Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
9
, No
.
5
,
Octo
ber
201
9
, pp.
3607
~3
614
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v9
i
5
.
pp3607
-
36
14
3607
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Outage
probabil
ity anal
ysis of E
H r
elay
-
assisted n
on
-
or
th
ogonal
mu
ltip
le a
cc
ess (
NO
M
A) systems
over blo
ck raylei
gh
fading c
hann
el
Tan N.
N
guye
n
1
, Mi
nh
Tr
an
2
, Van
-
D
uc Ph
an
3
, H
oang
-
N
am
Nguy
e
n
4
,
Thanh
-
Lo
ng
Nguy
e
n
5
1
W
ire
le
ss
Com
muni
cations
Rese
a
rch
Group,
Facult
y
of El
ec
tr
ic
a
l a
nd
Elec
tron
ic
s
E
ngine
er
ing,
Ton
Duc Thang
Univer
sit
y
,
Vie
t
nam
2
Optoel
e
ct
roni
cs
Resea
r
ch
Group
,
Facu
lty
of Elec
tri
c
al
and El
e
ct
r
onic
s E
ng
ine
e
rin
g,
Ton
Duc Thang
Un
ive
rsit
y
,
Vie
t
nam
3
Cent
er
of Excel
le
nc
e
for
Autom
at
ion
and
Prec
isi
on
Mec
han
ical E
ngine
er
ing, Nguy
en
T
at T
hanh
U
nive
rsit
y
,
Vi
et
na
m
4
Modeli
ng
Evolutiona
r
y
A
lgori
t
hm
s Sim
ula
ti
on
and
Artif
ic
i
al Int
el
li
g
ence, Fac
ul
t
y
of
E
lectr
i
ca
l
a
nd
Elec
tron
ic
s
Engi
ne
eri
ng,
To
n
Duc
Th
ang
Un
ive
rsit
y
,
Vi
et
na
m
5
Cent
er
for
Infor
m
at
ion
T
ec
hnolo
g
y
,
Ho Chi
Min
h
Cit
y
Un
ive
rsit
y
of
Food Indust
r
y
,
Vie
tna
m
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
J
a
n
21
, 2
01
9
Re
vised
A
pr
5
,
201
9
Accepte
d
Apr
13
, 201
9
Non
-
orthogona
l
m
ult
ipl
e
a
ccess
(NO
MA
)
has
bee
n
ide
nti
f
ie
d
as
a
prom
ising
m
ult
ipl
e
a
ccess
te
chni
qu
e
for
the
fift
h
gene
r
at
ion
(5G)
m
obil
e
net
works
due
to
it
s
superior
sp
ec
tr
al
eff
i
ci
en
c
y
.
In
th
is
pap
er,
we
in
troduc
e
and
inve
stigat
e
a
Non
-
orthogon
al
m
ult
ip
le
ac
c
e
ss
(
NOMA
)
of
ene
rg
y
har
v
esti
ng
(EH)
rela
y
assisted
s
y
stem
over
Bloc
k
R
a
y
leigh
Fading
Channe
l
.
To
eva
lu
at
e
the
per
form
anc
e
of
the
proposed
s
y
stem,
the
integr
al
expr
ession
of
the
outa
ge
proba
bil
i
t
y
is
a
naly
z
ed
and
de
rive
d
.
Num
eri
cal
r
esult
s
conf
ir
m
tha
t
ou
r
der
ive
d
ana
l
y
t
ical
result
s
m
at
ch
well
with
th
e
Monte
Carl
o
si
m
ula
ti
ons
i
n
conne
c
ti
on
wi
th al
l
poss
ibl
e
s
y
ste
m
par
amete
r
.
Ke
yw
or
d
s
:
Energy
harvest
ing
Mon
te
Ca
rlo
si
m
ula
ti
on
s
NO
M
A
Ou
ta
ge pr
obab
il
ity
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
:
Van
-
D
uc
P
han
Ce
nter of
Ex
c
e
ll
ence for
A
utom
at
ion
and Pre
ci
sion
Mec
han
i
cal
Enginee
rin
g
,
Nguyen
Tat T
ha
nh
U
niv
e
rsity
,
Ho Chi Mi
nh
Ci
ty
, V
ie
tnam
.
Em
a
il
:
pv
duc
@n
tt
.e
du.vn
Ho
a
ng
-
Nam
N
gu
ye
n,
Mod
el
in
g
E
voluti
on
a
ry
Algor
it
h
m
s Si
m
ulatio
n an
d A
rtific
ia
l In
te
ll
igence
,
Faculty
of Elec
tric
al
&
Ele
ct
r
on
ic
s
Engine
er
ing
,
To
n Du
c
Th
a
ng
Un
i
ver
sit
y,
Ho Chi Mi
nh
Ci
ty
, V
ie
tnam
.
Em
a
il
:
ng
uyen
ho
a
ngnam
@td
tu.edu.
vn
1.
INTROD
U
CTION
In
the
histo
ry
of
wi
reless
co
m
m
un
ic
at
ion
s
from
the
first
gen
e
rati
on
(1G
)
to
4G,
the
m
ulti
ple
acce
ss
schem
e
has
be
en
the
key
t
echnolo
gy
to
disti
nguish
dif
fer
e
nt
wireles
s
syst
e
m
s.
It
is
well
know
n
tha
t
fr
e
qu
e
ncy
-
div
i
sion
m
ulti
ple
a
ccess
(FDMA
)
for
1G,
ti
m
e
-
div
isi
on
m
ultip
le
acce
ss
(T
D
MA)
m
os
tl
y
f
or
2G,
cod
e
-
div
isi
on m
ul
ti
ple
acce
ss
(C
DMA)
f
or 3
G
,
a
nd o
rth
og
on
al
freq
ue
ncy
-
di
vision
m
ultip
le
acce
ss
(
OFDMA
)
for
4G
are
pr
i
m
aril
y
or
th
ogon
al
m
ulti
ple
acce
ss
(
OMA
)
schem
es.
