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
.
6
,
Decem
ber
201
9,
pp. 5
351~
5359
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v9
i
6
.
pp5351
-
53
59
5351
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
A
comp
arativ
e st
ud
y o
n h
andove
r probabil
ity anal
ysis
for futur
e
HetNet
s
S. Neera
j
a
,
A. Abhi
shikt
ha
Depa
rtment
o
f
E
le
c
troni
cs
and
C
om
m
unic
at
ion
E
ngine
er
ing,
GIT
AM
(Dee
m
ed
to be
Univer
si
t
y
)
,
I
ndia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
N
ov
8
,
201
8
Re
vised
Ju
l
1
0
,
201
9
Accepte
d
J
ul
1
7
, 2
01
9
The
ne
ed
of
wir
el
ess
serv
ices
in
cre
asing
d
a
y
b
y
da
y
du
e
to
th
e
a
dvanc
ement
s
in
th
e
field
of
wire
le
ss
t
ec
hn
olog
y
towa
rds
5G
for
insta
nt
tra
nsferri
n
g
the
m
ai
ls,
m
essa
ges
and
v
ide
o
c
alling
without
an
y
int
err
up
ti
on.
In
LT
E
an
d
5G
wire
le
ss
n
et
wor
ks,
m
aj
or
t
ask
i
s
to
p
rovide
se
a
m
le
ss
conne
c
ti
o
n
an
y
wher
e,
an
y
ti
m
e
when
th
e
user
m
a
y
roa
m
among
Hete
roge
neous
W
ire
l
e
ss
Networks
(HetNe
ts).
To
achie
ve
prope
r
m
obil
ity
m
ana
geme
nt
among
He
tNe
ts,
handof
f
or
hadove
r
is
r
e
quire
d.
Hando
ve
r
Probabil
i
t
y
is
one
of
the
m
etri
c
to
esti
m
ate
the
h
andove
r
p
er
form
anc
e,
which
is
a
p
roba
bi
li
t
y
of
Mobile
Node
to
handov
er
the
pre
s
ent
conn
ec
t
ion
from
the
cur
ren
t
base
st
at
i
on
to
ano
the
r
b
a
se
stat
ion
or
enode
B.
In
this
pape
r,
hando
f
f
proba
bilit
y
an
aly
s
is
i
s
done
f
or
m
ult
ipl
e
HetNe
ts
b
ase
d
o
n
Hand
over
Alg
orit
hm
.
To
esti
m
at
e
th
is
al
gori
th
m
,
bandwidt
h
is
conside
red
as
one
of
th
e
ke
y
p
a
ramet
er
.
A
compara
t
ive
ana
l
y
sis
of
handove
r
proba
bil
i
t
y
fo
r
two,
th
ree,
fou
r
and
f
iv
e
He
t
Nets
has
b
ee
n
per
form
ed.
Th
e
resul
ts
ca
n
demons
tra
te
th
at
the
v
ari
a
ti
on
of
handove
r
prob
a
bil
ity
w
it
h
respe
ct
to
tr
aff
ic
lo
ad,
thre
sh
old
and
bandw
idt
h.
I
t
is
obs
erv
ed
th
at,
as
th
e
num
ber
o
f
wire
le
ss
n
et
wo
rks
inc
r
ea
ses
ha
ndover
prob
abi
l
i
t
y
sligh
t
l
y
inc
re
ase
s
with
t
raf
fic
lo
ad.
Th
e
se
r
esult
s
ar
e
m
ore
signifi
c
ant
to
esti
m
ate
furthe
r
wrong
d
ec
ision
h
andovers
base
d
on
th
at
Quali
t
y
of
Serv
ic
e
(QoS
)
is
eva
lu
at
ed
in
pr
ac
t
ic
a
l
H
et
Nets
such
a
s
integr
at
ion
of
LTE,
W
i
-
Fi
and
W
iMAX
et
c.
Ke
yw
or
d
s
:
Hand
off/ha
ndover
H
an
dove
r
p
r
ob
abili
ty
HetNet
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
:
S.
Neer
a
j
a,
Dep
a
rtm
ent o
f El
ect
ro
nics
and C
omm
un
ic
ation
En
gin
ee
rin
g,
GI
T
AM
(D
ee
m
ed
to
be Un
i
ver
sit
y),
Gand
hi N
a
ga
r,
Rush
i
konda,
V
isa
kh
a
patnam
, A
nd
hr
a
Pr
a
des
h,
Pinc
od
e:
53
0045
, I
nd
ia
.
Em
a
il
:
gitam
ne
eraj
a
@g
m
ai
l.com
1.
INTROD
U
CTION
The
a
dvancem
ents
in
Wirele
ss
an
d
M
ob
il
e
Com
m
un
ic
at
i
on
te
ch
no
l
og
ie
s
are
dr
a
sti
cal
ly
m
ov
ing
towa
rd
s
t
he
fif
th
ge
ner
at
io
n
(
5G)
cel
lular
sy
stem
s.
W
it
h
th
e
te
chnolo
gy
a
dv
a
ncem
ents,
the
f
uture
ge
ne
rati
on
wireless
net
work
s
will
integ
ra
te
the
cu
rr
e
nt
wireless
net
work
s
s
uc
h
as
L
T
E,
WiM
A
X,
WLA
N
,
an
d
W
i
-
Fi
et
c
.
to
al
lo
w
t
he
use
r
t
o
pro
vid
e
a
nyw
her
e
,
a
nyti
m
e
best
c
omm
un
ic
at
io
n
ser
vi
ces
.
T
h
e
com
bi
nation
of
di
fferent
wireless
netw
orks
with
dif
fere
nt
acce
ss
te
c
hnologies
are
te
r
m
ed
as
hete
rog
eneous
wireles
s
netw
orks
(
He
tNet
s).
Con
se
quently
,
the
Mo
bile
N
odes
(MN’s)
al
ways
sel
ect
the
best
a
vaila
ble
netw
ork
for
ac
cess,
cal
le
d
t
he
“al
ways
best
co
n
n
ect
ed
''
netwo
r
k
[1,
2]
.
The
m
ulti
ple
MN’s
with
m
ulti
ple
interfac
es
will
able
to
choose
t
he
be
s
t
li
nk
a
m
on
g
al
l
a
va
il
able
can
did
at
es.
