TELKOM
NIKA
, Vol.13, No
.1, March 2
0
1
5
, pp. 181~1
9
2
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i1.365
181
Re
cei
v
ed Se
ptem
ber 3, 2014; Re
vi
sed
No
vem
ber 2
6
,
2014; Accep
t
ed De
cem
b
e
r
15, 2014
Critical Data Routing (CDR) for Intra Wireless Body
Sensor Networks
Jav
ed Iqbal
Bang
ash
1
, Abdul Hana
n Abdullah*
1
, Abdul Wahe
ed Kha
n
1
, Mohammad Abdur
Ra
zzaque
1
, Roha
na Yus
o
f
2
1
F
a
cult
y
of Co
mputin
g, Unive
r
siti T
e
knologi
Mala
ysi
a
(UT
M
), 81310 Sk
uda
i Johor, Mal
a
ys
ia
2
Universiti Ku
a
l
a Lum
pur Mal
a
y
s
ia, Instit
ute of Industrial T
e
chno
log
y
, Mal
a
ysi
a
T
e
l.: +
607-553-
876
1 F
a
x: +
6
0
7
-55
3
-88
2
2
*Corres
p
o
ndi
n
g
Author: emai
l
:
hanan
@utm.m
y
A
b
st
r
a
ct
T
he life critic
al
and re
al ti
me
me
dica
l ap
pl
i
c
at
ion
of W
i
rel
e
ss Body S
e
n
s
or Netw
orks (W
BSNs)
requ
ires the as
suranc
e of
the
de
ma
nde
d Qu
ality of Servic
e
(QoS) bot
h in
terms d
e
lay
an
d reli
abi
lity. T
h
i
s
pap
er pr
op
ose
s
Critica
l
D
a
ta
Rout
i
ng (
CDR)
that cate
gori
z
es the s
ens
ory
data
pack
e
ts
as critica
l
a
n
d
non-
critical
data
pa
ckets. Alon
g w
i
t
h t
he
heter
og
ene
ous
natur
e
d
d
a
ta, it a
l
so
addr
esses th
e
hig
h
a
nd
dyn
a
m
i
c
path
loss
and
the te
mp
erat
ure ris
e
iss
u
e
s
caus
ed
by
postura
l
mov
e
me
nt of th
e h
u
man
bo
dy a
n
d
electro
m
agn
eti
c
w
a
ves abs
o
r
ption r
e
sp
ecti
vely. T
he s
i
mulati
on r
e
sults
show
that th
e pro
pos
ed
C
DR
sche
m
e
achi
ev
es its des
ig
ned
obj
ective
of forw
arding th
e
crit
ical
data
packe
ts w
i
thin certai
n time l
i
mits an
d
w
i
th highest rel
i
abi
lity w
h
ile re
duci
ng the te
m
peratur
e rise of
the in-bo
d
y se
nsor no
des.
Keywo
r
k
s
:
w
ir
eless b
ody se
n
s
or netw
o
rks (W
BSNs), routi
ng, QoS, critical, temperatur
e
,
path loss
1.
Introduc
tion
In WBS
N
–
used
for contineo
use a
nd
remote
h
ealthcare
mo
nitoring
[1], the tiny
,
lightwei
ght, cost effe
ctive
and l
o
w-po
wer Bio
-
Me
dical Sen
s
o
r
De
vices (BMS
Ds) are
deploy
ed o
n
and/or in
sid
e
the human
b
ody to sen
s
e
and an
alyz
e t
he vital sign
data of the h
u
man bo
dy. It
It
has a three-tiered architec
t
u
re in
order t
o
send the
s
e
ns
ory data to
the final
des
t
ination [2],[3]. In
the firs
t tier i.e. Intra-WBS
N
, the tiny B
M
SDs s
end
t
he vital
sign
data to
on
-bo
d
y ba
se
stati
on
calle
d as Bo
d
y
Coordinato
r
(BC). In the
se
con
d
ti
er i.e. Inter WBS
N
, the BCs
are re
spo
n
sibl
e
to
forwa
r
d the
received vital-sig
n
data towa
rd
s
the sin
k
(s) u
s
ing
other BCs
and/or
regul
a
r
infrast
r
ucture like wireless
l
o
cal
area net
work.
Fi
nally
,
in the thi
r
d tier, it’s the
responsi
b
ility of the
sin
k
(s) to
se
n
d
the re
ceive
d
vital-sig
n
d
a
ta to the fin
a
l de
stination
whi
c
h
coul
d
be a p
h
ysi
c
ia
n,
health-ca
r
e
sever and/o
r
e
m
erg
e
n
c
y co
ntrol ro
om, using reg
u
la
r infrast
r
u
c
ture
su
ch a
s
intern
et.
As
WBSN
de
als
with th
e
h
u
man
body
a
nd d
ue to
structure, n
a
ture an
d b
ehavi
o
r
of the
human b
ody, it faces so
m
e
uniqu
e ch
al
lenge
s alo
ng
with the tradit
i
onal con
s
trai
nts of Wirele
ss
Senso
r
Networks
(WS
Ns). A range of
BMSDs,
like
blood p
r
e
ssure, Ele
c
troe
nce
phal
ograp
hy
(EEG), tem
p
eratu
r
e, Ele
c
troca
r
di
ograp
hy (ECG
) an
d many
more, are
de
plo
y
ed for different
appli
c
ation
s
,
whi
c
h m
a
ke
WBSN h
e
terogen
eou
s in
nature.
Du
e to its h
e
tero
g
eneo
us
nature it
gene
rate
s dif
f
erent
categ
o
r
ies of data,
whe
r
e
different QoS p
a
rameters a
r
e
among
the
key
requi
rem
ents.
