Indonesi
an
Journa
l
of El
ect
r
ic
al Eng
inee
r
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
No.
3
,
Septem
ber
2021
, pp.
1493
~
1500
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.
v
23
.i
3
.
pp
1493
-
1500
1493
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
Eva
lu
ation
of
dif
ferent
q
uantiz
ation
res
olu
tion
lev
els
on
th
e
BER
pe
rfo
rmance
of
m
assive
MI
MO
s
yst
ems
un
der
diffe
rent
operatin
g
s
cena
rios
Hayder
Kh
aleel
AL
-
Q
aysi,
Tahreer
M
ahmo
od
,
Kh
alid
Awa
ad
Hum
ood
Depa
rtment
of
E
le
c
troni
c
Eng
ineeri
ng,
Coll
ege
of
Engi
n
ee
ring
,
U
nive
rsit
y
of
Di
y
a
la
,
Ira
q
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
hist
or
y:
Re
cei
ved
M
ay
1
,
2021
Re
vised
Ju
l
7
,
2021
Accepte
d
J
ul
14
,
2021
The
m
assive
MIM
O
sy
stem
is
one
of
the
m
a
in
technologies
in
the
fi
fth
gene
ra
ti
on
(5G)
of
te
l
ec
om
m
unic
ation
s
y
st
ems
,
al
so
rec
ogn
ized
as
a
high
l
y
la
rge
-
sc
ale
s
y
ste
m
.
Constant
l
y
in
m
assive
MIMO
sy
st
ems
,
the
base
stat
io
n
(BS)
is
provide
d
with
a
la
rg
e
num
ber
of
ant
ennas
,
and
thi
s
la
rg
e
num
ber
of
ant
enn
as
ne
ed
high
-
quan
ti
z
a
ti
on
r
esolut
ion
le
v
el
s
an
al
o
g
-
to
-
digi
t
al
conve
rt
ers
(AD
Cs).
In
thi
s
situ
a
ti
on,
the
r
e
wil
l
be
m
ore
power
consum
pti
on
and
har
dware
c
osts.
Thi
s
pape
r
pre
sents
the
sim
ula
ti
on
per
for
m
anc
e
of
a
suggested
m
et
hod
to
inve
stigate
and
anal
y
z
e
the
eff
e
ct
s
of
diffe
ren
t
quant
i
za
t
ion
re
soluti
on
le
v
el
s
of
AD
C
s
on
the
bit
e
rror
rat
e
(BER
)
per
form
anc
e
of
m
assive
MIM
O
sy
st
em
under
diffe
ren
t
oper
a
ti
n
g
sce
nar
ios
using
MA
TL
AB
software
.
Th
e
result
s
show
that
th
e
SNR
ex
ceeds
12
dB
ac
coun
ts
for
onl
y
0
.
001%
of
BE
R
signal
s
when
the
num
ber
of
a
nte
nnas
60
with
low
quant
i
za
t
ion
a
2
bit
s’
le
ve
ls
AD
Cs,
ap
proximate
l
y
.
Bu
t
when
the
ant
enn
a
num
ber
rises
to
300,
th
e
SNR
exc
ee
ds
12
dB
ac
count
s
for
al
m
ost
0.
01%
of
BER
tra
nsm
it
te
d
signa
ls.
Com
par
ably
with
the
BER
per
form
ance
of
high
qu
ant
i
z
at
ion
,
4
bit
s
-
qu
ant
i
za
t
ion
r
esolut
ion
l
evels
AD
Cs
with
th
e
sam
e
diffe
ren
t
a
nte
nnas
hav
e
a
s
li
ght
degr
adatio
n.
The
r
efo
re
,
th
e
num
ber
of
ant
enn
as
is
a
v
er
y
important
inf
lu
enc
e
fa
ct
or
.
Ke
yw
or
ds:
BER
perf
or
m
a
nce
Ma
ssive
MIM
O
Qu
a
ntiza
ti
on
reso
l
ution
le
ve
ls
AD
Cs
RF
chai
n
This
is
an
open
acc
ess
arti
cl
e
un
der
the
CC
BY
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Hayde
r
Kh
al
ee
l
AL
-
Qaysi
Dep
a
rtm
ent
of
Ele
ct
ro
nic
En
gi
neer
in
g
Coll
ege
of
En
gi
neer
in
g,
U
nive
rsity
of
Diya
la
Ba
quba
h,
Diya
la
Prov
i
nce,
Ir
a
q
Em
a
il
:
hay.k
ha
.82@
uodiya
la
.
edu.iq
1.
INTROD
U
CTION
Ma
ssive
m
ult
i
-
in
pu
t
m
ulti
-
ou
tp
ut
(MIMO
)
is
a
hig
hly
la
rg
e
-
scal
e
physi
cal
layer
syst
e
m
.
The
Ma
ssive
MIM
O
is
co
ns
ide
re
d
one
of
the
r
el
ia
ble
te
chn
ol
og
ie
s
in
fifth
-
gen
e
rati
on
(
5G)
te
le
com
m
un
ic
at
ion
syst
e
m
s
and
be
yond
to
im
pr
ov
e
sp
ect
ral
ef
fici
ency,
e
nerg
y
eff
ic
ie
ncy,
t
hro
ughput,
sec
ur
it
y
an
d
rob
ust
ness,
hard
war
e
c
omplexit
y,
cost,
a
nd
si
gnal
proc
essing
[1
]
-
[
9
].
The
m
ai
n
idea
in
the
m
assive
MIM
O
syst
em
is
to
us
e
a
la
r
ge
nu
m
ber
of
tra
nsm
itti
ng
a
nten
na
s
(
)
in
t
he
ba
se
sta
ti
on
(B
S)
t
han
the
num
ber
of
recei
ving
anten
nas
(
),
w
hich
re
pr
ese
nt
the
us
ers
,
to
pro
vid
e
sim
ult
aneous
ser
vice
within
the
c
overa
ge
a
rea
with
relat
ively
si
m
p
le
sign
al
s
proc
essing
[6
]
,
[
7
]
,
[
10
].
H
owe
ve
r,
the
us
e
of
a
la
rg
e
num
ber
of
ante
nnas
in
the
BS
of
m
assive
MIM
O
syst
e
m
increases
the
hardw
a
re
c
os
t
an
d
com
plexity
,
a
nd
powe
r
co
nsum
ption
of
the
ci
rcu
it
com
po
ne
nts
in
the
rad
i
o
fr
e
quency
(RF
)
c
ha
in
that
c
onsis
ts
of
sig
nal
m
i
xer
s
,
powe
r
a
m
pl
ifie
rs,
ded
i
cat
ed
filt
ers,
a
nd
hi
gh
-
quantiz
at
ion
reso
l
ution
le
ve
l
analo
g
-
to
-
digi
ta
l
con
ve
rters
(ADCs)
[
11
]
,
[
12
].
Co
ns
e
qu
e
ntly
,
the
increa
sed
hard
war
e
c
os
t
s
an
d
po
wer
consum
ption
s
represe
nt
a
substanti
al
obsta
cl
e
in
the
pr
a
ct
ic
al
i
m
ple
m
entat
io
n
of
the
m
assive
MIM
O
sys
te
m
s
[
13
].
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
2
02
1
:
14
93
-
15
00
1494
The
powe
r
c
onsu
m
ed
by
high
-
qua
ntiza
ti
on
r
esolutio
n
le
vel
AD
Cs
re
pr
ese
nts
the
esse
ntial
par
t
of
the
total
power
c
onsu
m
ed
by
the
oth
er
ci
rcu
it
c
om
po
ne
nts
of
the
RF
chai
n
[
14
]
,
[
1
5
].
In
paral
le
l,
the
util
izati
on
of
high
-
quanti
zat
ion
res
ol
ution
le
ve
l
A
DC
s
co
ntribute
s
s
ign
ific
a
ntly
to
increasin
g
t
he
hard
war
e
c
ost
s
and
com
plexiti
es
and
reduci
ng
qu
a
ntiza
ti
on
error
s
in
the
m
assive
MIM
O
syst
e
m
s
[1
6
]
-
[
1
8
].