Ha
ndover
or
hand
off
is
a
pr
ocess
w
hi
ch
seam
le
ssl
y
trans
fer
the
ongo
i
ng
conve
rsati
on
f
r
om
on
e
netw
ork
to
a
no
t
her
ne
twork
wit
ho
ut
interr
up
ti
on
in
order
to
sup
port
the
high
m
ob
il
it
y
and
to
pro
vid
e
the
Q
oS
t
o
th
e
en
d
us
e
rs
durin
g
c
omm
un
i
cat
ion
.
Ha
ndof
fs
ca
n
be
div
i
ded
int
o
tw
o
t
ypes
:
1)
a
hand
ov
e
r
t
hat
oc
cu
r
betw
een
th
e
sim
i
la
r
ty
pes
of
netw
orks
is
cal
le
d
a
Horizo
ntal
Ha
ndoff
e
.g.
L
TE
to
LT
E
and
2)
a
hand
over
hand
off
tha
t
occ
ur
betwee
n
dissim
il
ar
typ
es
of
net
wor
ks
is
cal
le
d
a
Ve
rtic
al
Ha
ndoff
[3
]
e.
g.
LTE
t
o
W
iM
A
X.
Ve
rtic
al
Ha
ndoff
he
lps
M
N
t
o
sel
ect
the
best
netw
ork
a
m
on
g
HetNets
for
pro
vid
i
ng
de
sired
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.
6
,
Dece
m
ber
2
01
9
:
5351
-
5359
5352
QoS.
A
n
acc
urat
e
handoff
al
go
rithm
design
i
s
m
ajo
r
go
al
of
a
networ
k
des
ign
e
r
in
ord
er
t
o
ens
ure
MN’
s
can
m
ai
ntain an
ac
ceptable
QoS.
In
t
he
li
te
rature,
m
os
t
of
the
hand
off
al
gorit
hm
s
are
desig
ned
bet
ween
sim
il
ar
networks
w
hich
a
re
base
d
on
Re
ce
ived
Sig
nal
St
rength
(RSS
)
e.g
GS
M
t
o
G
SM
or
CDM
A
to
CDM
A
[
4,
5].
Var
io
us
ve
rtic
al
hand
off
al
gorithm
s
are
obse
rv
e
d
a
nd
desi
gn
e
d
base
d
on
a
rea
of
a
ppli
ca
ti
on
an
d
use
r
requirem
ents
[6
]
.
RSS
-
base
d
ve
rtic
al
ha
ndoff
al
gorithm
is
relat
ively
m
at
ur
e
an
d
eas
y
to
reali
ze,
howe
ver,
t
he
la
ck
of
com
pr
ehe
ns
i
ve
consi
de
rati
on
of
dif
fer
e
nt
net
work
pa
ram
et
e
rs
an
d
e
nd
use
r
requirem
ents
li
m
it
s
i
ts
per
f
or
m
ance
and
ap
plica
ti
on
in h
et
er
ogen
eous
en
vir
onm
ent. O
ther
tha
n
RSS
ba
sed
,
va
rio
us
v
erti
cal
ha
ndoff
al
gorith
m
s
are
pr
ese
nted
b
a
se
d on Sig
nal to
In
te
r
fer
e
nce
Noise Ra
ti
o [7
]
, c
os
t f
unct
ion a
nd addit
ive
we
igh
ti
ng
functi
on etc.
In
desi
gn
i
ng
cost
f
unct
ion
base
d
ver
ti
cal
ha
ndoff
al
gorithm
[8
]
,
m
ulti
ple
syst
e
m
perform
ance
par
am
et
ers
an
d
en
d
-
us
er
c
harac
te
risti
cs
are
t
aken
int
o
acc
ount,
the
refo
re,
the
al
go
rithm
perform
s
bet
te
r
th
a
n
conve
ntion
al
R
SS
base
d
ha
ndoff
al
gorithm
.
The
draw
bac
k
of
this
al
gorith
m
is
the
la
ck
of
ada
ptati
on
ca
pa
bili
ty
for
on
ce
de
fine
d,
t
he
f
orm
of
the
c
os
t
f
unct
io
n
is
fixe
d
f
or
al
l
the
us
e
rs
a
nd
for
al
l
the
se
r
vice
ty
pes,
he
nce
,
it
is
diff
ic
ult t
o
a
da
pt
the al
gorith
m
b
ased on dif
fer
e
nt user
re
quirem
ents, especi
al
ly
f
or
dif
f
eren
t se
r
vice tr
aff
ic
s.
In
a
dd
it
ive
wei
gh
ti
ng
ba
sed
ve
rtic
al
ha
ndoff
al
gorithm
[9
]
,
par
am
et
ers
af
f
ect
ing
ha
ndoff
perform
ance
are
ch
os
e
n
as
decisi
on
fact
ors
su
c
h
as
t
hr
es
ho
l
d,
e
nd
to
e
n
d
delay
et
c.
Di
ff
e
ren
t
factors
are
assig
ne
d
diff
e
ren
t
weig
hts
wh
ic
h
cou
l
d
be
dy
na
m
ic
al
l
y
adjuste
d
acc
ordi
ng
to
netw
ork
sta
te
a
nd
var
io
us
us
e
r
se
rv
ic
es
,
t
herefo
re,
it
is
po
ssible
to
gu
a
ran
te
e
dif
f
eren
t
us
er
QoS
requirem
ent
and
im
pr
ove
the
perform
ance
of
the
w
hole
net
work
m
eanw
hile.
Hand
off
al
gori
thm
s
based
on
t
he
R
SS
does
not
m
eet
the
requirem
ent
of
up
per
la
ye
r
ap
plica
ti
on
s
w
here
traff
ic
is
m
or
e
an
d
se
ns
it
ive
with
ti
m
e
so
ha
ndoff
is
im
ple
m
ented
ba
sed
on
the
ba
ndwidt
h
as
the
m
ai
n
par
am
et
er.
Ha
ndoff
al
gorith
m
based
on
ba
ndwidt
h
for
t
wo
an
d
t
hr
ee
heter
og
e
ne
ou
s
wireless
netw
orks
are
avail
able
in
the
li
te
ratur
e
[10],
[
11
]
.
