Int
ern
at
i
onal
Journ
al of Inf
orm
at
ic
s
and
Co
m
munic
at
i
on
Tec
hn
olog
y (IJ
-
I
CT)
Vo
l.
8
,
No.
3
,
D
ecembe
r
201
9
, pp.
184
~
190
IS
S
N:
22
52
-
8776
, DO
I: 10
.11
591/ij
i
ct
.v
8i
3
.
pp
184
-
190
184
Journ
al h
om
e
page
:
http:
//
ia
escore.c
om/j
ourn
als/i
ndex.
ph
p/IJI
C
T
Perform
ance ana
lysis of
CIC m
ulti
rate filt
er in c
og
nitive r
adio
for effici
ent pulse
shaping
in wir
eless dom
ain
A.
S.
Kang
,
Visha
l S
ha
r
ma
,
Renu Vi
g
Depa
rt
m
ent
o
f
E
CE,
UIE
T,
Panj
a
b
Univer
sity
,
Ind
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
hist
or
y:
Re
cei
ved
M
a
y
5
, 2
018
Re
vised
M
a
r
6
,
201
9
Accepte
d Nov
13
, 201
9
In
th
is
p
ape
r
,
the
Perform
ance
Analysis
of
C
asc
a
de
In
te
gr
at
or
Co
mb
Fil
te
r
in
cont
ex
t
of
Fi
lt
e
r
Bank
Mult
ica
rrie
r
Tra
nsmiss
i
on
has
be
en
pr
ese
nte
d
for
Cognit
ive
r
adi
o
.
A
bene
fit
of
the
chose
n
techniqu
e
is
that,
a
CIC
f
il
te
r
ca
n
b
e
designe
d
with
a
sl
ight
adj
ust
me
nt
in
par
amete
rs
of
intere
st
.
The
en
ti
re
per
forma
n
ce
of
the
fi
lters
desig
ned
is
an
al
y
ze
d
and
ev
al
ua
te
d
b
y
ana
l
yzi
ng
Norma
lized
Am
pli
tude
ver
sus
Norma
lized
Frequ
enc
y
plo
ts
at
typi
ca
l
K
and
N
val
ues.
Th
e
ro
ll
off
fa
ct
or
pla
ys
a
signif
icant
rol
e
in
pe
rform
an
c
e
ana
lysis
o
f
CIC
fil
t
er.
The
result
s
show
n
are
a
useful
advance
for
rf
d
esig
n
engi
ne
ers
working
in
th
e
dom
ai
n
of
mul
ti
r
ate
signa
l
proc
essing
i
n
wire
l
ess
com
munica
ti
on.
To
ensure
the
acce
pt
abl
e
per
for
ma
nc
e
of
Enha
n
ce
d
FB
MC,
com
put
at
ion
al
c
ompl
exity,
and
t
ran
smiss
ion
burs
t
l
engt
h
ne
ed
to
be
r
educ
ed
.
The
eff
ect
of
Sto
p
band
attenuati
on
on
the
edg
es
of
Magnit
ude
an
d
Freque
ncy
responses
has
bee
n
studi
ed
und
er
constraints
su
ch
as
Lp
,
K
an
d
M
during
diffe
ren
t
simu
lat
ion
runs
.
Ke
yw
or
d
s
:
CIC
Cognit
ive
r
adi
o
Fil
te
r
M
ulti
car
rier
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
A.
S. Ka
ng
,
Dep
a
rtme
nt of
ECE,
UI
ET
,
Panjab
U
niv
e
rs
it
y,
Cha
ndiga
r
h
,
I
nd
ia
.
Emai
l:
aska
ng_85@ya
hoo.co
.in
1.
INTROD
U
CTION
Diff
e
re
nt
puls
e
sh
a
pes
an
d
var
i
ou
s
ot
her
performa
nce
pa
rameters
of
i
nterest
s
uc
h
as
filt
er
le
ngth
,
delay,
ha
rdware
com
pu
ta
ti
on
al
comp
le
xity
and
c
os
t
et
c
w
hich
a
ff
ect
s
th
e
performa
nce
eff
ic
ie
nc
y
of
FBMC
M
ulti
car
rier
C
om
m
unic
at
ion
due
to
inc
rea
sed
s
pectral
e
f
fici
ency
an
d
r
edu
ce
d
disto
rtion
ef
fect
wit
h
highe
r
stopba
nd
at
te
nuat
ion
possi
bly
achieva
ble
un
der
dif
fer
e
nt
pe
rsisti
ng
c
onstra
ints
of
overla
ppin
g
fact
or
K,
numb
e
r
of
f
re
qu
e
nc
y
c
hannels
M,
with
s
uitable
enc
odin
g
a
nd
mod
ul
at
ion
te
c
hn
i
que
ap
plied
the
re
in
to
at
ta
in
Per
fect
Sign
al
Re
c
on
st
ru
ct
io
n,
with
ul
ti
mate
ly
fu
lfil
ing
t
he
N
yquist
crit
eria
of
e
ff
i
ci
ent
pu
lse
s
ha
ping.
