.  This  is 
illustrated  in  Fig.  II— 1 . 


W(t,  t)  k 


Transmitted 

s(t)  - 

Time- 

varying 

Attenuator 

Received 

Signal 

Delay 

Line 

Signal 

Fig.  11-1.  Ideal  Delay  Line 

The  delay  line  is  a  linear,  time-varying  filter.  The  output  y(t)  of  such  a  filter 
is  related  to  the  input  x(t)  by* 


*cf.  Laning  and  Battin  "Random  Processes  in  Automatic  Control,"  McGraw- 
Hill,  1956,  p.  226. 

32 


s(  t)  W  (t,  t)  c!t  , 


(H— 8) 


f(t) 


where  W  (t,  t)  is  the  time-varying  impulse  response.  We  may  think  of  it  as  a 
curve  in  the  t  -  t  plane  which,  for  the  case  of  constant  velocity,  is  a  straight 
line,  as  indicated  in  Fig.  EL— 2. 


Fig.  11-2.  Constant  Velocity  Case 


t 


5 


Here  t  is  the  "  input  time"  and  t  is  the  "response  time. "  From  Eq.  (II— 7) ,  it 
is  apparent  that  the  equation  of  this  line  is  given  by 


W(t,  t) 

For  convenience  let  R  = 
that  (H-9)  becomes 


Recognizing  that  W(t,  t)  = 


(II— 9) 

0  for  t  <  r,  we  find 


33 


But,  1  + 


2V  _  c+V 
c-V  “  c-V 


Letting 


.  2R  .  c+V 

b  =  —  and  a  =  — — 

c  c-V 


we  obtain 

f(t)  =  s 

The  Fourier  transform  of  f(t)  is  given  by 


(¥) 


F(w) 


■  s  r •-m « 

—  oo 

-  i.  f"  8  tS  e‘iwt 
27 r  J  a 


dt 


.  ±  r 

2tt  J 


8(u)  e'ia,b  e-lu(au) 


du 


=  a  e 


-jcob 


-  r 

27 r  J 


s(u)  e 


-j(acu)u 


du 


(H-ll) 


(11-12) 


(11-13) 


We  recognize  the  last  line  of  (11-13)  as  equal  to 

* 

F(to)  =  e_;iajb  a  S(aco)  .  (H-14) 

Thus,  the  signal  returned  from  the  target  is  related  to  the  transmitted  signal 
s(t)  by 


x(t)  =  k  s 


(v). 


(11-15) 


and  its  Fourier  transform  X(co)  is  related  to  S(a>)  by 

X(w)  =  k  a  e~^wb  S(aw)  .  (11-16) 

Equations  (11-15)  and  (11-16)  are  physically  plausible,  as  we  shall  now  show. 
Suppose  we  send  two  impulses,  one  at  t  =  0  and  one  at  t  =  r;  i.  e. ,  let 

s(t)  =  6(t)  +  6(t  -  t)  .  (H-17) 


34 


Then 


x(t)  =  k 


[•(?)•*(?--)] 
-■f(- ?)•*(■-?- 


c+V 

c-V 


(11-18) 


Thus,  the  received  signal  is  attenuated,  delayed  and  stretched  (since  the  target 
is  receding). 

Next  consider  a  C\V  signal;  i.  e.  ,  let 


S(co)  =  6(cd  -  cr^)  +  <5(co  +  cr^) 
from  (II— 1G)  we  obtain,  for  the  received  signal, 

X(oj)  =  ka  e  [6  (acd  -  cd^)  +  6  (acd  +  cd^)] 


(11-19) 


=  ka  e 


-jedb 


Cd  . 


0 


Cd 


0 


6  I  co  —  —  I  +  6  I  cd  +  — 


Cd, 


The  new  carrier  frequency  is  therefore  equal  to  —  ,  where  a  = 


Now, 


c+V 

c-V 


1  + 


2V 

c-V 


(11-20) 

c+V 
c^V  ’ 

(11-21) 


for  V  <  —  this  can  be  expanded  in  a  power  series,  i.e.  , 

tl 


(11-22) 


2  V 


For  V  <  <  c  this  becomes  approximately  equal  to  1  -  —  ,  so  that  the  new 
frequency  is 


"o  /.  2V 

T  o  y  T 


2V  \ 


(11-23) 


which  we  recognize  as  the  familiar  "doppler-shifted"  frequency. 