I
n t
his
pa
per,
P
roba
bili
ty
of
ha
ndoff
an
al
ysi
s
is
done
f
or
m
ulti
ple
HetNets
(tw
o,
th
ree,
f
our
an
d
five
)
ba
se
d
on
ha
ndover
al
gorithm
by
con
si
de
ri
ng
ba
ndwi
dth
as
20M
Hz
w
hic
h
is
s
ui
ta
ble
for
e
valuati
ng
the
pe
rfor
m
anc
e
of
cu
rr
e
nt
wi
reless
Het
Nets
su
c
h
as
inte
gr
a
ti
on
of
L
TE
a
nd
W
iM
AX,
LT
E
to
Wi
-
Fi
et
c.
F
rom
the
resu
lt
s,
the
hand
ov
e
r
pro
ba
bili
ty
com
par
at
ive
analy
s
is
is
car
ried
ou
t
for
m
ulti
ple
(
two,
t
hr
ee,
four a
nd
five) Het
Nets
.
2.
ANALY
SIS
O
F PROB
ABIL
ITY O
F H
A
N
DOVER
Pr
oba
bili
ty
o
f han
dove
r
is
the
proba
bili
ty
o
f M
N
is m
ov
in
g
ou
t
of its e
xisti
ng
netw
ork. Cons
i
der fi
ve
netw
orks
N
1,
N2,
N3,
N4
a
nd
N
5.
A
naly
ti
cal
m
od
el
s
f
or
H
an
dove
r
a
r
e
desi
gn
e
d
for
these
netw
ork
s
us
i
ng
Ma
rkov M
odel
as s
how
n
in
Fi
gure
1
.
He
re,
N
j
N
i
(
t
)
C
:
MN
is c
hangi
ng
f
ro
m
n
et
w
ork
N
i t
o netw
ork Nj at
tim
e t
.
Ni
(
t
)
S
:
MN
co
ntin
ue
s
at
curre
nt
ne
twork
N
i at
ti
m
e t
N
j
N
i
P
: Pr
ob
a
bili
ty
of
MN
tra
nsfer
ring
f
ro
m
n
et
w
ork
N
i t
o netw
ork Nj at
tim
e t
N
i
N
i
P
: Pr
ob
a
bili
ty
o
f
MN
c
on
ti
nu
i
ng
at
netw
ork N
i at
tim
e t
The
c
ha
ng
e
in
band
width
pr
oc
edure
of
net
w
ork
i,
i=
N
(wh
ere
N
is
the
no
of
netw
orks)
is
de
velo
pe
d
as
an
M/
M/
Bi
proce
ss
[
11
]
.
Wh
ere
Bi
,
t
he
m
axi
m
u
m
band
width
an
d
b
i
is
the
a
va
il
able
band
widt
h
of
the
netw
ork
.
T
he
ar
rival
rate
of
requirem
ent
for
cha
nnel
s
is
denoted
with
a
Po
iss
on
distrib
ution
with
par
a
m
et
er
i
,
N
i
.
a
nd
t
he
ser
vi
c
e
rate,
at
w
hich
t
he
cha
nnel
is
cl
ea
red,
wil
l
obey
a
n
e
xponentia
l
distrib
ut
ion
wit
h
par
am
et
er
i
,
N
i
.
i
,
y
,
N
i
.
y
0
,
1
,
.
.
.
.
B
i
Co
ns
ti
tute
the
pro
bab
il
it
y
of
occ
upie
d
cha
nnel
s
is
y.
Fr
om
queu
i
ng
theo
ry r
es
ult, t
he follo
wing e
qu
at
io
ns h
old
i
,
y
i
y
,
y
B
i
ot
he
r
w
i
s
e
,
0
y
i
i
,
y
i
,
0
y
!
,
0
y
B
i
(
1)
wh
e
re
is t
raffi
c load.
Bi
i
,
0
y
i
y0
1
,
i
1
,
2
,
3
,
4
,
5
y!
(
2)
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
A com
pa
r
ati
ve
stud
y
on a
ndov
er p
r
obabil
it
y a
naly
sis f
or f
ut
ur
e
HetN
et
s (S.
Neeraja)
5353
The
a
naly
ti
cal
m
od
el
of
ha
nd
ov
e
r
f
or
tw
o
w
irel
ess
n
et
w
ork
s
as
sho
wn
i
n
Fig
ure
.
1
(a
)
i
s
ex
plaine
d
by
C.
Chi,
et
.a
l.
[
10
].
Co
ns
id
er
,
the
pro
bab
i
li
ty
of
handov
er
f
or
th
ree
net
works
i.e.
Fig
ure
.
1
(b
).
In
order
t
o
fin
d
the
MN
m
ov
ing
bet
we
en
di
ff
e
ren
t
node
s
by
us
in
g
three
sta
te
Ma
rko
v
desi
gn
and
t
he
var
ia
bl
es
p,
q
represe
nt
the
pro
ba
bili
ty
of
MN
s
witc
hing
f
ro
m
N
1
to
N
2
a
nd
from
N
2
t
o
N
1.
V
aria
bles
s
a
nd
r
re
pr
ese
nt
the
prob
a
bili
ty
o
f
MN
m
ov
ing f
ro
m
N
3 t
o N
2
a
nd
N2 to
N3
a
nd the
var
ia
bles
u
a
nd t
re
present
the
pro
ba
bili
ty
of
MN
s
hiftin
g
from
N
1
t
o
N
3
a
nd
N3
t
o
N
1
.
T
he
pr
ob
a
bili
ti
es
(
1
-
p
-
u),
(
1
-
q
-
r
)
a
nd
(1
-
s
-
t
)
re
pr
ese
nt
the pr
ob
a
bili
ty
o
f
MN
sta
y at
N1,
N2 and
N
3 are
r
e
pr
ese
nte
d by
P
N1
,P
N2
a
nd P
N3
.
(a)
(b
)
(c)
(d
)
Fig
ure
1. A
nal
yt
ic
al
m
od
e
ls
f
or h
a
ndover
b
e
tween
heter
oge
neous
w
irel
ess
netw
orks
,
(a
)
Ma
rko
v
M
od
el
f
or
t
wo w
i
reless
netw
ork
s
, (b
)
Mar
kov
Mod
el
for
t
hr
e
e w
irel
ess
n
et
works
,
(c
)
Ma
rko
v
M
od
el
f
or
f
our w
irel
ess n
et
works
, (d
)
Mar
kov
Mod
el
for fi
ve wirel
ess
netw
orks
Algori
th
m
:
Prob
a
bili
ty
o
f Ha
ndover
A
l
gorithm
b
ased
on B
andwidt
h
Let
av
ai
la
ble
ba
ndwidt
h of ne
twork
b1, b
2, b3 a
nd th
reshol
d G
th
.