The
pow
e
r
consum
ption
in
the
A
DCs
de
creases
e
xpone
ntial
ly
with
quantiz
at
ion
re
sol
ution
le
vels
(
quanti
zat
io
n
bit
s)
an
d
li
near
ly
with
s
a
m
pling
f
requ
encies
(sam
pling
rates)
[1
9
].
Partic
ularly
,
AD
Cs
represe
nt
the
m
ai
n
par
ts
of
receiver
s
in
the
m
assive
MIM
O
syst
em
s
to
c
onve
rt
the
analo
g
si
gn
al
s
into
dig
it
al
sig
nals
th
rou
gh
s
a
m
pling
and
qua
ntiza
ti
on
processes
a
nd
pr
e
par
i
ng
the
m
for
s
ub
se
que
nt
di
gital
sign
a
l
processi
ng
op
erati
on
s
[
20
].
In
t
he
li
te
rature,
m
any
inv
est
igati
on
s
w
ork
cond
ucted
to
s
tud
y
the
e
ff
ect
of
dif
fer
e
nt
quantiz
at
io
n
reso
l
ution
le
ve
ls
of
AD
Cs
on
t
he
perf
orm
ance
of
m
assive
M
IMO
syst
em
s
un
de
r
diff
e
re
nt
operati
ng
scenari
os
,
s
uc
h
as:
usi
ng
a
diff
e
re
nt
num
ber
of
a
nten
nas,
var
i
ou
s
m
od
ul
at
ion
schem
es,
an
d
sig
nal
sy
m
bo
ls
detect
or
s
.
In
[1
4
],
the
a
utho
r
s
in
vestigat
ed
the
ef
fect
of
3
-
to
-
12
bits
-
qua
ntiza
ti
on
res
olu
ti
on
le
vels
A
DCs
on
the
upli
nk
e
nergy
ef
fici
e
ncy
(
EE)
in
the
m
assive
MIM
O
syst
e
m
,
with
tim
e
div
isi
on
dupl
exin
g
(
TD
D)
m
od
e
and
a
diff
e
rent
nu
m
ber
of
a
nten
nas
at
the
BS
us
ing
li
ne
ar
zer
o
-
forcin
g
(ZF)
a
nd
ze
r
o
-
forcin
g
su
c
c
essive
interfe
ren
ce
ca
ncell
at
ion
(
ZF
SI
C)
si
gn
al
sy
m
bo
ls
detect
ors,
with
both
pe
rf
e
ct
a
nd
im
per
fect
c
hannel
sta
te
inf
or
m
at
ion
(C
SI
)
.
T
heir
res
ul
ts
showe
d
that
the
c
on
s
um
ed
powe
r
is
si
gn
i
f
ic
ant
in
t
he
m
assive
MIM
O
s
yste
m
wh
e
n
us
in
g
l
ow
(
3bit
s)
-
quan
ti
zat
ion
res
olu
t
ion
le
vels
ADC
s
due
to
the
i
ncr
ease
in
qua
ntiza
ti
on
e
rrors
,
w
hich
needs
to
c
om
pen
sat
e
with
a
la
rg
e
num
ber
of
ante
nn
a
s
at
the
BS.
In
ad
diti
on
,
t
he
us
e
of
the
pe
rf
ect
CS
I
-
bas
e
d
li
near
ZFS
IC
sign
al
sym
bo
ls
detect
or
is
abl
e
to
i
m
pr
ov
e
t
he
EE
of
the
s
yst
e
m
m
or
e
th
an
the
pe
rf
ect
CSI
-
base
d
li
near
ZF
sign
al
sym
bo
ls
detect
or.
T
he
aut
hors
i
n
[1
6
]
in
vestigat
e
d
the
r
ole
of
usi
ng
low
-
qua
ntiza
ti
on
reso
l
ution
le
ve
ls
AD
Cs
wit
h
the
li
near
m
ini
m
u
m
m
ea
n
s
quare
d
e
rror
(LMM
SE
)
and
no
nlinear
la
tt
ic
e
reducti
on
-
base
d
su
cce
ssive
in
te
rf
ere
nce
ca
nc
el
la
ti
on
(LR
-
S
I
C)
sign
al
sym
bo
ls
detect
ors
in
i
m
pr
ov
i
ng
t
he
E
E
in
the
m
assive
MIM
O
syst
em
s
unde
r
dif
fer
e
nt
operati
ng
sc
enar
i
os
.
T
heir
resu
lt
s
s
how
ed
that
the
bit
er
r
or
rate
(BER)
pe
r
form
ance
usi
ng
the
LR
-
SI
C
si
gn
al
sy
m
bo
ls
detect
or
s
-
ba
sed
recei
ver
s
is
bette
r
than
t
hat
of
LMM
SE
sign
al
sym
bo
ls
detect
or
s
-
base
d
receive
r
s
an
d
that
the
EE
of
the
syst
em
dep
e
nds
on
the
nu
m
ber
of
ant
enn
a
s,
cel
l
siz
e,
and
sign
al
pr
ocessin
g
eff
ic
ie
ncy
em
plo
ye
d
at
the
BS.
In
a
dd
it
io
n,
the
a
uthors
s
uggeste
d
that
f
or
high
EE,
sign
al
sym
bo
ls
detect
or
s
sh
oul
d
be
care
fu
ll
y
sel
ect
ed
and
acc
ordin
g
to
a
n
ap
pro
pr
ia
t
e
op
erati
ng
sce
nar
i
o.
In
orde
r
to
im
pr
ove
EE
,
they
al
so
sug
gested
us
in
g
m
edium
(6
-
to
-
7
bits)
-
quantiz
at
io
n
res
olu
ti
on
le
vels
AD
Cs
to
achie
ve
a
tra
de
-
off
betwee
n
powe
r
c
onsu
m
ed
a
nd
no
nline
ar
distor
ti
on.
More
ov
e
r,
S
ha
o
et
a
l
,
[21]
propose
d
to
incr
ease
the
up
li
nk
EE
of
the
m
as
sive
MIM
O
syst
e
m
s
un
der
diff
e
re
nt
op
e
ra
ti
ng
scena
rios,
based
on
the
us
e
of
low
(1
bit)
-
qu
a
ntiza
ti
on
reso
l
ution
l
evels
A
DCs,
it
erati
ve
detect
ion
a
nd
decodin
g
(
IDD)
te
ch
nique,
and
li
near
-
low
res
olu
ti
on
awar
e
MM
SE
(LRA
-
MM
SE
)
signa
l
sy
m
bo
ls
detec
tor.
Thei
r
res
ul
ts
yi
el
ded
a
hi
gh
EE
gain
a
nd
im
pr
ov
e
d
BER
perform
a
nce
for
the
m
assiv
e
MIM
O
syst
em
.
In
[
22
],
the
a
uthors
i
nv
e
sti
ga
te
d
the
EE
in
the
m
assive
MIM
O
syst
em
us
in
g
ar
bitrar
y
(1,
2,
and
i
nf
init
y
bi
ts)
-
qu
a
ntiza
ti
on
res
olu
ti
on
le
vels
A
DCs,
s
uperim
po
sed
pilots
(SP)
te
ch
nolo
gy,
an
d
L
MM
SE
and
m
axi
m
u
m
rati
o
com
bin
ing
(MRC
)
sig
nal
sy
m
bo
ls
de
te
ct
or
s
un
der
diff
e
re
nt
op
e
ra
ti
ng
scena
rios.
Their
resu
lt
s
sho
wed
that
the
pilots
need
ed
to
sp
e
ci
fy
a
gr
eat
er
powe
r
i
n
the
m
assive
MIM
O
syst
e
m
with
hig
he
r
qu
a
ntiza
ti
on
r
esolutio
n
le
ve
ls
A
DCs
or
with
a
la
rger
num
ber
of
a
nten
nas
at
t
he
BS.
