TELK
OMNIKA
,
V
ol.
11,
No
.
12,
December
2013,
pp
.
7102
7109
e-ISSN:
2087-278X
7102
External
Biomedical
De
vice
Rela
ying
Bod
y
Sensor
Netw
ork
sc
heme
Xidong
Zhang*
1,2
,
Heng
Zhang
3
,
Guixia
Kang
1
,
Ping
Zhang
1
,
and
Hui
Li
2
1
K
e
y
Labor
ator
y
of
Univ
ersal
Wireless
Comm
unication,
Ministr
y
of
Education,
Beijing
Univ
ersity
of
P
osts
and
T
elecomm
unications
10th
Xitucheng
Road,
Beijing,
P
.R.
China
2
Depar
tment
of
Aero
Equipment,
Institute
of
Ar
m
y
A
viation
9th
T
aihu
Road,
Beijing,
P
.R.
China
3
China
Equipment
System
Engineer
ing
Cor
por
ation
20th
Fuxing
Road,
Beijing,
P
.R.
China
*
corresponding
author
,
e-mail:
st0r
m@126.com
Abstract
Biological
par
ameters
acquir
ing
de
vices
are
v
er
y
impor
tant
in
Ubiquitous
Medical/Health
Care
sys-
tem
as
the
y
are
responsib
le
f
or
gather
ing
ph
ysical
par
ameters
f
or
fur
ther
analyz
e
.
It
is
natur
al
that
the
y
f
or
m
a
wireless
netw
or
k
when
the
y
tr
ansmit
the
collected
data,
and
the
netw
or
k
is
ref
erred
to
as
BSN.
Energy
ef-
ficiency
is
vital
in
wireless
Body
Sensor
Netw
or
k
(BSN)
because
the
replacement
of
po
w
er
resources
of
the
de
vices
implanted
in
human
body
is
e
xtremely
costly
.
Consider
ing
the
diff
erence
betw
een
In-body
channel
and
special
channel,
a
no
v
el
BSN
tr
ansmission
scheme
,
Coop
BSN,
is
proposed
in
this
paper
.
In
Coop
BSN,
The
tr
ansmission
of
Inter
nal
Biomedical
De
vices
is
divided
into
tw
o
steps
as
the
e
xter
nal
de
vices
are
used
as
rela
ys
.
The
first
step
is
the
IBD-EEBD
tr
ansmission,
in
which
IBDs
tr
ansmit
the
data
the
y
collected
with
a
modified
cooper
ativ
e
algor
ithm.
The
second
step
is
EEBD-Coordinator
tr
ansmission,
in
which
m
ultiuser
div
ersity
is
e
xploited.
With
the
d
ivision
of
the
tr
ansmission,
the
o
v
er
all
tr
ansmission
energy
efficiency
is
op-
timiz
ed.
Theoretical
analyz
e
and
e
xper
imental
results
sho
w
ed
that,
the
scheme
can
prolong
the
lif
e
time
of
the
sensing
de
vices
,
espe
cially
de
vices
implanted
in
human
body
in
BSN,
and
can
impro
v
e
the
f
easibility
of
BSN.
K
e
yw
or
ds:
Ubiquitous
Netw
or
k,
BSN,
Rela
y
,
P
o
w
er
Sa
ving,
Cooper
ativ
e
T
r
ansmission,
Multiuser
Div
ersity
Cop
yright
c
2013
Univer
sitas
Ahmad
Dahlan.
All
rights
reser
ved.
1.
Intr
oduction
Ubiquitous
Medical/Health
Care
Application
is
one
of
the
impor
tant
br
anches
of
the
Ubiq-
uitous
netw
or
k
[1].
The
obtaining
of
biological
par
ameters
is
the
base
of
the
entire
Ubiquitous
Medical/Health
Care
system.
Basically
,
the
biological
par
ame
ters
are
acquired
b
y
the
de
vices
that
are
placed
near
patients
or
e
v
en
implanted
in
their
bodies
.
Natur
al
ly
,
all
the
de
vices
are
considered
to
be
organiz
ed
in
the
f
or
m
of
wireless
netw
or
k.
Theref
ore
,
on
the
basis
of
Wire-
less
S
ensor
Netw
or
k
(WSN),
the
concept
of
wireless
body
sensor
netw
or
k
(BSN
or
WBSN)
is
proposed,
which
ha
v
e
recently
dr
a
wn
increasing
attention
in
both
academic
and
industr
ial
areas
.
BSN
is
a
netw
or
k
that
consist
of
all
the
Biomedical
De
vices
(BD)
near
human
body
or
implanted
in
human
body
[2].
Those
BDs
can
acquire
the
f
ollo
wing
inf
or
mation:
some
of
the
impor
tant
ph
ysical
par
ameters
such
as
temper
ature
,
b
lood
glucose
,
b
lood
pressure
and
Electro-
cardiog
r
aph
(ECG);
body
activity
or
motion
of
the
patient;
the
en
vironmental
condition
that
the
patient
is
in.
After
some
necessar
y
processes
,
the
BDs
tr
ansmit
the
inf
or
mation
to
the
nearb
y
Coordinator
that
is
responsib
le
f
or
br
inging
t
he
inf
or
mation
to
ser
v
ers
f
or
stor
age
and
fur
ther
pro-
cess
.
The
designing
of
BDs
and
tr
ansmission
technology
tur
n
out
to
be
v
er
y
impor
tant
in
BSN
research.