An
i
ntere
sti
ng
CIC
Fil
te
r
a
nal
ys
is
is
ap
pro
pri
at
e
for
rf
wire
le
ss
sy
ste
m
de
sign
en
gin
ee
rs
to
pav
e
the
wa
y
for
t
he
s
ucce
ss
of
nex
t
ge
ner
at
io
n wi
reless c
ommu
nicat
ion sys
te
ms.
M
ore S
pectral
Re
so
luti
on, Re
du
ct
io
n
i
n
out
of
ba
nd
emissi
on
an
d
e
nh
a
nc
ed bit
r
at
e b
eca
us
e
of
lesse
r
gu
a
r
d
ba
nds
w
it
h
no
c
yclic
prefix
mak
es
F
BMC
to
assi
gn
di
ff
e
ren
t
s
ubcarriers
t
o
a
vaila
ble
unsync
hron
iz
e
d
us
ers
in
a
dyna
mic
ma
nn
e
r
[1]
.
Fit
er
Ba
nk
Mult
ic
arr
ie
r
reli
es
up
on
a
protot
yp
e
filt
er
with
good
ti
me
-
fr
e
quenc
y
pro
du
ct
.
T
he
c
onfine
ment
of
FBMC
sig
nal i
n b
oth t
ime a
nd
fr
e
quenc
y d
oma
in te
nds to
i
ncr
ease
it
s
rob
us
tness
to
ti
ming
e
rror
and
offset
in
f
reque
ncy.
A
uniq
ue
al
gorith
m
with
un
ifie
d
structu
re
f
or
F
BMC
Tra
ns
cei
ver
has
been
propose
d
[1].
T
he c
a
pa
bili
ty
of
FBMC
to
pro
vid
e
imp
r
ov
e
ment
in
fre
qu
e
nc
y sel
ect
i
vity
by
ma
ki
ng
us
e
of
longer
a
nd
s
pe
ct
rall
y
ef
fici
en
t
filt
er
makes
TMUX
desi
gn
wh
ic
h
c
ompri
ses
A
nalysis
(
Dow
ns
am
plin
g)
an
d
Synthesis
Fil
te
r
Ba
nk(
U
ps
am
pling).T
he
rise
and
fall
in
the
sign
al
sam
pli
ng
rate
is
relat
ed
to
the
c
omp
le
x
to
real
and
real
to
co
mp
le
x
co
nv
e
rsion
of
in
pu
t
sig
nal,
with
pre
an
d
po
s
t
OQAM
s
ub
channel
pr
oce
ssing.
The
stu
dy
relat
ed
t
o
t
he
FB
MC
subc
hannel
proce
ssi
ng,
de
pends
upon
c
ompli
cat
ed
filt
er
desig
n
pr
oce
ss
that
corres
ponds
to
varyin
g
filt
er
l
eng
t
hs
,
has
bee
n
sho
wn
he
re.
The
Ov
e
rlap
pin
g
facto
r,
K
ta
ken
is
i
n
the
s
ui
ta
ble
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Inf & C
ommu
n Tec
hn
ol
IS
S
N:
22
52
-
8776
Perf
orma
nce
analysis
of CIC
mu
lt
irate f
il
te
r
in cog
niti
ve ra
dio
f
or
eff
ic
ie
nt
p
ulse
sha
ping
…
(
A
. S. K
ang
)
185
ref
e
ren
ce
ra
nge
(2
-
20
)
[
2].
It
def
i
nes
the
ma
ximum
numb
e
r
of
m
ulti
carrie
r
symb
ols
wh
i
c
h
ove
rlap
eac
h
ot
her
in ti
me do
main
.
The
protot
yp
e
f
il
te
r
is
anal
yz
ed
to
f
ulfill
the
Nea
r
Pe
rf
ect
Recon
st
ru
ct
io
n
cha
racteri
sti
cs.
T
he
Pe
rf
ect
Re
con
st
ru
ct
io
n
is
on
l
y
achie
ved
i
n
case
of
ideal
tran
sm
issi
on
c
hannel
wh
ic
h
impli
e
s
that
interfe
r
ences
or
i
gin
at
in
g
fro
m
filt
er
ba
nk
are
t
oo
small
wh
e
n
co
mp
a
re
d
t
o
t
he
resid
ua
l
interfe
re
nce
s
due
to
tra
nsmi
ssion
channel.
T
he
C
ascade
d In
te
rc
onnecti
on
prot
otype
filt
ers
c
ould
provi
de hig
hest st
opba
nd
at
te
nu
at
io
n
t
ha
n
t
heir
equ
al
le
ngth
P
R
co
un
te
rp
a
rts
.
T
he
best
wa
y
t
o
desi
gn
t
he
NP
R
prototype
filt
er
is
t
o
directl
y
opti
mize
the
coeffic
ie
nts
of
the
filt
er
im
pulse
respo
ns
e.
The
only
li
mit
at
ion
of
this
te
chn
i
qu
e
is
t
hat
the
num
ber
of
filt
er
coeffic
ie
nts
in
crease
s
harply
wh
e
n
filt
er
bank
is
desig
ne
d
with
numb
e
r
of
s
ub
c
ha
nn
el
s
hi
gher
tha
n
256
[3
,
4].