I
n
t
hese
c
onve
ntion
al
m
ulti
ple
acce
ss
schem
es, d
iff
e
r
ent u
se
rs
are al
locat
ed
to
or
t
hogo
nal r
es
ourc
es in eit
her
the
tim
e, f
req
ue
nc
y, or
c
od
e
dom
ai
n
in
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
3
6
0
7
-
3
6
1
4
3608
order
t
o
av
oid
or
al
le
viate
interf
us
e
inte
rference
.
I
n
this
way,
m
ult
iplexin
g
gai
n
can
be
ac
hieve
d
with
reasona
ble
co
m
plexit
y.
Howev
e
r,
t
he
fas
t
gr
owth
of
m
o
bile
I
nter
net
has
pr
op
el
le
d
1000
-
fo
l
d
data
traff
i
c
increase
by
20
20
f
or
5G.
He
nce,
the
sp
e
ct
r
al
eff
ic
ie
ncy
beco
m
es
on
e
of
the
key
chall
eng
e
s
to
handle
su
ch
exp
l
os
ive
data
traff
ic
.
Mo
reover
,
due
t
o
the
rap
i
d
dev
el
op
m
ent
of
t
he
In
t
ern
et
of
Thi
ng
s
(IoT
),
5G
n
e
eds
to
su
pp
or
t
m
assiv
e
co
nnect
ivit
y
of
us
e
rs
an
d/or
de
vices
t
o
m
e
et
the
dem
and
for
lo
w
la
te
nc
y,
low
-
c
os
t
de
vices,
and
div
e
rse
se
r
vice
ty
pes.
T
o
sat
isfy
these
re
qu
i
rem
ents,
enh
ance
d
te
ch
nol
og
ie
s
a
re
nece
ssary.
S
o
fa
r,
s
om
e
po
te
ntial
can
did
at
es
hav
e
be
e
n
pr
opos
e
d
to
address
chall
e
ng
e
s
of
5G,
s
uch
as
m
assive
MIM
O
,
m
ill
i
m
et
er
wav
e
com
m
un
ic
at
ion
s
,
ultra
-
de
ns
e
netw
ork
,
an
d
non
-
ort
hogo
nal
m
ul
ti
ple
acce
ss
(
NO
M
A).
F
rom
that
analy
sis,
N
on
-
or
t
hogonal
m
ulti
ple
acce
ss
(N
OM
A)
has
been
i
den
ti
fie
d
as
a
prom
isi
ng
m
ulti
ple
a
ccess
te
chn
iq
ue
f
or
the
fifth
ge
ne
rati
on
(
5G)
m
ob
il
e
netw
ork
s
du
e
to
it
s
su
pe
rio
r
sp
ect
r
al
eff
ic
ie
ncy
[1
-
6]
.
In
c
ontrast
to
t
rad
it
io
nal
wate
r
-
fill
ing
po
wer
al
locat
ion
stra
te
gy,
N
OMA
al
locat
es
m
or
e
powe
r
to
the
us
er
s
with
w
or
se
c
ha
nn
el
co
ndit
io
ns
,
w
hic
h
re
s
ul
ts
in
a
bette
r
trade
off
betwe
en
the
syst
e
m
thr
oughput
an
d
us
e
r
fairn
e
ss.
F
r
om
the
pre
vious
researc
hes,
the
i
m
pact
of
us
e
r
pai
rin
g
on
dow
nlin
k
N
OM
A
syst
em
s
has
bee
n
char
act
e
rized
i
n
[
7],
t
he
wor
k
in
[
8]
has
s
tud
ie
d
the
po
wer
al
l
ocati
on
w
it
h
m
ax
-
m
i
n
fai
rn
e
ss
crit
erio
n.
An
upli
nk
N
O
MA
schem
e
with
j
oi
nt
powe
r
and
s
ubcar
rier
al
locat
ion
s
ha
s
been
pro
pose
d
in
[
9],
w
he
r
e
the
perform
ance
of
bo
t
h
li
nk
-
le
ve
l
and
syst
e
m
le
vel
has
bee
n
inv
est
igate
d
.
I
n
[10],
a
co
operati
on
-
ba
sed
NOM
A
schem
e
fo
r
c
oor
din
at
e
d
dire
ct
an
d
r
el
ay
tr
ansm
issi
on
s
ha
s
bee
n
i
ntr
oduced
.
A
div
e
rs
it
y
-
or
ie
nte
d
de
te
ct
ion
m
echan
ism
fo
r
the
co
operati
ve
relay
ing
syst
e
m
us
ing
N
O
MA
has
bee
n
pro
posed
in
[11
]
.
The
perform
ance
of
transm
it
antenn
a
sel
ect
io
n
for
NO
M
A
assis
te
d
m
ulti
ple
-
inp
ut
-
m
ulti
ple
-
ou
tp
ut
(M
IMO
)
relay
net
wor
ks
ha
ve
been
e
xam
ined
in
[12].
In
s
pir
ed
by
us
e
r
colla
borati
on,
a
co
op
e
rati
ve
N
O
MA
transm
issio
n
sc
hem
e
has
bee
n
pro
po
se
d
i
n
[13].
The
m
ai
n
obj
e
ct
ive
of
this
a
r
ti
cl
e
is
to
pro
pose
an
d
in
vesti
gate
a
N
on
-
O
r
thogonal
Mult
iple
Access
(NOMA
)
E
H
relay
assist
ed
syst
em
ov
er
Bl
ock
Ra
yl
ei
gh
Fa
ding
Chan
nel.
Fir
stl
y,
we
pro
pos
ed
an
d
inv
est
igate
d
a
Non
-
O
rth
ogon
al
Mult
iple
A
ccess
(
N
OMA
)
E
H
relay
-
ass
ist
ed
syst
em
ov
er
Bl
ock
Ra
yl
ei
gh
Fading
C
ha
nnel
.