The
BDs
can
be
divided
into
tw
o
categor
ies
according
to
the
place
the
y
are
deplo
y
ed.
The
first
one
is
Inter
nal
Biomedical
De
vices
(IBD),
which
includes
implantab
le
or
s
w
allo
w
ab
le
biomedical
de
vices
,
such
as
implantab
le
b
lood
glucose
monitor
and
pill-shaped
microcamer
as
.
Receiv
ed
J
une
28,
2013;
Re
vised
A
ugust
12,
2013;
Accepted
A
ugust
23,
2013
Evaluation Warning : The document was created with Spire.PDF for Python.
TELK
OMNIKA
e-ISSN:
2087-278X
7103
The
second
one
is
Exter
nal
Biomedical
De
vices
(EBD),
which
includes
the
biomedical
de
vices
that
acquire
ph
ysical/en
vironmental
par
ameters
or
human
motion
inf
or
mation
from
the
outside
of
human
body
,
such
as
b
lood
pressure
tr
ansducer
,
b
lood
o
xygen
satur
ation
sensor
,
acceler
ation
tr
ansducer
and
so
on.
Those
BDs
are
all
placed
near/in
human
body
,
thereb
y
their
wireless
tr
ansmissions
are
g
reatly
aff
ected
b
y
human
body
.
The
con
v
entional
wireless
channel
models
cannot
be
directly
used
in
the
analyzing
of
BSN.
In
the
liter
ature
of
BSN,
there
are
three
diff
erent
types
of
comm
u-
nication
conditions
.
First
one
is
Off-body
Comm
unication,
in
which
the
de
vices
that
are
placed
on
the
surf
ace
of
human
body
comm
unicate
with
the
de
vices
that
are
some
distance
a
w
a
y
.
Spatial
channel
is
mainly
concer
ned
in
Off-body
Comm
unication.
Second
one
is
On-body
Comm
uni-
cation,
the
comm
unicating
peers
are
both
placed
on
the
surf
ace
of
human
body
.
Signals
are
tr
ansmitted
through
human
skin
and
superficial
la
y
er
tissues
,
kno
wn
as
On-body
Channel.
Third
one
is
Intr
a-Body
Comm
unication
(IBC).
In
IBC
,
one
comm
unicating
par
t
is
implanted
de
vice
,
and
the
other
par
t
is
either
inside
or
upon
human
body
.
Signals
are
tr
ansmitted
through
human
inter
nal
tissues
,
which
is
called
In-body
Channel.
There
ha
v
e
already
been
some
researches
on
all
the
abo
v
e
mentioned
channels
.
[3]
proposed
a
m
ultila
y
er
mathematical
model
using
v
olume
con
ductor
theor
y
f
or
galv
anic
coupling
IBC
on
a
human
limb
with
consider
ation
on
the
inhomogeneous
proper
ties
of
human
tissues
.
[4]
de
v
eloped
a
preliminar
y
tw
o
dimensional
model
f
or
IBC
.
In
[5],
human
tissue
is
modeled
as
the
combination
of
resistors
and
capacitors
.
[6]
discussed
a
ne
w
wireless
implantab
le
BSN
that
oper
ates
in
medical
implant
comm
unication
ser
vice
(MICS)
frequency
band.
As
human
body
ha
v
e
g
reat
eff
ect
on
wireless
signals
,
the
channel
quality
near
human
body
ma
y
v
ar
y
dr
amatically
.
Se
v
ere
channel
impair
ment
can
be
mitigated
through
the
use
of
div
ersity
.
As
an
impor
tant
div
ersity
method,
cooper
ativ
e
tr
ansmission
can
be
e
xploited
in
BSN.
Among
the
MI
MO
cooper
ativ
e
str
ategies
,
oppor
tunistic
rela
y
selection
is
an
outage-optimal
and
lo
w-comple
xity
str
ategy
[7].
Recently
,
one
of
the
oppor
tunistic
rela
y
selection
method,
m
ultiuser
div
ersity
(MUD)
based
rela
y
selection
has
attr
acted
significant
attention.
In
MUD
,
div
ersity
is
attained
through
the
help
of
rela
ys
.
By
letting
only
the
user
with
the
highest
instantaneous
signal-to-noise
r
atio
(SNR)
tr
ansmit
at
a
giv
en
time
,
MUD
gain
can
be
collected
in
the
f
or
m
of
impro
v
ed
outage
perf
or
mance
or
increased
total
throughput.
[8]
de
v
eloped
and
analyz
ed
a
dis-
tr
ib
uted
method
to
select
the
best
rela
y
that
required
no
topology
inf
or
mation
and
w
as
based
on
local
measurem
ents
of
the
instantaneous
channel
conditions
.
The
scheme
achie
v
ed
the
same
div
ersity-m
ultiple
xing
tr
adeoff
as
achie
v
ed
b
y
more
comple
x
protocols
in
which
coordination
and
distr
ib
uted
space-time
coding
f
or
M
nodes
is
required.
In
[9],
the
perf
or
mances
of
amplify-and-
f
orw
ard
(AF)
and
decode-and-f
orw
ard
(DF)
in
oppor
tunistic
rela
y
are
analyz
ed
separ
ately
.
The
oppor
tunistic
rela
y
selection
in
[9]
w
ere
done
in
initiativ
e
manner
or
passiv
e
manner
.