The
ove
rlap
ping
facto
r
K=4
a
nd
l
ocali
zed
m
od
e
ba
nd
al
loc
at
ion
of
M
s=1
60
s
ubcar
riers
has
bee
n
us
e
d
in
th
e
cor
e
of the
FB
M
C s
ys
te
m
[
5].
2.
COGNITI
VE
RADIO
N
ET
WORK
FBMC
Tra
ns
c
ei
ver
f
or
m
ulti
us
e
r
asy
nc
hro
nous
tra
ns
miss
ion
on
the
f
ra
gm
e
nt
ed
s
pectrum
plays
a
major
ro
le
i
n
s
pectr
um
se
ns
in
g
in
wireless
c
ogniti
ve
ra
dio
netw
ork
th
ough
it
al
so
de
pe
nds
up
on
the
c
ha
ll
eng
es
and
de
sig
n
tra
de
-
offs
w
hich
ar
e
re
qu
ire
d.
T
he
intel
li
gen
ce
i
n
wireless
c
ognit
ive
ra
dio
net
w
ork
is
i
ntr
oduc
ed
by
sp
ect
r
um
se
nsi
ng
the
band
width
siz
e
on
m
ulti
ple
c
ha
nn
el
s
with
va
ry
i
ng
sig
nal
s
tren
gth
s
a
nd
diff
e
re
nt
transmissi
on
te
chn
i
qu
e
s
[6
,
7].
G
ood
s
pectral
co
ntainme
nt
i
s
ma
nd
at
ory
f
or
getti
ng
rid
of
the
distor
ti
on
e
ff
ect
from
as
yn
c
hro
nous
si
gn
al
s
in
adjace
nt
ba
nds.
Ther
e
is
al
w
a
ys
a
requireme
nt
to
minimi
ze
th
e
delay
of
prot
otype
filt
er
as
the
sel
ect
ion
of
num
ber
of
c
oe
ff
ic
ie
nts
is
a
trade
-
off
between
delay
a
nd
filt
er
stop
ban
d
at
te
nu
at
io
n
[
8
,
9].
2.1
.
M
ulti
m
ode tr
an
smissi
on i
n f
il
ter
b
ank pr
ocessin
g
The
Fil
te
r
Ba
nk
Processi
ng
is
ba
sed
on
mu
lt
iple
acce
s
s
te
c
hniq
ue
w
hich
e
na
bles
t
he
m
ob
il
e
te
rmi
nals
to
tra
ns
mit
at
a
ti
me
in
r
eve
rs
e
li
nk
at
d
if
fere
nt
ope
rati
on modes,
na
mely
FBMC,
FB
-
F
BMC
an
d
al
s
o
in
seri
al
carrier
m
odulat
ion
just
to
pro
vi
de i
nde
pe
nd
e
nt
upli
nk
si
gn
al
mu
lt
iplexes
with
m
ore
ba
ndw
idth
e
ff
ic
ie
ncy.
T
he
rev
e
rse
li
nk
c
an
sup
port
bot
h
sin
gle
car
rier
an
d
m
ulti
car
rier
m
od
e
s
by
util
isi
ng
the
bette
r
sel
ect
ivit
y
of
fr
e
qu
e
nc
y
sa
m
pling
-
based
pr
oto
ty
pe
fi
lt
er
desig
n.
A
lo
w
er
cal
culat
io
n
com
plexity
ba
sed
li
nea
r
e
qu
al
iz
er
structu
re,
is
ca
pab
le
of
fr
e
que
ncy
sel
ect
ive
pe
r
s
ub
band
pro
cessi
ng
to
facil
it
at
e
eff
ic
ie
nt
c
hannel
c
ompe
nsa
ti
on
[10].
I
nfact
,
th
e
op
ti
miza
ti
on
of
Mod
i
fied
D
FT
Fil
te
r
Ba
nk
Mult
ic
arr
ie
r
M
od
ulati
on
Syst
em
is
ve
ry
use
fu
l
for
the d
e
sig
n of
modu
la
ti
on s
yst
ems in
t
he nex
t gen
e
rati
on
ne
tworks
[
11
,
12]
.
3.
FREQ
UEN
C
Y
S
Y
NCHR
O
NIZATI
ON F
OR FB
M
C
T
RANS
MISSI
ON
The
hi
gh
e
r
st
opba
nd
at
te
nuat
i
on
of
the
Fil
te
r
ba
nk
al
lo
ws
c
hannel
s
el
ect
ion
filt
ering
a
nd
Narrow
band
In
te
r
fer
e
nce
re
du
ct
io
n
at
the
r
ecei
ver
a
nalysi
s
ba
nk,
with
ou
t
pr
e
-
proce
ssin
g,
besides
the
anti
-
al
ia
sing
filt
ering
determi
ned
by
the
sam
plin
g
rate
at
anal
ys
is
filt
er
ba
nk.
