DIGITAL COMMUNICATION (18EC61)

MODULE- 1
BANDASS SIGNAL TO EQUIVALENT LOW PASS

HILBERT TRANSFORM
iovolvu the
Fousieu, haplace and z-trarmo
change o Time - Dpmain Teprerentation
signal tB the Frequency-domain sepruentation
the signal
The Tesultng siqnalk ame eauivalent repr eaent
- aioo the Same sinal in terma oy.Frcquen ys
S- doain and Z-dom ain trano n doenot
The HiLbeut
’ In. Contrast, otheu
involve change o dormain, Unltke many
ís not an eauivalent eprue.
’ Hilbert TTannom frey a
the aiol, rathes it s ënt
-ntation but n the Sane Domin,
disent"sipal evaluating the
is Usel T
Fowr ler TocnsHomSignalIt "Providea mathema
-
Teaueny content
Vtical batis Aororo a
analyzin
analyzin and ddtgning
daigning aton e
'selective lterS jor th Sepor
Frequency - Content
Signal on the bauisbasis o theIr requeny
HiLbevt l3ayorm met hod o Separating
on Phase Selecivity,
signau bas ed stonal.
between pe tinent
phhe shijt phoue angles o all'the Componenta
’. When the shtted"
shiyted by t40; the
qives Sional
a ane Kre HPlbet
'rditing tunction o, trme is Known
eltin9 the signal.
also Called a Quadsatune hilter a it
’ It s
Provid t90° phare shft.
HiLbert tranorner is a ta0° phae Shi ter
in tirne dormain without charqing Anpltude
vaue
 (t) Hilbeot
Tranfovn ht) Output
Input
Siqnal Signal
’ Hibcvt tranHoym o a signal t(t) is d enoted
by )
’ let x(t) be aa jourien tiouns Hormable stgnal with
its 4ourter transöym X (4)
’ The Hibert Taun}orn
2t t) be deyined oe
ou24) =L4) *ht)
t) a input stgna
it, 4) =
(t). d
(t-)

’ Hilbnt Tranon is a linear junction. So,

ts Invesse Hlbnt tranm is qiven by
L() = =L(- X()_d
(t-)
-’ From dejinition, ilbert Tras orm tt) is
the Con voution o (4) with the tirne unction
hit) = ITt.
(t) = (t) * TTt
wKT. Convoution in time domain ís coual
to prouct in Fre, Domai n.
Appuying FT tb cauato vn
FT : -jSqn4)
Note
wKT, .' W2 nf
Sqnlt)< jw DualityPrpny
Using Duaity proputy 9()G)
Tt
FT
ET ,-jSqn(f) kSgn(-)=
Sqnt)
 where Sqni) is the Sign wn jun Chon, dyned
equeny oman a
( 1 ; f>0
Sqn()=o: f-
-1 ; f<0
o
Foun ien t rans form o tata) is qtven by
bt
X ) = j Sqn(f). x(9)
-
’ Equahon (4) statee that the qiven jo tr
jormable siqnal xlt), we may obtatn the
tr auna jornn t)
Fourler Tranjorm oy fts lbent tranwhoe jrequency
t) thvough LTI Sytem
by pauing
Tiaporse is
+ve Spectum ey a t |+ ve jacq
’ T} we multiply-j and -ve spettrum o x(t) [ it, -Ve
o x4)J by theHitbert txan
tj, we obtain
$Teo o X)J by Frequenay Domain lie. (4)].
o} (t) in
H()|
4 +900

(a)

(b)

fq: Magnitude Te ponse

b] phase Tuponse o Hilbeot Tramjorm
Appications
In the Genyatton ooy ss B modulation
’Repretentation o Band pa siqnal mathematHcally.
 Hilbet Tromjom D Basic Sianals
Time Funcion iLbent Tranornm
m) Cd (2rfct) mt) sin2ntct)
mlt) sin (2rfet) - m(t) cd (2 rtt)
c (2Tft) Sin(2uft)
4 Sio (21ft) co l2T4et)
5) Sint I- cot
t
6. Tect)
Lt+y2

8
.+t2

- TSt)

* Propertieu oy Hilbert Troform

:A Signal lt) andt it tilb ent tranaBm
signa!
A()'have the Same
Spectru.
magnitud e