[10]
pro-
posed
an
efficient
scheme
f
or
the
combined
use
of
cooper
ativ
e
div
ersity
and
m
ultiuser
div
ersity
consider
ing
DF
oppor
tunistic
rela
ying
str
ategy
.
There
are
f
e
w
researches
concer
ning
cooper
ativ
e
comm
unication
in
BSN
f
or
no
w
.
[11]
e
xploited
cooper
ativ
e
tr
ansmission
scheme
in
the
comm
unication
betw
een
IBDs
and
Coordinator
in
BSN.
The
str
ategy
increased
the
energy
efficiency
compar
ing
with
single
node
tr
ansmission,
b
ut
the
diff
erence
betw
een
In-body
channel
and
spatial
channel
w
as
not
fully
considered.
As
to
the
cooper
ativ
e
str
ategies
with
oppor
tunistic
rela
y
selection,
there
are
e
v
en
less
researches
in
the
field
of
BSN.
In
this
paper
,
w
e
propose
a
BSN
cooper
ativ
e
tr
ansmission
scheme
based
on
the
rela
y
of
EBD
in
BSN
(Coop
BSN).
Aiming
at
maximiz
e
the
lif
etime
of
IBDs
and
minimiz
e
the
total
po
w
er
consumption,
Coop
BSN
e
xploited
the
f
eatures
of
IBDs
and
EBDs
,
and
separ
ate
the
tr
ansmission
on
the
In-body
channel
and
spatial
channel.
The
remainder
of
the
paper
is
organiz
ed
as
f
ollo
ws:
Section
2
introduces
the
netw
or
k
architecture
and
descr
ibes
the
oppor
tunistic
cooper
ativ
e
tr
ansmission
procedure
of
Coop
BSN.
In
Section
3,
the
outr
age
perf
or
mance
of
the
system
is
ana
lyz
ed
.
Sim
ulation
results
and
perf
or-
mance
compar
ation
are
giv
en
in
section
4.
Section
5
concludes
the
paper
.
Exter
nal
Biomedical
De
vice
Rela
ying
Body
Sensor
Netw
or
k
scheme
(Xidong
Zhang)
Evaluation Warning : The document was created with Spire.PDF for Python.
7104
e-ISSN:
2087-278X
2.
Ar
c
hitecture
and
Data
T
ransmission
of
Coop
BSN
As
long
as
tw
o
diff
erent
channels
,
In-body
channel
and
Spatial
channel,
are
both
in
v
olv
ed
in
the
tr
ansmission
of
BSN,
simple
netw
or
k
architecture
cannot
optimiz
e
the
po
w
er
efficiency
of
BDs
,
especially
IBDs
.
In
this
paper
,
in
order
to
maximiz
e
the
po
w
er
efficiency
of
BDs
,
w
e
propose
a
no
v
el
BSN
architecture
to
separ
ate
the
tr
ansmission
on
In-body
channel
and
Spatial
channel
b
y
e
xploiting
the
rela
y
of
EBDs
.
IBDs
are
implanted
in
the
body
of
pati
ents
,
theref
ore
it
is
pr
ice
y
to
ha
v
e
their
po
w
er
source
replaced.
Fur
ther
more
,
the
har
m
caused
b
y
IBD
wireless
tr
ansmission
m
ust
be
considered.
F
or
these
reasons
,
the
tr
ansmission
po
w
er
of
IBDs
m
ust
be
str
ictly
restr
icted.
As
to
EBD
,
the
restr
ic-
tions
on
their
tr
ansmission
are
m
uch
less
,
and
their
po
w
er
sources
c
a
n
be
easily
replaced,
so
it’
s
proper
f
or
them
to
bear
more
tr
ansmission
b
urdens
.
In
Coop
BSN
scheme
,
the
data
tr
ansmission
of
IBD
is
composed
of
tw
o
hops
,
In-body
channel
tr
ansmission
and
Spacial
channel
tr
ansmission.
In-body
channel
tr
ansmission
is
oper-
ated
on
the
MICS
de
fined
frequency
band,
while
Spacial
channel
tr
ansmission
is
oper
ated
on
2.4GHz
frequency
band.
Enhanced
Exter
nal
Biomedical
De
vices
(EEBD)
is
introduced
in
Coop
BSN.
Compar
ing
with
ordinar
y
EBD
,
EEBD
is
equipped
with
a
module
that
suppor
ts
the
comm
u-
nication
on
MICS
defined
frequency
band,
which
enab
les
it
to
receiv
e
the
signal
tr
ansmitted
b
y
IBD
.
EEBDs
can
then
be
used
to
f
orw
ard
the
data
from
IBD
to
the
Coordinator
.
Figure
1.
Coop
BSN
architecture
.
Coop
BSN
architecture
is
sho
wn
in
Figure
1.
Where
h
I
I
(
i;
j
)
denotes
the
channels
be-
tw
een
the
i
th
and
j
th
IBD
,
h
I
E
(
i;
n
)
denotes
the
channels
b
etw
een
the
i
th
IBD
and
n
th
EEBD
.
h
E
(
n
)
denotes
the
channels
betw
een
the
n
th
EEBD
and
the
Coordinator
.
There
are
tw
o
steps
in
the
tr
ansmission
in
Coop
BSN.
The
first
step
is
IBD-EEBD
tr
ansmission
which
is
based
on
a
mod-
ified
decode
and
f
orw
ard
(DF)
cooper
ativ
e
tr
ansmission.
The
second
step
is
EEBD-Coordinator
tr
ansmission
which
is
based
on
MUD
.