Similarly, due to
sali
ne-wate
r na
ture, the h
u
man tissu
e
s
ab
so
rb t
he
electroma
gne
tic wave
s
carrying th
e
informatio
n
in wirele
ss commu
ni
cation
s. Thi
s
electroma
gne
tic wave
s ab
sorption al
on
g with the e
n
e
rgy con
s
um
ption du
ring t
he ope
ratio
n
s
of
the BMSDs result in
temp
eratu
r
e
rise
of the BM
SDs. In case of
the impla
n
te
d BMSDs, this
temperature
rise might affect and/
or damage the human tissues i
f
remain for l
ong time [4],[5]. In
addition
to th
ese, th
e p
o
st
ural
moveme
nt of
the
hu
man b
ody al
ong
with th
e
ele
c
trom
agn
etic
wave
s ab
so
rption re
sult
s in dynami
c
an
d high p
a
th lo
ss
while
exch
angin
g
inform
ation with oth
e
r
implanted BM
SDs an
d/or B
C
. Due to this high and
dyn
a
mic path lo
ss the co
nventional path lo
ss
model
s of wireless commu
nicatio
n
are n
o
t applicable
for intra WBS
N
s.
Different
peo
ple have
trie
d to ad
dre
ss
these
chall
enge
s of
WBSNs a
nd
p
r
opo
se
d
variou
s routin
g proto
c
ol
s. .
The high an
d
dynamic pat
h loss i
s
sue o
f
intra WBSN due to postu
ral
movement of
human bo
dy has be
en a
ddre
s
sed in [
6
]–[9]. In [6]
Quwaide
r
an
d Biswa
s
ha
ve
prop
osed a routing sch
e
m
e
by partition
ing the sens
or field into
different pa
rtitions an
d use
s
store a
nd floo
d mech
ani
sm
to route the sen
s
o
r
y data towards BCl
While in [7]
th
e
s
a
me
au
th
or
s
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 1, March 2
015 : 181 – 1
9
2
182
have call
ed i
n
tra WBS
N
a
s
Del
a
y Tole
rant Ne
twork (DT
N
) usi
ng store and
fo
rward
ap
pro
a
c
h
.
Similarly, in [
8
]
the
same
partitionin
g
a
ppro
a
ch i
s
b
e
ing
used a
s
in [6]
a
nd u
s
e
s
store
an
d
forwa
r
d me
chani
sm. in [8]
,
the authors have u
s
ed
Line-of
-Si
ght
(LoS) a
nd
Non
-
Lin
e
-of
-
Sight
(NL
o
S) me
ch
anism to forward the gen
erated a
nd/
or receive
d
dat
a packet
s
toward BC. All
the
aforem
ention
ed scheme
s
use
no
n real
time
routin
g
m
e
ch
ani
sm by
storin
g the
p
a
c
kets
at
seve
ral
interme
d
iate
node
s which
make
s the
m
unre
a
listi
c fo
r critical medi
cal ap
plicatio
ns. Fu
rtherm
o
re,
they do not consi
der the h
e
terog
ene
ou
s natured
d
a
ta
and the thermal effects of
the implante
d
BMSDs
.
Tempe
r
atu
r
e Aware Routin
g
Algo
rithm (TARA)
[5]
i
s
the first
ro
uting p
r
oto
c
ol t
hat aim
s
to decrea
s
e
the therma
l effects of the implan
ted BMSDs.
Each nod
e
obse
r
ves t
h
e
comm
uni
cati
on a
c
tivities
of its neig
h
b
o
r no
de
s
an
d estimate
s t
heir tem
perature. A nei
gh
bor
node having temperature
rise
great
e
r
than
ce
rtain t
h
re
shol
d i
s
d
e
cla
r
ed
hot
spot nod
e, wh
ich
increa
se
s the
delay and en
ergy co
nsum
ption. The sh
ortco
m
ing of
TARA has b
e
en add
re
ssed
in
Lea
st
Tem
p
e
r
ature Rise (L
TR)
[10]
whe
r
e a
hop
-cou
n
t
is
asso
ciate
d
with
ea
ch
d
a
ta pa
cket. T
he
data pa
cket is di
sca
rd
ed if the
hop
-cou
nt rea
c
he
s b
e
yond the th
resh
old value,
which result
s in
low p
a
cket
d
e
livery ratio.
Another t
e
mp
eratu
r
e
a
w
a
r
e ro
uting
sch
e
me n
a
med
as Ad
aptive L
east
Tempe
r
atu
r
e
Rise (ALT
R)
prop
osed in [
10]
,
whe
r
e in
stead of d
r
op
ping the d
a
ta
packets it u
s
es
Shorte
st Ho
p Algorithm
(SHA) to se
nd the
data
packets to
wards B
C
. Least Total Route
T
e
mp
er
a
t
ur
e (
L
T
R
T
)
pr
es
e
n
t
ed
in
[1
1
]
is desi
gne
d to focus
o
n
the entire
route in
stead
of
individual no
des a
nd hig
h
hop-co
unt. All the
afo
r
eme
n
tione
d
temperatu
r
e
aware routi
n
g
scheme
s
a
d
d
r
ess the th
ermal effects
of the im
plante
d
BMSDs
whi
l
e com
p
letely
ignore the Q
o
S
para
m
eters d
e
mand
ed by hetero
gen
eo
us natu
r
ed d
a
te and the high and dyna
mic path loss of
Intra WBSNs, due to whi
c
h the
s
e a
r
e
not suit
able
and practi
cally impleme
n
table for Int
r
a
WBSNs
.
In [12]
,
the a
u
thors divide
the vital sig
n
data into
four c
a
tegories
.
(1)
Critical Dat
a
(CD) –
need to be transmitted
with least delay
and high
est
reliability. (2) Delay Sen
s
itive Data (DS
D
)
–
requi
re l
e
a
s
t
delay an
d
ca
n tolerate so
me pa
ck
ets l
o
ss. (3
)
Relia
bility Sensitive Data
(RSD) –
need
high
est
reliability a
nd
can
a
c
cept
some d
e
lays
.