Prco: The Signal at) be jounin tran 4or mable

with 4ories trosgm XH)
Let tt) be its hilbert trontorm be
the four Yen trantorn oy 2(t). Then
X() =-jsqnlf)- x()
t)1=1-ijsgnt4)-x(4)|I
IG)| = 1x(f)|
Hence proved
 Propesty &: C t) is the tilbert tranyorm x)
then the Hilbent traorm a t) is
-I(t)
arglx(4)]- -arg{a}|
(t) ) - (t)
H)- -jsnt)
obtain
Cascading two ideal two- port olevicu to
double tlbert.Tountm
)= xt)*ht)
Taking FT above cquation,
X4)= x). HE) E:HG)-jsgnlf)
x() = -j Sqn Cf)- X(4)
Hilbot Tnorm y i() is qven by
ie, it)= t ) * ht) We
ta KYng owrin trarsyrm we gt,
}(t)-(-jsgne). x(9Jjsqnt)]
o2
W5T, Sqn(t). Sonlf) =4 and j-1
+)= -x(f)
we qet, ilbert trantun
Taking
ie, 2t) = - Xt)
Henee prove d
signal x(t) and its tilbert tranBor
Propurty 3: A
Ovex the entiye ime
}) bre
bre orthogonal
nterval (-o, oo)
(xt) t)dt- D
 TO Prove that t) and ) oe
we hove to Prove that orthogoral ,

Proof: UsTng parse vat'k theo1 en, E,: E

bo

(o(t)- t) dt = ( x(4)-x) df
bo

wKT, fo aa real vaued signal,

Rewrittinq ta,uy be qt
jA)-)
(t) dt = [ xt) -+) df
OKT, }() JSqn (t)x(f)
)= jSqn(-)" x(-4) = jsgn(f)·x(-)
Substituting the Value ol x (-4) in é we gt
to
[X4) (t) at - x(+) j Sgnlf) x(-+)d)
-Do

= jx(4)·x (-+) Sgn () dt

jjix4)1.sqn(4)dt

WKT, x(4) x(-t)|X4 whi ch is even func¡y

7 Sqnt) is odd uncion. Hence Inteqration
un cion Ovt the imits - o tD o is Zero
Hence' proved.
Additional Proputies
alko
1. The tilbnt Tran}orm areal stgnal is oko
Teal
The Hrlbert Trans form oq an wen u
is odd and odd unctton is e
3. The mag ntt ude pe cta oy a signal t ) s
ta T t) ae tdenti cal
2(4) is
4. T|( the HT ) x(t) then HTet4) is -xlt)
,Hlbnt Tramorm o Low Pau Stgnal
The below qune depicta Fowin
Four tianm
a lowpoun signal' xt), whde frequeny ontnt
extendh jom -w tD + w x)

ftg: Spectum qy siqnal xt)

Hilbert Transyo thts sfgnal yido
Q Arpgng
ne
Signal t) whdbe founiettao
m'(E)is
Shown as qolons X4)
Jxo)

x(0)

H : Spe ct umo tZlbent tronjorm

t).
* PRE -ENVELOPES
a siqnal is deined
The Hilbert Trou joTm oy ajTeaienuu,
hor both peitive and nqatve
Can modiy the
we Can reauen ontnt
eal valued sanal Such
that oll negative
a
eueny Component s are Complety
poitive
with the aid o Pre-Envelope Concept
elim fnated. isa Complex
’ The pre - ehvelope % a signal
siqnal with either oniy p&itfve eaueniu
Valued Ant dercted as'<)4Yt)
ony neq ative jrecyuenciea. by a
t is a Complex ignal qenexated
adding a siqnal withiA htlbert tramorm
Quadibte.
 The Complr valued signal pre- envelpe is tormall,
detned a

,)= (t) +j }t)

where Z(t)- postive pre -envelope o} l+)

x(t) a real Vaued stq nal

Aplying We

x,t)= x4)+jlisgnt4).x 4]
- x(t)+ San() -x(4) (:ji-)
X4) = x(t) [1+ sqn(f))
We krno w thot