2.1.
IBD-EEBD
Cooperative
T
ransmission
IBDs
are
implanted
in
human
body
.
Compar
ing
with
ordinar
y
WSN
nodes
,
the
y
ha
v
e
m
uch
smaller
siz
es
,
lo
w
er
oper
ational
capability
and
less
po
w
er
supply
.
Theref
ore
,
in
the
design
of
IBD-EEBD
tr
ansmission
algor
ithm
which
is
oper
ated
on
In-body
channel,
it
is
required
that
the
energy
consumption
should
be
small
enough
and
the
algor
ithm
should
be
simple
enough
while
tr
ansmission
v
er
acity
is
guar
anteed.
In
this
paper
,
a
modified
DF
cooper
ativ
e
tr
ansmission
algor
ithm
is
designed.
Nor
mally
,
there
are
only
a
f
e
w
IBDs
in
human
body
,
so
the
y
are
all
par
tners
with
each
other
in
the
cooper
ativ
e
tr
ansmission.
The
time
slot
allocation
scheme
of
IBD-EEBD
DF
cooper
ativ
e
tr
ansmission
is
demonstr
ated
in
Figure
2.
It
can
be
seen
from
the
figure
that,
in
each
cycle
,
e
v
er
y
IBD
has
a
g
reat
n
umber
of
time
being
in
sleep
mode
.
In
Data
Acquisition
phase
,
all
IBDs
acquire
biological
par
ameters
with
their
sensing
components
and
store
them
tempor
ar
ily
.
Later
,
each
IBD
is
assigned
a
time
slot
to
tr
ansmit
the
acquired
data.
So
the
data
tr
ansmissions
of
IBDs
are
in
a
Time
Division
Multiple
Address
(TDMA)
manner
,
data
collision
can
be
a
v
oided
hence
.
In
Data
T
r
ansmission
phase
,
each
IBD
is
in
activ
e
mode
,
and
tr
ansmit
data
in
the
assigned
time
slot.
While
the
other
time
in
data
tr
ansmission
phase
,
all
nodes
receiv
e
and
store
the
data
TELK
OMNIKA
V
ol.
11,
No
.
12,
December
2013
:
7102
7109
Evaluation Warning : The document was created with Spire.PDF for Python.
TELK
OMNIKA
e-ISSN:
2087-278X
7105
Figure
2.
W
or
kflo
w
of
IBD-EEBD
Cooper
ativ
e
T
r
ansmission.
tr
ansmitted
b
y
other
IBDs
.
As
the
tr
ansmission
po
w
er
of
IBDs
are
limited,
an
EEBD
that
is
f
ar
from
an
IBD
ma
y
f
ail
to
receiv
e
the
data
tr
ansmitted
b
y
the
IBD
.
In
order
to
mak
e
sure
that
all
EEBDs
can
receiv
e
the
data
tr
ansmitted
b
y
each
IBD
,
EEBD
F
eedbac
k
phase
and
Data
F
orw
ard
phase
is
settled.
In
EEBD
f
eedbac
k
phase
,
all
IBDs
are
still
in
activ
e
mode
and
ready
to
receiv
e
the
ac
kno
wledgement
from
EEBDs
.
All
EEBDs
contain
the
sequence
n
umbers
the
y
receiv
ed
from
IBDs
in
the
f
eedbac
k
signals
.
The
IBDs
that
receiv
es
a
f
eedbac
k
signal
then
compare
the
sequence
n
umbers
with
the
data
from
all
IBDs
stored
in
itself
,
to
recon
that
if
the
EEBD
receiv
ed
all
the
dat
a
tr
ansmitted
b
y
IBDs
.
If
all
the
EEBDs
ha
v
e
receiv
ed
all
the
IBD
data,
all
IBDs
enter
sleep
mode
in
adv
ance
,
ot
herwise
,
it
comes
to
Data
F
orw
ard
phase
.
In
Data
F
orw
ard
phase
,
the
IBDs
that
deter
mined
that
an
EEBD
did
not
receiv
e
all
IBD
data
f
orw
ard
the
data
the
y
stored
in
Data
T
r
ansmission
phase
in
an
allocated
time
slot,
and
sleep
in
the
other
time
slots
.
The
abo
v
e
procedure
is
a
modified
DF
cooper
ativ
e
tr
ansmission,
which
can
be
ref
erred
to
as
Dela
y
ed
Decode-and-F
orw
ard
(DDF)
cooper
ativ
e
tr
ansmission.
With
the
oper
ation
in
the
first
step
,
it
is
guar
anteed
that
all
EEBDs
receiv
e
all
the
IBD
data.
2.2.
EEBD-Coor
dinator
MUD
T
ransmission
MUD
is
pro
v
ed
to
be
ab
le
to
impro
v
e
the
tr
ansmission
perf
or
mance
when
there
are
more
than
one
users
,
and
is
easy
to
implement.
In
BSN,
all
EEBDs
are
considered
as
users
,
and
Coordinator
is
considered
as
base
station.
All
EEBDs
tr
ansmit
ref
erence
signals
to
Coordinator
per
iodically
,
and
Coordinator
e
v
aluate
the
channel
conditions
betw
een
itself
and
all
the
EEBDs
according
to
the
ref
erence
signals
.
The
Coordinator
broadcast
the
data
tr
ansmission
per
mission
per
iodically
,
which
appoints
the
EEBD
that
has
the
best
channel
condition
can
proceed
with
data
tr
ansmitting.