If
a
MN is
at N
1
the
n
If
b2
-
b1 ≥
G
th
, th
e
MN
m
oves to
N
2
Othe
rw
ise
sta
ys at
N1.
Else
if a
MN
is
at N
1
the
n
If
b3
-
b1 ≥
G
th
, th
e
MN
m
oves to
N
3
Othe
rw
ise
sta
ys at
N1.
Si
m
il
arly
, f
or
a
MN
is at
N
2
a
nd N3.
2.1.
Mathem
ati
cal
ca
lc
ula
tio
ns
of
MN
st
ay
i
ng
an
d
mo
v
in
g
p
robabil
itie
s
:
Con
si
der,
thre
e
wireless
ne
twork
s
as
s
how
n
in
Fi
g
ure
.
1
(b
).
T
he
pr
ob
a
bili
ty
of
m
ob
ile
node
sta
yi
ng
at
N1 is cal
culat
e
d
as
,
N
1
N
2
N
1
N
3
N
1
N
2
N
2
N
1
N
1
N
3
N
3
N
1
P
N1
PP
P
P
P
P
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.
6
,
Dece
m
ber
2
01
9
:
5351
-
5359
5354
P
q
t
p
q
t
u
N1
(
3)
Si
m
il
arly
, f
or
N2
MN
stay
in
g
at
N2
is
P
p
s
p
q
s
r
N2
(
4)
MN stay
ing at
N3 is
P
r
u
r
u
s
t
N3
(
5)
The
MN
m
ov
es
from
it
s
pr
esent
net
work
w
hen
t
he
band
w
idth
of
t
he
cu
r
ren
t
netw
ork
i
s
le
ss
than
the
oth
e
r
netw
orks
ba
ndwi
dth
at
th
reshold
G
th
.F
or
eac
h
ne
twork
band
width
cha
nges
dy
nam
ic
ally
so
c
on
si
de
r
the n
et
wor
k
sta
ble stat
e su
c
h t
hat h
a
ndove
r d
oes n
ot affect
ed by t
he
ti
m
e p
e
rio
d
t.
MN m
ov
in
g
f
r
om
N
2
to
N1 i
s
th
P
P
b
1
b
2
N
1
N
2
G
(6)
MN m
ov
in
g
f
r
om
N
1
to
N
2
i
s
th
P
P
b
2
b
1
N
2
N
1
G
(7)
MN m
ov
in
g
f
r
om
N
3
to
N1 i
s
th
P
P
b
1
b
3
N
1
N
3
G
(
8)
MN m
ov
in
g
f
r
om
N
1
to
N
3
i
s
th
P
P
b
3
b
1
N
3
N
1
G
(
9)
MN m
ov
in
g
f
r
om
N
2
to
N
3
i
s
th
P
P
b
3
b
2
N
3
N
2
G
(
10
)
MN m
ov
in
g
f
r
om
N
3
to
N
2
i
s
th
P
P
b
2
b
3
N
2
N
3
G
(
11)
MN
sta
yi
ng at
N1 is
P
P
P
1
p
u
1
N
1
N
1
N
2
N
1
N
3
N
1
(
12)
MN
sta
yi
ng
at
N
2
is
P
P
P
1
q
r
1
N
2
N
2
N
1
N
2
N
3
N
2
(
13)
MN
sta
yi
ng
at
N
3
is
P
P
P
1
s
t
1
N
3
N
3
N
1
N
3
N
2
N
3
(
1
4)
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
A com
pa
r
ati
ve
stud
y
on a
ndov
er p
r
obabil
it
y a
naly
sis f
or f
ut
ur
e
HetN
et
s (S.
Neeraja)
5355
B
i
G
th
th
i
,
y
y0
P
b
i
G
,i
1
,
2
,3
(
15)
W
it
h
in
de
pe
ndent ass
um
ption
s of
b1, b2,
b3
we have
i
,
B
i
j
j
,
B
j
i
G
th
th
Bi
Bj
P
P
b
i
b
j
N
i
N
j
i0
j
0
,
i
j
ij
w
h
e
r
e
i
j
,
i
j
1
,
2
,
3
.
G
(
16
)
2.2.
Prob
ab
il
ity
of
hand
ov
er
ca
l
culati
on
Assum
e,
m
axim
u
m
ba
ndwidt
h
of
the
net
work
s
B1=B
2=B3
=B
an
d
T
raffic
loa
d
for
t
hr
ee
netw
orks
is
assigne
d
a
s,
1
2
3
.
Ther
e
f
or
e,
N
i
N
j
Q
,
i
j
,
i
j
1
,
2
,
3
P
.
Pr
oba
bili
ty
o
f han
dove
r
is,
P
P
P
P
P
P
P
P
P
P
P
P
P
H
(
)
(
)
(
)
(
)
(
)
(
)
3
*
Q
N
1
N
2
N
1
N
2
N
1
N
2
N
3
N
1
N
3
N
1
N
3
N
1
N
2
N
3
N
2
N
3
N
2
N
3
(17)
The
a
bove
pro
cedure
is
re
pea
te
d
f
or
de
vel
oping
pro
ba
bili
ty
of
hand
over
a
lgorit
hm
fo
r
f
our
an
d
five
netw
orks
w
hich
a
re
s
how
n
i
n
Fig
ur
e
1(
c
)
a
nd
Fig
ure
1(
d
).
F
or
exam
ple,
co
ns
ide
r
m
ulti
ple
net
works
(
N=
5),
then
t
he
c
orres
pondin
g
m
ov
in
g
a
nd
sta
yi
ng proba
bili
ti
es o
f
t
he
MN
are
calc
ulate
d
as
,
Mov
i
n
g Pr
oba
bili
ty
o
f
MN
fr
om
Cur
re
nt N
e
twork
t
o neig
hbori
ng
netw
ork
s is,
N
i
N
j
th
P
b
i
b
j
,
f
o
r
(
i
j
)
G
P
(
18)
W
it
h
in
de
pe
ndent ass
um
ption
s of
b1, b2,
b3,
b4, b
5
the
n,
th
i
,
,
B
i
j
j
,
,
B
j
i
th
Bi
Bj
P
P
b
i
b
j
N
i
N
j
G
i0
j
0
,
i
j
ij
G
f
o
r
(
i
j
)
,
i
j
1
,
2
,
3
,
4
,
5
(
19)
for(
i
≠ j), i =
j
= 1
,
2,
3,
4, 5
The
sta
yi
n
g
P
r
ob
a
bili
ti
es o
f
the MN
is cal
cu
la
te
d
as,
N
i
N
i
N
j
N
i
1
f
o
r
(
i
j
)
,
i
j
1
,
2
,
3
,
4
,
5
PP
(
20)
N
j
N
j
N
i
N
j
1
,
f
o
r
(
i
j
)
,
i
j
1
,
2
,
3
,
4
,
5
PP
(
21
)
Si
m
il
arly
, assu
m
e, Bandwidt
h = B
1 = B
2 =
B3 = B
4
= B
5 = B
.