Hen
ce
,
the
Fil
te
r
Ba
nk
prov
i
des
go
od
fr
e
quenc
y
sel
ect
ion
for
t
he
desire
d
sp
e
ct
ra
l
co
mpo
ne
nts,
so
it
is
de
sirable
t
o
t
hink
of
a
FBMC
receiver
wh
e
re
al
l
the
base
band
sig
na
l
proce
ssin
g
functi
ons
are
performe
d
in
fr
e
qu
e
nc
y
do
main.
T
he
ent
ire
c
hannel
pa
rameter
est
imat
ion
is don
e a
fter a
naly
sis fil
te
r
bank a
t a le
sser
rate [13].
Pr
a
ct
ic
al
l
y
,
the t
ran
s
mis
sion
s
ys
te
ms a
r
e p
eak
powe
r
li
mit
ed
as
the
y
sho
w
non
-
li
near
c
ha
racteri
sti
cs
whic
h
ca
us
e
sp
e
c
tral
wi
den
i
ng
of
the
tra
ns
mit
sig
nal.
The
cu
rr
e
nt
c
om
m
unic
at
ion
s
ys
te
ms
nee
d
to
rest
rict
the
po
wer
s
pectr
al
de
ns
it
y
of
the
t
ran
s
mit
si
gn
al
t
o
correct
l
y
s
peci
fied
s
pectr
um
mask
[14
,
15]
4.
ALGO
RITH
M
The
desig
n
of
Ca
scade
d
In
t
egr
at
or
C
omb
Fil
te
r
-
P
ro
t
otype
Fil
te
r
f
or
plo
tt
ing
A
m
plit
ud
e
w.r.t
Normali
zed
Frequ
e
nc
y
Re
s
pons
e
has bee
n
s
how
n here
to
e
xamine
the
ef
f
ect
o
f
d
if
fere
nt
p
a
rameters
.
K=4
;
N=64;N
1=
16
;
N2
=
4
;L=
20;f=
0:0
.
005:
1;o
me
ga=p
i
*f
;
f
m=0.0
8;
I1
=
(sin(o
mega
*N
/
2)./si
n(ome
ga*N
1/2))/N
2;
I2
=
(sin(o
mega
*N1/2)./si
n(
ome
ga/2
))
/N
1;
G=sin
(p
i
*f*
N1/2).
/(
N
1*sin(pi
*f
/
2));
al
ph
a=
0.22*
pi*
f
m;G
q
=(si
n((
pi*f+al
pha)
*N1/2)./si
n((
pi*f
+al
ph
a
)/2));
Gq
=
G
q.*(
si
n((
pi*f
-
al
pha
)*N
1/2)./si
n((
pi*f
-
al
ph
a
)/2));G
q=
(Gq*1.0
125)/
(
N1
^
2)
;
GM
C
3=
Gq.*G;
I_
s
h_m=a
bs
(
(
3*I
1.
^(
2*K
)
-
2*I
1.
^(
3*K
)).*(I
2.^
L))
;
I_
s
h_m
1=abs(
(
3*I
1.
^(
2*K
)
-
2*I
1.
^(
3*K
)).*(
GM
C
3.
^L
))
;
%fil
te
r
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2252
-
8776
In
t J
Inf & C
ommu
n Tec
hn
ol
,
V
ol.
8
,
No.
3
,
Dec
20
1
9
:
184
–
190
186
Figure
(2)
plo
t(
f,20*lo
g1
0(I_
s
h_m
),f,
20*log10
(I_s
h_m
1))
axis(
[0,0.2
,
-
500,5]);g
rid o
n;hold
on;
G=(
si
n(ome
ga
*N
/
2)./si
n(ome
ga/2
))
/N;
GK
=
G.
^K;
F
igure(
2)pl
ot(f,20*
l
og10(a
bs(
GK)),'k'
,'Line
Width'
,2)
I_
s
h=a
bs
(
3*G
.
^(2*K
)
-
2*G
.^(
3*K));pl
ot(f,20
*log10
(I_s
h)
,'
--
')
5.
RESU
LT
S
AND DI
SCUS
S
ION
Norma
li
sed
amplit
ude
ver
sus
norma
li
s
ed
subc
arr
ie
r
fre
qu
enc
y
plot
betwee
n
K
and
Alph
a
as
show
n
in
Figure 1
-
3.
Com
puta
ti
on
of
norm
al
i
ze
d
am
p
li
tud
e
,
stopband
a
ttenu
at
ion
oc
cur
ing
at
var
i
able ov
erl
ap
ping fa
ct
or
K
an
d
roll
off
factor
as
sho
wn
in
Ta
b
le
1
.
Var
i
ou
s
M
at
la
b
Pl
ots
obta
ine
d
betwee
n
K
,
Alpha
a
nd
Sto
pb
a
nd
Atte
nu
a
ti
on
f
or
CIC Pr
oto
t
yp
e
Fil
te
r
Desig
n h
ave
been s
how
n
in
Fig
ure
4
-
11
dep
ic
t t
he
a
na
lyti
cal
co
m
pa
rison
done.
Figure
1
.
Norma
li
sed
am
pl
it
ude
ver
sus
norma
l
ise
d
subca
rri
er
fr
eq
uenc
y
p
lot a
t
K=
4,
Alpha
=0.