In
a
dd
it
ion
,
SP
te
chnolo
gy
is
su
pe
rio
r
to
ti
m
e
-
m
ulti
plexed
pilots
(TP)
t
echnolo
gy
for
low
(1
bit)
-
quantiz
at
io
n
r
es
olu
ti
on
le
vels
A
DCs,
a
nd
SP
te
ch
no
l
ogy
is
m
or
e
s
uitable
w
he
n
us
in
g
higher
-
qua
ntiza
ti
on
res
ol
ution
le
vels
A
DCs
an
d
the
use
d
num
ber
of
a
nten
nas
in
the
BS
is
sm
al
l.
Fu
rt
her
m
or
e,
t
he
auth
ors
in
[2
3
]
inv
est
igat
ed
the
eff
ect
s
of
us
in
g
low
(1
-
to
-
3
bits)
-
qu
a
ntiza
ti
on
reso
l
ution
le
ve
ls
AD
Cs
an
d
the
num
ber
of
anten
nas
avai
la
ble
at
the
base
sta
ti
on
us
i
ng
LMM
SE
s
ign
al
sy
m
bo
ls
detect
or
in
i
m
pr
ovi
ng
the
upli
nk
sp
ect
r
um
eff
iciency
(S
E)
of
the
m
assive
MIM
O
syst
e
m
unde
r
diff
e
re
nt
opera
ti
ng
sce
nar
i
os
.
Their
res
ults
pro
ved
to
im
pr
ove
SE
by
m
or
e
t
han
90%
us
ing
low
(3
bits)
-
qu
a
ntiza
ti
on
re
so
luti
on
le
vels
AD
Cs.
In
ad
di
ti
on
,
the
SE
im
pr
ov
e
d
wh
e
n
us
ing
1
bit
-
qu
antiz
at
ion
res
ol
ution
le
vels
AD
Cs
with
an
inc
rea
se
in
the
num
ber
of
anten
nas
at
the
base
sta
ti
on
.
In
[2
4
],
the
auth
or
s
pro
pose
d
a
novel
hybr
i
d
i
nd
e
pe
nd
e
nt
c
om
po
nen
t
a
naly
sis
(HICA
)
sig
nal
sym
bo
ls
de
te
ct
or
in
orde
r
to
im
pr
ove
the
SE
and
reli
abili
ty
of
t
he
MIM
O
syst
e
m
un
de
r
diff
e
re
nt
opera
ti
ng
sce
nar
i
os
.
Their
res
ults
s
howe
d
that
t
he
HI
C
A
sign
al
sym
bo
ls
detect
or
sho
w
ed
i
m
pr
ovem
e
nt
in
the
BER
and
MSE
perf
or
m
ances,
but
it
was
ineff
ic
ie
nt
and
fail
ed
to
reduc
e
the
pea
k
to
a
ver
a
ge
power
r
at
e
(P
A
PR)
rat
e
com
par
ed
to
oth
e
r
sta
te
-
of
-
t
he
-
a
rt
sig
nal
sym
bo
ls
detect
or
s
s
uch
as
ICA
a
nd
wa
velet
de
no
isi
ng
ICA
(
WDICA
).
In
[2
5
],
the
au
thors
inv
e
sti
ga
te
d
the
j
oi
nt
eff
ect
s
of
the
low
(1
bit)
-
qu
a
ntiza
ti
on
res
ol
ution
le
vels
AD
Cs
an
d
the
hard
war
e
im
pairm
ents
in
the
m
assive
MIM
O
syst
em
,
us
in
g
a
novel
pro
posed
al
te
rn
at
in
g
directi
on
m
et
ho
d
of
m
ulti
pli
ers
(
A
DMM)
-
base
d
sig
nal
s
ym
bo
ls
detect
or,
qua
dr
at
ur
e
phase
-
s
hift
ke
yi
ng
(Q
P
S
K)
m
od
ul
at
ion
,
a
nd
different
nu
m
ber
s
of
a
nten
nas
at
the
BS.
T
he
re
su
lt
s
of
thei
r
inv
e
sti
gation
s
s
how
e
d
that
the
AD
M
M
-
base
d
sig
na
l
sy
m
bo
ls
detect
or
achie
ve
d
bette
r
pe
rfor
m
ance
tha
n
the
oth
e
r
sta
te
-
of
-
t
he
-
a
rt
li
near
si
gn
al
sym
bo
ls
detect
ors
in
te
rm
s
of
i
m
pr
o
ving
BER
pe
rfor
m
ance
a
nd
the
gai
n
with
an
i
ncr
ease
in
the
nu
m
ber
of
ante
nn
a
s
a
nd
us
e
rs
in
the
m
assive
MIM
O
syst
em
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Evalu
atio
n and
m
e
as
ure
me
nt
of the effe
ct
s
of d
if
fe
rent
quan
ti
za
ti
on
re
so
l
ution
…
(
H
ay
der K
ha
le
el
AL
-
Qaysi
)
1495
In
this
pap
e
r,
t
he
m
ai
n
ob
j
ect
ive
is
to
evalu
at
e
and
m
easure
the
influ
e
nce
s
of
diff
e
re
nt
quantiz
at
io
n
reso
l
ution
le
ve
ls
of
analo
g
-
to
-
dig
it
al
conv
erters
on
the
BER
eff
ect
ive
ness
in
the
m
assive
MIM
O
syst
e
m
unde
r
dif
fer
e
nt
op
e
rati
ng
sce
na
rios
us
i
ng
the
LMM
SE
sign
a
l
sy
m
bo
ls
detect
or
.
T
he
rem
a
ind
e
r
of
this
pa
per
is
arr
a
ng
e
d:
T
he
desc
riptio
ns
of
t
he
qu
a
ntize
d
m
assive
MIM
O
syst
em
e
qu
at
io
ns
with
the
L
MM
SE
s
ign
a
l
sy
m
bo
ls
detect
or
are
prese
nt
ed
in
Sect
ion
2.
Ma
tl
ab
sim
ula
ti
on
res
ults
that
exp
la
in
the
infl
uen
c
e
of
qu
a
ntiza
ti
on
re
so
luti
on
le
vels
analog
-
to
-
di
gital
con
ve
rters
on
the
BER
ef
fici
ency
in
the
m
assive
m
ult
i
ple
-
input,
m
ulti
ple
-
outp
ut
syst
em
are
pre
sente
d
in
Sect
io
n
3.
Fi
nally
,
Sect
io
n
4
c
on
cl
ud
e
s
thi
s
pa
per.
2.
QUANTIZ
ED
M
AS
SI
VE
M
IMO
WITH
LMM
SE
SI
GNAL
S
YM
B
OL
S
DETE
CT
O
R
First,
in
this
s
ect
ion
,
we
pr
e
sent
the
quant
iz
ed
m
assive
MIM
O
syst
em
m
od
el
,
w
her
e
dif
fer
e
nt
-
qu
a
ntiza
ti
on
re
so
luti
on
le
vels
AD
Cs
a
re
us
ed
in
the
receive
r
at
the
BS.
T
he
n
we
prese
nt
the
LMM
SE
det
ect
or
to
detect
t
he
tr
ansm
itted
sig
na
l
sy
m
bo
ls
vec
tor
by
the
us
e
rs
at
the
BS
recei
ver.
2.1.
Q
uantiz
ed
ma
ssive
MI
MO
mo
del
w
ith
ADCs
We
co
ns
ide
r
that
the
up
li
nk
of
the
m
assive
m
ulti
us
er
(MU)
MIM
O
s
yst
e
m
,
as
sh
own
in
Figure
1,
has
a
total
M
anten
nas
at
the
BS
and
sim
ultaneo
us
ly
ser
ves
K
sin
gle
-
a
nten
na
us
e
rs,
w
here
M
≫
K
≫
1
.
Th
us
,
the
cha
nnel
m
a
trix
(
H
)
co
nnect
ing
t
he
c
hannel
coeffic
ie
nts
bet
ween
K
si
ng
le
-
anten
na
us
ers
a
nd
M
a
nten
nas
at
the
BS
is
giv
e
n
b
y
(1)
[7
]
,
[
2
6
]
,
[
2
7
]
,
H
=
[
h
11
⋯
h
1j
⋮
⋱
⋮
h
i1
⋯
h
ij
]
(1)
wh
e
re
H
∊
ℂ
M
×
K
a
nd
the
el
em
ent
h
ij
re
pr
ese
nt
t
he
c
ha
nn
el
at
te
nuat
ion
c
oeffici
ent
from
the
ith
r
ecei
ving
anten
na
to
t
he
j
th
tra
ns
m
it
te
r
anten
na.