(4)
Reg
u
lar Data (RD)
– do
not d
e
man
d
f
o
r
any QoS pa
rameter. It uses two si
nks – prim
ary a
nd seconda
ry, for each
p
a
tient. Each
data
packet is forwarded to
wa
rds b
o
th sin
k
s, whi
c
h re
sults in increa
se net
work traffic. DMQ
o
S
prop
osed i
n
[
13]
u
s
e
s
the
same
catego
ries
of data
as in [12]
that
ai
ms to
provid
e
the de
man
d
ed
QoS paramet
er ba
sed o
n
the nature of data pa
ck
ets.
It uses hop
-b
y-hop whe
r
e the so
urce no
de
is co
mpletely
depen
dent u
pon a
single
node in te
rm
s of laten
c
y and/or reliabilit
y. This locali
zed
hop-by-h
op a
ppro
a
ch d
o
e
s
not
en
sure
the su
cce
ssf
ul delive
r
y of the data
pa
ckets.
QPRD [
14]
and QPRR [15]
are two
other QoS a
w
are routin
g scheme
s
de
signed to displ
any the patient’s
vital information. QPRD cl
assifies the g
enerated
traff
i
c into DS
D a
nd ND whil
e QPRR
divide
s it
into RSD an
d ND. All the aforeme
n
tioned QoS
-
aw
are routing schem
es
con
s
ider inter
WB
SNs
comm
uni
cati
on a
n
d
comp
letely igno
re
the uni
que
c
hallen
ges i.e.
high
a
nd
dynamic p
a
th
loss
and tempe
r
at
ure ri
se i
s
sue
s
of intra WB
SNs.
To the b
e
st o
f
the autho
r’s kno
w
le
dge,
TMQoS [16]
and RAR
[17
]
are the
Qo
S-aware
routing
sche
mes in the
e
x
isting literat
ure, that
is d
e
sig
ned fo
r intra WBS
N
s.
TMQoS con
s
ide
r
s
thermal
effect
s of
the i
m
pla
n
ted BMS
D
s
along
with
th
e hete
r
o
gene
ous natu
r
e
d
data p
a
cket
s. It
divides t
he p
a
tient’s vital i
n
formatio
n in
to four
cla
sses
sam
e
a
s
i
n
[12],[13]. TMQoS p
e
rfo
r
ms
well in orde
r to provide th
e deman
ded
QoS paramet
ers
as
comp
ared to oth
e
r state-of-th
e
-art
scheme
s
giv
en in [5], [11]
.
While ro
utin
g the criti
c
al
data pa
ck
ets,
it send
s two
copie
s
of ea
ch
data p
a
cket
simultan
eou
sl
y, one to
the
neig
hbo
r
no
de h
a
ve le
ast delay
while
other to the
one
having highest
reliability. Redundant
data packets delivery results
in hi
gh
net
work traffic,
more
netwo
rk
co
n
gestio
n
and t
hus
red
u
ci
ng
delivery succ
e
ss
ratio. Similarly, it also ca
uses m
o
re
energy con
s
u
m
ption a
nd
high temp
eratu
r
e
rise of th
e
implanted
BMSDs. F
u
rth
e
rmo
r
e, it i
s
also
not payin
g a
n
y attention
towards the
high
and
dy
namic p
a
th l
o
ss of
intra
WBSNs.
RAR
con
s
id
ers the
temperatu
r
e
rise i
s
sue of the
implante
d
BMSDs al
ong
with dynami
c
and high
pat
h
loss of int
r
a
WBSNs a
nd
the divids the
data p
a
cket
s co
ntaining
th
e vital sig
n
s i
n
formatio
n in
to
reliability
constrained data packets and normal
dat
a packets. It does not
consider the
critical
data that need to be transmitted with
least delay and highest reliability.
To the be
st of the autho
rs’ kn
owl
edg
e
,
no
su
ch
scheme exi
s
t in the litere
c
ture that
ensure
s
th
e
provisi
on
of the d
e
man
d
e
d
QoS
pa
ra
meters
a
nd consi
ders
th
e high and
dyn
a
mic
on a
nd i
n
-b
o
d
y path l
o
ss whil
e mai
n
taining
the
te
mperature
of
the im
plante
d
BMSDs
at an
accepta
b
le le
vel for ci
ritcal
data p
a
cket
s. In th
is pa
pe
r, we
have
propo
s
ed
Critical Data
Routing
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Critical Data
Routing (CDR) for Intra Wire
le
ss Bo
dy
Senso
r
Network
(Javed Iq
bal B.)
183
(CDR) fo
r In
tra
WBSNs t
hat aim
s
to
provide
be
st
ro
utes for critical
data
that ne
ed to
be
transmitted
within
certai
n time fram
e and hi
ghe
st po
ssi
ble reliability. Beside
s late
ncy
and
reliability, it also ad
dresse
s
the path l
o
ss
and tem
p
eratu
r
e issu
e
s
of the impl
anted BMSDs. It
focuses on end-to-end
path latency, reliability and temper
ature without mai
n
tain end-to-end
path, whe
r
e each interme
d
iate sen
s
o
r
node is in
volved in rout
e sele
ction p
r
ocess. Its use
s
modula
r
ba
sed app
roa
c
h
whe
r
e ea
ch module
h
a
s a
ssi
gne
d
its duty. We evaluated
and
comp
ared ou
r pro
p
o
s
ed CDR
scheme i
n
terms of
de
livery failure ratio due to high and dyna
mi
c
path lo
ss, d
e
livery succe
s
s ratio,
on
-time de
live
r
y su
cce
s
s rati
o, ene
rgy
co
nsum
ption
a
n
d
temperature
rise with oth
e
r state-of-th
e
-art schem
es.
The re
st of this pa
per
ha
s bee
n organ
ized a
s
follo
w: Section 2
pre
s
ent
s the
resea
r
ch
method u
s
ed,
while the propo
sed Critical Data Ro
uting (CDR) is
being di
scussed in Section
3.
In Section 4, we present re
su
lts a
nd di
scussion a
nd finally Section
5 con
c
lu
de
s this pa
per.