When
EEBDs
need
to
tr
ansmit
data
to
the
Coordinator
,
the
y
receiv
e
the
data
tr
ans-
mission
per
mission.
Then
the
appointed
EEBD
combines
the
data
it
acquired
with
the
IBD
data
and
tr
ansmit
it
to
the
Coordinator
.
The
w
or
kflo
w
of
EEBD-Coordinator
MUD
tr
ansmission
is
sho
wn
in
Figure
3.
After
this
procedure
,
the
second
step
is
finished.
Figure
3.
W
or
kflo
w
of
EEBD-Coordinator
MUD
tr
ansmission.
3.
Outa
g
e
Pr
obability
Anal
ysis
3.1.
Outa
g
e
Pr
obability
in
IBD-EEBD
DDF
Cooperative
T
ransmission
In
IBD-EEBD
DDF
cooper
ativ
e
tr
ansmission,
the
In-body
channel
path
loss
model
is
[5]:
Exter
nal
Biomedical
De
vice
Rela
ying
Body
Sensor
Netw
or
k
scheme
(Xidong
Zhang)
Evaluation Warning : The document was created with Spire.PDF for Python.
7106
e-ISSN:
2087-278X
T
(
d;
f
)
=
P
R
X
P
T
X
=
(
K
f
)
2
(1
+
f
=
2
f
0
)
2
(1
+
f
=
2
f
0
)
P
A
G
(1)
Where
K
=
3
10
9
,
f
0
=
10
7
=D
,
P
=
3
;
D
>
0
:
8
4
;
el
se
,
D
is
In-body
channel
length,
A
G
is
the
siz
e
of
ref
erence
g
round
of
human
surf
ace
.
Commonly
,
its
v
alue
is
A
G
=
0
:
0025
m
2
.
f
is
the
signal
frequency
.
Substituting
the
MICS
defined
frequency
f
=402MHz
into
(1)
and
tr
imming
it,
w
e
ha
v
e
T
0
(
d
)
=
P
T
X
P
R
X
=
277
:
78
(1
+
20
:
1
d
)
d
>
0
:
8
277
:
78(1
+
20
:
1
d
)
2
el
se
(2)
Con
v
ersing
its
unit,
w
e
ha
v
e
the
path
loss
model
of
In-body
channel:
P
I
L
(
d
)
=
10
lg
(
T
0
(
d
))
(3)
It
is
e
vident
that
the
path
loss
is
mainly
decided
b
y
the
distance
betw
een
tr
ansmitter
and
receiv
er
.
Assume
that
there
are
M
IBDs
and
N
EEBDs
in
BSN.
Assume
the
tr
ansmitting
po
w
er
of
IBD
is
P
I
,
the
Additiv
e
White
Gaussian
Noise
(A
WGN)
on
the
In-body
channel
is
N
I
0
(both
in
dBm).
When
an
IBD
broadcasts
its
data,
the
Signal
Noise
Ratio
(SNR)
of
the
receiv
ed
signal
of
another
IBD
is:
ii
=
P
I
P
I
L
(
d
ii
)
N
I
0
(4)
Assume
the
SNR
threshold
of
receiv
ed
signal
of
IBD
is
I
,
then
the
distance
threshold
betw
een
an
IBD
and
its
par
tner
is:
d
thI
I
=
1
:
79
10
4
e
1
10
(
P
I
I
N
I
0
)
(5)
Let
D
(
I
)
denotes
the
collection
of
decodab
le
par
tner
IBDs
.
As
to
IBD
m
,
the
probability
that
its
collection
of
decodab
le
par
tner
IBDs
e
xists
is:
Pr
f
D
m
(
I
)
g
=
Pr
f
d
I
mn
<
d
thI
I
g
(6)
Where
d
I
mn
is
the
distance
betw
een
IBD
m
and
all
the
other
IBDs
.
In
MICS
,
it
is
defined
that
the
maxim
um
tr
ansmitting
po
w
er
is
P
I
max
=25mW
.
In
our
al-
gor
ithm,
w
e
define
a
maxim
um
IBD
tr
ansmitting
po
w
er
P
0
I
max
=15mW
(or
11.7dBm).
It
can
be
calculated
that,
when
the
sensitivity
of
IBD
receiv
er
is
under
-74dBm,
d
thI
I
=0.94m.
Gener
ally
speaking,
most
IBDs
are
implanted
in
the
upper
par
t
of
the
body
,
theref
ore
the
distances
be-
tw
een
IBDs
usually
f
all
in
the
distance
threshold.
Under
such
circumstance
,
w
e
ha
v
e
D
m
(
I
)
=
f
I
n
;
n
=
1
;
2
;
:
:
:
;
M
1
g
and
Pr
f
D
m
(
I
)
g
=
1
.
Similar
ly
,
when
the
par
tner
IBDs
f
orw
ard
the
data
from
an
IBD
to
EEBDs
,
the
SNR
of
the
receiv
ed
signal
of
EEBD
is:
ie
=
P
I
P
I
L
(
d
ie
)
N
I
0
(7)
Assume
the
SNR
t
hreshold
of
receiv
ed
signal
of
EEBD
is
E
,
then
the
distance
threshold
betw
een
an
IBD
and
ter
minal
EEBD
is:
d
thI
E
=
1
:
79
10
4
e
1
10
(
P
I
E
N
I
0
)
(8)
Let
D
(
E
)
denotes
the
collection
of
decodab
le
EEBDs
.