T
he
c
hannel
m
at
rix
(H)
in
the
m
ass
ive
MIM
O
syst
e
m
pr
ovide
s
us
wit
h
the
necessa
ry
knowle
dge
ab
out
the
cha
nn
el
sta
te
info
rm
at
i
on
(CSI)
an
d
thu
s
helps
s
olve
m
any
pr
oble
m
s
that
we
face
in
wir
el
ess
com
m
un
ic
at
ion
syst
em
s
su
c
h
as
path
losses,
fa
ding,
sh
a
dowing,
sc
at
te
ring
,
di
ffra
ct
ion
s,
and
pr
op
e
rtie
s
of
the
c
ha
nn
el
[27
]
,
[
28]
.
Figure
1.
U
plin
k
syst
em
m
od
el
of
qua
ntize
d
m
assive
MU
-
MIM
O
The
i
nfor
m
at
ion
sig
nals
sym
bo
ls
(
a
K
)
tra
ns
m
it
t
ed
by
al
l
the
K
sing
le
-
ante
nna
us
e
rs
a
re
e
ncoded
fir
st
by
the
c
ha
nn
e
l
encode
rs
a
nd
the
n
m
odulate
d
acco
r
ding
to
the
gi
ven
m
od
ulati
on
sc
hem
es
to
the
sign
al
sy
m
bo
ls
(
x
K
).
Th
ese
m
od
ulate
d
sign
al
sym
bo
ls
(
x
K
)
are
tra
ns
m
itted
t
hroug
h
the
c
hannels.
T
he
M
ante
nn
a
s
at
the
BS
re
cei
ve
these
m
od
ul
at
ed
sig
nal
sy
m
bo
ls
(
x
K
)
c
orr
up
te
d
by
the
channel
e
ff
ect
s
an
d
no
ise
.
The
corrupted
an
d
unquantiz
ed
si
gn
al
sym
bo
ls
ve
ct
or
receive
d
in
the
base
sta
ti
on
is
gi
ven
by
(
2)
,
y
=
Hx
+
n
(2)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
2
02
1
:
14
93
-
15
00
1496
wh
e
re
x
=
[
x
1
,
x
2
,
x
3
,
…
…
.
,
x
K
]
T
is
the
vect
or
of
the
tra
ns
m
i
tt
ed
sign
al
sym
bo
ls
f
ro
m
al
l
use
rs,
y
=
[
y
1
,
y
2
,
y
3
,
…
…
.
,
y
M
]
T
i
s
the
vect
or
of
the
recei
ved
sign
al
sym
bo
ls
,
an
d
n
is
the
a
dd
it
ive
w
hite
Gau
s
sia
n
nose
vecto
r.
T
he
m
od
el
giv
e
n
in
(
2)
re
prese
nts
the
m
assive
MIM
O
syst
e
m
mo
del
th
rou
gh
wh
ic
h
it
is
po
s
sible
to
fin
d
the
cha
nn
el
m
at
rix
(H)
ba
sed
on
the
pro
vid
e
d
in
f
or
m
ation
a
bo
ut
t
he
x
an
d
y
si
gn
al
s
ym
bo
ls
vect
or
s
.
G
e
n
e
r
a
l
l
y
,
t
h
e
s
t
r
u
c
t
u
r
e
of
A
D
C
s
c
o
n
s
i
s
t
s
of
f
i
l
t
e
r
s
,
s
a
m
p
l
i
n
g
a
n
d
h
o
l
d
c
i
r
c
u
i
t
s
,
a
n
d
q
u
a
n
t
i
z
e
r
s
.
P
a
r
t
i
c
u
l
a
r
l
y
,
t
he
q
u
a
n
t
i
z
e
r
s
r
e
p
r
e
s
e
n
t
t
h
e
c
o
r
e
s
of
A
D
C
s
t
h
a
t
p
e
r
f
o
r
m
t
h
e
q
u
a
n
t
i
z
a
t
i
o
n
p
r
o
c
e
s
s
e
s
for
t
h
e
r
e
c
e
i
v
e
d
s
i
g
n
a
l
s
y
m
b
ol
s
v
e
c
t
o
r
(
y
)
g
i
v
e
n
in
(2)
i
n
t
o
the
q
u
a
n
t
i
z
e
d
s
i
g
n
a
l
s
y
m
b
o
l
s
v
e
r
s
i
o
n
v
e
c
t
o
r
(r)
[1
4
]
,
[
1
8
].
As
s
h
o
w
n
in
F
i
g
u
r
e
1,
by
u
t
i
l
i
z
a
t
i
o
n
of
two
b
b
i
t
-
q
u
a
n
t
i
z
a
t
i
o
n
r
e
s
o
l
u
t
i
on
l
e
v
e
l
s
A
D
C
s
for
t
h
e
r
e
a
l
a
nd
i
m
a
g
i
n
a
r
y
p
a
r
t
s
of
t
h
e
c
om
pl
e
x
r
e
c
e
i
v
e
d
s
i
g
n
a
l
s
y
m
b
ol
s
v
e
c
t
o
r
(
y
)
,
t
h
e
o
u
t
c
om
e
q
u
a
n
t
i
z
e
d
s
i
g
n
a
l
s
y
m
b
ol
s
v
e
c
t
o
r
(r)
is
g
i
v
e
n
by
(3),
r
=
(
y
)
=
[
ℛ
e
(
y
)
+
j
Ι
m
(
y
)
]
(3)
wh
e
re
(.)
de
s
cribes
the
qu
a
ntiza
ti
on
pr
oc
ess
for
the
re
al
(
ℛ
e
(
y
)
)
and
im
aginar
y
(
Ι
m
(
y
)
)
pa
rts
of
the
com
plex
receiv
ed
si
gn
al
sym
bo
ls
vector
(y)
,
and
it
is
a
non
-
li
near
process.
The
cha
racteri
sti
cs
of
3
bits
reg
ular
m
id
as
cend
i
ng
qu
a
ntiza
ti
on
proces
s
es
are
sh
ow
n
in
Fig
ur
e
2.
Assum
ing
that
(.)
re
presenti
ng
b
bits
re
gula
r
m
id
ascend
i
ng
qu
a
ntiz
at
ion
proce
sse
s
ac
hieve
d
by
the
qu
a
ntize
r
with
N
=
2
b
res
olu
ti
on
le
vels
of
t
he
A
D
Cs.
C
on
se
que
nt
ly
,
the
quanti
zer
pe
rfor
m
s
the
co
nversi
on
process
to
the
real
pa
rt
of
the
input
sig
nal
sy
m
bo
ls
vect
or
(y)
into
a
re
al
evaluated
outp
ut
sig
nal
sy
m
bo
ls
vecto
r
(r),
w
he
re
r
=
[
r
1
,
r
2
,
r
3
,
…
…
,
r
]
a
nd
s
=
1
,
2
,
3
,
…
,
N
.
T
he
en
dpoi
nts
of
the
in
put
inte
r
val
(
y
s
)
are
giv
e
n
by
(
4),
y
s
=
{
−
∞
for
s
=
1
(
−
N
2
⁄
−
1
+
s
)
∆
f
or
s
=
2
,
3
,
4
,
…
…
,
N
+
∞
for
s
=
N
+
1
(4)
wh
e
re
∆
re
prese
nts
the
per
i
od
siz
e
of
qu
a
ntiz
at
ion
process
.
The
pro
duced
sign
al
sym
bo
ls
val
ues
(
r
s
)
of
the
qu
a
ntize
r
a
re
gi
ven
by
(5),
r
s
=
(
−
N
2
−
1
2
+
s
)
∆
for
s
=
2
,
3
,
4
,
…
…
,
N
(5)
Figure
2.
Stai
r
case
re
pr
ese
nta
ti
on
of
a
th
ree
bits
re
gu
la
r
m
i
d
asce
ndin
g
quantiz
at
ion
proc
esses
2.2.