35 


APPENDIX  III 


ANALYSIS  OF  A  TIME-DILATED  SIGNAL  AND 
APPROXIMATIONS  TO  THE  FRESNEL  INTEGRAL  EXPRESSIONS 

We  start  with  Eq.  (37  of  the  text,  which  reads 

—  2 

ka  i a  C  2  n  +  C4  n  1 

y(t)  =  2  Re  -  e]  e  dS2  . 

2n  J  W" 

~2 

The  exponent  inside  the  integral  can  be  written  as 

j[o3n*c4a2]  =  jc4  | n + 
so  that  (HI-1)  becomes 


c. 


j  Iff  - 


kn. 

y(t)  =  2  Re,  —  e 


w 

4c- 1  ~2 


I 


W" 

2 


ln+  2-r 


dS2  . 


Let 


3  1  IT  a 

c,  (H+  - — 

4  *  2  C4 


.  -  h 

7 r  I  2c 


dn  =  JJL '  da 
V2C4 


(HI-1) 


(III- 2) 


(HI— 3) 


(HI— 4) 


37 


Thus,  we  obtain 


y(t) 


j 


ka 

=  —  Re  e 

7T 


2 

a 


This  can  be  written  as 


y(t) 


(m-5) 


(m-6) 


where  Z(u  )  and  Z(u  )  are  the  complex  Fresnel  integrals,  with  u  and  u  given 

1  C*  X  Li 

by 


(HI-7) 


We  must  take  the  real  part  of  the  product  of  several  complex  quantities.  For 
convenience,  let  us  list  all  the  other  symbols  appearing  in  (HI-6)  in  terms  of 
the  more  elementary  parameters;  i.  e. , 


38 


c+V  2  V  2c 

a  ~  c-V  ’  ^  _  c-V  ’  X  ""  c-V  ’ 


a  =  (t'  +yci)  co”  +yXc2  (coJ')  , 


C3  =  v  +Tci  +  2  7^c2  uq  > 


C4  =  yXC2  ’ 


1  L  _  w\ 

W  1  0  2  y 

T  (  W\ 
W  (^0  +  2  j 


for  FM  up  , 


_T_ 

2W 


for  FM  down  , 

for  FM  up  , 


2W 


for  FM  down  , 


co" 

0 

W" 

2 


/  c  \  _  W  /  V  \ 

^0  l  c+vJ  2  (c+v) 

(cTv)  -  “o  (cTv) 


w 

2 


2R 


t'  =  t - T 


( III—  8) 


39 


We  note  that 


is  imaginary  for  FM  up  and  real  for  FM  down;  i.  e. , 


for  FM  up 


for  FM  down 


(HI- 9) 


where  j  = 


Now  let 


ka 

2 

c 


(t')2  +  2  y\c  V 

4  yXc2 


2y\c2 


2y\c( 


(III- 10) 


Using  these  symbols,  we  have,  for  FM  up, 

{cos  x  +  j  sin  x  i 

- - - -  [Z(jx3)-Z(jx4)]J 


(III-ll) 


40 


Similarly,  we  obtain,  for  FM  down, 


y(t)  =  x  Re 


cos  +  j  sin  x 


0  ™  )' 


-  [Z(x3)  -  Z(x4)]|  . 


(Ill- 12) 


All  the  x's  are  now  real  quantities.  The  complex  Fresnel  integral  Z  (x)  can  be 
expressed  (for  real  x)  as  follows: 


Z  (x)  =  C  (x)  +  j  S  (x) 
Z  ( jx)  =  S  (x)  +  j  C  (x) 


(III— 13) 


where 


f\X  2  rx  2 

(o'  [a 

C  (x)  =  \  cos  7r  —  do1  ;  S  (x)  =  \  sin  7 r  —  do:  . 

^  n  *  2 


(m-i4) 


These  are  tabulated  functions.  Using  (III— 13) ,  we  obtain 
x 

y(t)  =  —  ][c(x3)  "  c(x4)l  cos  X1  +  [  s(x3)  "  s(x4)l  sin  X1  [  (HI— 15) 
2 

for  FM  up,  and 

x 

y(t)  =  —  }tC(x3)  -  C(x4)l  cos  Xx  "  tS(x3)  "  S<x4)l  sin  xj  }  •  (HI- 16 ) 
2 

for  FM  down.  (Note  that  x  ,  x  and  x  have  different  values  for  FM  up  and  FM 

1  u  Tr 

down. )  We  can  express  y(t)  in  the  following  form: 


y(t)  =  (  3T  J  R  cos  (xx  -  p)  . 