2. Researc
h
Method
2.1. Net
w
o
r
k Model
Con
n
e
c
tivity
grap
h can be
used to mo
d
e
l
the differe
nt implanted
BMSDs a
nd
on-b
ody
BC with re
pla
c
ea
ble po
we
r source, a
s
gi
ven in equati
on (1
).
)
,
(
E
V
G
(1)
Whe
r
e
V
corre
s
p
ond
s to the set of
N
implanted B
M
SDs
and
BC, i.e.
BC
U
s
s
s
s
V
n
.
,.........
,
,
3
2
1
while
E
co
rresp
ond
s to
M
possible
wirele
ss lin
ks betwe
en any
two BMS
D
s
and/or a
BM
SD a
nd B
C
i.
e.
m
e
e
e
e
E
.
,.........
,
,
3
2
1
. T
h
e ass
u
mp
tio
n
s
made
ar
e
:
all BMSDs e
x
erci
se th
e
same a
nd lo
w tran
smi
ssio
n
power du
rin
g
communi
ca
tion with
oth
e
r
BMSDs an
d BC. Secondly
,
the BMSDs might play t
he role of sou
r
ce nod
e as well as forward
i
ng
node
s.
2.2. Classifi
cation o
f
Da
ta
Con
s
id
erin
g the critical
m
edical ap
plications
of WB
SNs, the
ob
serve
d
patie
nt’s vital
informatio
n can be
categ
o
rized a
s
Cri
t
ical Data
(
C
D) a
nd
Non
-
Critic
al D
a
ta
(N
CD
). The
CD
packet
s
need to be transm
itted within
certai
n time f
r
ame
wi
th
highest possi
bl
e reli
ability while
NCD pa
ckets do not dema
nd any su
ch
QoS paramet
ers.
3. Proposed Critical Data
Ro
uting (
C
D
R
)
for Intr
a WBS
N
s
Figure 1 given below illustrates the network arch
itecture of the prop
osed CDR, which i
s
a cross-layered mod
u
la
r
based a
ppro
a
ch. It’s the
job of
MAC Re
cei
v
er
to
rec
e
ive the data
and/or
Hello
packets fro
m
neigh
bor
node
s an
d/o
r
BC an
d forwa
r
d the
m
toward
Packets
Cla
ssif
i
e
r
wh
ere th
e in
co
ming pa
cket
s a
r
e
cla
ssif
i
ed a
s
data
and
Hello
Packets. After
cla
ssifi
cation,
the Hello a
nd data pa
ckets a
r
e rout
ed to routing
module an
d
data packet
s
cla
ssifie
r
re
spectively. Up
on re
ceiving
the data pa
ckets eith
er from pa
ckets
cla
ssifie
r
or f
r
om
uppe
r layers,
the
Dat
a
P
a
ck
et
s
Clas
sif
i
er
cl
assifie
s
them a
s
CD
and NCD an
d sen
d
s th
e
m
to
QoS-a
w
a
r
e
n
e
xt-hop
sel
e
ctor. On
the
other han
d, the
MAC Transmitter
fo
r
w
ar
ds
th
e d
a
t
a a
n
d
/
or
Hello
pa
ckets towards
other BMS
D
s
and/or BC.
The d
e
tailed
description
of the re
mai
n
ing
module
s
is gi
ven in the followin
g
su
b-se
ction
s
.
3.1. Dela
y
Estimator
At any no
de
n
i
, the e
s
tima
tion an
d
cal
c
ulation
of the
delay i
s
don
e by d
e
lay e
s
timato
r
usin
g equatio
n (2) given b
e
low, where
QD
ni
is the q
ueue del
ay and
TD
i,j
is the
delay cau
s
e
d
by
transmitting the pa
cket
P
from nod
e
n
i
to node
n
j
over
link
L
i,j
. Proce
s
sing
and p
r
opa
gat
ion
delays
a
r
e other
delay
s experie
nced by
a packet
P
,
but the
s
e
are
negli
g
ibly
small an
d
ca
n
be
ignored.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 1, March 2
015 : 181 – 1
9
2
184
j
i
ni
ni
TD
QD
ND
,
(2)
Queu
e Delay
QD
ni
at node
n
i
is the time for whi
c
h the packet
P
rem
a
ins in qu
eue
before
transmissio
n
and i
s
co
nsi
d
ered
only for
CD
pa
cket
s
i
n
our
pro
p
o
s
ed CDR. It can be
cal
c
ula
t
ed
by using Exp
onentially Weighted M
o
ving Aver
ag
e (EWMA) form
ula sa
me a
s
in [18]
,
given in
equatio
n (3
). The value of
smoothi
ng fa
ctor
con
s
tant
α
rang
es b
e
twee
n 0 an
d 1
and we ch
oo
se
α
= 0.2
i
n
ou
r simul
a
tion
same a
s
in [13],[14],[16]. I
n
itially
QD
ni
i
s
the
que
ue
delay expe
rie
n
ce
d
by firs
t CD pac
ket.
ni
ni
ni
QD
QD
QD
1
(3)
Tran
smi
ssi
on
Delay
TD
i,j
is the time for
whi
c
h the
pa
cket
P
rem
a
i
n
in the MA
C layer of
node
n
i
before either
su
ccessfully tran
smitted to no
de
n
j
over li
nk
Li,j
o
r
dro
pped. it ca
n
be
cal
c
ulate
d
using the form
u
l
a given in [1
4]
as in e
qua
tion (4
), whe
r
e
DR
bits
repre
s
ent
s the dat
a
rate in bits
,
SP
bits
repre
s
ents th
e si
ze
of pa
ckets i
n
bits
and
NP
is the
num
ber
of tran
smitted
pack
e
t
s
in time interval
δ
t
.