L
MMSE
sign
al
s
ymbol
s
detec
t
or
As
a
res
ult
of
the
interuse
r
interfe
ren
ce
a
m
on
g
the
us
e
r
s,
the
sig
nal
sy
m
bo
ls
detect
or
is
use
d
to
evaluat
e
the
ve
ct
or
of
the
t
ransm
it
te
d
sign
al
sy
m
bo
ls
(
x
)
from
the
vector
of
t
he
recei
ved
sign
al
sym
bo
ls
(
y
)
.
In
this
wor
k,
we
hav
e
a
dopt
ed
the
LMM
S
E
sig
nal
sym
bo
ls
detect
or
fro
m
the
oth
e
r
s
ta
te
-
of
-
the
-
art
detect
or
s
because
of
its
a
dv
a
ntage
s,
i
nclud
i
ng
[7
]
,
[
1
6
],
It
has
lo
wer
co
m
pu
ta
ti
on
al
co
m
plexit
y.
It
has
noise
im
pro
vem
ent.
It
outpe
rfor
m
s
the
ZF
sig
nal
s
ym
bo
l
detect
or
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Evalu
atio
n and
m
e
as
ure
me
nt
of the effe
ct
s
of d
if
fe
rent
quan
ti
za
ti
on
re
so
l
ution
…
(
H
ay
der K
ha
le
el
AL
-
Qaysi
)
1497
Im
ple
m
entation
the
li
nea
r
sig
nal
detect
or
by
us
in
g
M×
K
ma
trix
(
A
)
to
the
receive
d
sig
na
l
sy
m
bo
ls
vecto
r
(y
),
t
he
transm
itted
sig
nal
sym
bo
ls
ve
ct
or
is
giv
e
n
by
(6),
x
̂
=
A
H
y
=
A
H
(
Hx
+
n
)
(6)
wh
e
re
x
̂
=
[
x
̂
1
,
x
̂
2
,
x
̂
3
,
…
…
.
,
x
̂
K
]
T
is
K
×
1
si
gnal
cha
racters
ve
ct
or
co
ns
ist
s
of
data
in
form
at
ion
st
ream
of
K
sing
le
-
a
nten
a
us
er
.
Em
plo
yi
ng
the
LMM
S
E
sig
nal
sym
bo
ls
detect
or
by
def
i
ning
t
he
detect
io
n
m
atr
ix
data
stream
(A)
as
gi
ven
by
(7),
A
=
H
[
H
+
n
2
x
2
I
K
]
−
1
(7)
wh
e
re
x
2
a
nd
n
2
are
the
tra
ns
m
itted
ori
gi
nal
sig
nal
a
nd
t
he
receive
d
noise
sig
na
l
var
ia
nces
corres
pondin
gly
.
For
a
qua
ntize
d
m
assive
MU
-
MIM
O
syst
e
m
,
we
hav
e
the
Kth
el
em
ent
of
x
̂
as
gi
ven
by
(
8)
,
x
̂
=
A
H
r
=
A
H
(
y
)
=
A
H
(
Hx
+
n
)
=
a
K
H
h
K
x
K
+
∑
a
K
H
h
i
x
i
K
i
≠
k
+
a
K
H
n
(8)
wh
e
re
the
first
te
rm
in
(8)
re
presents
the
de
s
ired
sig
nal,
the
second
te
rm
r
epr
ese
nts
the
inter
us
er
inter
fe
ren
ce
,
and
the
t
hir
d
te
rm
rep
rese
nts
t
he
no
ise
.
3.
SIMULATI
O
N
RESU
LT
S
In
t
his
sect
io
n
of
the
pa
per,
MATLAB
sim
ulati
on
s
are
a
ppli
ed
to
ex
plai
n
the
ef
fects
of
di
ff
e
ren
t
qu
a
ntiza
ti
on
re
so
luti
on
le
vels
of
a
nalo
g
-
to
-
dig
it
al
co
nvert
er
(
A
DCs)
on
the
BER
pe
rfor
m
ance
in
m
assive
m
ul
ti
ple
input
m
ul
ti
ple
ou
t
p
ut
(MIM
O)
syst
em
s
un
der
di
ff
e
ren
t
operati
ng
scenari
os
.
3.1.
BER
per
fo
rm
an
ce
usin
g
di
ff
eren
t
sc
h
emes
of
m
odu
lation
Figures
3
a
nd
4
show
the
BE
R
per
f
or
m
ances
us
in
g
two
di
ff
e
ren
t
schem
es
of
quad
ratu
r
e
a
m
plit
ud
e
m
od
ulati
on
,
4
QAM
a
nd
16
QA
M,
an
d
dif
f
eren
t
qu
antiz
at
ion
res
olu
ti
on
l
evels
of
AD
Cs
an
d
LMM
SE
sign
a
l
sy
m
bo
ls
detect
or
ha
ve
been
us
e
d
at
the
B
S.
Sim
ulati
on
s
res
ults
are
im
plem
ented
with
80
ante
nn
as
at
the
transm
itter
an
d
ser
ving
5
ante
nn
a
s
as
a
us
e
r’s.
T
he
ef
fects
of
dif
fer
e
nt
quantiz
at
ion
resol
ution
le
vels
of
A
DCs
on
the
BER
pe
rfor
m
ance
are
sh
ow
n
as
a
sig
nal
to
noise
rat
io
(SNR)
get
bi
gg
e
r,
the
re
su
l
t
dem
on
strat
e
t
hat
the
BER
de
gr
a
dation
ex
pone
ntial
ly
in
a
sim
i
la
r
way
for
bo
t
h
schem
es
of
m
od
ulati
on
s
w
hen
usi
ng
the
sam
e
qu
a
ntiza
ti
on
re
so
luti
on
le
vels
of
A
DCs.
Ne
ve
rtheless,
the
r
esults
il
lustrate
that
by
i
ncr
ea
sing
the
quanti
zat
ion
reso
l
ution
le
ve
ls
of
A
DCs,
the
BER
pe
rfor
m
ance
beco
m
e
s
m
or
e
ideal
.
F
i
g
u
r
e
3.
B
E
R
p
e
r
f
o
r
m
a
n
c
e
u
s
i
n
g
d
i
f
f
e
r
e
n
t
q
u
a
n
t
i
z
a
t
i
o
n
r
e
s
o
l
u
t
i
o
n
l
e
v
e
l
s
of
A
D
C
s
,
=
8
0
,
=
5,
a
n
d
4
Q
A
M
In
orde
r
to
stu
dy
highe
r
-
orde
r
m
od
ulati
on
s
,
we
obse
rv
e
t
hat
with
us
in
g
2
an
d
3
bits
-
qu
a
ntiza
ti
on
reso
l
ution
le
ve
ls
of
AD
Cs
,
the
BER
sh
ow
i
ng
degrade
s,
wh
il
e
us
in
g
the
conditi
on
4
bits
-
qu
a
nt
iz
at
ion
reso
l
ution
le
ve
ls,
the
BER
beco
m
e
s
acce
ptab
le
.
Furthe
rm
or
e,
ha
ving
5
bits
-
qua
ntiza
ti
on
reso
l
ution
le
ve
ls
and
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
2
02
1
:
14
93
-
15
00
1498
ov
e
r
head,
the
BER
eff
ect
ive
ness
is
ve
ry
cl
ose
to
the
in
finity
bits
-
qua
ntiza
ti
on
r
esol
ution
le
vels.
T
he
ou
t
com
es
sh
ow
wh
e
n
us
i
ng
highe
r
orde
r
of
m
od
ulati
on
(4
Q
AM
a
nd
16
Q
AM),
t
he
BER
increas
es
corres
pondin
gly
for
sp
eci
fic
quanti
zat
ion
res
olu
ti
on
le
vels
of
A
DCs.
Co
ns
eq
ue
ntly
,
hig
he
r
quantiz
at
io
n
res
olu
ti
on
le
vels
of
A
DCs
are
ob
li
gat
or
y
for
a
dv
a
nce
d
orde
r
m
od
ulati
ons
to
at
ta
in
the
equ
i
valent
perf
or
m
ance.