(Ill— 1 7) 


41 


where 


R  = 


[C(Xg)  -  C(X4)]2  +  [S(Xg)  -  Stx^]2 


(m-i8) 


for  both  FM  up  and  FM  down,  and 


i  rs<*3)-s(x4) 


S(x3)  -  S(x4) 


for  FM  up  , 


for  FM  down  . 


(Ill- 19) 


These  expressions  are  exact  (within  the  limits  of  our  approximations),  and 
can  be  evaluated  with  the  aid  of  available  tables  of  Fresnel  integrals.  However, 
we  shall  try  to  obtain  approximate  closed  form  representations  for  R  and  0  in 
the  vicinity  of  the  peak  of  the  envelope. 

Let  us  first  consider  the  envelope  R.  We  wish  to  determine  its  maximum. 
From  the  analysis  leading  to  Eq.  (16)  in  the  text,  we  expect  this  maximum  to 
occur  near  c^  =  0.  Let  us  postulate  that  the  peak  actually  occurs  at  that  point. 

A  necessary  condition  for  this  to  be  so  is  that 


when 


c3  =  0  =  t'+yc1  +  2y\  c2  a," 


(HI-20) 


t-  T  -T+yc1  +  2yXc2  co- 


42 


Let  us  check  whether  this  condition  is  satisfied.  From  (III— 18)  we  have, 


W  =  -k  {2[c(x3>  -  C(x4)1  S  [c(x3>  -  C(X4)I 
+  2  [S(x3)  -  S(x4)]  j  (S(x3)-S(x4)]|  . 

Equation  (III— 2 1 )  will  be  equal  to  zero  if  both 

i  [C(x3)  -  C(x4)]  =  0,  and 

Tt  lS(X3>  -  S(x4)1  =°' 


(III— 2 1 ) 


(III— 22) 


provided  that  R  does  not  equal  zero  at  the  same  time.  The  Fresnel  integrals  in 
(III— 22 )  are  all  of  the  form 


f  (a)  da  =  F  [x  (t)]  -  F  (0)  , 


(III— 23 ) 


where  F  is  the  indefinite  integral  of  f.  Differentiating  (III— 23)  with  respect  to  t, 
we  have 

f  (a)  da  =  f[x(t)]  .  (Ill— 24) 

In  our  case,  x(t)  is  equal  to  either  x  or  x  .  Referring  to  (III— 10),  we  see  that 

(III— 25) 


x„  = 


kl  C3  +  k2 


x.  = 


k,  C  -  k 

13  2 


43 


where  k  and  k  are  constants,  and  c  is  of  the  form  c  =  t  +  another  constant. 
12  3  3 

Hence, 


dx  dx , 

3  4 

dT  =  kl  dT 


(IH-26) 


Thus,  the  — in  (III-24)  is  equal  to  k, ,  a  constant.  Now,  the  function  f  (a)  in 
dt 


’  dt  '  7  ?  v  i  /  2v 

our  case  is  either  equal  to  cos  I-  a  J  or  sin  l  -  a  J  .  Combining 


all  this 


information.  (Ill— 22)  becomes 


dT  lC<X3>-C(X4)]  -  k! 


dt  tS(X3-S<X4)! 


[cos  (*  X2^ 

!-  :\1 

1) 

C0S  (2  x3j 

I  -  cos  Ah 

1 

[sin  (1  4) 

- si"  (1  <)]  j 

(111-27) 


We  wish  to  find  out  whether  both  these  equations  in  (III— 2 7)  are  identically  zero 
when  c  =  0.  Substituting  (III— 25)  into  (III— 27),  we  obtain  first 

*3 


dt  1C<X3>-C(V!  =  k 


1  [C°S  T  (kl  °3  +  k2  +  “l  k2  C3) 

—  (k2  c2  +  k^  -  2k  k  A] 

2  \  1  3  2  1  2  3/J 


-  cos 


(III— 28 ) 


Let 


y  =  —  (2k  k  c  ) 
J  2  '  1  2  37 


(III— 29) 


44 


Making  use  of  the  trig-identity 

cos  (x  ±  y)  =  cos  x  cos  y  +  sin  x  sin  y  , 

Eq.  (Ill— 28)  becomes 

s  [C(V-C<X4>]  =ki[-2 

Now,  since  y  contains  c  as  a  factor,  we 

o 

Hence,  sin  y  =  0  when  c  =  0,  and 

O 

^  [C(xg)  -  C(x4)l  =  0.  when  c3  =  0  .  (III-31) 