NP
z
SP
DR
TD
NP
z
bits
bits
j
i
1
,
1
(4)
3.2. Reliability
Estimator
The
responsi
bility of the reliability esti
ma
tor i
s
to estimate and
calcul
ate the
LR
i,j
–
the
average reli
ability of the li
nk
L
i,j
from
n
ode
s
n
i
to no
de
n
j
. Is
NP
s
u
ccessfu
l
is the numbe
r of th
e
packet
s
that
are
su
ccessf
ul tran
smitted
and
NP
total
is the total packets t
r
an
smitted, then
P
average
given in equation (5) is the
av
erage
probability of the successf
ul
transmi
ssi
on ov
er li
nk
L
i,j
duri
ng
time interval
δ
t
. Windo
w Mea
n
with
Exponentiall
y Weighte
d
Moving Average (WME
WMA)
formula
sam
e
as in [1
9]
,
can be
used t
o
cal
c
ul
ate
LR
i,j
i.e. the av
erage lin
k reli
ability of link
L
i,
j
betwe
en the
transmitting n
ode
n
i
and th
e re
ceiving
n
ode
n
j
, give
n
in equ
ation (6). The val
u
e
of
the weig
hting
factor
β
ran
g
e
s bet
wee
n
0
and 1 and
we cho
o
se
β
= 0.4
in our si
mulation, sa
me
as
in [13],[15],[16].
total
successful
average
N
N
P
(5)
average
j
i
j
i
P
LR
LR
1
,
,
(6)
3.3. Path Lo
ss Estimato
r
The
estimati
on a
nd
cal
c
u
l
ation the
pat
h lo
ss
PL
i,j
of
the
wirele
ss link
L
i,j
between
th
e
transmitting node
n
i
an
d receiving no
d
e
n
j
is ca
rrie
d
out by path loss estim
a
tor. The de
rived
versio
n of Frii
s form
ula [20]
,
kno
w
as
se
mi-empi
r
i
c
al formul
a [21]
,
can be u
s
e
d
to model
PL
i,j
as
a function of t
he dista
n
ce
d
i,
j
betwee
n
the
transmitting
node
n
i
a
nd t
he re
ceiving
node
n
j
, give
n in
equatio
n (7
), whe
r
e
PL
0
de
noted the ref
e
ren
c
e p
a
th loss at the ref
e
ren
c
e di
stan
ce
d
0
.
0
,
0
,
log
10
d
d
n
PL
PL
j
i
j
i
(7)
Ho
wever, du
e to dynamic
postu
ral mov
e
ments
of the
human bo
dy, the path loss of intra
WBSNs is dy
namic i
n
natu
r
e. Zero-m
ea
n Gau
s
sian
random va
riab
le
X
σ
with sta
ndard deviati
on
σ
ca
n be u
s
e
d
to formulate
equation (7)
as
in eq
uatio
n (8). Th
e qu
ality of the lin
k
L
i,j
, denoted
by
LQ
i,j
, betwe
e
n
the tran
smi
tting node
n
i
and
receiving
n
j
can be
calcul
ated u
s
i
ng equ
ation
(9)
derived from
equatio
n (8
)
same
as in [2
1]
with tran
smitting power of
P
trans
, path loss of
PL
i,j
from
equatio
n
(8) and
thresho
l
d level
of
LQ
thre
. Furthermore, the transmitting n
ode
n
i
can
only
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Critical Data
Routing (CDR) for Intra Wire
le
ss Bo
dy
Senso
r
Network
(Javed Iq
bal B.)
185
comm
uni
cate
with th
e receiving no
de
n
j
if the q
ualit
y of the lin
k
L
i,j
between
them, de
note
d
by
LQ
i,j
, is great
er th
an
or eq
ual to th
e
pre-defin
e
th
re
shol
d level. I
n
ou
r
pro
p
o
s
ed
CDR
sch
e
me,
we use
the
whole body pa
th
loss mod
e
l
propo
sed
in
[21]
,
whi
c
h
covers the
different
path
lo
ss
model
s propo
sed for int
r
a
WBSNs.
X
d
d
n
PL
PL
j
i
j
i
0
,
0
,
log
10
(8)
2
2
1
2
1
,
,
thre
j
i
Trans
j
i
LQ
PL
P
erf
LQ
(9)
3.4. Tempera
t
ure Es
timator
The jo
b of th
e tempe
r
atu
r
e e
s
timator i
s
to e
s
timate
and
cal
c
ulate
the temp
erat
ure
ri
se
occurre
d
at
any implante
d
BMSD
n
i
.
As di
scusse
d ea
rlier, the
human
tissues
ab
sorb t
he
electroma
gne
tic wave
s du
ring the wi
rele
ss
co
mm
uni
cation amo
ng
the implante
d
BMSDs. Th
e
rate at whi
c
h
the human t
i
ssue, with d
ensity of
ρ
a
nd ele
c
tric
condu
ctivity of
σ
, abso
r
b the
electroma
gne
tic wave
s, ha
ving indu
ced
electri
c
field o
f
E
, per unit weight as
[5]:
2
E
SAR
(10
)
The
se
con
d
reason that
causes
th
e te
mperature
ri
se of the im
pl
anted BMS
D
s, is th
e
energy con
s
umption requ
ired to ca
rry
out
the different operation
s
, denote
d
b
y
P
c
, which
can
measured by
dividing the
power
con
s
u
m
ed by volu
me of the implanted BMS
D
. Finally, the rate
at which the temperature
of
the implan
ted BMSDs rise
s ca
n be cal
c
ulate
d
by using Pen
n
e
s
Bioheat form
ula [22]
,
given in equ
ation
(11
)
. Table
(1) p
r
ov
ide
s
the explan
atio
n of the different
para
m
eters o
f
equation (1
1
)
while thei
r value
s
are o
b
tained fro
m
[23].
p
c
b
C
P
SAR
T
T
b
T
K
dt
dT
2
(11)
Figure 1. Net
w
ork Archite
c
ture of the Propo
sed
CDR
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 1, March 2
015 : 181 – 1
9
2
186
Table 1. De
scriptio
n of the Paramete
rs
use
d
in Equat
ion (11
)
Parameter Description
dT/dt
Rate at
w
h
ich ris
e
temperatu
r
e
occur
s
T
K
2
TR due to
tissue’s thermal
conductivity
b
T
T
b
TR due to
blood
perfusion
SAR
TR due to
electro
m
agnetic
w
a
ve absor
p
tion
P
c
TR due to
po
w
e
r
consumed b
y
nodes’ operation
Mass density
C
p
Tissue’s
specific
heat
3.5. Routing
Module
The three
sub-m
odul
es
of the routin
g modul
e a
r
e: routing ta
ble con
s
tru
c
tor, Hell
o
packet
s
constructor and routing t
abl
e.