F
i
g
u
r
e
4.
B
E
R
p
e
r
f
o
r
m
a
n
c
e
u
s
i
n
g
d
i
f
f
e
r
e
n
t
q
u
a
n
t
i
z
a
t
i
o
n
r
e
s
o
l
u
t
i
o
n
l
e
v
e
l
s
of
A
D
C
s
,
=
8
0
,
=
5,
a
n
d
16
QAM
3.2.
BER
per
fo
rm
an
ce
usin
g
di
ff
eren
t
nu
mber
of
anten
na
s
Furthe
r
sim
ul
at
ion
is
acc
ompli
sh
e
d
to
dete
rm
ine
BER
per
f
or
m
ance
as
a
functi
on
of
qu
a
ntiza
ti
on
reso
l
ution
le
ve
ls
of
AD
Cs
with
a
diff
e
re
nt
nu
m
ber
of
BS
ante
nn
as
,
a
nd
LMM
SE
si
gnal
sy
m
bo
ls
detect
or
hav
e
bee
n
us
e
d.
We
exam
ine
a
con
sta
nt
nu
m
ber
of
us
e
rs
as
a
receiver
=
5,
an
d
the
num
ber
of
ante
nn
as
at
the
BS
will
be
ta
ken
from
60
to
300
ante
nna
s
.
Fig
ures
5
an
d
6
ex
plain
t
ha
t
increasin
g
t
he
num
ber
of
ant
enn
a
s
at
the
tra
ns
m
itt
er
will
decr
eas
e
the
BER
re
gressi
on
cau
sed
by
the
low
qu
a
ntiza
ti
on
res
olu
ti
on
le
vels
of
AD
C
s
.
Figure
5
il
lustrate
a
2
bits
-
quantiz
at
io
n
res
olu
ti
on
le
vels
AD
Cs
with
th
r
ee
differe
nt
nu
m
ber
of
the
a
nt
enna
(60,
120
,
an
d
300)
has
an
obvi
ous
cha
ng
e
with
the
incr
ease
the
nu
m
ber
of
a
nten
nas
.
Com
par
ably
,
with
F
i
g
u
r
e
6,
t
h
e
B
E
R
p
e
r
f
o
r
m
a
n
c
e
of
4
b
i
t
s
-
q
u
a
n
t
i
z
a
t
i
o
n
r
e
s
o
l
u
t
i
o
n
l
e
v
e
l
s
A
D
C
s
w
i
t
h
t
h
e
s
a
m
e
di
f
f
e
r
e
n
t
a
nt
e
n
n
a
s
h
a
v
e
a
s
l
i
g
h
t
d
e
g
r
a
d
a
t
i
o
n
.
W
h
e
r
e
f
o
r
e
,
t
h
e
h
i
gh
e
r
n
u
m
b
e
r
of
a
n
t
e
n
n
a
s
at
t
h
e
BS
is
prop
o
s
e
d
to
r
e
c
o
v
e
r
t
h
e
B
E
R
p
e
r
f
o
r
m
a
n
c
e
r
e
g
r
e
s
s
i
o
n
in
t
h
e
s
i
t
u
a
t
i
o
n
of
h
a
v
i
n
g
c
o
n
s
t
a
n
t
l
o
w
q
u
a
n
t
i
z
a
t
i
o
n
r
e
s
o
l
u
t
i
o
n
l
e
v
e
l
s
A
D
C
s
.
Figure
5.
BER
perform
ance
usi
ng
2
bit
-
qu
a
nt
iz
at
ion
res
olu
t
ion
le
vels
ADC
s
an
d
4
QAM
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Evalu
atio
n and
m
e
as
ure
me
nt
of the effe
ct
s
of d
if
fe
rent
quan
ti
za
ti
on
re
so
l
ution
…
(
H
ay
der K
ha
le
el
AL
-
Qaysi
)
1499
Figure
6.
BER
perform
ance
usi
ng
4
bit
-
q
ua
nt
iz
at
ion
res
olu
t
ion
le
vels
ADC
s
an
d
4
QAM
4.
CONCL
US
I
O
N
In
this
w
ork,
we
eval
uated
and
m
easur
e
d
the
eff
ect
s
of
diff
e
re
nt
quant
iz
at
ion
res
olu
ti
on
le
vels
of
AD
Cs
on
t
he
BER
per
f
orm
a
nce
of
m
assive
MIM
O
syst
em
s
un
de
r
di
ff
e
ren
t
ope
rati
ng
scenari
os
.
T
he
two
ty
pical
m
od
ula
ti
on
s,
4
Q
AM
and
16
Q
AM
ha
ve
bee
n
us
e
d
with
LMM
SE
sign
al
sym
bo
ls
detect
or
to
ev
al
uate
the
syst
e
m
.
The
resu
lt
s
dem
on
strat
ed
t
hat
the
perform
ance
of
BER
in
m
as
sive
MIM
O
te
chnolo
gy
is
affe
ct
ed
directl
y
by
the
qu
a
ntiza
ti
on
r
esolutio
n
le
vel
s
of
AD
Cs
,
t
he
nu
m
ber
of
anten
nas
,
a
nd
m
od
ulati
on
sc
hem
es.
Lo
w
(1
bit
a
nd
2
bits)
-
qua
ntiza
ti
on
re
so
l
ution
le
vels
A
DCs
can
reali
ze
appr
opriat
e
BER
perfor
m
ances
com
par
ed
to
the
high
qua
ntiza
ti
on
res
olu
ti
on
le
vels
A
D
Cs
wh
e
n
the
nu
m
ber
of
ant
enn
a
s
at
the
BS
high.
More
o
ve
r,
we
ob
s
er
ve
that
us
ing
4
bits
-
quan
ti
zat
ion
reso
l
ut
ion
le
vels
ADC
s
in
16
Q
AM
,
the
S
NR
of
a
rou
nd
15.5
dB
is
obli
gato
ry
to
at
ta
in
ed
the
B
ER
of
10
−
3
,
w
her
ea
s
usi
ng
4
bits
-
quanti
zat
ion
r
esol
ution
le
vels
AD
C
s
in
4
QAM
re
quire
only
6
dB
SN
R.
We
c
on
c
lud
e
th
at
incre
asi
ng
the
m
od
ulati
on
or
der
will
sign
ific
ant
ly
aff
ect
the
BER
perf
or
m
ance
of
hi
gh
-
qua
ntiza
ti
on
r
esol
ution
le
vels
AD
Cs
wh
il
e
sli
ghtl
y
aff
ect
e
d
the
BER
perform
ance
w
hen
low
-
quanti
zat
ion
reso
l
ution
le
vels
ha
ve
been
use
d.
REFERE
NCE
S
[1]
Y.
M.
Ta
b
ra
an
d
B.
M.
Sabb
ar,
"H
y
brid
MVDR
-
LMS
Bea
m
form
ing
for
Mass
ive
MIM
O,"
Indone
sian
Journal
of
El
e
ct
rica
l
Enginee
ring
and
Computer
Sci
e
nce
(
IJE
ECS)
,
vol.
16,
no
.
2,
pp.
715
-
723,
2019
,
do
i:
10.
11591/ijeecs.
v16.
i2.
pp715
-
72
3
.
[
2
]
P.
S
u
n
i
t
a
,
R
o
s
a
l
i
n
S
a
m
a
n
t
a
r
a
y,
P
r
a
d
yu
m
n
a
K
u
.
M
o
h
a
p
a
t
r
a
,
R.
N.
P
a
n
d
a
a
n
d
P
a
d
m
a
S
a
h
u
,
"A
N
e
w
C
o
m
p
l
e
x
i
t
y
R
e
d
u
c
t
i
o
n
M
e
t
h
o
d
s
of
V
-
B
L
A
S
T
M
I
M
O
S
ys
t
e
m
in
a
C
o
m
m
u
n
i
c
a
t
i
o
n
C
h
a
n
n
e
l
,
"
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
of
I
n
f
o
r
m
a
t
i
cs
and
C
o
m
m
u
n
i
c
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
(IJ
-
I
C
T
)
,
v
o
l
.