Similarly,  using  (III— 29)  again  and  the  trig-identity  sin  (x  ±  y)  =  sin  x  cos  y 
±  cos  x  sin  y,  we  obtain,  for  the  second  equation  in  (III— 2 7) , 

^  [S(x3)  -  S(x4)]  =  k1  [2  cos  x  sin  y]  ,  (HI-32) 

which  again  contains  sin  y  as  a  factor,  yielding 

[S  (x  )  -  S  (x  )]  =  0,  when  c  =  0  .  (HI-33) 

dt  3  4  3 


sin  x  sin  y]  .  (Ill— 30) 

see  that  y  =  0  when  c  =  0. 


Thus  we  see  that  (III— 22)  is  satisfied.  It  can  readliy  be  checked  that  R  is  not 


identically  zero  when  c  equals  zero.  Hence,  we  conclude  that 

O 


dR  , 

— —  =  0  when  c 
dt  f 


=  0  ^i.  e.  , 


2R 


when  t  =  —  +  T-  yCj  -  2  y  X  c^  c u"  J .  (Ill— 34 ) 


') 


Let  us,  therefore,  think  of  c  as  a  shifted  time  variable;  i.  e.  ,  let 

O 


t"  =  c  =  t  - 


2R 


C  -T+yc  -2yXc 


(III— 35 ) 


45 


Note  that  Eq.  (Ill— 34)  is  satisfied  regardless  of  the  target  velocity  V.  When 

/  2  V ' 

V  =  0,  y  =  0  fe 


<) 


^eince  y  =  — —  |  ,  which  serves  as  a  check. 

Let  us,  therefore,  accept  our  postulate  as  being  true;  namely,  that  the 
envelope  R  has  its  peak  when  t"  =  0. 

When  t"  =  c  =  0,  the  arguments  x  and  x  of  the  Fresnal  integrals  are 


(from  III— 10)  equal  to 


From  (HI-8)  we  see  that 


W" 

2 


_  /y\T  /W  "\ 

V  7rw  y  2  /  * 

/yXT  (  w»\ 

V  7TW  \  2  f 


which,  for  V  <  <  c,  reduces  approximately  to 


Similarly,  we  have 


so 


A 

=  V7 


W" 

w 

2 

«  -  -c 

Y 

2V 

c-V 

2V 

c-V 

|x4l 

when  c  = 
3 

/4vt 

C7T  W 

’w 

—  —  oo 

2  0 

v 

0  c 


2V 

c 


_ 

TW  /2V\  f  /~T~ 
7 r  \  c  /  7rW 


(HI-36) 


(III— 37) 


46 


Letting  B  =  —  and  f 

B  2tt  0 


—  ,  this  becomes 

27 r 


(III— 38) 


For  the  particular  radar  of  interest,  is  approximately  1300  megacycles,  and 
there  are  two  pulse-compression  systems  to  be  considered: 

System  No.  1: 

T  =  1  millisecond,  B  =  1  megacycle  ; 

System  No.  2: 

T  =  2  milliseconds,  B  =  5  megacycles  . 

4 

The  range  of  radial  velocities  we  consider  here  are  0  ^  V  ^10  m/sec.  Let  us 
check  if  x  has  a  maximum  or  a  minimum  in  this  region;  i.  e.  ,  using  (III— 38) , 
we  solve  for 


dx 

dV 


0 


d_ 

dV 


[a  V1/2  -  b  V3/2' 


— a .  ,  _  5  b  Vv"  =  0. 
2  vH7  2 


For  V  /  0,  we  obtain  a  -  3b V 


V  = 


3b 


which  simplifies  to 


V 


Be 

6fo 


(in-39) 


47 


4 

For  System  No.  1,  this  corresponds  to  V  =  3.85  x  10  m/sec,  which  fall6 
outside  our  region  of  interest.  For  System  No.  2,  the  value  is  even  larger. 

4 

Thus,  since  there  are  no  local  maxima  or  minima  between  0  ^  V  ^  10  m/sec, 

we  can  compute  the  extreme  values  of  x  by  taking  the  end  points.  For  V  =  0, 

4 

we  simply  get  zero.  When  V  =  10  m/sec,  we  obtain 


x  w  0.  237  for  System  No.  1  ; 
x  «  0.  745  for  System  No.  2  . 