Sqnlt)=o;t:0
Subsitutin
-1; f<0..
the valua oj Sqn ) in e e
we
qet x) f>o
x()= xto)
c3)
where x(o) is the value ol x t) at the
oTiqin f=0.
Equation (3) Cle ory Shows that the pre-enve lp
the Siqnal xt) hau no freaueney Content to
all neqative 4v equenciea
Hence pie-Envelope jor pöitve treas
wil not contain Cuny negative freaen cia
We may detenmine pre-en velope Aos poitive
jrequena te, x, (t) fn One o) tw0 eaivalent
PrOceduses
4. Terne -Domain Prceduse Gfven the signal
We
éay (i) to Compute pre-en veope
2. Frequenuy - Domatn Procedue
- Frstly determie Fourtn tiamyom the
Siqnal ()" ic, x(4) to ind. Pre -envelo pe or
Then Use eq, (3) ydornain ie, X, (4)
positive requeniu Io quen
Then obtain invenke touiu tram $or m g
X )to Obtain
4t)-2 )enpGj2 nft) ·dt
Depending on the decstption oy the signal
o theproCedlreto obtotn
ma Sele t Ony
we 'poliive
pre-eovelope 4or jreauenctea,
lope jor Negattve FYequencia,
Pre- enve tran}ma.
Conside a Teal valued.wrir
x be it jortur
-ble siqnal ). Such that
toarsos m. neqative Tequenies
The Pre- envelope hor
dyined 2 t) = x(t)-j(t)
The tuwo pre-envelopu 4t)& t) ane
Simply Complv con ugate each other
_t)= t)
fng Four ten trayorm oy éal4) we
qct
Tak
x() = X() -jx (f)
=
x(4) -jl-jSant4) x+)]
- xt) - Sqnt) x 4)
(5)
Aplying the valuu o} Sgnlt) to éq,(s ) we
Pre-inVelofe 4or neqattve'i1eajencie a.
f>o
X_cf) =
 Frorn ea(6), it is cleor that Pre-envelop e
4oY negaive equen ciea donot Contatn any potive
'peauchta o the spectrum .

’ The Sum oy X,t) and t ) is exacthy twice

the ntgInal signal (t).

xt) = )+-()
’ The Spechum ol pre - envelope ilt)
4(4) s nongro
ony sor positive' eauen ia, hence the ue ay
o+a the Subsiipt
The Spetsum oy pre-envlope t) is non
8vn0 ony 4or neqatve jreauenciea, hence the
use oy -
Pre-envelope oy a Lonw Pau sinal
Constdn a low pas sianau x4) wit îta
ounier tranyom x() aa chown below

Spe coun o pre-tnvelope 2t(t) and t)

Shown below
X+() aX_(4)
X (o) ax(o)

tw

Spe ctoun o K,4) Spectrum t)

(o<4<w) fwsteo)
 ignalk ng ove& ANGN Chaun nels.

Geometric epreyetaim signals.

Text1 7.1,4.2,43,7:4

A 4ypical Diyitd (omMuniculim sysuw

Memn1 x)
Soe Vector Si

Mi3
Aceires Copls
Noi
(AN&N) Sitn and
Thi's is chanre
dsove k in
ConveitB
asumd l
zet uf
n e Pro duue each edw
A
n u e Si M
S ida
Nnumsey CmtnmO nys
a ector

nten
Comrepmird Si S(t)

M oSE<T

M
Sontte KI5 phajet si
at+ s Cosids a

m e n Stnnbls.
Let M, Ma, 3 . m arenls
. A

in nhake. md SonR ce
t M Spmbols asa emip Sas
emis M Symbak ewe T SecCmds.
 is
pr' rbab kkies PCM:)
i 1, 2, M -O
P(M:)=L
= . .

Sone prenli la vecft sowce

Meast
Shech oduces a e t r q numles

S
S: i I,2, M -
Sia

SiN
ve N dimsim q vectw N SM

T inpt l6 Ikk m i d is thes S (em )

Th mo dddr cms kuel a sistinet inëli eujj

S i a S; CE) hnnii T Bec mds

SCt)dt, i= 1, 2 . M -

a dashnt

Cvey menni
S0nice ontp M , /2l, 2, - M,
w e u g y Sign S;(t) g mendi T e cnnds, han-
mHed thogh the ehan ra.

the chans is a u e e hae S Sneu'et

f t u e t BN,

Sitd S; (t) CoTld hoie , Hs oise is addiKe,

e man stnk masy ansi
6awsj an neik p Cem Ct)

ta Aacejed s i n l t ) is

zCt)= S; Gt) + »CE)

 L t 6awsiamm
h o i NC)

Tra |l
Pecaivod
hit si) (
S t)

ANGN Channel

T Channl P

A 6GN ekanra
Mod

howe ohe o pene g e i , eei h

inevtasy mke occamimw Y
s
e Prba L: G S m enTÝ Pe
t*)
PePm;) PCm =oliant) + Pn;) PCA= ||o
trurnil
gml 1 Cnd'h
 bjee metrf'cPepre stariar S i g a d s .