22
Figure
2
.
Norma
li
z
ed
a
mpl
i
tude
ver
sus
norma
l
ize
d
subca
rri
er
fr
eq
uenc
y
Plo
t at
K=
4,
Alpha
=0.
78
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Inf & C
ommu
n Tec
hn
ol
IS
S
N:
22
52
-
8776
Perf
orma
nce
analysis
of CIC
mu
lt
irate f
il
te
r
in cog
niti
ve ra
dio
f
or
eff
ic
ie
nt
p
ulse
sha
ping
…
(
A
. S. K
ang
)
187
Fi
gure
3
.
Norma
li
sed
am
pl
it
ude
ver
sus
norma
l
ize
d
fre
quuen
cy
p
lo
t
for
CIC
propos
ed
proto
type
at
K=11.
4,
Alph
a=
0
.
22
Ta
bl
e
1
.
Compu
t
at
ion
of
nor
ma
l
i
ze
d
am
pl
it
ud
e
(d
B),
stopband
a
tt
e
nuat
ion
occ
ur
ing
at
var
ia
bl
e
ov
erlappi
ng
f
ac
to
r
K
and
roll
off
fa
ct
or
Ov
erlapp
in
g
Factor
K
Ro
ll off
fa
cto
r
Alp
h
a
No
rm
aliz
ed
Am
p
litu
d
e(dB
)
Sto
p
b
an
d
Attenu
at
io
n
starting
at
a No
rm
ali
zed
Frequ
en
cy
4
0
.22
-
1
0
to
-
341db
0
.06
8
2
(Norm
al T
r
en
d
)
4
0
.78
-
1
0
to
-
348db
0
.06
9
1
E
arly sid
e lob
e tail
d
ecay
1
1
.4
0
.22
-
1
0
to
-
462db
0
.03
5
2
(Faster
Sid
e lob
e tail
d
ecay)
Figure
4.
O
ver
l
app
i
ng f
act
or
K vers
us
sto
pband att
enuat
ion
f
or
C
IC
filt
er
desi
gn
Figure
5. N
umber
of
ta
ps
per
channel
ver
s
us
r
oll
off fact
or
al
ph
a
f
or
C
IC
f
il
te
r
desig
n
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2252
-
8776
In
t J
Inf & C
ommu
n Tec
hn
ol
,
V
ol.
8
,
No.
3
,
Dec
20
1
9
:
184
–
190
188
Figure
6
.
N
umber
of
ta
ps
L
pe
r
c
hannel
ver
s
us
overlap
ping
f
act
or
K for C
IC
filt
er
desig
n
Figure
7
.
N
umber
of
ta
ps
L
pe
r
c
hannel
ver
s
us
st
opba
nd att
enu
at
io
n
f
or
C
I
C
filt
er
Figure
8
.
M
=
N
/2 v
e
rs
us
st
opba
nd att
enuat
ion
A
s
for
C
IC
filt
er
desig
n
Figure
9. M
nu
mb
e
r
of
Subc
ha
nn
el
s
v
e
rs
us
L
num
ber
of
ta
ps
per cha
nnel
for
CIC
filt
er
de
sign
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Inf & C
ommu
n Tec
hn
ol
IS
S
N:
22
52
-
8776
Perf
orma
nce
analysis
of CIC
mu
lt
irate f
il
te
r
in cog
niti
ve ra
dio
f
or
eff
ic
ie
nt
p
ulse
sha
ping
…
(
A
. S. K
ang
)
189
Figure
10
.
M
s
ub
c
ha
nn
el
s
v
e
r
su
s
overla
pp
i
ng
facto
r
K f
or
CIC
filt
er
desi
gn
Figure
11. M
s
ub
c
ha
nn
el
s
v
e
r
su
s
ro
ll
off fact
or alp
ha
f
or CI
C
filt
er d
esi
gn
6.
CONCL
US
I
O
N
The
st
udy
done
has
it
s
im
pac
t
on
t
he
Pe
rformance
E
nha
nc
ement
of
Fil
te
r
Ba
nk
Mult
ic
ar
rier
S
ys
te
m
base
d
Co
gnit
ive
Ra
di
o
unde
r
the
e
ff
ect
of
ub
i
quit
ou
s
an
d
pe
r
vasive
ra
dio
c
ommu
nic
at
ion
.
T
he
res
ults
of
pro
po
se
d
C
IC
Fil
te
r
desi
gn
ar
e
co
mp
a
red
to
tho
s
e
ob
ta
ine
d
by
We
ssel
L
ubberh
uizen
.
It
ha
s
bee
n
obser
ve
d
t
hat
K=8
wit
h
R
oll
off
fact
or
=
0.4
gi
ves
opti
mum
re
su
lt
s
i
n
te
rms
of
normali
zed
a
mp
li
tu
de
of
-
10db
t
o
be
yond
-
500db
with
sto
pb
a
nd
at
te
nuat
ion
occ
urrin
g
a
t
a
nor
mali
zed
fr
e
qu
e
nc
y
of
0.