I
n
the
ana
ly
sis
process,
we
a
naly
ze
a
nd
de
rive
the
integ
ral
ex
pr
ession
of
t
he
ou
ta
ge
pro
bab
il
it
y
.
Finall
y,
the
Mo
nt
e
Ca
rlo
Sim
ul
at
ion
is
use
d
f
or
validat
in
g
t
he
analy
ti
cal
analy
sis
in
co
nnect
ion
with
al
l
po
ssi
bl
e
syst
e
m
par
am
et
ers.
The
re
su
lt
s
sho
w
tha
t
the
analy
ti
cal
and
sim
ulati
on
res
ults
agr
ee
well
with
eac
h
ot
he
r
in
c
on
ta
ct
with
al
l
possible
s
yst
e
m
par
am
e
t
ers.
T
he
m
ai
n
con
t
rib
ution
s
of
this researc
h
can
be
fo
c
us
e
d
on
as t
he follo
wings:
-
We
pro
pose
a
nd
in
vestigat
e
a
Non
-
O
rth
ogonal
Mult
iple
A
c
cess
(NOM
A
)
EH
relay
assist
ed
syst
e
m
over
Bl
ock
Ray
le
ig
h
Fa
ding C
ha
nnel
.
-
The
i
ntegral
fo
rm
ex
pr
essi
on
of the
outa
ge p
roba
bili
ty
is analy
zed and deri
ved.
-
The
i
nf
lue
nce
of all
possible
syst
e
m
p
aram
e
te
rs
on t
he ou
ta
ge pr
obabili
ty
is in
vestigat
ed
and d
is
c
us
se
d.
-
All res
ults are
ver
ifie
d by the
Mon
te
Ca
rlo
si
m
ula
ti
on
.
The
rest
of
this
m
anu
scri
pt
is
organ
iz
e
d
as
fo
ll
ows.
I
n
se
ct
ion
2,
we
pr
esent
the
pro
pose
d
syst
e
m
m
od
el
and
ana
ly
sis
of
the
ou
ta
ge
pro
ba
bili
t
y
of
the
pro
posed
syst
em
mo
del.
N
um
eric
al
resu
lt
s
an
d
so
m
e
discuss
i
on
are
dr
a
w
n
in
sect
io
n
3.
Sect
ion 4
con
cl
ud
e
s this
m
anu
scri
pt.
2.
SY
STE
M MO
DEL A
N
D
O
UTAG
E
P
ROB
ABILIT
Y A
NA
L
YS
I
S
As
s
how
n
in
Fig
ure
1,
w
e
pro
po
se
d
a
si
m
ple
coop
erati
ve
net
wor
k
w
he
re
a
s
ource
node,
com
m
un
ic
at
es
with
a
destin
at
ion
node
,
D,
via
the
helpi
ng
of
a
dec
od
e
-
an
d
-
f
orward
(D
F
)
relay
node
,
R.
More
ov
e
r,
the
so
urce
S
ca
n
a
lso
directl
y
co
m
m
un
ic
at
e
the
destinat
io
n
D.
He
re,
we
a
ssum
e
t
hat
the
sy
stem
s
op
e
rate
in
half
-
duplex
m
od
e
,
i.e
,..,
the
relay
cannot
tra
ns
m
it
and
rec
ei
ve
a
sym
bo
l
at
the
sam
e
tim
e.
We
denote t
he c
hannel c
oeffic
ie
nts
,,
S
R
S
D
R
D
h
h
h
of
the
li
nk
betwee
n
S
, R
, and D,
r
es
pec
ti
vely
.
Fig
ure
1.
Syst
em
m
od
el
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ou
t
ag
e
pro
babi
li
ty
a
na
ly
sis
of
EH rel
ay
-
assi
ste
d
non
-
or
t
hogona
l m
ulti
ple
access
…
(
Tan
N. Ng
uyen
)
3609
The
fad
i
ng
ga
ins
in
al
l
in
vo
l
ved
li
nks
are
ass
um
ed
to
f
ollo
w
the
Ra
yl
ei
gh
distr
ibu
ti
on
with
the pr
ob
a
bili
ty
d
en
sit
y functi
on
(P
D
F) as in
[14
]
.
2
1
(
)
e
x
p
(
)
k
h
kk
x
fx
(1)
Wh
e
re
{
,
,
}
k
S
R
S
D
R
D
.
Fr
om
(1
)
t
he
c
um
ulati
ve
distribu
ti
on
f
unct
io
n
(C
DF
)
ca
n
be
ob
ta
ine
d
as
2
(
)
1
e
x
p
(
)
k
h
k
x
Fx
,
k
is
the m
ean o
f
th
e
ra
ndom
var
ia
ble
2
k
h
.
As
in
m
any
previo
us
publica
ti
on
s,
we
ass
um
e
that
the
source
a
nd
destin
at
ion,
as
well
as
the
relay
,
know the
ch
a
nnel
g
ai
ns
. T
he
energy
harvesti
ng and i
nfor
m
at
ion
processi
ng f
or t
his pr
opose
d
m
od
el
syst
em
as
sh
ow
n
in
Fig
ure
2.
I
n
this
protoc
ol,
the
tra
ns
m
issi
on
is
di
vid
ed
i
nto
bloc
ks
of
le
ng
t
h
T,
w
hich
c
ons
ist
s
of
three
tim
e
slot
s.
In
the
first
ti
m
e
slot
α
T
(α
is
the
tim
e
swi
tc
hin
g
facto
r,
01
)
,
the
relay
harvests
ene
rg
y
from
the
sourc
e
no
de
S
n
.
I
n
the
sec
ond
inter
val
ti
m
e
(1
-
α
)
T
/2,
t
he
s
ource
S
tra
nsfers
the
inf
orm
ation
to
R
a
nd
D
at
th
e
sam
e
tim
e.