The ipatante omemc epreenaliä S i g s

is loepre stn any set Memery sitnls SG) s
basis uncham,
n Camhrak m s g orthondmas
Nue NM
N otST+ Symb
dwali
S;t)= S jCt) i1,3,

T Sir,2, M
Here NM
(E)Þ;Ct)Jt -I, N

St =S
and N vdnd basic
me i ms 6,ct), Pat). ¥ t
ostho ndmd set
vms

4 i
Ksonackex deHe

prestn hd Ni, *t g cvemk

vie ed N

ector: dinokd by S
N di'mensinnnd Signad
 N ekmens g ectr S; opeLrkrg o4
inpd,
G5ien
Sitnad s,CA) 2erdts
en

ohstinutf
S1 ni enet
da) sit S; (+)

Sta

S Sche me Viened

a SttlkTsizas
PCt)

thnimad
S ectw

N Numscs
Set

SE) is1, 2, M , 0peLale ip

Al, tven sigt ,

SiN m em
e Cam Cndensa coeffpatns Si1, Si2,

-
t)
S;(t)
banE q N-pvsdhutt

vieed
Cm mm fnpt
am3
 edr
Relim betwen lenatk ect
nd
Sitnal enugH.

vectr Sf in N-dinensind
et n Considek an arbtaty ,

hs neto
knclidkau space smmd k kq

2
(S;, S;) =
S's;
St
= S 1 , Si2 S Sia

Si + S -
N 22

M
1,2,
j
Co-Co -o
dnates
S
sam S; os,'he q
t rectsinst, tHam hi
t S amd S a
s
twecfors S; o m d s
egk
batwean
avms S sTS
iner pro duct S i
Cos 1 ; )Clengt f) ls:||S
CLengtk S
-
** SCt) is
sigras S(t)
ene E amoia td
No
2
SCt)dt
N
Sso)(3h ) dt
-

ovdevin cm),
 N

F;
T
ing rthezend prpuh t)Ct) d itk

e 2
2
F Stj
thw
CorTeapmdins vectr S

NoN Euckdean ohxfance etweun S md ect

S;
NoN,
pestkr sid s; ( ) md S; (t) is
T
drj 1 - l | - L (Sk-7)J-o
al
dij ct) -sjcEs) Je
 Schmid+ Orthognalsatim Pro edue.
ram
SC)
Aeal vadntd enuH zal s, (t), SaCt).
each g ohwalim T Secnds YepresCn l
N

Si(t) = Si;C t ) , o s t s T i1,2, . M

hene co - efeu'emt espmdim deyiad dy

S-S,ct) tjce)dt, J1, 2,2 N

amd v d n a d basis mek P,Ct), ACt)-

T í-JN
PCt)p;t) = 1, -

,1t
em O Cam Cpmndad

s,CE)= S P,C+) +Sa t ) + +SIN PNlt) ) t:|

4
Salt) S2, (t) +S2a ta Ct)+ +Sa l(+)
SaCt)= 1 f ,CE) +Sgaalt) + +gn n(t) A3

tSmN Pnlt) J 4N
SmCt) Smi , lt) tSma fa(Y) +E

Suryanarayana NK
Dept of ECE
SVIT, Bengaluru 64
 to ind orlkantmnd set g basjswmuhs.

sT 1 Pt Si; =0, exapt S in em 4 |

S,(t) =
S, P, Ct)
PC+) = S,Ct
S Shae S = JE, eruTIGSiz
S, Ct)
t n d S , Syuae cm 4 bolk si'das amd

Secmds.
inliyrali on T

S'sce)d+ - t)d4 -
hVing unit magnifrk

S (s2(E)JE -

Step 2 Cmsi der em 42.