026
i
nd
ic
at
in
g
faster
si
de
lo
be
ta
il
decay
ha
pp
e
ni
ng
ir
res
pecti
ve
o
f pu
lse
s
ha
pin
g
c
onstrai
nts
. D
iffe
re
nt simu
l
at
ion
ru
ns
h
a
ve
b
een
done
t
o
obta
in
the
res
ults
for
CIC
filt
er
desi
gn
i
n
te
rms
of
Stopba
nd
Atte
nu
at
io
n
As,
Overlap
ping
Fact
or
K,
L
ta
ps
pe
r
cha
nnel
a
nd Roll
off
fa
ct
or
Alpha
(0
<
α<1).
7.
IMP
AC
T
OF
STUDY
A
ND
FUTU
RE S
C
OPE
The
pr
e
sent
w
ork
ca
n
be
e
xtend
e
d
for
furthe
r
anal
ys
is
on
M
AC
la
yer
pr
oto
c
ol
m
od
el
usi
ng
NS2
a
nd
adap
ti
ve
sig
nal
proces
sin
g
a
ppr
oac
h,
ca
n
be
ap
plied
t
o
highli
gh
t
the
e
ff
ec
t
of
am
biguit
y
s
urfaces
of
differen
t
protot
yp
e
filt
er
s
co
rr
es
pondin
g
to
diff
e
ren
t
pulse
s
hap
e
s
a
nd
var
i
ou
s
oth
er
pe
rformance
pa
rameters
of
i
nt
erest
su
c
h
as
filt
er
le
ng
t
h,
delay,
ha
rdwar
e
c
omp
utati
on
al
c
omp
le
xity
an
d
co
st
et
c
wh
ic
h
a
ffec
ts
the
perf
ormanc
e
eff
ic
ie
nc
y
of
F
BM
C
M
ulti
car
rier
Co
mm
unic
at
ion
du
e
to
i
nc
reased
spe
ct
ra
l
eff
ic
ie
nc
y
a
nd
re
duce
d
dist
or
ti
on
eff
ect
with
hi
gh
e
r
sto
pb
a
nd
at
te
nu
at
io
n
po
s
sibly
ac
hievab
le
un
der
diff
e
re
nt
pe
rsi
sti
ng
c
onstrai
nts
of
ov
e
rlap
ping
fa
ct
or
K,
num
be
r
of
fr
e
quenc
y
cha
nn
el
s
M,
with
s
uitab
le
e
ncodin
g
a
nd
modu
la
ti
on
te
c
hn
i
qu
e
app
li
ed
the
rein
to
ac
hie
ve
Pe
r
fect
Re
co
ns
tr
uc
ti
on
of
the
sig
nal,
with
ulti
m
at
e
fu
l
filme
nt
of
N
yquist
c
rite
ria
o
f
eff
ic
ie
nt
pulse
sh
a
ping.
A
n
int
eresti
ng
CIC
Fi
lt
er
anal
ys
is
is
appr
opriat
e
f
or
rf
wi
reless
s
ys
t
em
desig
n
e
ngi
neer
s
to
pa
ve
the
wa
y
f
or
the
s
ucce
ss
of
ne
xt
gen
e
rati
on
wireless
com
munica
ti
on
s
ys
te
ms
[16
,
17].The
a
naly
sis
of
CIC
M
ulti
rate
Fil
te
r
in
Co
gnit
ive
Ra
dio
f
or
E
ff
ic
ie
nt
P
ulse
S
hap
i
ng
has
bee
n
do
ne
w
hich
s
hows
it
s
si
gn
i
ficance
as
ref
le
ct
ed
in
t
he
earli
er
w
ork
done
f
or
t
he
ubiq
uitous
-
perv
asi
ve
co
mputi
ng
thr
ough
matl
ab
base
d
sim
ul
at
ion
s
in the wirel
ess
domain
[
18
-
22
].
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2252
-
8776
In
t J
Inf & C
ommu
n Tec
hn
ol
,
V
ol.
8
,
No.
3
,
Dec
20
1
9
:
184
–
190
190
ACKN
OWLE
DGE
MENT
The
hel
p
rende
red
by
P
rof.
Dr.
Re
nu
Vi
g i
s
duly
ack
nowle
dged
.
Sh
e
was i
ns
tr
um
e
ntal
in
gu
i
ding
a
nd
impleme
nting
t
he
w
ork
pla
n
of
t
he
present
re
search
done
.
T
he
a
ut
hors
are
t
hank
fu
l
to
Re
s
earch
P
r
omoti
on
Ce
ll
,
Panjab
Unive
rs
it
y
Cha
nd
i
garh
for
mo
ti
vati
on.
REFERE
NCE
S
[1]
Yun
Cui,
Zh
ife
n
g
Zh
ao
et
a
l,
"
A
n
Eff
ic
i
ent
Fil
te
r
Banks
Based
M
ult
icarri
er
Sys
tem
in
Cogn
it
iv
e
Radi
o
Networks,
"
Radi
o
Engi
ne
eri
ng
,
v
ol
.
19
,
n
o.