Fi
nally
,
the
rem
ai
nin
g
tim
e
slot
(1
-
α
)
T
/2
is
use
d
f
or
i
nfor
m
at
i
on
tra
ns
fe
rr
i
ng
fro
m
the r
el
ay
node t
o
the
destinat
i
on.
Fig
ure
2
.
EH
a
nd I
T
pr
ocesse
s
The har
veste
d ener
gy at the
re
la
y ca
n
be
e
xpress
ed
as
in
[
14
]
2
h
s
S
R
E
T
P
h
(2)
Wh
e
re
01
is
ene
r
gy con
ver
si
on
eff
ic
ie
ncy
,
an
d
01
is
a
ti
m
e
-
switch
in
g
factor.
The
a
ver
a
ge
tr
ansm
it
p
ow
er
a
t t
he
relay
ca
n be
giv
e
n by
2
(
1
)
/
2
h
R
s
S
R
E
P
P
h
T
(3)
Wh
e
re
2
1
The
c
omm
un
ic
at
ion
p
r
ocess
f
or
coo
per
at
ive r
el
ay
syst
e
m
s
con
sist
s of
the
r
e
m
ai
nin
g
tw
o
consecuti
ve
tim
e
s
lots.
Durin
g
the
sec
ond
tim
e
slot,
the
so
urce
S
will
transm
i
t
a
sy
m
bo
l
1
a
with
the
powe
r
P
s
to
bot
h
the r
el
ay
R a
nd
d
est
inati
on
D.
Ther
e
f
or
e,
the
receive
d
sig
nal
at R an
d D ca
n be e
xpresse
d as, r
e
sp
ect
ivel
y.
1
1
,1
,1
,
R
a
S
S
R
R
D
a
S
S
D
D
y
P
h
a
n
y
P
h
a
n
(4)
Wh
e
re
,
RD
nn
are
the
add
it
ive
wh
it
e
Gau
s
sia
n
noise
(AWGN)
with
zero
m
ean
and
var
ia
nce
0
N
,
an
d
2
1
1
a
,
is ex
pectat
ion
op
e
rato
r.
The recei
ve
d
si
gn
al
t
o no
ise
r
a
ti
o
(SNR)
for
the
sym
bo
l
1
a
at
R
and
D
is
giv
e
n by,
res
pecti
vel
y.
1
1
2
2
,0
0
2
2
,0
0
,
S
S
R
S
R
a
S
R
S
S
D
S
D
a
S
D
Ph
h
N
Ph
h
N
(5)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
3
6
0
7
-
3
6
1
4
3610
Wh
e
re
0
0
s
P
N
.
In
the
t
hir
d
ti
m
e
slot,
R
wil
l
fo
r
ward
the
decode
d
sym
bo
l
1
a
with
the
power
P
R
to
D,
and
S
will
transm
it
ano
th
er
sym
bo
l
2
a
with
2
2
1
a
an
d
the
power
P
s
t
o
D
.
Hen
ce
,
the
r
e
cei
ved
si
gn
al
at
D
is
giv
e
n by
12
D
R
R
D
s
S
D
D
y
P
h
a
P
h
a
n
(6)
Apply
su
cces
sive
interfe
re
nce
cancel
la
ti
on
(S
IC
)
-
base
d
N
OMA
sc
hem
e,
D
firstly
deco
de
s
the
sy
m
bo
l
1
a
by
disposin
g
of
the
sym
bo
l
2
a
as
a
no
ise
te
rm
.
A
lso,
1
a
is
rej
ect
ed
f
ro
m
y
D
by
us
in
g
SI
C
to
decode
2
a
.
S
o,
t
he
r
ecei
ve
d SNR
s for
sym
bo
l
1
a
and
2
a
are r
es
pecti
vely
obtai
n
ed
as
1
2
2
2
0
,
22
00
1
R
R
D
S
R
R
D
Da
s
S
D
S
D
P
h
h
h
P
h
N
h
(7)
2
2
,0
D
a
S
D
h
(8)
Finall
y, we
c
an
obtai
n
the
e
nd to
e
nd S
NR fo
r
the
tran
sm
i
tt
e
d
sym
bo
l
1
a
as th
e f
ollow
i
ng
1
1
1
2
,
,
,
m
i
n
,
e
e
a
D
a
S
R
a
(9)
Using
the
sel
ect
ing
c
om
bin
ing
(
SC)
te
c
hn
i
que
at
the
recei
ver,
so
the
t
otal
ou
ta
ge
pr
ob
a
bili
ty
of
1
a
is
giv
e
n by
1
1
1
1
,
,
,
22
22
0
00
2
0
0
00
0
0
00
0
P
r
m
a
x
m
i
n
,
,
P
r
m
a
x
m
i
n
,
,
1
P
r
m
a
x
m
i
n
,
,
1
P
r
m
i
n
,
,
1
a
D
a
SR
a
SD
a
th
SR
R
D
SR
SD
th
SD
th
th
OP
hh
hh
h
XY
XZ
Z
XY
X
Z
0
00
0
0
0
0
0
0
0
0
0
0
0
0
0
0
P
r
or
1
P
r
or
1
P
r
,
P
r
,
1
Pr
1
th
th
th
th
th
th
th
th
th
th
th
th
th
th
Z
XY
X
and
Z
Z
XY
and
Z
X
and
Z
Z
XY
Z
X
Z
Z
XY
and
X
a
Z
0
00
0
0
0
0
0
(
1
)
00
0
0
0
(
1
)
0
0
0
(
)
(
)
(
)
(
)
(
)
(
)
th
th
th
th
th
th
Z
X
th
th
Z
X
Y
X
Z
Z
X
Z
X
Y
nd
Z
f
z
dz
f
x
dx
f
y
dy
F
F
f
z
dz
f
x
dx
f
y
dy
(10)
Wh
e
re
we
de
note
2
2
2
2
,
,
,
2
1
R
S
R
R
D
S
D
t
h
X
h
Y
h
Z
h
, a
nd
R
is
th
e
source
r
at
e.