SCt) = Sa o,Ct) +Saa taCt) 42
nu TSem
wwlkply k sideu by éct) nd ingin o

Sa Ct) ,ct) d+
) 0, ifj

S2 SaCt) f,ct) dt
 Nou et aCt) = Salt) - Sa #,C+)

ere ga (t) =
Saa Palt)

Oe T etnds.

Ce)Ct) dt - 9 Cd+

Saa 9 CJ+
rm e m ()
.a(t) = ) -

'Ce)dt
Canhnming h s ud a, he
M , in gener N
d i-
,Ct)=S; (t)-2 Á; (t) , j=1, 2-- i-

SSi S; (t) f,ce)d+, j-l.a.. . i-i

-(
P:(t) = (t)
j-, 2- . .
N

t)dt
Suryanarayana NK
Suryanarayana NK
Dept of ECE
Dept of64
SVIT, Bengaluru ECE
SVIT, Bengaluru 64
 Summa

,Ct)= shre =Ce)dt5

VE
Cni du inlovmedinli 4mekm 9. lt)
j-2--1-l
(t)= S; (4)- Sj CE)
have N

t)

2, 3- M-1,
T
- Sa, Sa3, Set
E; 4:Ct)dt

Suryanarayana NK
Dept of ECE
SVIT, Bengaluru 64
z

&

suryanarayana NK,
Dept of ECE ,SVIT
W
 suryanarayana NK,
Dept of ECE ,SVIT

O
S2
suryanarayana NK,
Dept of ECE ,SVIT

N
 suryanarayana NK,
Dept of ECE ,SVIT
E

N
w
 w
w

O
suryanarayana NK,
Ao Dept of ECE ,SVIT

M
(2) ihen S,Ct), SaCt) SsC+) obtuin ortha goralisim
,

wncta wing Grm Schud+ Proudwe

4S, Ct) Ss C4)
3
4 S2Ct)
3 1
2. 2 3

T 3clands
S,CE), S2CE). SCt) 4w
- -

encry si
peenfed
N stTT
S;(+) Sj #;ct), 20,, 2. -. M

S,(+) SiP Ct) T

S sfCtd+
,C+) S,Ct)
=

SI Lo

S-se 4 -e-
s, C+) o sts 4
+)

step 2 r i- et w
d falt).
SgCt) = Sa P,+) +Saa fat)
et gCt) = Saa a lt)
T
Sa2 9Ct) d +

bt aCt) S t ) - Sa P C+)
Sa C+) SC+) ,Ct)d+
3 .sSo
bu

/ x Z =3
4 S2 Ct) #,4)
Sa3|

No
No, , (+) =
S, Ct)
-

Sa1P, C9
Satt)
3-3 ,

S2,t)

+74
432Ct)
32
O
NoH 3

S31 3.d+4x2 4

Ss Ct)
3
S32 (-k)-
itt)
S323

HoH Ct) = g t) - S31f,C4) - g2f2t)

- + 0
3
402
-t
sCt)4 3-- 0

oi (t): i s
S33
Ssafalt) hot exist

alt)

s,Ct) 6f

SCt) 3 - 3 2 $3
(3) wi r
3win Schnudt rlIha gtvakisafon P a dy

detek mun ortho ndmal bsis tnchms t r h

4 SaCt) st)
45t)

T 2T
.
Soluim
N
So st$T
sl S;Ct) 4, i
i,2 -. M
JI
i1,2. . M
Sij jl,2 N

Cme
Ct) 6CE) Jt 1, ij

or i,em0
S,Ct) S,, P,CE)
PCt)=,(E)
'sCt)dt+

,Ct)= S4) S T 12 J
TIs
3 SCt)
S
VT T
StST
P) olkwj-
 O is
sl a: fr ia3, c m

Sa(t)= S21 #+) + Saaflt)

T
dCt)S2aLO| 2Ct) 5()

SaC PaCt)=2lt)

S2
2Tl

Ct)=SaCt) - S2, P, CE)

Ts
T S2Ct)
Sa
o2Tl
4S21C) Sa1,Ct)14)
T
92Ct) S,sts

T aTl oStT3
2T
Sa29ct)d+
Ts aGt)» 1CE)
Te 7stsT
ostSTl
Sfep 3

4 ) - SCE)-,/,ct)-f4)