4,
pp
.
479
-
487
,
20
10
.
[2]
PH
YDYAS
,
"
P
HY
DYAS
-
physi
ca
l
la
yer
for
dynam
i
c
spec
tr
um
acce
ss
and
cogni
ti
v
e
rad
i
o,
”
ht
tp:
ww
w.i
ct
-
phydyas.
org/201
3
.
[3]
Mauric
e
G
B
ellange
r
,
"
Spe
ci
fi
c
ation
and
Des
ign
of
a
Proto
type
Fi
lt
er
for
Filt
er
B
ank
B
ase
d
Mult
icarri
er
Tra
nsmiss
ion,
"
Proc.
I
EEE
In
te
r
nati
onal
Con
fe
re
nce
,
Paris,
Franc
e,
pp
.
2417
-
2420
,
2001
.
[4]
Peim
an
Amini,
Behrouz
Farha
n
g
-
Borouje
ny,
"
A
Compa
r
ison
of
Alternat
ive
Filt
er
B
ank
Mul
ti
c
a
rrie
r
Methods
fo
r
Cognit
ive Ra
d
io
Sys
te
ms,
"
Proc.Soft
ware
De
fi
ne
d
Radi
o
Technic
al
Conf
ere
nce,
S
DR
2006
,
pp.
11
7
-
123
,
2006
.
[5]
Dami
en
Roque
,
Cyrille
Siclet
,
"
A
Perform
ance
Compa
r
ison
of
FB
MC
Modula
ti
on
Sch
em
es
with
Short
Perfe
c
t
Rec
onstruc
ti
on
-
Filt
ers,
"
Proc.
19
th
Int
ernati
onal
Confe
renc
e
on
Telecomm
unic
ations
,
ICT 2012
,
L
eba
non,
pag
es
6.
[6]
Marius
Caus,
A
na
I.
Pere
z
-
Neir
a,
"
A
Subopt
imal
Pow
er
Al
locati
on
Algorit
h
m
for
FB
MC/OQ
AM
,
"
Proc.
IE
E
E
Inte
rnational
Co
nfe
renc
e
on
Ac
o
ustic
s,
Sp
eech
a
nd
Signal P
roc
e
ss
ing,
ICASSP
,
p
p.
2660
-
2664
,
20
11
.
[7]
Y.
Medja
hd
i,
M
.
T
err
e
,
"
Inte
r
Cel
l
In
te
rf
ere
n
c
e
Analysis
for
OF
DM
/
FB
MC
Sys
te
ms,
"
Proc.IEE
E
In
te
rnat
io
nal
Confe
renc
e
200
9
,
pp
.
598
-
602
,
2
009
.
[8]
Dirk
S.
W
al
dhau
ser
L
eona
rdo
G.
Bal
tar
et
a
l,
"
Co
mpa
rison
of
Fi
lter
B
ank
b
ase
d
Multi
carrie
r
Sys
t
em
s
with
OF
DM
,
"
Proc.
IE
EE Int
er
nati
onal
Con
fe
re
nce
,
A
PCCA
S
20
06
,
pag
es
5
,
200
6
.
[9]
Te
ro
Ih
al
a
ine
n
,
Ari
Vihola
in
en
e
t
al,
"
Filt
er
Bank
Based
Multi
mo
de
Multi
p
le
Ac
c
ess
Schem
e
for
Wi
re
le
ss
Uplink,
"
Proc.
17
th
Europ
ean
Signa
l
Pro
c
essing
Confe
ren
ce
,
Glasgow,
S
c
otl
and
,
pp.
1354
-
1358
,
2009
.
[10]
Mohamm
ed
Ab
o
-
Za
hhad
,
"
Curr
ent
sta
te
and
Future
Dire
ctions
of
Multi
ra
te
Fi
lter
Banks
and
th
ei
r
app
lications,
"
Digit
al
S
ignal P
roce
ss
ing
,
v
ol
.
1
3,
pp
.
495
-
518
,
2
003
.
[11]
Alphan
Sahin
,
Ismai
l
Guvenc
,
"
A
Survey
on
Multi
c
arr
i
er
Com
m
unic
a
ti
ons:
Prot
otype
Fil
te
rs,
Latti
c
e
Struc
ture
s
a
nd
Impl
ement
at
ion
As
pec
ts,
"
IE
EE
Comm
unic
ati
ons Surve
ys
&
Tuto
rials
,
v
ol
.
16
,
n
o.
3,
pp
.
1312
-
133
8
,
2014
.
[12]
Pila
r
M
art
in
–
Ma
rti
n,
Rober
t
Br
e
govic
et
a
l,
"
A
Gene
rali
z
ed
Window
Approac
h
for
Designing
Tra
nsmult
iplexe
rs,
"
IEE
E
Tr
ansacti
o
ns on
Circuits Sy
stems
,
II
,
v
ol
.
54
,
no
.
7,
pp
.
631
-
635,
2007
.
[13]
El
ef
the
rios
Kofi
dis
and
Di
mi
tr
i
os
Katse
li
s,
"
Im
prove
d
Int
erf
er
e
nce
Approxi
ma
t
ion
Method
for
Prea
mb
le
B
ase
d
Channe
l
Estimation i
n
FB
MC/OQ
AM
,
"
Proc.