(IH-40) 


Asymptotic  series  expansions  of  the  Fresnel  integrals  C(x)  and  S(x)  exist 
both  in  ascending  powers  of  x  (for  x  <  1)  and  in  descending  powers  of  s  (for  x  >  1). 
The  former  are  useful  for  x  <  .<  1  and  the  latter  for  x  >  >  1,  since  a  few  terms 
of  the  series  then  give  us  a  good  approximation. 

From  the  above  estimates,  we  see  that  at  the  peak  of  the  envelope  the 

4 

argument  of  the  Fresnel  integrals  is  close  to  one  when  V  =  10  m/sec.  Thus, 
asymptotic  expansions  appear  to  be  unsuitable  in  the  immediate  vicinity  of  the 
peak  for  large  velocities. 

We  could,  of  course,  consider  smaller  velocities.  Suppose  we  require 

that 

x  =  0.  01  <  <  1  , 

(where  x  =  x  =  |x  |  when  c  =  0). 

O  4  O 

Using  Eq.  (Ill— 10),  we  find  that  this  corresponds  to  a  radial  velocity  of 
approximately 

V  »  10.5  m/sec  for  System  No.  1,  and 

V  «  1.5  m/sec  for  System  No.  2  . 


48 


However,  for  such  low  velocities,  the  analysis  of  a  simple  Doppler- 
shifted  signal  (as  done  in  Appendix  I)  is  expected  to  be  quite  adequate.  It  does 
not  seem  worthwhile  to  go  through  the  necessary  approximations. 

Let  us  consider  whether  we  can  infer  anything  about  the  nature  of  y  (t)  as 
we  move  away  from  the  peak  of  the  envelope.  Going  back  to  (III-10)  and  substi¬ 
tuting  t"  for  c  ,  we  have 

O 


(M.  ATT 

’  tM  u 

(  — 1 

|x4,  V,w 

2y\c  t  ! 

f  2  J 

(III— 41 ) 


(Here  we  imply  that  +  goes  with  x  ,  -  with  x  ) 

O  4 

Let  us  rewrite  (III— 41 )  as 


(III— 42) 


We  have  previously  (III— 40)  found  that  the  product  k  k  is  of  the  order  one  when 

1  O 

t"  =  0.  We  would  now  like  to  know  how  large  t"  has  to  be  such  that  |x  |  or 

O 

lx  I  are  of  the  order  100  (so  that  we  might  use  asymptotic  expansions  for 

i  4> 

large  x). 

We  let 


t" 

k„ 


=  100  k 


(III— 43) 


Substituting  for  k  and  k  ,  we  have 
Z  u 


|t"J  =  100 

This  is  approximately 

|t"|  a  200 


[?  (*)  -  (*)] 


l2yxcl 


(III— 44) 


[2  0  cJ[Vc/wJ 


49 


Using  V  =  10  m/sec,  we  have 


1 1"  |  «  6psec  for  System  No.  1; 

1 1"  |  «  12psec  for  System  No.  2. 


(HI-45) 


The  following  asymptotic  approximations*  are  expected  to  be  quite  good  for 
large  values  of  x  (say,  x  ^  100). 

2  V 

1  sin  7rx  /2  I 

2  +  7rx  ’  r 

>  (m-46) 

2  i 

1  COS  7TX  /2  | 

2  7TX  '  f 

2  . 

Letting  o:  =  n  x  /2  and  using  the  approximations  in  (HI-46),  we  have 


C(x) 

S(x) 


C(x3)  -  C(x4) 
S  (xg)  -  C  (x4) 


(III— 47) 


(Although  we  are  using  equal  signs,  it  is  understood  that  these  are  only  approxi¬ 
mations  . ) 

Substituting  (III— 47)  into  (III— 1 8) ,  we  have 


R 


3  4 


x„  x, 
3  4 


(sin  a  sin  a  +  cos  a  cos  a. 
3  4  3  4 


>] 


cos  (a 

O 


(III— 48) 


*c.f.  Watson  "Theory  of  Bessel  Functions,"  p.  545. 


50 


Let  us  rewrite  (III— 42)  as 


where 


|A|  <<|a| 


Thus,  we  have 


X3  A  +  A  A 


i  ^1  -  — J  (ignoring  higher  order  terms). 