Let
’s us to de
note
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ou
t
ag
e
pro
babi
li
ty
a
na
ly
sis
of
EH rel
ay
-
assi
ste
d
non
-
or
t
hogona
l m
ulti
ple
access
…
(
Tan
N. Ng
uyen
)
3611
0
0
0
0
0
0
(
1
)
0
1
0
0
0
0
0
0
0
0
00
0
0
0
(
1
)
(
)
(
)
(
)
(
)
(
)
(
1
)
(
)
1
e
xp
(
)
1
(
)
(
)
(
)
th
th
th
th
th
th
Z
X
th
Z
X
Y
Z
Y
X
th
ZX
RD
Z
X
Z
SR
Z
P
f
z
dz
f
x
dx
f
y
dy
f
z
dz
F
f
x
dx
X
Z
f
z
dz
f
x
dx
X
f
z
dz
f
x
dx
f
z
dz
0
0
0
0
0
0
00
00
(
1
)
e
xp
1
(
1
)
(
)
e
xp
th
th
R
D
SR
th
th
ZZ
SR
R
D
SR
Zx
dx
x
Zx
F
f
z
dz
dx
x
(1
1)
Apply eq
(3.
32
4,
1)
of
t
he
ta
bl
e
of
inte
gr
al
[15], (1
1) can
be
rewrit
te
n
as
t
he
f
ollow
i
ng
0
0
00
11
0
0
0
0
00
1
0
0
0
0
(
1
)
(
1
)
(
)
4
2
1
(
1
)
(
1
)
2
e
x
p
2
th
th
t
h
t
h
t
h
ZZ
R
D
S
R
R
D
S
R
t
h
t
h
t
h
Z
S
D
R
D
S
R
S
D
R
D
S
R
zz
P
F
f
z
K
d
z
z
z
z
F
K
d
z
(1
2
)
wh
e
re
()
v
K
is t
he
m
od
i
fied
Be
ssel
functi
on
of
t
he se
co
nd k
i
nd and
v
th
or
der
.
And
0
0
0
0
0
0
0
0
0
(
1
)
2
0
0
0
0
0
0
0
0
0
0
00
0
(
)
(
)
(
)
1
(
1
)
(
)
(
)
(
)
e
xp
1(
(
)
e
xp
th
th
th
th
th
th
th
th
Z
X
Z
X
Y
th
Z
X
Z
S
R
R
D
S
R
t
h
t
h
t
h
Z
X
Z
SR
P
f
z
dz
f
x
dx
f
y
dy
Zx
f
z
dz
f
x
dx
f
z
dz
dx
x
Z
F
F
f
z
dz
0
0
0
1)
th
R
D
S
R
x
dx
x
(1
3
)
Re
place (
1
2
),
(1
3
)
i
nto
(1
0
),
we have
0
0
1
00
00
1
00
0
0
0
00
1
(
1
)
(
1
)
1
2
e
xp
2
1
(
1
)
e
xp
e
xp
th
th
SD
th
th
t
h
t
h
a
S
D
R
D
S
R
S
D
R
D
S
R
th
S
R
S
D
S
D
R
D
S
R
z
z
z
O
P
e
K
dz
z
Z
x
dz
dx
x
(14)
Si
m
il
arity, t
he e
nd to
e
nd S
N
R for
t
he
tra
nsm
itted
sym
bo
l
2
a
can
be
cal
c
ula
te
d
as
2
1
2
2
,
,
,
m
i
n
,
e
e
a
D
a
D
a
(15)
Ther
e
f
or
e,
the
total
o
uta
ge pr
ob
a
bili
ty
of
2
a
ca
n
be
c
om
pu
te
d
as
2
2
1
2
1
2
2
,
,
,
,
,
22
2
0
0
2
0
P
r
P
r
m
in
,
1
P
r
,
1
P
r
,
1
a
e
e
a
t
h
D
a
D
a
t
h
D
a
t
h
D
a
t
h
S
R
R
D
t
h
S
D
SD
OP
hh
h
h
(16)
In
orde
r
to
fi
nd
the
total
ou
ta
ge
pr
ob
a
bili
ty
of
2
a
,
we
de
no
te
that
2
2
2
,
S
R
R
D
S
D
V
h
h
W
h
,
(
16
)
ca
n
be
rewrit
te
n
as
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
3
6
0
7
-
3
6
1
4
3612
2
0
0
0
0
0
0
0
0
0
1
1
1
Pr
,
1
Pr
,
1
1
(
)
(
)
th
th
th
th
a
t
h
t
h
WV
w
W
V
O
P
W
V
W
W
f
w
d
w
f
v
d
v
(
17
)
Util
iz
ing
the
r
e
su
lt
in [1
6]
, t
he
CDF
of
V
ca
n
be
s
how
n
as
th
e
(
18)
:
1
(
)
1
2
2
V
S
R
R
D
S
R
R
D
vv
F
v
K
(18)
wh
e
re
()
v
K
is t
he
m
od
i
fied
Be
ssel
functi
on
of
t
he se
co
nd k
i
nd and
v
th
or
der
.
Fr
om
(
17
),
(
18
),
we ha
ve:
2
0
w
00
1
00
11
2
12
SD
th
t
h
t
h
a
S
D
S
R
R
D
S
R
R
D
ww
O
P
e
K
d
w
(19)
3.
RESU
LT
S
AND DI
SCUS
S
ION
Fo
r
validat
io
n,
the
correct
ne
ss
of
the
de
rive
d
syst
e
m
per
f
orm
ance
expressio
ns
,
as
well
as
inv
est
igati
on
of
the
ef
fect
of
v
ari
ou
s p
aram
et
ers
on
the
syst
e
m
per
form
a
nce,
a
set
of
Mo
nte
Ca
rl
o
sim
ulati
ons
are
c
onduct
ed
and
prese
nted
i
n
this
sect
ion
[
17
-
21
]
.