8,
n
o
.
1,
pp.
29
-
38,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
i
c
t
.
v
8
i
1
.
p
p
2
9
-
38
.
[3]
O.
Agboje
,
Ns
i
kan
Nkordeh
,
U
za
iru
e
Stan
ley
I
dia
ke
,
Olol
ade
Olado
y
in,
Kenn
ed
y
Okokpuji
e
and
Ibin
abo
Bo
b
-
Manue
l
,
"M
IMO
Channe
ls:
O
pti
m
iz
ing
Throu
ghput
and
R
ed
uci
ng
Outag
e
by
In
creasing
Multi
pl
exi
ng
Ga
in
,
"
TEL
KOMNIKA
Tele
communic
a
t
ion,
Computing
,
El
e
ct
ronics
and
Control
,
vo
l.
1
8,
no.
1,
pp
.
41
9
-
426,
2020
,
doi
:
10.
12928/telkomnika.
v18
i1.
8720
.
[4]
M.
K.
Hus
sein
a
nd
N.
N.
Kh
amiss
,
"Inte
gra
t
ing
Mill
imete
r
W
av
e
with
H
y
br
id
Pr
ec
oding
Multi
us
er
Mass
ive
MIM
O
for
5G
Com
m
un
ic
a
ti
on,
"
TEL
KOMNIKA
Tel
ec
om
municat
i
on
,
Co
mputing,
El
e
ct
ro
nic
s
and
Con
trol
,
vo
l.
18,
no.
1,
pp.
90
-
98
,
2020
,
doi:
10.
12928
/t
e
lkomnika.
v18i1
.
13674
.
[5]
A.
Nalba
nd
,
M.
Sarva
g
y
a,
and
M.
R.
Ahm
ed,
"
Pow
er
Saving
a
nd
Optimal
H
y
b
rid
Prec
oding
in
Mill
imet
er
W
a
ve
Mass
ive
MIM
O
Sy
st
ems
for
5G
.
"
TEL
KOMNIK
A
Tele
communic
ati
on,
Computin
g,
El
e
ct
ronics
a
nd
Control
,
vol.
18,
no
.
6,
pp
.
28
42
-
2851,
2020
,
doi:
10
.
12928/te
lkomnika.
v18i6
.
15952
.
[6]
L.
Van
d
er
Perre
,
L.
Li
u
,
and
E.
G.
La
rss
on,
"Eff
ic
i
ent
DSP
and
Circ
uit
Ar
chi
t
ec
t
ure
s
for
Mass
ive
MIM
O:
Stat
e
of
the
Art
and
Fut
ure
Dire
ct
ions
,
"
IEEE
Tr
ansactions
on
Signal
P
roce
ss
ing
,
vol
.
66,
no.
18,
pp.
4717
-
4736,
Sept
.
2018
,
doi
:
10
.
11
09/T
SP
.
2018.
28
58190
.
[7]
M.
A.
Albre
em,
M.
Juntti
and
S.
Shahabuddi
n,
"M
assive
M
IMO
Dete
c
ti
on
Te
chni
qu
es:
A
Surve
y
,
"
IEEE
Comm
unic
ati
ons
Surve
ys
&
Tuto
rials
,
vol
.
21
,
no
.
4,
pp
.
3109
-
31
32,
2019
,
doi
:
10
.
1109/COMS
T.
2
019.
2935810
.
[8]
C.
Stergi
ou,
E.
P.
Kos
ta
s,
B.
B.
Gupta
and
Y.
Ishibashi,
"S
ec
ur
ity
,
Priva
c
y
&
Eff
iciency
of
Sus
ta
ina
bl
e
Clou
d
Com
puti
ng
for
Big
Data
&
IoT,"
Sustainabl
e
Computing:
Informatic
s
and
Syste
m
s
,
vol.
19,
pp.
17
4
-
184,
2018
,
doi
:
10.
1016/j.sus
com.2018.
06.
003
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
3
,
Se
ptem
ber
2
02
1
:
14
93
-
15
00
1500
[9]
R.
Yang,
Shuqi
Xi
,
Qibo
C
ai
,
Z
hiz
hou
Chen
,
Xi
aoha
ng
W
ang
,
Gui
Liu
,
"A
Co
m
pac
t
Plan
ar
D
ual
-
Band
Multi
p
le
-
Input
and
Multi
ple
-
Output
An
tenna
with
High
Is
ola
ti
on
for
5G
a
nd
4G
Applications
,
"
Mic
rom
ac
hine
s
,
vo
l.
12
,
n
o.
5,
pp
.
1
-
8,
Ma
y
2
021,
doi
:
10
.
33
90/mi12050544
.
[10]
P.
Varz
aka
s,
"A
ver
age
Ch
anne
l
Capa
c
ity
for
Ra
y
l
ei
gh
Fading
S
pre
ad
Spectrum
MIM
O
Sy
st
ems
,
"
Inte
rnational
Journal
of
Comm
unic
ati
on
System
s
,
vol.
19,
no.
10,
pp
.
108
1
-
10
87,
De
c.
2006
,
d
oi:
10
.
1002/dac.784
.
[11]
C.
W
ang,
C.
W
en,
S.
Jin,
and
S.
Tsa
i,
"F
ini
te
-
A
lpha
be
t
Prec
od
i
ng
for
Mass
ive
MU
-
MIMO
with
Low
-
Resoluti
on
DA
Cs,"
IEE
E
Tr
ansacti
ons
on
Wireless
Comm
unic
ati
ons
,
vo
l.
17
,
no
.
7,
p
p.
4706
-
4720
,
Jul
y
2018
,
do
i:
10.
1109/t
wc
.
201
8.
2830343
.
[12]
S.
Jac
obss
on,
Giuseppe
Durisi
,
Mika
el
Co
ldrey
,
Ulf
Gus
ta
vss
on
,
Christoph
S
tude
r
,
"Through
put
Anal
y
sis
of
Mass
ive
MIM
O
Uplink
wi
th
Lo
w
-
Resolut
ion
A
DCs
,
"
IEE
E
Tr
ansacti
ons
on
W
i
rele
ss
Comm
u
nic
ati
ons
,
vol
.
16
,
no.
6,
pp
.
4038
-
4051,
June
2017
,
doi
:
10
.
1109/TW
C.
2017.
2691318
.
[13]
J.
Choi,
Y.
Cho
,
and
B.
L.
Eva
n
s,
"Q
uant
ized
M
assive
MIM
O
Sy
stems
with
Mu
lt
icell
Coordin
ated
Be
amform
ing
and
Pow
er
Con
trol
,
"
I
EEE
Tr
ansacti
ons
on
C
omm
unic
ati
ons
,
vol.
69,
no
.
2,
pp.
946
-
961,
Feb.
2021
,
doi:
10.
1109/T
COM
M.2020.
303668
9
.
[14]
T.
Li
u,
Jun
Ton
g,
Qinghua
Guo,
Jiang
ta
o
Xi
,
Yangua
ng
Yu
a
nd
Zhi
t
ao
Xi
ao
,
"Ene
rg
y
Eff
ici
ency
of
Mass
ive
MIM
O
Sy
stems
with
Low
-
Reso
lut
ion
AD
Cs
an
d
Succe
ss
ive
In
te
rfe
r
ence
Can
c
el
l
at
ion
,
"
IEEE
Tra
nsacti
ons
on
Wirel
ess
Comm
unic
ati
ons
,
vol
.
1
8,
no
.
8,
pp
.
398
7
-
4002,
Aug.
20
19
,
doi
:
10
.
1109
/T
W
C.
2019.
292
0129
.
[15]
J.
Chen,
S.
Che
n,
Y.
Qi
,
and
S.
Fu,
"Int
el
l
ige
n
t
Mass
ive
MIM
O
Antenna
Se
lecti
on
Us
ing
Mo
nte
Ca
rlo
Tree
Sear
ch,
"
I
EE
E
Tr
ansacti
ons
on
Signal
Pro
ce
ss
ing
,
vol
.
6
7,
no.
20,
pp
.
5380
-
5390,
Oct.
2019
,
do
i
:
10.