As
to
IBD
p
,
the
probability
that
its
collection
of
decodab
le
ter
minal
EEBDs
e
xists
is:
Pr
f
D
p
(
E
)
g
=
Pr
f
d
I
pq
<
d
thI
E
g
(9)
Where
d
I
pq
is
the
distance
betw
een
IBD
p
and
EEBD
q
.
TELK
OMNIKA
V
ol.
11,
No
.
12,
December
2013
:
7102
7109
Evaluation Warning : The document was created with Spire.PDF for Python.
TELK
OMNIKA
e-ISSN:
2087-278X
7107
When
P
0
I
max
=15mW
,
it
can
also
be
calculated
that,
when
the
sensitivity
of
EEBD
receiv
er
is
under
-79dBm,
dthIE=1.55m.
Then
the
minim
um
distance
betw
een
IBDs
and
EEBDs
usually
f
all
in
the
distance
threshold.
Under
such
circumstance
,
w
e
ha
v
e
D
p
(
E
)
=
f
E
p
;
p
=
1
;
2
;
:
:
:
;
N
g
and
Pr
f
D
p
(
E
)
g
=
1
.
The
outage
probability
in
the
tw
o
hop
cooper
ativ
e
tr
ansmission
is:
P
O
U
T
=
1
Pr
(
D
m
(
I
)
D
p
(
E
))
(10)
Theref
ore
,
if
the
perf
or
mances
of
IBDs
and
E
EBDs
all
meet
the
requirement
mentioned
abo
v
e
,
w
e
can
ha
v
e
P
O
U
T
=
0
.
3.2.
Outa
g
e
Pr
obability
in
EEBD-Coor
dinator
MUD
T
ransmission
Assume
that
the
channels
betw
een
N
EEBDs
and
the
Coordinator
,
h
E
(
j
)
(
j
=
1
;
2
;
:
:
:
N
)
,
are
Ra
yleigh
f
ading
channels
and
are
independent,
then
the
bandwidth
capacity
is:
I
M
U
=
N
max
i
=1
log
2
1
+
j
h
E
(
i
)
j
2
(11)
Where
is
the
a
v
er
age
SNR
betw
een
all
the
EEBDs
and
the
Coordinator
.
Assume
that
the
SNR
threshold
of
receiv
ed
signal
of
the
Coordinator
is
,
then
the
outage
probability
of
MUD
tr
ansmission
is:
P
O
U
T
=
Pr
(
I
M
U
<
)
=
N
Q
i
=1
Pr
h
log
2
1
+
j
h
E
(
i
)
j
2
<
i
=
N
Q
i
=1
1
e
2
1
(12)
According
to
the
proper
ties
of
e
xponent
distr
ib
ution,
when
the
a
v
er
age
SNR
is
high
enough,
(12)
can
be
re
wr
itten
as:
P
O
U
T
=
2
1
N
(13)
It
can
be
seen
from
(13)
that,
the
outage
probability
of
MUD
tr
ansmission
v
ar
ies
in
v
ersely
with
’
s
N
-th
po
w
er
.
On
the
perspectiv
e
of
div
ersity
,
EEBD-Coordinator
MUD
tr
ansmission
has
the
same
div
ersity
deg
ree
with
more
comple
x
Space-Time
Code
(STC)
cooper
ativ
e
tr
ansmission.
4.
Sim
ulation
Results
Consider
a
BSN
sho
wn
in
Figure
1.
Assume
that
BPSK
(Binar
y
Phase
Shift
K
e
ying)
is
adopted
in
all
B
Ds’
tr
ansceiv
ers
,
and
the
tr
ansmission
r
ate
is
R
b
=10kbps
.
Assume
that
the
A
WGN
on
the
special
channel
has
mean
of
0dB
and
v
ar
iance
of
5.6dB
,
and
the
tr
ansmitting
po
w
er
of
IBD
is
P
0
I
max
=15mW
.
Assume
that
the
capacity
of
the
BD
energy
resource
is
100mAh,
and
the
po
w
er
consumption
of
BDs
in
the
sleep
mode
is
appro
ximately
0mA.
The
compare
of
energy
consumption
of
Coop
BSN
and
other
schemes
is
sho
wn
in
Figure
5.
In
DT
(direct
tr
ansmission)
scheme
,
an
IBD
directly
tr
ansmits
data
to
t
he
Coordinator
,
while
in
Amplify-and-f
orw
ard
(AF)
cooper
ativ
e
tr
ansmission
scheme
,
the
par
t
ner
IBDs
AF
the
data
from
an
IBD
.
Cooper
ativ
e
tr
ansmission
under
BSN
(BSN
CT)
is
the
scheme
proposed
in
[11]
which
considered
the
impact
of
In-body
channel.
It
can
be
seen
from
Figure
4
that,
if
cooper
ativ
e
tr
ans-
mission
is
used,
the
energy
efficiency
of
tr
ansmission
is
impro
v
ed
g
reatly
when
the
tr
ansmission
distance
is
not
v
er
y
shor
t.
BSN
CT
has
better
perf
or
mance
than
AF
as
it
jointly
optimiz
ed
mod-
ulation
inde
x
and
par
tner
n
umber
ha
ving
In-body
channel
considered,
b
ut
as
the
tr
ansmitting
technology
remains
the
same
while
In-body
channel
and
special
channel
are
diff
erent,
the
energy
efficiency
of
this
scheme
is
not
appealing.