The
responsi
bility of the r
outing tabl
e
constructor is to
con
s
tru
c
t an
d peri
odi
cally update the
routing ta
ble
base
d
on t
he inform
atio
n obtaine
d from
variou
s pa
ra
meters estim
a
tors a
nd oth
e
r BMSD
s in neigh
borhoo
d
.
After obtain
i
ng the req
u
ired
informatio
n
e
quation
s
(12
)
,
(13), and (1
4) c
an
used t
o
cal
c
ul
ate th
e end
-to-end
path del
ay
PD
i,j
,
reliability
PR
i,
j
, and temperatu
r
e
PT
i,j
from the transmitting
n
ode
n
i
to
BC
thro
ugh t
he
interme
d
iate node
n
j
resp
ectively. The
routin
g tabl
e pa
ramete
rs are
given i
n
Figure 2,
while
Table (2) de
scribe
s the
s
e
param
eters. Once the
routing table i
s
co
nstructe
d or upd
ated
, it
provide
s
i
n
formation to
Hel
l
o pa
ckets
ge
nerato
r
to
ge
nerate
the
Hello pa
cket in
ord
e
r to
info
rm
other BMSDs in its neighb
orho
od.
ni
j
i
j
i
ND
PD
PD
,
,
(12
)
j
i
j
i
j
i
LR
PR
PR
,
,
,
(13
)
ni
j
i
j
i
NT
PT
PT
,
,
(14
)
3.6. QoS-A
w
are Ne
xt-Ho
p
Selector
The re
sp
on
si
bility of QoS-aware next-hop sele
ctor is to sele
ct
the desi
r
ed
next-hop
based o
n
the
deman
ded
Q
o
S paramete
r
s.
Up
on
re
cei
v
ing the data
packet
s
P
fro
m
data p
a
cke
t
s
cla
ssifie
r
m
o
dule,
our p
r
opo
sed
critical dat
a
ro
uting al
gorith
m
, given
belo
w
, sea
r
che
s
t
he
routing ta
ble
(
RT
) for
only
those
nod
es i
n
the nei
ghb
orho
od
who
s
e link
quality (
LQ
i,j
)
is
gr
eate
r
than or eq
ual
to the pre-de
fined thre
shol
d level (
LQ
thre
) and pla
c
e th
em in
NN
LQ
(li
nes: 2–
4). Th
e
data pa
cket
P
is di
scard
ed
immediately i
n
ca
se
of em
pty
NN
LQ
(lin
e
s
: 5-6). Othe
rwise, in case
of
P
belon
ging t
o
CD
, Delay
Aware Pro
c
e
dure
is call
ed
with in
puts
NN
PD
and
P
(li
nes:
7–8
). While
in ca
se of
P
belon
ging
to
NC
D
, the desired next hop
(
DN
H
) is the
node belo
ngi
ng to
NN
LQ
with
least e
nd-to
-end p
a
th te
mperature
(
PT
i,j
) (lin
e: 9).
Once del
ay awa
r
e
pro
c
edure is called, it
sele
cts only t
hose n
ode
s b
e
longi
ng to
NN
LQ
, w
h
os
e en
d
-
to-
e
nd
pa
th
d
e
l
a
y
(
PD
i,j
) is l
e
ss th
an
or
equal to req
u
i
red del
ay (
PD
req
) and pla
c
e them in
NN
PD
(lines: 10
–12). In ca
se
of empty
NN
PD
,
the data pa
cket
P
is discarded imme
diat
ely (lines: 1
3
–14). If there
is a si
ngle n
o
de in
NN
PD
then
that node i
s
selecte
d
a
s
DN
H
(line
s
: 1
5
–16).
Otherwi
se
reliability a
w
are p
r
o
c
ed
u
r
e i
s
called
wi
th
inputs
NN
PD
and
P
(line:
17).
Upo
n
ca
lling reliabilit
y awa
r
e p
r
o
c
edure, it sel
e
cts
only tho
s
e
node
s
who
s
e
end
-to-e
nd p
a
th reli
ability (
PR
i,j
) i
s
g
r
e
a
t
er than
or
eq
ual to the
req
u
ired
reli
abilit
y
(
PR
req
) a
nd
store the
m
in
NN
PR
(li
n
e
s
:
18–2
0). If no
ne of n
ode
s
fulfills the re
quire
d reliabil
i
ty
deman
d then
the
DN
H
is th
e node bel
on
ging to
NN
PD
having high
e
s
t
PR
i,j
(line
s
: 21–22
). If th
ere
is a singl
e en
try in
NN
PR
then that node
is sel
e
cte
d
as
DN
H
(line
s
: 23–2
4). Othe
rwi
s
e, the no
de
belon
ging to
NN
PR
with leas
t
PT
i,j
is selected as
DN
H
(line:
25
).
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Critical Data
Routing (CDR) for Intra Wire
le
ss Bo
dy
Senso
r
Network
(Javed Iq
bal B.)