Similarly, 


^  ^1  +  (ignoring  higher  order  terms), 


(in-49) 


Furthermore,  we  have 

2(i)(-|)s  (1  +  f)= 

2  2 

-  =  —  (again,  ignoring  higher  order  terms). 

X„  X  .Li 

3  e  A 


51 


Similarly, 


7T  /  2  2  \ 

a3  "  "4  "  2  V  3  X4  ) 


=  J  [A2  +  2  A  A  +  A2  -  (A2  -  2A  A  +  A2)] 

Li 


-  3  [4  A  A] 


=  277  A  A  . 

Substituting  the  above  approximations  into  (III-48),  we  have 

1/2 


B  -  1  fj.  ± 

"  ;  [a  a2 


cos  2  7r  A  A 


] 


=  [1  -  COS  2  7T  A  A]  1/2 

7T  A 


But  1  -  cos  2x  =  2  sin  x,  so  that  (III— 50)  becomes 

R  »  2A 

Referring  back  to  (III— 1 7) ,  the  whole  expression  for  the  envelope  E  is 


(sin  7 r  A  A\ 
7T  A  A  J 


(ni-50) 


(HI-51) 


E  =  —  2  A  —  A-A 

X  7T  A  A 

Li 


Substituting  our  elementary  parameters  back,  we  finally  obtain 


E  =  a 


(3)  ft) 


FW)J 


(HI-52) 


52 


This  is  a  good  approximation  for 


If,  ,100  (sp)  ,27xc2|  -  100  E:  (^)(^)  i  -  ,00  (2)  T 
Let  us  now  look  at  the  phase  angle  0  in  the  same  region.  We  had 


0  =  ±  tan 


1 


S  (Xg)  -  S  (X4) 

C(x3)  -  C(x4) 


](■* 


for  FM  up  ) 
for  FM  down  / 


Let 


Using  (III— 47) ,  we  have 


ip  =  ±  tan  0  . 


cos  a,  cos  « 


4>  = 


sin  a  sin  a. 


(Ill— 53) 


(III— 54) 


We  have,  as  before, 


*(i)  *  1 


With  A  <<  A,  let  us  ignore  all  terms  except  the  first.  Then  we  get 


ili  = 


cos  cv  -  cos  a 
4 _ 3_ 

sin  a  -  sin  a 
3  4 


(III— 55) 


53 


But 


Letting 


we  have 


7 r  A  2 

“3  =  2  (A  +  A) 

7T  .  2 

“4  =  2  (A  ■  A) 


7T  2  2 

“  (A  +  A  +  2  A  A)  ; 

£  (A2  +  A2  -  2  A  A)  . 


X 


y 


7T 


2 

7T 

2 


2  2 
(A  +  A  )  , 

(2  A  A)  , 


cos  (x  -  y)  -  cos  (x  +  y) 
sin  (x  +  y)  -  sin  (x  -  y) 


(III-56) 


Using  some  trig- identities,  we  obtain 


2  sin  x  sin  y 
2  cos  x  sin  y 


tan  x  . 


But  if>  =  ±  tan  0,  so  that  0  =  ±  x.  Substituting  for  x,  we  finally  have 


0  =  ± 


y  X  T 
2W 


N’ 


+  for  FM  up 
-  for  FM  down 


r 


(in-58) 


The  carrier  term  is  (from  III— 1 7)  equal  to  cos  (x^  -  0).  Letting  9  =  x^  -  0 
we  obtain,  after  a  fair  amount  of  algebra, 


e 


"o  »t»  + 


yXT 

2W 


f 


*  (c”) 


T 


(W”)21 

“J 


(III— 59) 


54 


where  the  upper  signs  are  for  FM  up,  and  the  lower  signs  for  FM  down.  Now, 
since  W"  <  <  u^'  by  about  three  orders  of  magnitude,  suppose  we  ignore  the 
second  term  in  the  brackets.  Substituting  for  y  and  X  ,  we  have 


6  ~  (jj. 


b . .  (S)  (S)  (A  <»;•' 


+  for  FM  up  | 
-  for  FM  down  i 


(III— 60) 


Again,  this  is  for 


t"  > 100 


55 


Security  Classification 


DOCUMENT  CONTROL  DATA  •  R&D 

( Security  c/eaa///caHon  0/  title  body  0/  ebatracf  end  indexing  ennotetion  muet  be  entered  when  the  overell  report  ie  c leaeilied) 

1  QRIGINATIN  G  ACTIVITY  ( Corporal .  tulhor) 

MITRE  Corporation 

Bedford,  Mass. 