For
ea
ch
sim
ulati
on
,
we
first
prov
i
de
the
gr
a
phs
from
the
analy
ti
cal
fo
rm
ulas.
Sec
ondly,
we
plo
t
the
sam
e
cur
ves
that
resu
lt
from
the
Mon
te
Ca
rlo
si
m
ulati
on
.
For
this
pur
po
se
,
we
ge
ner
at
e
10
6
ra
ndom
sa
m
ples
of
each
c
ha
nnel
gain
,
w
hich
are
Ra
yl
ei
gh
distrib
uted.
F
inall
y,
the an
al
yt
ic
al
cu
r
ve
a
nd the
sim
ula
ti
on
one
s
hould m
at
ch
tog
et
he
r
to
v
e
rif
y t
he
co
rr
ect
ne
ss of
our
a
naly
sis.
The ou
ta
ge pr
obabili
ty
o
f
t
he pr
opos
e
d
syst
e
m
v
ersus
tra
ns
m
it
ti
ng
S
NR γ
0
is plotte
d
i
n
Fi
g
ure
3
with
basic
syst
e
m
p
aram
et
ers
as
R
=0.5
bps,
η=
0.
8,
α
=0.4
5
an
d
0.85.
In
this
a
naly
sis,
the
transm
it
SN
R
γ
0
var
ie
s
from
-
10
dB
to
20
dB
co
ntinuo
us
ly
.
From
the
research
r
esults,
we
see
that
the
ou
ta
ge
pro
bab
il
it
y
of
the
pro
po
se
d
syst
em
decr
eases
sign
i
ficantl
y
with
the
incr
easi
ng
the
tran
sm
i
t
SN
R
γ
0
.
It
ca
n
be
obser
ve
d
that
the
higher
t
ran
sm
it
SN
R
γ
0
ca
us
e
s
a
hig
he
r
th
rough
pu
t
of
t
he
pro
po
se
d
syst
e
m
.
All
the
si
m
ulati
on
a
nd
an
al
yt
ic
al
curves m
at
ched
well
w
it
h
e
ac
h othe
r.
Fu
rt
her
m
or
e,
Figure
4
sho
ws
the
c
onne
ct
ion
bet
wee
n
the
syst
e
m
ou
ta
ge
pro
bab
il
it
y
and
tim
e
switc
hing
fact
or
α
with
the
m
ai
n
syst
e
m
par
am
et
ers
as
R
=0.5
bps,
γ
0
=
10
dB
and
η=
0.4,
0.8
.
As
show
n
in
Figure
4,
we
c
an
see
that
the
ou
ta
ge
pr
ob
a
bi
li
ty
of
the
syst
e
m
m
od
el
has
a
decr
ease
w
he
n
the
ti
m
e
swit
chin
g
factor
α
va
ries
from
0
to
1
in
the
co
nn
ect
io
n
with
the
fact
that
m
or
e
powe
r
is
us
ed
for
ha
rv
est
in
g
e
nerg
y
at
R
than
power
is
us
e
d
f
or
in
for
m
at
ion
transm
issi
on
betwee
n
D,
R
an
d
S.
Ag
ai
n
al
l
si
m
ulati
on
a
nd
a
na
ly
ti
cal
resu
lt
s a
gr
ee
wel
l wit
h
eac
h other.
Fig
ure
3
.
O
P
ve
rsu
s
tra
ns
m
it
SN
R γ
0
Fig
ure
4
.
O
P
ve
rsu
s
α
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ou
t
ag
e
pro
babi
li
ty
a
na
ly
sis
of
EH rel
ay
-
assi
ste
d
non
-
or
t
hogona
l m
ulti
ple
access
…
(
Tan
N. Ng
uyen
)
3613
More
ov
e
r,
t
he
functi
on
of
the
ou
ta
ge
pr
ob
a
bi
li
t
y
on
the
e
ne
rg
y
c
onversi
on
eff
ic
ie
ncy
η
a
s
show
n
in
Figure
5.
He
re
we
set
the
m
ain
syst
em
par
am
et
ers
as
R
-
0.
5
bp
s
,
α
=0
.5
a
nd
γ
0
=
5,
10
dB
.
Sim
il
ar
to
the
a
bove
cases,
the
outa
ge
probabil
it
y
of
t
he
m
od
el
s
yst
e
m
increases
cr
ucial
ly
wh
i
le
the
ene
r
gy
c
onve
rsion
e
ff
ic
ie
ncy
var
ie
s
from
0
t
o
1.
It
ca
n
be
obser
ve
d
that
th
e
m
or
e
eff
ic
ie
n
t
energy
c
onve
rsion
of
the
syst
e
m
the
le
ss
outa
ge
pro
bab
il
it
y.
I
n add
it
io
n
, t
he
a
naly
ti
cal
cu
rv
e
is the sam
e as
the sim
ulati
on
cu
r
ve
as
s
how
n
in
Fig
ure
5.
Finall
y,
the
ou
ta
ge
prob
a
bili
ty
ver
s
us
the
source
rate
R
as
sh
ow
n
i
n
Fi
gur
e
6
with
α
=0
.5,
η=
0.8
a
nd
γ
0
=
5,
10
dB.
Fr
om
the
resu
l
ts,
the
outa
ge
pro
bab
il
it
y
increases
sig
nific
antly
wh
e
n
the
so
urce
rate
i
nc
reases
from
0
to
7.
W
e
can
see
th
at
the
sim
ulatio
n
a
nd
th
e
analy
ti
cal
resu
lt
are
th
e
sa
m
e
with
al
l
values
of
the
s
ource
rate R.
Fig
ure
5
.
O
P
ve
rsu
s
η
Fig
ure
6
.
O
P
ve
rsu
s
R
4.
CONCL
US
I
O
N
In
this
pa
per
,
we
int
rod
uc
e
a
nd
in
vestigat
e
a
N
on
-
Or
th
og
on
al
Mult
iple
Access
(
NO
M
A)
EH
relay
assist
ed
syst
em
ov
er
Bl
ock
Ra
yl
e
igh
Fa
ding
Chan
nel.