Ss Ct)
d

S T T

Ct) 3 Ct)
:O
S33
T
Ss2 t ) S322C

ASsLt)
T T

S3O
Th

oTI
- o-1 o

s32 T
faces T

)
Th T T *T

ST St<T
T 9C+) o st $4T, t T
 $38 (T-2)
2T

SCt) SaCt) TYS+sST

, T , oStSTl3
S33

sl 4 No 1m avefr ms ve
4SCE)

s Tls

Ts T

SCt) =SCt) FSs lt)

S , Ct)=T )

faCt)
S Ct) TT
S CE) 9,C)tTfaCt)
 As t) 4C)

- 3

sl 4oris1, ind Su, P,(t)

V2
Su scd /
P,t)S,(t)
osts 2
S ookav i

sp ori-a, kd Sa, , Sza, #2(+) | , ct) is kuei

Sa S t#, (t) dE 2 t) Sa Ct) -

Sa, 4, C4)

S 9C.d Pa Ct)
S22
SLt)

Sa
9

-Kde+(-a)d+ 1 P C )

2
 s t s |

9t) 6,(t), ol i k

S22 .d+ -
0sts
3aCt)=aCt),
a Ct)= Sa
o,oloTi

sla i-s, S31, 32, ss, Pslt)|,Ct), CE) as

t)

S32 sC)Cd
2

s4)

9, (t)
-

S3af2C+)
,C4)S Ct)- S3, PCe)a)
°sts]
Ws ts
-1+ tl -
%Ctl
stss
-0o

t13
 isC4) stS|
Ta, °sts
StS3

Jlay4 +

1 ) - stSasts
t) S33

sl 4by obsctvivg
k waef s,

SyCt) =
S, Ct) + S2lt) +Sz +)
S+) st) Vag, +)
S2Ct) 1, (t)
.

Saie)
+As Ct)

9,CE)-/a)

s4)= 4)-9,L4)++
SCt)
givem S, (E) and St), wi 6.S. paadhe fin
t ) nd alt) amd skath fCt) 4 ¢2tt)

AS t) AS2Ct)
A

T T2
Solusn
SSasteie =/'e AJT

A sT
Pce)5)
S AT:

T
S2l+)
2At
Sa SC+) ,CE J+
0

T2 t
Tl2 T

T
Sa lt)
SA hvT
t
Th
S22 (Sa Ct)-Sa ce) dt| SaCt) AJTVF

Tl, 2

CA-A)JH+ |(0-A4
Tla
TaJ ' T * AT
PaCt) = Jalt) Sat)-Sa P,Ct)
S22 S22 A S2G+)
A

T
Sa4l5
stsT
P2Ct)
StST

,4)

t
Convesi
COnveysi am
m cndnuong AN,N chanrel inlo
ehaune.
a Ve ctor

(t)
Si t)
ceived sima
trursullad

nal NCt)
i bansian noix

lombinatm
is i n a
Set oM -eneuysitnns S;C+)
ndmd bw dnctm
N- lho
S ost< T
S(t) = t e , 2 , 3 . .. M

Shve coe t c ' t e p d i m n n

T M
S;(t);t)dt ta,a. N

I Aevdnd bai funetns ,(t),,t) 4)

m slks ndms wf
m

is
than Aa/ved Sinas zlt)
(t) = SCt) t NCE) -
ostS T
1 2 . M1M

are Ct) is sampk mtßm 4 k i i euwsian noig

pAces (E) 4 a mean é PSD No
 A

(U)
H

Type text here

2
 Sij t
Considet a ne o Sto chaske p cen X (t) ko
Let
*' (E) dalalid l6 Aatalved Sir t)
metkm
Samp N

z'(t) =
z(t- Yhe)
e m Swbsk han
Subsk : em
N

* (t) =
S; (t) + (E)
N
N
S (E) +wl£)
J

N
- s + ( t ) - s p é )- 2 ÁU)

(t) =
N(t)- ÁjCE)
hoi pro ces NCt)
nce x'(+ dapnds mby
m

and o t Sinad S (t)

m

't) = 't)

cam se li n
en , x(t)

xt) =
2 t)+ + *'(t)
j P em
 N

Ct) =
2 (t) + 'Ct)