2*.  REPORT  SECURITY  CLASSIFICATION 

UNCLASSIFIED 

26  GROUP 

n/a 

3  REPORT  TITLE 

Phase  and  Envelope  of  Linear  FM  Pulse-Compression  Signals  From  High-Velocity 
Targets 

4  DESCRIPTIVE  NOTES  (Type  0/  report  en<f  induelve  detee) 

5  AUTHORS  Ctaaf  n«ma,  ttret  neme,  inlttel) 

Ueberschaer,  M.H. 

6.  REPO  RT  DATE 

Nov  64 

7.  TOTAL  NO.  OF  PAGE*  7b.  NO  O P  REF* 

SB  0 

8«  CONTRACT  OR  GRANT  NO. 

AF19(628)2390 

b.  PROJECT  NO. 

750 

C- 

d 

•  a.  ORIGINATOR'*  REPORT  NUMBCRfSj 

TM-03916 

f  b.  OTHER  REPORT  NOfSj  (A  ny  other  numbere  thet  mey  be  eeeigned 
thle  report) 

ESD-TDR-64-128 

10.  A  VA  IL  ABILITY/LIMITATION  NOTICES 

Qualified  Requesters  May  Obtain  From  DDC. 

Aval  From  OTS. 

11  SUPPLEMENTARY  NOTES 

12.  SPONSORING  MILITARY  ACTIVITY 

Directorate  of  Radar  and  Optics,  ESD, 

L.G.  Hanscom  Field,  Bedford,  Mass. 

13.  ABSTRACT 


Equations  for  the  phase  and  envelope  of  the  output  signal  from  a 
linear  filter,  matched  to  the  transmitted  signal,  are  derived.  The  trans- 
mitted  signal  is  assumed  to  have  a  flat  band-limited  amplitude  spectrum 
and  a  linear  group  delay.  The  input  to  the  "matched  filter"  is  the  radar 
echo  returned  from  a  moving  target  whose  velocity  is  essentially  constant 
during  the  illumination  time.  It  is  shown  that  the  returned  signal  is  related 
to  the  transmitted  signal  by  a  time  dilation.  The  resulting  expressions  for  the 
phase  and  envelope  are  functions  which  involve  Fresnel  integrals.  Approximations 
for  these  expressions  are  worked  out.  They  are  shown  to  be  similiar  in  form 
to  those  which  are  obtained  when  the  returned  signal  is  assumed  to  be  trans- 
mitted  signal  by  a  Doppler  shift. 


DD  .MM.  1473 


Security  Classification 


Security  Classification 


14. 


KEY  WORDS 


LINK  A 


LINK  B 


.  ROLf 


WT 


LINK  C 


Raoar 

Targets 

Pulse  Compression 
Linear  FM 


INSTRUCTIONS 


1.  ORIGINATING  ACTIVITY:  Enter  the  name  and  addreaa 
of  the  contractor,  subcontractor,  grantee,  Department  of  De¬ 
fense  activity  or  other  organization  (corporate  author)  issuing 
the  report. 

2a.  REPORT  SECURITY  CLASSIFICATION;  Enter  the  over- 
aii  security  ciaasification  of  the  report.  Indicate  whether 
"Restricted  Data"  ia  included.  Marking  is  to  be  in  accord¬ 
ance  with  appropriate  security  regulations. 

26.  GROUP:  Automatic  downgrading  ia  specified  in  DoD  Di¬ 
rective  5200.10  and  Armed  Forces  Industrial  Manuai.  Enter 
the  group  number.  Also,  when  applicable,  show  that  optional 
markings  have  been  used  for  Group  3  and  Group  4  aa  author¬ 
ized. 

3.  REPORT  TITLE:  Enter  the  complete  report  title  In  all 
capital  letters*  Titles  in  aii  casea  should  be  unclassified* 

If  a  meaningful  title  cannot  be  selected  without  ciaasifica¬ 
tion,  show  title  classification  in  all  capitals  in  parentheaia 
immediately  following  the  title. 

4.  DESCRIPTIVE  NOTES:  If  appropriate,  enter  the  type  of 
report,  e.g. ,  interim,  progress,  summary,  annual,  or  final. 

Give  the  inclusive  dates  when  a  specific  reporting  period  ia 
covered. 

5.  AUTHOR(S):  Enter  the  name(s)  of  authors)  aa  shown  on 
or  in  the  report.  Entei  last  name,  first  name,  middle  initial. 