Firstl
y,
we
pro
po
se
d
an
d
inv
est
igate
d
a
Non
-
Or
t
hogonal
M
ulti
ple
Access
(NOMA
)
E
H
r
el
ay
-
assist
ed
s
yst
e
m
ov
er
Bl
ock
Ra
yl
ei
gh
Fading
C
ha
nn
e
l
.
In
the
analy
sis
proces
s,
we
analy
ze
and
de
rive
t
he
integral
e
xpres
sion
of
the
outa
ge
prob
a
bili
ty
.
Finall
y,
the
Mon
t
e
Ca
rlo
Sim
ulatio
n
is
us
e
d
f
or
validat
in
g
the
analy
ti
cal
analy
sis
in
con
necti
on
wit
h
al
l
po
ssible
syst
e
m
par
am
et
ers.
Th
e
resu
lt
s
show
that
the
analy
tical
and
sim
ula
ti
on
res
ults
agree
well
with
each
oth
e
r
in
co
n
ta
ct
with all
possi
ble syst
e
m
p
ara
m
et
ers.
REFERE
NCE
S
[1]
Chen
He,
Cha
o
Zhai,
Yonghui
L
i
and
Branka
Vuce
tic
,
"Coopera
t
ive
Strat
eg
ie
s
for
W
i
rel
ess
-
Pow
er
e
d
Com
m
unic
at
ions:
An Overview
,
"
IEE
E
Wire
le
ss
Comm
unic
ati
ons
,
vol
.
25
,
no
.
4
,
p
p.
112
-
19
,
08
20
18.
[2]
Yu
H.,
Lee
H.
,
Jeon
H.
,
"
W
hat
is
5G
?
Emerg
in
g
5G
Mobile
Se
rvic
es
and
Ne
twork
Requi
r
ements.
Sus
ta
in
abi
l
ity,
vol.
9,
no
.
10
,
pp
.
1848
,
Oct
2017
.
[3]
Sharm
a
V.,
Kar
m
aka
r
P.
,
"
A
N
ovel
Method
of
Opportunisti
c
W
ire
le
ss
Ene
rg
y
Harv
esti
ng
in
Cognit
iv
e
Rad
i
o
Networks
,"
2015
7th
Int
ernational
Confe
ren
c
e
on
Computational
Intelli
g
ence,
Comm
unic
a
tion
Syste
ms
and
Net
works
,
Jun
2
015
.
[4]
Bocc
ard
i,
Feder
ic
o,
Rober
t
He
at
h,
Angel
Lozano,
Thomas
Marz
etta
and
P
et
ar
Popovs
ki
,
"F
iv
e
Disruptive
Te
chno
log
y
Dir
e
ct
ions f
or
5G
,
"
I
EE
E
Comm
unications
Magazine
,
vol
.
52
,
no
.
2
,
p
p.
74
-
80
,
Feb
20
14
.
[5]
Dai
Li
ng
long,
Bic
hai
W
ang,
Y
ife
i
Yuan
,
Shua
ngfe
ng
Han,
Ch
ih
-
Li
n
I
and
Zh
aoc
heng
W
ang
,
"N
on
-
orthogona
l
Multi
ple
Acc
ess
for
5G:
So
lut
ions,
Challe
nges,
Opportun
it
ie
s
,
and
Fut
ure
Rese
arc
h
Tre
nds
,
"
I
EE
E
Comm
unic
ati
ons Magazine
,
vol
.
53,
no
.
9
,
pp
.
74
-
81
,
Sep
2015
.
[6]
Sait
o
Yu
y
a,
Yos
hihi
sa
Kishi
y
ama,
Anass
Benj
eb
bour,
Ta
k
ehi
ro
Naka
m
ura
,
Anxi
n
Li
,
and
Keni
c
hi
Higuchi
,
"N
o
n
-
Orthogonal
Mul
ti
ple
Acc
ess
(NO
MA
)
for
Cel
lu
la
r
Futur
e
Rad
io
Acc
ess
,
"
2013
IEE
E
77th
Ve
hi
cul
ar
Techno
log
y
Confe
renc
e
(
VTC Spring)
,
Jun
2
013.
[7]
Ding
Zhi
guo,
Pi
ngzhi
Fan
and
H.
Vince
nt
Poor
.
,
"Im
pac
t
of
Us
er
Pairi
ng
on
5
G
Nonorthogonal
Multi
p
le
-
Ac
c
ess
Dow
nli
nk
Tra
ns
m
issions
,
"
IEE
E
Tr
ansa
ct
ions o
n
Ve
h
ic
ular
Tech
nology
,
vol
.
65
,
no.
8
,
pp
.
6010
-
023
,
Aug 2016
.
[8]
Ti
m
othe
ou
Steli
os
and
Ioa
nnis
Krikidi
s
,
"F
ai
rn
ess
for
Non
-
Ort
hogonal
Multi
p
l
e
Acc
ess
in
5G
Sy
st
ems
,
"
IEEE
Signal
Proce
ss
in
g
Letters
,
vol
.
22
,
no
.
10
,
pp
.
164
7
-
651
,
Oct
2015
.
[9]
Al
-
Im
ari
Moham
m
ed,
Pei
Xiao
,
Muham
m
ad
A
li
Im
ran
and
Rahi
m
Ta
faz
ol
li
,
"
Uplink
Non
-
orthogona
l
Multi
pl
e
Acc
ess
for
5G
W
ire
le
ss
Networks
,
"
2014
11th
Inte
rnationa
l
S
ymposium
on
Wirel
ess
Comm
unic
ati
ons
S
yste
ms
(
ISWCS
)
,
Aug
2014.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
3
6
0
7
-
3
6
1
4
3614
[10]
Kim
Jung
-
Bin
a
nd
In
-
Ho
Lee
,
"
Non
-
Orthogona
l
Multi
ple
Acc
es
s
in
Coordina
te
d
Dire
ct
and
Re
lay
Tr
ansm
ission
,
"
IEE
E
Comm
unications
Letters
,
v
ol.