Traffic
L
oad
of f
i
ve netw
orks
is assum
ed
as,
1
2
3
4
5
.
The
n,
N
i
/
N
j
P
Q
,
i
j
,
i
j
1
,
2
,
3
,
4
,
5
NN
(
)
,
i
j
1
,
2
,
3
,
4
,
5
N
i
N
j
N
i
i
1
j
1
ij
PH
PP
(
22
)
Fr
om
(22
)
, t
he Proba
bili
ty
o
f Han
dove
r (P
H) f
or
t
wo,
th
ree,
f
our
a
nd f
i
ve
Het N
et
s
are
ca
lc
ulate
d.
F
or tw
o netw
orks
, P
roba
bili
ty
o
f
h
a
ndover
=
PH
=
Q
T
hr
ee
n
et
wor
ks
,
p
r
ob
a
bili
ty
o
f han
dove
r
=
P
H=3*
Q
F
our net
works
,
p
roba
bili
ty
o
f han
dove
r
=
P
H
=6*Q
F
ive
netw
orks, Pr
obabili
ty
o
f han
dove
r
=
P
H
=10*
Q
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.
6
,
Dece
m
ber
2
01
9
:
5351
-
5359
5356
3.
RESU
LT
S
A
ND AN
ALYSIS
The
pro
ba
bili
ty
of
ha
ndover
(P
H
)
analy
sis
f
or
m
ulti
ple
(
tw
o,
th
ree,
f
our
a
nd
fi
ve
)
HetNe
ts
has
been
ca
rr
ie
d o
ut
i
n
t
erm
s o
f
tra
ff
ic
load
,
th
reshold
and
band
widt
h.
3.1.
Prob
ab
il
ity
of
hand
ov
er
v
s
t
raffic l
oad
Fig
ure
2
a
nd
Table
1
repres
ents
the
va
riat
ion
of
pro
ba
bili
ty
of
ha
ndover
with
res
pect
t
o
tra
ff
ic
l
oa
d
for
di
ff
e
ren
t
th
reshold
value
s,
her
e
t
he
ba
nd
width
is
fixe
d
at
20
M
Hz.
F
or
five
netw
orks
for
a
fi
xed
t
hr
e
sh
ol
d
(G
th
)=
11,
at
t
he
tra
ff
ic
loa
d
(
)=1,
the
pro
ba
bili
ty
of
hand
over
is
0.8
1272
178
an
d
at
t
he
traff
ic
loa
d
(
)=
8
,
the
pro
bab
il
it
y
of
hand
ov
e
r
is
0.92
2491
97.
It
is
no
ti
ced
that
si
m
il
arl
y
fo
r
t
wo,
three
a
nd
f
our
net
wor
ks
,
at
fixed
thres
ho
l
d value
, th
e
pro
ba
bili
t
y of ha
ndover
sli
gh
tl
y i
ncr
eas
es w
it
h i
ncr
ea
s
e in tra
ff
ic
l
oad.
Fig
ure
2.
Pro
ba
bili
ty
o
f han
dove
r (P
H)
vs t
r
aff
ic
loa
d (
)
f
or d
if
fer
e
nt th
res
ho
l
ds
and d
i
ff
e
ren
t
ne
tworks
at ba
ndwi
dth
(
B
)
=
20
MHz
Table
1.
Prob
a
bili
ty
o
f h
an
do
ver (PH)
vs
t
ra
ff
ic
loa
d (
)
f
or
diff
e
re
nt thresh
old
s
and d
i
ff
e
ren
t
ne
tworks
at ba
ndwi
dth
(
B
)
=20 MHz
S
.
No
.
Tr
af
f
ic load
(
)
Fo
r
two
netwo
rks
at
G
th
=0
,
P
rob
ab
ility
o
f
han
d
o
v
er
(PH)
F
o
r
th
ree
n
et
wo
rks
Prob
ab
ility
o
f
han
d
o
v
er
(PH)
at
G
th
=2
F
o
r
f
o
u
r
n
etwo
rks
Prob
ab
ility
o
f
han
d
o
v
er
(PH)
at
G
th
=6
F
o
r
f
iv
e netwo
rks
Prob
ab
ility
o
f
han
d
o
v
er
(PH)
at
G
th
=
11
1
1
0
.29
8
7
2
0
7
3
0
.60
6
4
6
1
9
4
0
.73
6
2
6
9
2
3
0
.81
2
7
2
1
7
8
2
2
0
.29
8
7
4
0
9
8
0
.60
6
5
5
1
6
5
0
.73
6
7
4
5
4
5
0
.81
8
3
0
9
3
7
3
3
0
.29
8
7
5
9
3
5
0
.60
6
6
3
4
0
3
0
.73
7
1
9
9
1
8
0
.82
4
0
9
9
7
7
4
4
0
.29
8
7
7
9
6
5
0
.60
6
7
2
6
4
7
0
.73
7
7
3
2
8
6
0
.83
1
7
1
0
0
5
5
5
0
.29
8
8
0
3
7
0
0
.60
6
8
3
7
9
8
0
.73
8
4
1
0
8
1
0
.84
2
6
2
2
4
4
6
6
0
.2
9
8
8
3
3
1
0
0
.60
6
9
7
6
7
6
0
.73
9
3
0
1
8
1
0
.85
8
9
0
3
5
4
7
7
0
.29
8
8
6
9
5
6
0
.60
7
1
5
2
2
2
0
.74
0
4
9
4
3
2
0
.88
3
7
7
2
1
9
8
8
0
.29
8
9
1
5
2
3
0
.60
7
3
7
6
2
6
0
.74
2
1
1
0
6
6
0
.92
2
4
9
1
9
7
3.2.