19
th
European
S
ign
al
Proc
essing
C
onfe
renc
e
,
pp.
16
03
-
1607
,
2011
.
[14]
Tz
ung
Hw
ui
Lu
o,
Chih
Hao
Li
u
et
al,
"
Design
o
f
Channel
r
esil
i
e
nt
DF
T
Bank
Tr
ansc
ei
v
ers,
"
I
EEE
Tr
ansacti
ons
on
Signal
Proce
ss
in
g
,
pag
es
4
,
2005
.
[15]
T.
H.
St
it
z
,
Te
r
o
Iha
lainen
et
a
l,
"
Prac
ti
c
al
Iss
ues
in
Freque
n
c
y
Domai
n
Sync
hroniz
a
ti
on
for
Filt
er
B
ank
Bas
ed
Multi
carrie
r
Tr
a
nsmiss
ion,
"
Pro
c.
IE
EE Int
ernat
i
onal
Conf
ere
nce ISCCSP
2008
,
p
p.
411
-
416
,
2008
.
[16]
Guangyu
Wa
ng
,
W
ei
we
i
Zh
an
g
et
al,
"
Opti
m
iz
a
ti
on
and
I
m
ple
m
ent
a
ti
on
fo
r
the
Modifie
d
DF
T
Filt
er
Ba
nk
Multi
carrie
r
Mo
dula
ti
on
Sys
te
m
,
"
Journal
of
Co
mm
unic
ati
ons
,
v
ol.
8,
n
o.
10
,
pp
.
651
-
657
,
2013
.
[17]
[T.
H.
Stit
z
,
T
ero
Iha
la
in
en
et
a
l,
"
Miti
gation
of
Na
rrowband
Int
erf
e
ren
ce
in
Fil
te
r
B
ank
B
ase
d
Mul
ticar
ri
er
Sys
te
ms,
"
Proc.
I
EEE
In
t
S
ymp.
Circu
it
s an
d
Syste
ms
,
page
s
.
6
,
2006
.
[18]
A.
S.
Kang,
Vi
shal
Sharma,
R
enu
Vig,
"
Perfo
rma
nc
e
Ev
al
ua
tion
of
Prototype
FB
MC
Cognit
ive
Radi
o
und
er
constra
in
ts
of
L
number
of
Sen
sing
Samples
at
Vari
abl
e
SN
R
for
AWGN
and
Rayleigh
fad
in
g
Env
iron
me
n
t,
"
Inte
rnational
Jo
urnal
of Adv
an
c
ed
Sc
ie
n
ce
&
Te
chnol
ogy
(
IJ
AST
)
,
SE
RSC Australia
,
v
ol.
107
,
pp
.
1
-
20
,
2017
.
[19]
A.
S.
Kang
,
Vishal
Sharma,
Ren
u
Vig,
"
Perfor
m
anc
e
Eva
lu
at
ion
of
OF
DM
and
Prototype
FB
MC
-
OQ
AM
Cognit
i
ve
Radi
o
under
co
nstrai
nts
of
v
ar
ia
bl
e
Overl
apping
fac
tors
,
"
Int
ernati
onal
Journal
of
Eng
inee
ring
Re
search
&
Dev
el
opmen
t
,
v
o
l.
13(9),
pp.
71
-
91
,
2017
.
[20]
A.
S.
Kang
,
R
en
u
Vig,
"
Design
&
I
mpl
e
me
nt
at
i
on
of
Eff
ic
i
ent
Analysis
and
Sy
nthe
sis
QM
F
Ba
nk
for
Multi
ca
rr
ie
r
Cognit
ive
R
adi
o
Comm
u
nicati
on
,"
TEL
KOMNIKA
-
Telecomm
unication,
Computin
g,
Elec
troni
cs
&
Control
,
v
ol.
15
,
n
o.
2
,
pp
.
636
-
6
45
,
2017
.
[21]
A.
S.
Kang
,
Re
nu
Vig
,
"
Perform
ance
Analysis
of
Ne
ar
Perfe
c
t
R
ec
onstru
ct
ion
Fil
te
r
Bank
in
Cogni
ti
ve
R
adio
Envi
ronm
ent
,
"
I
nte
rnational
Jou
rnal
of Adv
an
ced Ne
twork
ing
an
d
Applications,
v
ol.
8(3)
,
pp.
3070
-
3083
,
2016
.
[22]
A.
S.
Kang
,
R
e
nu
Vig,
"
Perform
ance
Ana
lysis
of
FB
MC
-
CR
through
Opti
mi
z
at
ion
–
Tra
d
eof
f
bet
we
en
Var
io
us
Perform
ance
Pa
ram
e
te
rs
in
the
Cognit
ive
Rad
i
o
Environmen
t
,
"
T
ELECOMMUNICATIO
N
SYST
EMS,
Spring
er
,
vol.
63
,
p.
23
,
20
16.
Evaluation Warning : The document was created with Spire.PDF for Python.