187
Table 2. De
scriptio
n of the Paramete
rs
use
d
in Figure (2
)
Parameter Description
ID
Des
t
Destination (bod
y coordinato
r) I
D
Loc
Des
t
Coordinates of th
e bod
y
coo
r
dinat
or
ID
n
j
ID of the neighb
o
r
node n
j
LQ
i,
j
Link Qualit
y
bet
ween nodes n
i
, an
d n
j
PD
i,j
End-to-e
nd
path dela
y
from node n
i
to BC throu
gh no
de n
j
PR
i,j
End-
to-
e
nd path
r
e
liability
fr
om
node
n
i
to BC th
rough
node n
j
PT
i,j
End-to-e
nd path
temperatu
r
e fro
m
node n
i
t
o
BC thr
ough node n
j
Loc
n
j
Coordinates of th
e neighbor no
de
n
j
Figure 2. Rou
t
ing Table Pa
ramete
rs
3.7. QoS-A
w
are Que
u
es
Two ind
epe
n
dent queu
es
– Critical Dat
a
Queue
(CDQ
) and No
n-Critical Dat
a
Queue
(NCDQ),
are
used,
Once
the d
e
si
re
d
next hop
is
being
sele
cte
d
, the d
a
ta
p
a
cket i
s
sen
d
to
QoS-a
w
a
r
e q
ueue,
whe
r
e
CDQ is at hi
gher
prio
ri
ty than
NCDQ.
The CD p
a
ckets wait in CDQ
before
tra
n
smissi
on
s
whil
e NCD in
NCDQ. T
he
dat
a
pa
ckets
waiti
ng in
NCDQ
are t
r
an
sferre
d to
CDQ after sp
ecific time pe
riod in o
r
de
r to prevent the
m
from indefi
n
itely blockin
g
.
4. Results and Disc
uss
i
ons
Network Sim
u
lator ve
rsi
o
n-2
(NS2
) i
s
being u
s
e
d
to carry ou
t the simulat
i
on an
d
perfo
rman
ce
evaluation of
our pro
p
o
s
e
d
CDR sche
me. In our si
mulation, the
BMSDs ca
n
be
use
d
a
s
sou
r
ce
s an
d/or
rel
a
ying no
de
s. Some
of
the BMSDs gen
e
r
ate CD pa
ckets
while
oth
e
rs
gene
rate
NCD p
a
cket
s.
We have
taken
the ave
r
age
result
s by
ch
a
nging
the BM
SDs ge
nerating
CD an
d NDC packets. We
have compa
r
ed ou
r pro
p
o
se
d CDR scheme with T
M
QoS [16]
and
LTRT [11]
.
T
he network p
a
ram
e
ters used in our
sim
u
lation are gi
ven in Table (3).
Average
pa
cket lo
ss
ratio
again
s
t diffe
rent lin
k q
ual
ities by con
s
i
derin
g differe
nt data
generation rates i
s
shown
in Figu
re 3. I
t
can be
seen that
the proposed CDR re
sult
s in lower
packet lo
ss
ratio as
comp
ared to TM
Q
o
S [16] and LTRT [11], while LT
RT [1
1] sho
w
s
po
ore
r
perfo
rman
ce
among
all. T
he re
ason b
ehind th
eir p
oor p
e
rfo
r
ma
nce i
s
that t
hey com
p
let
e
ly
ignore the
hi
gh an
d dyn
a
m
ic p
a
th lo
ss issue
of
impl
anted BMS
D
s
which
we h
a
ve add
re
sse
d
in
our p
r
op
osed
CDR schem
e.
Critical Data Rou
t
ing Algorithm
Inpu
ts:
Data Pa
cket P, and RT
1.
for
each data
pa
cket P
do
2.
for
each no
de
n
i
RT
do
3.
if
LQ
i,j
> = L
Q
thre
the
n
4.
store node
n
i
into NN
LQ
5.
if
NN
LQ
= = NULL
th
en
6.
discard P immediatel
y
7.
else if
P
CD
8.
call Delay
Aw
a
r
e Proced
ur
e
9.
else
DN
H =
n
j
NN
LQ
w
i
th lea
s
t PT
i,j
Dela
y
A
w
are
Proced
ure
10.
for
each no
de
n
i
NN
LQ
do
11.
if
PD
i,
j
> = PD
re
q
th
e
n
12.
store node
n
i
into NN
PD
13.
if
NN
PD
= = NULL
th
en
14.
discard P immediatel
y
15.
else if
NN
PD
= = 1
then
16.
DNH = n
j
NH
PD
17.
else
call Reliability
A
w
a
r
e Pro
c
edure
Reliabili
t
y
A
w
are
Proced
ure
18.
for
each no
de
n
i
NN
PD
do
19.
if
PR
i,
j
> = PR
re
q
th
e
n
20.
store node
n
i
into NN
PR
21.
if
NN
PR
= = NULL
th
en
22.
DNH = n
j
NN
PD
w
i
th
highest PR
i,j
23.
else if
NN
PR
= = 1
then
24.
DNH = n
j
NH
PR
25
.
else
DN
H = n
j
NN
PR
w
i
th
least PT
i,
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 1, March 2
015 : 181 – 1
9
2
188
Table 3. Simulation Netwo
r
k Pa
ramete
rs
Deplo
y
ment
Area
3m x 2m
Number of
Node
s
14 BMSDs and 0
1
BC
Initial Energ
y
100 Joule
Buffer Size
60 Packets
Transmission Range
40 cm
Transmission Pow
e
r
1.8467e
-
12
Bi
t Error Rat
e
10
-
2
– 10
-
4
T
a
sks
Application Type
Event Driven
Propagation Mod
e
l
T
w
oRa
y
Grou
nd
Net
w
ork Inte
rfac
e T
y
p
e
WirelessPhy
Traffic T
y
pe
Constant Bit Rat
e
(CBR)
MAC
IEEE 802.15.4
Default Values
Simulation
Time
1000 Seconds
Figure 3. Average Pa
cket Loss Ratio Vs Lin
k
Quality
at Different Data Ge
neration Rate
s
Figure 4 sho
w
s the im
pa
ct of the data gene
rati
on
ra
tes over the
p
a
cket su
cce
s
s ratio by
con
s
id
erin
g d
i
fferent lin
k
q
ualities,
wh
ere the
pa
ck
et
su
cc
es
s rat
i
o
is de
cr
eain
g
slightly with
t
he
increa
se
in
d
a
ta ge
ne
ratio
n
for all
sch
e
m
es that
mi
g
h
t be
du
e to
con
g
e
s
tion.