Prob
ab
il
ity
of
hand
ov
er
v
s
thre
sho
l
d
Fig
ure
3
a
nd
T
able
2
re
presen
ts
the
var
ia
ti
on
of
pro
bab
il
it
y
of
ha
ndover
with
res
pect
t
o
th
reshold
f
or
diff
e
re
nt
tra
ff
i
c
loa
d
val
ues
a
nd
for
dif
fere
nt
net
wor
ks
,
here
the
band
widt
h
is
fixe
d
at
20MHz.
As
t
he
num
ber
of
netw
orks
i
nc
reases,
t
he
th
reshold
re
qu
i
r
e
m
ent
of
t
he
MN
va
ries
du
e
to
this
probabil
it
y
of
ha
nd
ov
e
r
is
est
i
m
at
ed
at
di
ff
e
ren
t
th
resho
ld
base
d
on
th
e
no.
of
netw
orks
.
For
fi
ve
ne
tworks
at
thr
e
sh
ol
d
(
G
th
)=
11,traf
fic
load
(
)=
8
the
pro
bab
il
it
y
of
hand
ov
e
r
is
0.
876525
43
a
nd
f
or
the
th
reshold
(
G
th
)=
15,
t
raffic
loa
d
(
)=8
the
pr
ob
a
bili
ty
of
ha
ndover
is
0.6
437302
2.
F
or
tw
o
net
works
at
t
hr
e
s
ho
l
d
(G
th
)=0,
traff
ic
loa
d
(
)=
8
,
the
pro
bab
il
it
y
of
hand
ov
e
r
is
0.566
70640
a
nd
f
or
the
th
res
hold
(
G
th
)=
4,
tr
aff
ic
loa
d
(
)=8
the
pro
bab
il
it
y
o
f
hand
ov
e
r
is
0.1
942773
3.It is o
bs
er
ve
d
that t
he
proba
bili
ty
o
f han
dove
r dec
reases
with i
nc
rease in
th
reshold.
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
A com
pa
r
ati
ve
stud
y
on a
ndov
er p
r
obabil
it
y a
naly
sis f
or f
ut
ur
e
HetN
et
s (S.
Neeraja)
5357
Fig
ure
3
.
Pro
ba
bili
ty
o
f han
dove
r (P
H)
vs
T
hr
es
hold
(
G
th
) for
diff
e
re
nt Traffic loa
d (
)
and d
i
ff
e
ren
t
ne
tworks
at
ba
ndwi
dth
(
B
)
=
20
MHz
Table
2.
Prob
a
bili
ty
o
f han
do
ver (PH)
vs t
hr
esh
old
(G
th
)
f
or
diff
e
ren
t
Tra
f
fic loa
d
(
)
and f
or d
i
ff
e
re
nt n
et
wor
ks
at
band
width (
B
)
=20
M
Hz
S.
No
Thresh
o
ld
(
G
th
)
F
o
r
two
n
etwo
rks
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH
)
At
traff
ic
lo
ad
=8
Thresh
o
ld
(
G
th
)
F
o
r
th
ree
n
etwo
rks
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH
)
At tr
af
f
ic
lo
ad
=8
Thresh
o
ld
(
G
th
)
F
o
r
f
o
u
r
n
etwo
rks
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH)
At
traff
ic load
=8
Thresh
o
ld
(G
th
)
F
o
r
f
iv
e
n
etwo
rks
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH)
At
traff
ic
lo
ad
=8
1
0
0
.56
6
7
0
6
4
0
2
0
.86
9
0
3
0
0
7
6
0
.87
3
4
4
3
6
3
11
0
.87
6
5
2
5
4
3
2
1
0
.38
3
2
6
1
5
4
3
0
.69
8
1
0
7
4
9
7
0
.77
4
6
4
7
3
1
12
0
.80
7
6
3
6
6
8
3
2
0
.28
9
6
7
6
6
9
4
0
.58
2
8
3
1
9
9
8
0
.69
4
8
8
2
4
3
13
0
.74
6
9
3
4
6
1
4
3
0
.23
2
7
0
2
5
0
5
0
.49
9
6
6
8
4
7
9
0
.62
8
9
7
8
7
8
14
0
.69
2
7
4
6
0
6
5
4
0
.19
4
2
7
7
3
3
6
0
.43
6
7
2
1
8
2
10
0
.57
3
4
6
4
2
7
15
0
.64
3
7
3
0
2
2
3.3.
Prob
ab
il
ity
of
hand
ov
er
v
s
. ban
dwi
dt
h
Figure
4
an
d
T
able
3
re
pr
ese
nt
s
the
va
riat
ion
of
pr
ob
a
bili
ty
of
ha
nd
ov
e
r
wi
th
r
espect
to
ba
ndwidt
h
f
or
diff
e
re
nt
thres
ho
l
d
val
ues,
he
re
the
tra
ff
ic
lo
ad
is
fixe
d
at
8.
For
fi
ve
net
w
o
r
ks
f
or
a
fi
xe
d
th
reshold
(
G
th
)=11,
at
the
traff
ic
lo
ad
(
)=8,
the
pro
ba
bili
ty
of
ha
ndover
is
0.7
692307
7
at
the
band
width
=
12
and
at
band
width
=20
t
he
prob
a
bi
li
t
y
of
hand
over
is
0.0
.87
652543.
It
is
noti
ced
that
sim
il
arl
y
fo
r
tw
o,
t
hre
e
an
d
f
our
networks
,
at
fixe
d
t
hr
es
hold
val
ue
of
c
orres
pondin
g
netw
ork
a
nd
t
raffic
loa
d,
th
e
pro
ba
bili
ty
of
ha
ndove
r
sl
igh
tl
y
increases
w
it
h
increase i
n ba
ndwi
dth
.
Figure
4
.
Pro
ba
bili
ty
o
f h
an
dove
r (P
H)
vs
b
andwidt
h (
B
) f
or d
if
fer
e
nt t
ra
ff
ic
loa
d (
)
and d
i
ff
e
ren
t
ne
tworks
at co
nst
ant Thres
ho
l
d (G
th
)
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.
6
,
Dece
m
ber
2
01
9
:
5351
-
5359
5358
Table
3.