Coop
BSN
scheme
made
use
of
the
f
eature
of
both
IBD
and
EEBD
.
EEBDs
f
orw
ard
the
data
from
IBD
,
thereb
y
the
tr
ansmission
on
In-body
channel
and
special
channel
is
separ
ated,
and
the
o
v
er
all
tr
ansmission
energy
efficiency
is
impro
v
ed.
Exter
nal
Biomedical
De
vice
Rela
ying
Body
Sensor
Netw
or
k
scheme
(Xidong
Zhang)
Evaluation Warning : The document was created with Spire.PDF for Python.
7108
e-ISSN:
2087-278X
0
100
200
300
400
0
1
2
3
4
x 10
−4
d
(m)
E
bt
(J)
BSN CT
DT
AF
Coop BSN
Figure
4.
Energy
consumption
compar
ison
betw
een
Coop
BSN
and
other
schemes
.
0
20
40
60
80
100
0
50
100
150
200
250
300
350
d
(m)
Life Time of BD (year)
DT
BSN CT
AF
Coop BSN EEBD
Coop BSN IBD
Figure
5.
BD
lif
e
time
compar
ison
betw
een
Coop
BSN
and
other
schemes
.
The
lif
e
time
of
BDs
in
Coop
BSN
and
other
schemes
are
compared
in
Figure
5.
The
concept
of
lif
e
time
is
that
when
all
the
po
w
er
is
used
to
tr
ansmit
data,
the
tim
e
dur
ation
bef
ore
the
po
w
er
supply
in
a
BD
is
depleted.
Assume
that
a
BD
tr
ansmits
data
f
or
800
seconds
in
one
da
y
.
It
can
be
seen
from
Figure
5
that,
when
the
tr
ansmission
distance
is
relativ
ely
shor
t,
BDs
in
DT
scheme
ha
v
e
longer
lif
e
time
,
b
ut
with
the
increase
of
the
tr
ansmission
distance
,
cooper
ativ
e
is
pro
v
ed
to
prolong
the
lif
e
time
of
BDs
.
BSN
CT
has
longer
lif
e
time
than
AF
,
b
ut
as
all
BDs
tr
ansmit
on
the
2.4GHz
frequency
band,
the
signals
f
ad
g
reatly
in
human
body
,
the
energy
efficiency
of
this
scheme
is
not
high
enough.
In
Coop
BSN
scheme
,
IBDs
only
ha
v
e
to
tr
ansmit
the
data
the
y
collected
to
EEBDs
cooper
ativ
ely
on
the
In-body
channel,
so
the
po
w
er
consumption
do
not
v
ar
y
with
the
dist
ance
of
the
Coordinator
.
Fur
ther
more
,
when
EEBDs
tr
ansmit
data
to
the
Coordinator
on
the
special
channel,
the
y
can
ha
v
e
MUD
gain,
hence
a
higher
tr
ansmission
energy
efficiency
is
gained.
F
rom
the
perspectiv
e
of
BSN
implement,
the
higher
energy
efficiency
of
BDs
in
Coop
BSN,
especially
IBDs
,
can
mak
e
the
dur
ation
of
IBDs
m
uch
longer
,
thereb
y
impro
v
e
the
f
easibility
of
BSN.
TELK
OMNIKA
V
ol.
11,
No
.
12,
December
2013
:
7102
7109
Evaluation Warning : The document was created with Spire.PDF for Python.
TELK
OMNIKA
e-ISSN:
2087-278X
7109
5.
Conc
lusion
In
this
paper
,
a
BSN
cooper
ativ
e
tr
ansmission
scheme
based
on
the
rela
y
of
EBD
is
proposed.
In
the
proposed
shceme
,
the
tr
ansmission
of
IBDs
is
divided
into
tw
o
steps
.
The
first
step
is
the
IBD-EEBD
tr
ansmission,
in
which
IBDs
tr
ansmit
the
data
the
y
collected
with
DDF
cooper
ativ
e
algor
ithm.
The
second
step
is
EEBD-Coordinator
tr
ansmission,
in
which
MUD
is
e
xploited.
With
the
division
of
the
tr
ansmission,
the
o
v
er
all
tr
ansmission
energy
efficiency
is
optimiz
ed.
Theoretical
analyz
e
and
e
xper
imental
results
sho
w
ed
that,
the
scheme
can
prolong
the
lif
e
time
of
BDs
,
especially
IBDs
in
BSN.
This
f
eature
can
help
to
impro
v
e
the
f
easibility
of
BSN.
Ref
erences
[1]
Zhang
Ping,
Miao
Jie
,
Hu
Zheng,
Tian
Hui,
”A
Sur
v
e
y
of
Ubiquitous
Netw
or
k,
”
Jour
nal
of
Beijing
Univ
ersity
of
P
osts
and
T
elecomm
unications
,
v
ol.
33,
pp
.
1-6,
Ma
y
2010.
[2]
Emil
Jo
v
ano
v
,
Car
men
C
.
Y
.
P
oon,
Guang-Zhong
Y
ang,
Y
uan-Ting
Zhang,
”Body
Sensor
Net-
w
or
ks:
F
rom
Theor
yto
Emerging
Applications
,
”
IEEE
T
r
ansactions
on
Inf
or
mation
T
echnology
in
Biomedicine
,
v
ol.
13,
pp
.
859-863,
J
une
2009.