1109/T
SP
.
20
19.
2940128
.
[16]
Z.
Xi
ao,
Jinca
n
Zha
o
,
T
i
anle
L
iu
,
L
ei
Geng
,
F.
Z
hang
and
Jun
To
ng
,
"On
the
En
er
g
y
Eff
i
ci
en
c
y
of
Mass
ive
MIMO
S
y
stems
with
Low
-
Resolut
ion
AD
Cs
and
La
tt
i
ce
Reduction
Ai
ded
Dete
c
tors,
"
Symmet
ry
,
vol
.
12,
no.
3,
pp
.
1
-
20,
2020,
doi
:
10
.
33
90/s
y
m
1203040
6
.
[17]
H.
W
ang,
W.
S
hih,
C.
W
en,
a
nd
S.
Jin,
"Reli
abl
e
OFDM
Re
ce
iv
er
with
Ultr
a
-
Low
Resolut
i
on
AD
C,
"
IEE
E
Tr
ansacti
ons
on
Comm
unic
ati
ons
,
vol
.
67
,
no
.
5,
p
p.
3566
-
3579
,
M
a
y
2019
,
doi
:
10
.
1109/T
COM
M.2019.
2894629
.
[18]
J.
Li
u,
Z.
Luo,
a
nd
X.
Xiong,
"Low
-
Resolut
ion
A
DCs
for
W
ire
le
ss
Comm
unic
at
ion:
A
Com
pre
h
ensive
Surve
y
,
"
in
IEE
E
Acce
ss
,
vo
l.
7,
pp
.
91291
-
9
1324,
2019
,
doi
:
10.
1109/ACCESS
.
2019.
2927891
.
[19]
J.
Zha
ng
,
Li
ng
l
ong
Dai
,
Ziy
an
He
,
Shi
Jin
,
and
Xu
Li
,
"P
erf
or
m
anc
e
Anal
y
sis
of
Mixed
-
ADC
Mas
sive
MIM
O
S
y
stems
Over
R
ic
i
an
Fading
Channe
ls,"
IE
EE
J
ournal
on
Sel
ec
t
ed
Areas
in
Comm
unic
ati
ons
,
v
ol.
35,
no.
6,
pp
.
1327
-
1338,
201
7
,
doi
:
10
.
1109/J
SA
C.
2017.
2687278
.
[20]
T.
W
aday
am
a
a
nd
S.
Ta
k
abe,
"
Quanti
z
er
Opti
m
iz
at
ion
B
ase
d
on
Neura
l
Quan
ti
z
erf
or
Sum
-
Pr
odu
ct
De
code
r
,
"
in
2018
IEE
E
Glob
al
Comm
unic
ations
Confe
renc
e
(
G
LO
BE
COM
)
,
Abu
Dhabi,
Unit
ed
Arab
Emirat
e
s,
2018
,
pp.
1
-
6
,
doi:
10
.
1109/GL
OCO
M.2018.
8647503
.
[21]
Z.
Shao,
R.
C.
de
La
m
are,
and
L.
T.
N.
La
nd
au,
"Ite
r
ative
D
et
e
ct
ion
and
De
codi
ng
for
Larg
e
-
Scale
Multi
p
le
-
Antenna
S
y
s
te
m
s
with
1
-
Bit
AD
Cs,"
IEE
E
Wireless
Comm
unic
ati
ons
Lett
ers
,
vo
l.
7,
no.
3,
pp
.
476
-
479,
June
2018
,
doi
:
10
.
1109/LW
C.
2017.
2787159
.
[22]
C.
Chen,
W.
Zh
ang,
and
X.
B
ao
,
"A
chi
eva
b
le
R
at
e
Anal
y
sis
on
the
Uplink
of
Mass
ive
MIM
O
w
it
h
Superimpos
ed
Pilot
s
and
Arbit
rar
y
-
Bit
AD
Cs,"
in
2019
IE
EE
19th
In
te
rnatio
nal
Conf
ere
nce
on
Comm
unic
a
ti
on
Te
chnol
og
y
(
ICCT
)
,
Xi'
an
,
C
hina
,
2019
,
pp.
6
74
-
678
,
doi
:
10
.
1109/ICCT
4680
5.
2019.
8947297
.
[23]
Y.
Dong
and
L.
Qiu,
"S
pectra
l
Eff
iciency
of
M
assive
MIM
O
Sy
stems
with
Lo
w
-
Resolu
ti
on
A
DCs
and
MM
S
E
Rec
e
ive
r
,
"
IE
EE
Comm
unications
Letters
,
vol.
21,
n
o.
8,
pp.
1
771
-
1774,
Au
g.
2017
,
doi
:
10.
1109/L
COM
M.2017.
269327
6
.
[24]
R.
Bh
anda
r
i
a
nd
S.
Jadha
v
,
"
Spect
ral
Eff
i
ci
en
t
B
li
nd
Ch
anne
l
Est
imati
o
n
T
ec
hniqu
e
f
or
MIM
O
-
OF
D
M
Com
m
unic
at
ions,
"
Inte
rnat
iona
l
Journal
of
Ad
va
nce
s
in
Applied
Sci
en
ce
s
(
IJA
AS
)
,
vol.
7,
no.
3,
p
p.
286
-
297,
201
8
,
doi:
10
.
11591/ij
aa
s.v7.i3.
pp286
-
297
.
[25]
Ö.
T.
Dem
ir
an
d
E.
Björnson,
"ADMM
-
Based
One
-
Bit
Qua
n
ti
z
ed
Signal
Det
ection
for
Mass
ive
MIM
O
Sy
st
ems
with
Hardware
Im
pai
rm
ent
s,"
in
ICASSP
2020
-
2020
IEE
E
Int
ernati
onal
Conf
ere
nce
on
Ac
ou
stic
s,
Speech
and
Signal
Proce
ss
in
g
(
ICASSP
)
,
Bar
ce
lon
a,
Spain,
2
020
,
pp
.
9120
-
9
124
,
doi
:
10
.
110
9/i
c
assp40776.2020.
9053984
.
[26]
S.
Jac
obss
on,
Giuseppe
Durisi,
Mikae
l
Coldr
e
y
,
Ulf
Gus
ta
vss
on
a
nd
Christoph
Studer
,
"O
ne
-
Bit
Mass
ive
MIM
O
:
Channe
l
Esti
m
ation
and
High
-
O
rde
r
Modulation
s,"
in
2015
IE
E
E
Inte
rnat
ional
Confe
renc
e
on
Comm
unic
ati
on
Workshop
(
ICC
W)
,
London,
UK,
2015
,
pp
.
1304
-
13
09
,
doi
:
10
.
1
109/ICCW.2015.
7247358
.
[27]
V.
Bhat
ia,
M.
R.
Tri
pat
h
y
,
and
P.
Ranj
an
,
"D
ee
p
Le
arn
ing
for
Mass
ive
MIM
O:
C
hal
l
enge
s
and
Future
Pros
pec
ts,
"
in
2020
IEE
E
9th
Inte
rnationa
l
Confe
renc
e
on
Comm
unic
ati
on
Syste
ms
and
Net
work
Technol
ogie
s
(
CSNT
)
,
Gw
al
ior,
In
dia,
pp.
26
-
31
,
2020
,
doi:
10.
1109
/CSN
T48778.
2020.
9115783
.
[28]
L.
V.
Ngu
y
en
,
D
u
y
Trong
Ngo
,
Nghi
H.
Tra
n
,
A
.
L
ee
Sw
indl
ehu
rst
and
Du
y
H
.
N.
Ngu
y
en
,
"S
uper
vised
and
Se
m
i
-
Supervised
Lear
ning
for
MIM
O
Bli
nd
Det
ec
t
io
n
with
Low
-
Res
olut
ion
AD
Cs,"
IEE
E
Tr
ansact
ions
o
n
Wirel
es
s
Comm
unic
ati
ons
,
vol
.
19
,
no
.
4,
p
p.
2427
-
2442
,
A
pril
2020
,
do
i:
1
0.
1109/t
wc
.
2020
.
2964661
.
Evaluation Warning : The document was created with Spire.PDF for Python.