Prob
a
bili
ty
o
f han
do
ver (PH)
vs
b
a
ndwidt
h (B)
for
c
on
sta
nt t
hres
ho
l
d (G
t
h)
=
9,1
0,11
at
t
raf
fic l
oad (
)=8
S.No
Ban
d
wid
th
(
B
)
in
MHz
F
o
r
two
n
etwo
rks
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH
)
at
traff
ic load
=8
and
th
resh
o
ld
(
G
th
)
=0
F
o
r
th
ree
n
et
wo
r
k
s
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH
)
at
traff
ic load
=8
and
th
resh
o
ld
(
G
th
)
=2
F
o
r
f
o
u
r
n
etwo
rks
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH
)
at
traff
ic load
=8
and
th
resh
o
ld
(
G
th
)
=6
Fo
r
f
iv
e netwo
rk
s
Prob
ab
ility
o
f
h
an
d
o
v
er
(PH
)
at
traff
ic load
=8
and
th
resh
o
ld
(
G
th
)
=1
1
1
12
0
.56
3
8
8
7
4
6
0
.85
4
6
0
2
9
5
0
.83
3
0
1
9
5
4
0
.76
9
2
3
0
7
7
2
13
0
.56
3
4
8
2
6
7
0
.85
6
0
7
6
4
9
0
.84
0
2
5
6
1
1
0
.79
5
0
0
8
2
8
3
14
0
.56
3
4
2
7
9
0
0
.85
7
7
0
1
8
0
0
.84
6
5
1
5
4
7
0
.81
3
9
5
8
0
7
4
15
0
.56
3
6
7
9
3
4
0
.85
9
5
1
8
2
6
0
.85
2
1
1
4
0
1
0
.82
8
9
8
6
6
0
5
16
0
.56
4
1
5
1
2
8
0
.86
1
4
6
9
8
8
0
.85
7
2
0
8
2
7
0
.84
1
4
4
9
3
2
6
17
0
.56
4
7
5
2
6
8
0
.86
3
4
6
6
6
3
0
.86
1
8
6
3
9
6
0
.85
2
0
7
7
4
3
7
18
0
.56
5
4
0
9
1
5
0
.86
5
4
2
5
8
8
0
.86
6
1
0
8
3
3
0
.86
1
2
9
9
4
0
8
19
0
.56
6
0
7
0
1
0
0
.86
7
2
9
0
4
8
0
.86
9
9
6
0
4
7
0
.86
9
3
8
6
4
5
9
20
0
.56
6
7
0
6
4
0
0
.86
9
0
3
0
0
7
0
.87
3
4
4
3
6
3
0
.87
6
5
2
5
4
3
4.
CONCL
US
I
O
N
In
this pa
per,
t
he
m
at
he
m
at
ical
m
od
el
of
pr
ob
a
bili
ty
of
ha
ndover
f
or
m
ulti
ple
heter
ogen
eous n
et
w
orks
is
dev
el
op
e
d.
Th
e
analy
sis
is
carried
out
by
c
om
par
ing
t
he
pro
bab
il
it
y
of
hand
ov
e
r
of
m
ulti
ple
heter
ogeneou
s
wireless
netw
orks
in
te
rm
s
of
pro
bab
il
it
y
of
hand
ov
e
r
vs
.
tr
aff
ic
loa
d,
probabil
it
y
of
hand
ov
e
r
vs
.
th
reshold
a
nd
pro
bab
il
it
y
of
hand
ov
e
r
vs
.
ba
ndwidt
h.
Fro
m
the
res
ults,
i
t
is
obser
ve
d
t
hat
inc
rease
i
n
traff
ic
l
oad
wit
h
fixe
d
thres
ho
l
d
the
re
is
no
m
uch
va
ri
at
ion
in
pro
ba
bi
li
t
y
of
hand
ov
er
an
d
al
s
o
obs
erv
e
d
t
hat
with
increase
i
n
th
re
sh
ol
d
the
prob
a
bili
ty
of
ha
ndover
is
dec
rease
d.
Fur
ther
,
it
is
al
s
o
ob
s
er
ved
that
f
or
m
ulti
ple
HetNets
,
at
fixe
d
traff
i
c
load,
the
pro
ba
bili
ty
of
ha
nd
ov
e
r
in
creases
with
t
he
inc
re
ase
in
nu
m
ber
of
net
wor
ks
at
diff
e
re
nt
thre
sh
ol
d
values
.
The
pro
po
s
ed
pro
bab
il
it
y
of
ha
dove
r
analy
sis
is
further
hel
pful
f
or
m
ini
m
izing
the
w
ron
g
de
ci
sion
hand
ov
e
r
pro
ba
bili
ti
es
in
pr
a
ct
ic
al
HetNets
su
c
h as
inte
gr
a
ti
on
of
L
TE,
W
i
-
Fi
a
nd
W
i
MAX e
tc
.
f
or
i
m
pr
ov
i
ng
their
QoS.
ACKN
OWLE
DGE
MENTS
This
wor
k
was
sup
ported
un
de
r
Ea
rly
Ca
ree
r
Re
sea
rch
A
w
ard
by
t
he
SE
RB
,
De
par
tm
e
nt
of
Scie
nce
and
Tec
hnolog
y
(D
S
T),
Ne
w
Delhi,
I
nd
ia
(
Gr
a
nt
no
:
ECR
/2016
/
001804
dated:
13
-
FEB
-
20
17).
The
a
ut
hors
would l
ike to
t
hank
t
he
m
anag
em
ent f
or en
c
oura
ging to
wards
researc
h
e
nvir
on
m
ent.
REFERE
NCE
S
[1]
S.
Bhosale
and
R
.
Daruwal
a.,
“
Multi
-
criter
ia
V
ert
i
ca
l
Handoff
Dec
ision
Algor
ithm
Us
ing
Hier
ar
ch
y
Model
ing
a
nd
Additi
ve
W
ei
gh
t
ing
in
an
Inte
gr
at
ed
W
LAN/W
i
MA
X/UM
TS
E
nvironment
-
A
Case
Stud
y
,”
K
SII
Tr
ansacti
ons
on
Inte
rnet
&
Infor
mation
Syst
ems,
vol.
8
,
2014
.
[2]
E
.
Gus
ta
fss
on
a
nd
A
Jons
son,
Eri
csson
Rese
ar
ch
.
,
“
Alwa
y
s
B
est
Conne
ct
ed
,”
IEEE
W
irel
ess
Comm
unic
ati
on
s
,
vol.