Figures 5
(
a)
an
d
5(b) illustrate the average pack
et success
ratio against
diffe
rent
demanded reliabilities by
con
s
id
erin
g d
i
fferent link
q
ualities
at lo
w an
d hig
h
d
a
ta gen
eratio
n rate
s respe
c
tively. Similarly,
the impa
ct of the time co
n
s
traint o
n
the
aver
ag
e on
-time packet
delivery ratio
by con
s
ide
r
i
ng
different lin
k qualitie
s, whe
n
the data is
gene
rated
at
low an
d hig
h
rates is
sh
own
in Figures 6
(
a)
and 6(b) re
sp
ectively.
It
is clea
r
from Fi
gure
s
4,
5,
a
nd 6,
that the
propo
se
d
CDR sch
e
me
out-
perfo
rms th
e other
state-of
-the-art
s
whil
e TMQo
S [16
]
shows b
e
tter perfo
rma
n
ce as
comp
are
d
to LTRT [11].
Figure 4. Average Pa
cket Success Ratio Vs Data
G
e
neratio
n Rate
at Different L
i
nk Qu
alities
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Critical Data
Routing (CDR) for Intra Wire
le
ss Bo
dy
Senso
r
Network
(Javed Iq
bal B.)
189
(a)
(b)
Figure 5. Average Su
cce
s
s Ratio Vs De
mande
d Reli
ability Con
s
id
ering
Differe
n
t
Link Qualitit
es
at (a) Lo
w Da
ta Gene
ration
Rate and (b)
High
Data Ge
neratio
n Rate
(a)
(b)
Figure 6. Average O
n
-Tim
e Packet Deli
very Rati
o Vs Requi
red TT
L Con
s
id
erin
g Different Li
nk
Qualitites at (a) Lo
w Data
Gene
ration
Rate and (b
) Hi
gh Data G
e
n
e
ration
Rate
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 1, March 2
015 : 181 – 1
9
2
190
The
rea
s
o
n
b
ehind
the p
o
o
r
pe
rforman
c
e of LT
RT
[11
]
is that it
s d
e
s
ign
ed
obje
c
tive is to
minimize the
thermal
effect
s of the i
n
-b
o
d
y BMSDs
a
nd it doe
s n
o
t con
s
id
er th
e
path lo
ss i
s
su
e
and the hete
r
ogen
eou
s nat
ured d
a
ta of intra WBS
N
s.
Similarly, TMQoS [16] co
mpletely igno
res
the dyna
mic
and
high
path
loss i
s
sue
of
the impl
a
n
ted
BMSDs. Fu
rtherm
o
re,
it se
nds two
copie
s
of the ea
ch
CD pa
cket, whi
c
h
re
sults in
high
con
g
e
s
tion that may
cau
s
e
pa
cket losse
s
. On t
he
other han
d, our propo
se
d
CDR
ad
dresse
s
the
s
e
sh
ort-comi
n
g
s
by co
nsi
derin
g hi
gh
and
dynamic p
a
th
loss a
nd tem
peratu
r
e ri
se i
s
sue
s
alon
g with heteroge
neuo
u nature
d
data.
Figures 7
an
d 8
show the
impa
ct of
th
e dat
a
ge
ne
ration
rate
s ov
er th
e ave
r
ag
e en
ergy
con
s
um
ption
and
ave
r
ag
e temp
eratu
r
e ri
se
respe
c
tively. It ca
n be
o
b
se
rv
ed from th
e
both
figure
s
that the high dat
a gene
ration
rates
result in high en
ergy co
nsum
ption and hi
gh
temperature
rise. Fro
m
Fig
u
re
s 7 and
8
,
it is cl
ear th
at LTRT [11]
con
s
ume
s
more e
nergy as
comp
ared to
other two b
u
t out-pe
r
forms the oth
e
rs
in terms of
temperature
rise
be
cau
s
e its
focu
s to the
redu
ce the te
mperature
ri
se of t
he impl
a
n
ted BMSDs
and it do
es n
o
t con
s
id
er th
eir
energy co
nst
r
aint. It can a
l
so b
e
seen t
he TMQ
o
S [1
6] con
s
um
es
less en
ergy t
han LT
RT [1
1]
but more a
s
comp
ared to CDR. Furth
e
rmore, it re
sul
t
s in high tem
peratu
r
e ri
se
among all, a
s
it
sen
d
s t
w
o
copie
s
of ea
ch CD pa
cket
, which resul
t
s in hig
h
en
ergy
con
s
um
ption an
d hi
gh
temperature
rise of the imp
l
anted BMSDs.
Figure 7. Average En
ergy
Con
s
um
ption
Vs Data G
e
n
e
ration
Rate
at Different Li
nk Qu
alitites
Figure 8. Average T
e
mpe
r
ature Ri
se V
s
Data Ge
n
e
ra
tion Rate at Different Link
Qualitites
5. Conclusion
Critical Data
Routin
g (CDR)
schem
e fo
r intra
WBSNs ha
s b
een
p
r
opo
se
d in thi
s
pa
per.
The sen
s
o
r
y data pa
cket
s are bei
ng
categ
o
ri
zed
as
CD a
nd
NCD p
a
cket
s. The d
e
si
g
ned
obje
c
tive of t
he p
r
o
p
o
s
ed
CDR i
s
to fo
rward th
e
CD pa
ckets with
in certai
n tim
e
limit a
nd
wi
th
highest reli
ability while the NCD pa
cket
s in such a
way that resu
lts in lower
temperature
rise.
Evaluation Warning : The document was created with Spire.PDF for Python.