If  military,  show  rank  and  branch  of  service.  The  name  of 
the  principal  author  is  an  absolute  minimum  requirement. 

6.  REPORT  DATE:  Enter  the  date  of  the  report  as  day, 
month,  year;  or  month,  year.  If  more  than  one  date  appears 
on  the  report,  uae  date  of  publication. 

la.  TOTAL  NUMBER  OF  PAGES:  The  total  page  count 
should  follow  normal  pagination  procedures,  i.  e. ,  enter  the 
number  of  pages  containing  information. 

76.  NUMBER  OF  REFERENCES:  Enter  the  total  number  of 
references  cited  in  the  report. 

8a.  CONTRACT  OR  GRANT  NUMBER:  If  appropriate,  enter 
the  applicable  number  of  the  contract  or  grant  under  which 
the  report  was  written. 

86,  8c,  &  8d.  PROJECT  NUMBER:  Enter  the  appropriate 
military  department  identification,  such  as  project  number, 
subproject  number,  system  numbers,  task  number,  etc. 

9a.  ORIGINATOR'S  REPORT  NUMBER(S):  Enter  the  offi¬ 
cial  report  number  by  which  the  document  will  be  identified 
and  controlled  by  the  originating  activity.  This  number  must 
be  unique  to  this  report. 

96.  OTHER  REPORT  NUMBER(S):  If  the  report  has  been 
assigned  any  other  report  numbers  (either  by  the  originator 
or  by  the  sponsor),  also  enter  this  number(s). 

10.  AVAILABILITY/LIMITATION  NOTICES:  Enter  any  lim¬ 
itations  on  further  dissemination  of  the  report,  other  than  those 


imposed  by  aecurity  claaaification,  uaing  standard  statements 
such  as: 

(1)  "Qualified  requeatera  may  obtain  copiea  of  thia 
report  from  DDC." 

(2)  "Foreign  announcement  and  diaseminatlon  of  this 
report  by  DDC  is  not  authorized." 

(3)  "U.  S.  Government  agenciea  may  obtain  copies  of 
this  report  directly  from  DDC.  Other  qualified  DDC 
usera  ahall  request  through 

if 


(4)  MU.  S.  military  agenciea  may  obtain  copiea  of  thia 
report  directly  from  DDC  Other  qualified  uaers 
ahaii  request  through 

ii 


(5)  "All  distribution  of  thia  report  is  controlled.  Qual¬ 
ified  DDC  users  ahaii  request  through 

ii 


If  the  report  has  been  furnished  tc  the  Office  of  Technical 
Services,  Department  of  Commerce,  for  sale  to  the  public,  indi¬ 
cate  this  fact  and  enter  the  price,  if  known. 

11.  SUPPLEMENTARY  NOTES:  Use  for  additional  explana¬ 
tory  notes. 

12.  SPONSORING  MILITARY  ACTIVITY:  Enter  the  name  of 
the  departmental  project  office  or  laboratory  sponsoring  (pay¬ 
ing  for)  the  research  and  development.  Include  address. 

13.  ABSTRACT:  Enter  an  abstract  giving  a  brief  and  factual 
summary  of  the  document  indicative  of  the  report,  even  though 
it  may  also  appear  elsewhere  in  the  body  of  the  technical  re¬ 
port.  If  additional  space  is  required,  a  continuation  sheet  shall 
be  attached. 

It  is  highly  desirable  that  the  abstract  of  classified  reports 
be  unclassified.  Each  paragraph  of  the  abstract  shall  end  with 
an  indication  of  the  military  security  classification  of  the  in¬ 
formation  in  the  paragraph,  represented  as  (TS).  (S),  (C),  or  (V) 

There  is  no  limitation  cn  the  length  of  the  abstract.  How¬ 
ever,  the  suggested  length  is  from  150  to  225  words. 

14.  KEY  WORDS:  Key  words  are  technically  meaningful  terms 
or  short  phrases  that  characterize  a  report  and  may  be  used  as 
index  entries  for  cataloging  the  report.  Key  words  must  be 
selected  so  that  no  security  classification  is  required.  Identi¬ 
fiers,  such  as  equipment  model  designation,  trade  name,  military 
project  code  name,  geographic  location,  may  be  used  as  key 
words  but  will  be  followed  by  an  indication  of  technical  con¬ 
text.  The  assignment  of  links,  rules,  and  weights  is  optional 


Security  Classification