[3]
Sio
Hang
Pun,
Y
ue
Ming
Gao
,
P
engUn
Mak,
Mang
I
V
ai,
Min
Du,
”Quasi-Static
Modeling
of
Human
Limbf
or
Intr
a-Body
Comm
unications
With
Exper
iments
,
”
IEEE
T
r
ansactions
on
Inf
or-
mation
T
echnology
in
Biomedicine
,
v
ol.
15,
pp
.
870-876,
J
une
2011.
[4]
Y
ue-ming
Gao
,
Si
o
Hang
Pun,
Min
Du,
Mang
I
V
ai,
P
eng
Un
Mak,
”A
Preliminar
y
T
w
o
Di-
mensional
Model
f
or
Intr
a-body
Comm
unication
of
Body
Sensor
Netw
or
ks
.
,
”
in
Inter
national
Conf
erence
on
Intelligent
Sensors
,
Sensor
Netw
or
ks
and
Inf
or
mation
Processing,
2008.
ISS-
NIP
2008.
,
2008,
pp
.
273-278.
[5]
Cho
N.,
Y
oo
J
.,
Song
S
.
J
.,
Lee
J
.,
Jeon
S
.,
Y
oo
H.
J
.,
”The
Human
Body
Char
acter
istics
as
a
Signal
T
r
ansmission
Medium
f
or
Intr
a-body
Comm
unication,
”
IEEE
T
r
ansaction
on
Micro
w
a
v
e
Theor
y
and
T
echniques
,
v
ol.
55,
pp
.
1080-1086,
Ma
y
2005.
[6]
Qiang
F
ang,
Shuenn
Y
uh
Lee
,
Hans
P
er
mana,
Kamr
an
Ghorbani,
Irena
Cosic.,
”De
v
eloping
a
Wireless
Implantab
le
Body
Sensor
Netw
or
k
in
MICS
Band,
”
IEEE
T
r
ansactions
on
Inf
or
mation
T
echnology
in
Biomedicine
,
v
ol.
15,
pp
.
567-576,
Apr
.
2011.
[7]
R.
Knopp
,
E.
Humb
let,
”Inf
or
mation
Capacity
and
Ppo
w
er
Control
in
Single
Cell
Multiuser
Comm
unications
,
”
in
IEEE
Inter
national
Conf
erence
on
Comm
unications
,
1995.
ICC’
95.
,
1995,
pp
.
331-335.
[8]
Bletsas
A.,
Kllisti
A.,
Reed
D
.
et
al.,
”A
Simple
Cooper
ativ
e
Div
ersity
Method
Based
on
Netw
or
k
P
athselection,
”
IEEE
Jour
nal
on
Select
ed
Areas
in
Comm
unications
.
,
v
ol.
24,
pp
.
659-672,
Mar
.
2006.
[9]
Bletsas
A.,
Hyundong
S
.,
Win
M.
Z.,
”Cooper
ativ
e
Comm
unieations
With
outage-Optimal
Op-
por
tunistic
Rela
ying,
”
IEEE
T
r
ansactions
on
Wireless
Comm
unications
,
v
ol.
6,
pp
.
3450-3460,
Sep
.
2007.
[10]
Haiy
ang
Ding,
Jianhua
Ge
,
Daniel
Bene
vides
da
Costa,
”A
Ne
w
Efficient
Lo
w-Comple
xity
Scheme
f
or
Multi-Source
Multi-Rela
y
Cooper
ativ
e
Netw
or
ks
,
”
IEEE
T
r
ansactions
on
V
ehicular
T
echnology
,
v
ol.
60,
pp
.
716-722,
Dec.
2011.
[11]
Xiangyi
Li,
Guixia
Kang,
Y
if
an
Zhang,
Xidong
Zhang,
Longf
eng
Chen,
Lichen
Lee
,
”Chronic
Disease
Management
System
withBody-implanted
Medical
De
vices
Based
onWireless
Sen-
sor
Netw
or
ks
,
”
in
Inter
national
Conf
erence
on
E-Health
Netw
or
king,
Application
a
nd
Ser
vices
,
2012.
Healthcom
2012.
,
2012,
pp
.
510-513.
[12]
M.
S
.
W
egm
ueller
,
A.
K
uhn,
J
.
F
roehlich,
M.
Ober
le
,
N.
F
elber
,
N.
K
uste
r
,
W
.
Fichtner
,
”An
attempt
to
model
the
human
body
as
a
comm
unication
channel,
”
IEEE
tr
ansactions
on
bio-
medical
engineer
ing
,
v
ol.
54,
pp
.
1851-1857,
Oct.
2007.
[13]
Ramakr
ishna
J
anas
w
am
y
.
Radio
w
a
v
e
Propagation
and
Smar
t
Antennas
f
or
Wireless
Com-
m
unications
,
2nd
edition.
Kluw
er
Academic
Pub
lishers
,
2002.
[14]
H
Huo
,
W
Shen,
Y
Xu,
et
al,
”The
eff
ect
of
human
activities
on
2.
4
GHz
r
adio
propagation
at
home
en
vironment,
”
in
Procceedings
of
2
nd
IEEE
Inter
national
Conf
erence
on
Broadband
Netw
or
k
and
Multimedia
T
echnology
,
2009,
pp
.
95-99.
Exter
nal
Biomedical
De
vice
Rela
ying
Body
Sensor
Netw
or
k
scheme
(Xidong
Zhang)
Evaluation Warning : The document was created with Spire.PDF for Python.