Ceived sima C se Aaprenfed 67 ectr

nne
form
Ct)
(t) = Lx +'Ct)

statisth( chasacteaize-tim q Corelut ohpuls

O o obtnin van L X
t x() deno Stockaske p Cess, a Sampk untßm
h i t k is epreKnib ecaied sind xtt) .

vatallk ho sSannph van

Let
LtiX; olens Aondm

j 1, 2 - N,COTnItr ontpt.
epreknl by ,

Let N; dmeLmdmm vnu'ash vepre n d

orreltor in
Sumph vnns pevuud b
re to SAEGanaiaun mpment NCt).

andam vntasu hsae mLan , a channel

fR preces hCE) repnund NCt) hu
N e
kar mLan.
mLan.
Opiwwm uC ves i ng Cokeret dalechnm.

ML deosing ( Maasmnm Lhlkhwd daodhing).

Het w Cmide M pouisk Sitrads S, Ct), Sa t)..

SmCt) truns melt k e n psbat; ky /n
T Secmds

Coweltrs
C mm m

hns, el p t ect
is
as Ct) copd
tamk C&rltws, dina dsaui nd .

i mhip betwean 0Swavali m eetr md

2
Paaid nnd pein
Noik e c t hN.

ecr

S Sinad e c t w
 Sind delethun stlimn

to hmnk amld s
em
in n
Ny tt ianvd mi'nmme presasky
M
Cem"
e u i m mksns

eutmbe x.
M, Ke
Snpp kus'm mk:n Mad M a

e r e i n dausan is
awer nLy gf

eMj,L) |- P(M; at
l z) -

p tmunm oleunm l n l .

Set M,
P(M t | x ) >P (Me ant|)
a E Í , ka1, 2,3. - .
M

This i Lnlhd MAP (Maai'mum ostevi om polnkls

Tk r ( Im)
trr)
F pve brlty ranlsmto M
 m;)
(dibnd ahkh dem
ww. Landom esseLvalian ectw
Xm M am ma'lld Syol Mg

candti nW sakK dea

fxC) ww.k m .

n Aashi i Ank g du'hm,in texns

L(M)
St M;if L(M) is mari mum Jr Eí

Vhw, MRmnm hteLhnod decoder Lamputy o

keih vod
Mmag Sm bo/s amdC
Compases i m amd dectdes
in 4av

p h i c d m t l i ni a deui nle

pale
tw

is Z prk and ind.

e t a f l burwin veace
M deuim ugm 1, m

bevwli vectr x hes in vem z if LME)

amuM ki.
 Deu'siM
bnd

P&*3

- JE
E

Pg n 4

hstane X-Sis ma Mum or c ff

k--(- N

Yt feNm nneud m if is indapemdat

N
ise k. Secane teum n inner nesnet q 0&vnm
vectw Z md r u ml hd iW Se . t i v d tLm

N
:p S2
me MaNmm akhwd daunm Anl Astald
o5wd veutw x Ls iu rgm ,
S AIN MWM
(%-E)
 Correlaahi an Peces'u

ANGN Chunel, Shan trawmlie Sitrads a

emalJkaly, opiwmm Auui vw ieh minim
awun ikly ews CPe) s wrelk m uaiv.
Jhih sis5 g tho
pS.
Pt a Sm g Coe lndvs shav ontys s'*

Pste Sital ansSim deiodey

atcutr
dlenne hdd
KCE), ast <T, 6 M
Peuu'y

- + .

+
t)
mal

eutr MAX
ktelhod
rlkonsmw
Iner t
nis nnthm Susnt 4
C
Max-Likehhnd deods
peauli. an *' pva duce

(M) d n e fe
Matthad Fl etu i ves
Cortk 4,t) 4 )

4 ) tsel

idv
p
4

LTI WIE impud

, t ) = J x( h,Ct-)47

viz, s'stsT, (plon T7 t)

Ct)=z( A;(T-t)Jt

No Cinnder deleut» 6d m n Co elasa

h;CTt)= ; t) stST, jl,2-- M

 impod m darind
e my eareM (nmdrkm
ImpnNL
M
hCt) f(T-6) 9stT, j-ha-

Raue TACT-
xL)

Mathad
Snpk
atT
Jl

fCt) i 5 impad
M
ep hlt) s ies
r Ost T
hCE) = ( T - t )

s
Calld
mretm
mathid l t a i n plag
l t Aeerie