10
,
pp
.
49
-
5
5,
2003
.
[3]
B
.
Ma
,
et
al
.,
“
Modeli
ng
and
A
naly
s
is f
or
Ver
tical Handoff
B
ase
d
on
the Decisi
on
Tr
ee i
n
a
Het
ero
gene
ous
Vehi
cl
e
Network,
”
IEEE
Acce
ss
,
vo
l. 5, p
p.
8812
-
8824
,
2
017.
[4]
S.
Nee
ra
ja
and
G.
S
.
Rao
,
“
A
C
om
par
at
ive
stud
y
on
Handoff
Al
gorit
hm
s
for
GS
M
and
CDM
A
Cel
lular
Ne
tworks,
”
Inte
rnational
Jo
urnal
of El
e
ct
ri
c
al
and
Comput
er
Engi
n
ee
ring
(
IJE
CE)
,
vo
l
.
7
,
pp
.
1219
-
1227,
201
1.
[5]
S
.
Nee
r
aja,
e
t
al.
,
“
Anal
y
sis
o
f
A
dapt
iv
e
H
y
ster
esis
Based
H
oriz
o
nta
l
Handoff
Al
gorit
hm
for
GS
M,
”
In
te
rnation
al
Journal
of
Inno
vat
i
ve
Re
search
in
Elec
tric
al
,
El
e
ct
ronics,
Instrum
ent
ati
on
an
d
Control
Enginee
ring
,
vol.
1
,
pp.
433
-
437
,
20
13.
[6]
R
.
S
.
Chang
and
S
.
J
.
L
eu
,
“
Handoff
Orderi
ng
U
sing
Signal
Str
e
ngth
for
Mult
im
edi
a
Com
m
unicati
ons
in
W
ire
l
e
s
s
Networks,
”
I
EEE
Tr
ansacti
ons
on
Wireless
Comm
unic
ati
ons
, v
ol.
3
,
pp
.
1526
-
1
532,
2004
.
[7]
K
.
Yang,
et
al
.,
“
Multi
-
Dim
ensiona
l
Adapt
ive
SIN
R
Based
Vert
i
c
al
Handoff
for
H
et
ero
g
ene
ous
W
i
rel
ess
Networks,
”
IEE
E
Comm
unications
Letters
,
v
ol.
12
,
pp
.
438
-
4
40,
2008
.
[8]
F
.
Zhu
and
J
.
McNai
r,
“
Optimiz
at
ions
for
Ver
ti
c
al
H
andof
f
De
cis
ion
Algorit
hm
s,
”
I
EE
E
Wire
le
ss
Comm
unic
ati
on
s
and
Net
work
ing Confe
renc
e
,
pp.
867
-
872,
2004
.
[9]
D
.
He,
et
a
l
.
,
“
A
Sim
ple
and
R
obust
Vert
i
ca
l
Handoff
Algorithm
for
Hete
rog
e
neousW
ire
le
ss
Mobile
Ne
tworks,
”
Wirel
ess P
ersonal
Comm
unic
at
i
ons
,
vol. 59, pp.
361
-
373
,
2011
.
[10]
C.
Ch
i,
e
t
al.
,
“
Modeli
ng
and
Anal
y
s
is
of
Hand
over
Algor
it
hm
s
,
”
IE
EE
Global
Tele
communic
a
t
ions
Confe
r
enc
e
,
pp.
4473
-
4477
,
2007.
[11]
S.
Nee
ra
ja
and
G.
S
.
Rao
,
“
Perf
orm
anc
e
Anal
y
s
is
of
Bandwid
th
Based
Handoff
Algorit
hm
for
4
G
Hete
roge
n
eou
s
W
ire
le
ss
Ne
tw
orks
Based
On
W
DH
OP
,
”
ARPN
Journal
o
f
Enginee
ring
and
App
li
ed
Sc
ie
n
ce
s
,
vo
l.
12,
pp
.
96
-
106
,
2017.
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
A com
pa
r
ati
ve
stud
y
on a
ndov
er p
r
obabil
it
y a
naly
sis f
or f
ut
ur
e
HetN
et
s (S.
Neeraja)
5359
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
S.
Nee
raj
a
,
s
he
is
pre
sently
wor
king
as
a
As
sist
ant
Profess
or
in
the
Depa
rtment
of
El
e
ct
ron
ic
s
a
nd
Com
m
unic
at
ions
Engi
neering,
GI
TAM
(Dee
m
ed
to
be
Univer
si
t
y
)
.
She
recei
v
e
d
Ph.
D
degr
ee
f
rom
Andhra
unive
rsit
y
in
the
y
e
ar
201
3.
She
has
over
1
3
y
ea
rs
of
T
each
ing
Expe
r
ie
nc
e.
She
has
publi
she
d
m
ore
tha
n
30
res
ea
rch
p
ape
rs
in
v
ari
ous
r
eput
ed
In
te
rna
ti
ona
l/
Nat
io
nal
journals/Con
fer
ences.
H
er
ar
e
a
of
Interest
is
W
ire
l
ess
an
d
Mobile
Co
m
m
unic
at
ions,
W
ire
l
ess
Sensor
Networks
and
CDM
A/MIMO/OFD
M
A
W
ire
les
s
Com
m
unic
at
ions.
Curre
n
tly
,
s
he
is
hand
li
ng
a
DS
T
-
SERB
rese
arc
h
proje
c
t.
A.
Ab
his
hik
t
h
a
,
s
he
is
pr
ese
ntly
working
a
s
a
JRF
in
th
e
Depa
r
tme
nt
of
El
e
ct
ron
ic
s
and
Com
m
unic
at
ions
Engi
nee
r
ing,
G
ITAM
(Dee
m
ed
to
be
Univer
si
t
y
)
under
DS
T
-
SERB
rese
arc
h
project
.
She
has
rec
e
ive
d
the
B.
Tech
deg
ree
from
Vignan
Nirula
Eng
ine
e
r
ing
Coll
eg
e
and
M.E
from
AN
ITS
Engi
ne
eri
ng
Col
le
ge
.
Her
ar
ea
s
of
in
te
r
ests
ar
e
W
ir
el
ess
Co
m
m
unic
at
ions
a
nd
W
ireless
Se
nso
r
Networks.
Evaluation Warning : The document was created with Spire.PDF for Python.