19
,
no
.
11
,
pp
.
2037
-
040
,
Nov
2015
.
[11]
Xu
Min,
Fei
Ji,
Miaowen
W
en
a
nd
W
ei
Duan
,
"
Novel
Recei
v
er
Design
for
the
C
ooper
ative
Re
lay
ing
S
y
stem
W
ith
Non
-
Orthogonal
Multi
pl
e
Ac
ce
ss
,
"
IE
EE
Comm
u
nic
ati
ons
Letters
,
vol
.
20
,
no
.
8
,
p
p.
1679
-
682
,
Au
g
2016
.
[12]
Kizi
li
rm
ak
R
efik
Cagl
ar
,
"N
on
-
Orthogonal
Mul
ti
ple
Acc
ess
(NO
MA
)
for
5G
Networks
,
"
Tow
ards
5G
Wirel
ess
Net
works
-
A Ph
ysic
al
Layer Per
spec
tive
,
De
c
20
16.
[13]
Li
u
Yuanwei
,
Z
higuo
Ding,
Mage
d
El
kashl
an
a
nd
H.
Vince
nt
Poor
,
"Coopera
t
iv
e
Non
-
orthogona
l
Multi
pl
e
Acc
e
ss
W
it
h
Sim
ult
an
eous
W
ire
le
ss
Inform
at
ion
and
Pow
er
Tr
ansfe
r
,
"
I
EE
E
Journal
on
Sel
e
ct
ed
Areas
in
Comm
unic
ati
ons
,
vol
.
34
,
no
.
4
,
p
p.
938
-
53
,
Apr
2
016
.
[14]
Nasir
Ali
A.
,
Xiang
y
un
Zhou
,
Salman
Durra
ni
and
Rodn
e
y
A.
Kenne
d
y
,
"Relay
ing
Proto
col
s
for
W
irele
ss
Ene
rg
y
Harve
st
i
ng
and
Inform
ation
Proce
ss
ing
,
"
IEE
E
Tr
ansactions
on
Wirel
ess
Comm
unic
ati
on
s
,
vol.
12
,
no.
7
,
pp.
3622
-
636
,
Ju
l
2013
.
[15]
Ta
bl
e
of
Inte
gr
als
,
Serie
s
,
an
d
Pr
oduct
s
,
2015
.
[16]
Zhong
Caijun,
S
hi
Jin,
Ka
i
-
Kit
W
ong
and
Matthew
R.
Mck
a
y
,
"Ergodi
c
Mutu
al
Inform
at
ion
An
aly
s
is
for
Mult
i
-
Ke
y
hol
e
MIM
O
Channe
ls
,
"
IEEE
Tr
ansacti
ons on
Wireless
Comm
unic
ati
ons
,
vol
.
10,
no
.
6
,
pp
.
17
54
-
763
,
2011
.
[17]
Ngu
y
en
Ta
n
N
.
,
et
al
.
,
"
On
the
Perform
anc
e
of
Pow
er
Split
ti
ng
Ene
rg
y
Harv
ested
W
ire
le
ss
Full
-
Duplex
Rela
y
in
g
Network
with
Im
per
fec
t
CS
I
over
Diss
imila
r
Channe
ls
,
"
Se
cu
rity
and
Comm
u
nic
ati
on
Ne
tworks
2018
,
pp.
1
-
11,
Dec
2018
.
[18]
Ngu
y
en
Ta
n
N
.
,
et
a
l.
,
"P
erf
or
m
anc
e
Enh
ance
m
ent
for
En
erg
y
Harv
esti
ng
B
ase
d
Two
-
wa
y
Rel
a
y
Proto
col
s
in
W
ire
le
ss
Ad
-
hoc
Networks
with
Parti
al
and
Full
Rel
a
y
Sel
ec
t
ion
Methods.
"
Ad
Hoc
Net
works
,
vol
.
84
,
pp.
178
-
87
,
Mar
2019
.
[19]
Ngu
y
en
T
an,
et
al.
,
"H
y
br
id
TSR
–
PS
R
Alte
rna
t
e
Ene
rg
y
Ha
rve
s
ti
ng
Rel
a
y
Ne
twork
over
Ric
ia
n
Fading
Channe
ls
:
Outage
Probab
il
i
t
y
and
SER Ana
l
y
sis
,
"
S
ensors
,
v
ol.
18
,
no
.
11
,
pp
.
3839
,
Nov
201
8
.
[20]
Ngu
y
en
Ta
n
,
Tr
an
Quang
Minh
,
Phuong
Tr
an
and
Miroslav
Vozňá
k
,
"Ene
rg
y
Harve
sting
ove
r
Rician
Fad
in
g
Channe
l:
A
Perf
orm
anc
e
Anal
y
s
is
for
Half
-
Duplex
Bidi
re
ct
ion
al
Sensor
Networks
under
Hardware
Im
pai
rm
ent
s
,
"
Sensors
,
vol.
18,
no.
6
,
pp.
1781
,
Jun
2018
.
[21]
Ngu
y
en
Ta
n
N.
,
Phuong
T.
Tr
an
,
Tra
n
Hoang
Quang
Minh,
Mir
oslav
Vo
zna
k
an
d
Luka
s
Sevci
k
,
"Two
-
W
a
y
Hal
f
Duplex
Dec
od
e
and
Forw
ard
Re
lay
ing
Ne
twork
with
Hardware
I
m
pai
rm
ent
over
Ric
ia
n
Fading
C
hanne
l
:
S
y
st
em
Perform
anc
e
An
aly
s
is
,
"
E
lektron
ik
a
Ir
Elek
trot
echnik
a
,
vol
.
24
,
n
o.
2
,
Apr
2018
.
Evaluation Warning : The document was created with Spire.PDF for Python.