TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol.12, No.6, Jun
e
201
4, pp. 4679 ~ 4
6
8
4
DOI: 10.115
9
1
/telkomni
ka.
v
12i6.545
1
4679
Re
cei
v
ed
De
cem
ber 2
9
, 2013; Re
vi
sed
March 7, 201
4; Acce
pted
March 20, 20
14
A Method for Designing Complex Machine Olfaction
System
Hua Fa
ng
1,2
, Minglei Shu*
1,2
, Yunxiang Liu
1,2
, Wanjun Yu
1,2
, Wenju Li
1,2
1
Shang
hai i
n
sti
t
ute of technol
og
y, Schoo
l of
computer sci
en
ce and i
n
forma
tion en
gi
neer
in
g
2
Shang
hai i
n
sti
t
ute of technol
og
y, Rese
arch
center of mach
ine o
l
factio
n
No.10
0
Hai
q
u
a
n
Roa
d
, Shan
g
hai cit
y
, 201
41
8, P.R.China
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: shumin
gle
i
@
s
it.edu.cn
A
b
st
r
a
ct
In rece
nt ye
ar
s, vario
u
s
gas
sensors
h
a
ve
bee
n w
i
d
e
ly
u
s
ed to
d
e
tectin
g o
d
o
u
rs, an
d
mach
ine
olfactio
n syste
m
bas
ed
on s
ensor
arrays
a
dopt k
i
nds
of i
n
telli
ge
nt
instr
u
ments and
a
d
vanc
ed e-n
o
s
e
s.
While
m
a
c
h
ine olfaction syst
em
m
e
ets
the information sharing problem
an
d is diffic
u
lt to flexibly
and
unifor
m
ly
assi
gn these
devi
c
es. In the pa
per a te
mp
l
a
te
technol
ogy h
a
s bee
n ap
pli
ed to the co
mple
x
platfor
m
of m
a
chine olfaction system
. The simulation sensor array templa
te receives field odor data
or
simulat
e
s the data via recor
ders in
data
b
a
s
e, then trans
mits to the pla
tform w
h
ich co
nsists of severa
l
distrib
u
ted
mo
nitori
ng su
bsys
tems. Each su
bsystem
ma
tch
e
s a set of gas
sensors arr
a
y, and h
a
s functio
n
s
of logging data, communica
ting and sim
u
lating
industry application.
The
data from the
subsystem
and the
prepr
ocesse
d
data ar
e se
nt to the w
eb s
e
rv
er center
an
d
stored i
n
t
he
d
a
tabas
es. T
h
e
n
sens
ors p
e
rfor
m
ana
lysis by s
e
veral l
a
yer
alg
o
rith
ms. And t
he exc
h
a
ngi
ng
alg
o
rith
ms co
nvert the fie
l
d
odor
data to
g
a
s
co
n
c
en
tra
t
i
o
n
s
wi
th
p
p
m
va
l
ues, th
e
e
x
p
e
r
t syste
m
s o
r
re
cog
n
i
tio
n
al
g
o
r
i
t
hm
s an
a
l
y
z
e th
e p
p
m
va
lu
e
s
and
show
the a
ppl
i
c
ation r
e
sults.
All d
a
ta of e
a
c
h
lay
e
r are
stor
ed i
n
se
rv
er da
tabase syst
ems, and e
a
ch
la
yer
alg
o
rith
ms can
be upd
ated a
n
d
saved.
F
i
na
ll
y, the supporti
ng pl
atform tha
t
appli
ed to in
d
u
strial
mo
nitori
n
g
systems, w
a
s
deve
l
op
ed w
i
th
a kind of indu
strial
config
ura
t
ion
softw
ares, w
eb MIS
and
databas
es, an
d
was utili
z
e
d to
reali
z
e monitor
to the envir
onm
ental systems by the simulation tem
p
late.
Ke
y
w
ords
: machi
ne olf
a
ctio
n, gas sens
or array,
simul
a
tio
n
template, su
pporti
ng pl
atfor
m
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Since t
he
ori
g
inal
definitio
n of el
ect
r
oni
c n
o
se
(e-no
s
e) by P
r
of.
J.W.
Gardne
r in 1
994
[1], electro
n
ic nose i
s
an
essential
co
mpone
nt of
a
r
tificial olfa
cti
on in
strum
e
n
t. It is clea
r t
hat
pattern
re
cog
n
ition, sig
nal
and d
a
ta p
r
oce
s
sing
of machi
ne olfa
ction. A sen
s
or array i
s
an
essential p
a
rt of the elect
r
oni
c
olfactio
n pro
c
e
s
s. G
a
s
sen
s
o
r
h
a
s be
en d
e
volope
d to ap
ply
accordingly i
n
vario
u
s fiel
ds
as an
in
strument o
r
a
system. Ma
rta
re
porte
d o
n
an o
p
tical
se
nso
r
for remote
re
sidu
al d
r
y sa
lt monitori
ng,
with
appli
c
at
ions in
roa
d
winter mai
n
te
nan
ce [2].
Th
e
prop
osed se
nso
r
sy
stem
has a
com
p
act de
sign a
nd sh
ort re
spon
se time, cap
able of b
e
ing
mounted
o
n
-
boa
rd
main
tenan
ce ve
hicle
s
fo
r
real
-time
ro
ad
con
d
itio
n a
s
sessm
e
nt.
Nan
o
mate
rial
s of
conj
uga
ted polyme
r
s are
found
to have
su
pe
rior
pe
rform
a
nce
rel
a
tive to
conve
n
tional
material
s d
u
e
to mu
ch la
rge
r
expo
se
d
su
rface are
a
[3].
Machin
e olfactio
n were
desi
gne
d to solve medical probl
em
s [4-6] and envir
o
n
ment monito
ring [7-9]. Th
e system in the
pape
r [10] prese
n
ted a wi
rele
ss mo
nito
r and cont
rol
system usi
n
g mobile co
mmuni
cation
and
inclu
ded the remote mea
s
u
r
eme
n
t units, GPRS
network, Internet an
d a serve
r
co
mputer.
The a
bove
re
sea
r
che
s
e
m
pha
size
sen
s
or d
e
velop
m
e
n
t and
ap
plication a
nd ig
n
o
re th
at
informatio
nization plan
nin
g
of sen
s
o
r
s and ap
plic
ations i
s
mo
re
important a
nd more po
p
u
lar.
The p
ape
r fo
cu
se
s o
n
technolo
g
ies of
sup
portin
g
pl
atform of m
a
chin
e olfa
ctio
n that p
r
ovid
e
sign
al, data
pro
c
e
ssi
ng
solution
s an
d
appli
c
at
ion
s
. We
have
stru
ctured th
e content
s i
n
to
followin
g
se
ctions: we de
sign firstly the
stru
cture
of sup
portin
g
pl
atform of ma
chin
e olfactio
n;
se
con
d
ly, the templates of
sen
s
o
r
array
s
and
appl
i
c
a
t
ions ad
opt to solve the
complex probl
em
of various o
dour mo
nitoring enviro
n
m
ents; thir
dl
y, we select
the environ
mental syste
m
appli
c
ation
s
a
i
ming to simul
a
tion templat
e
of machin
e olfaction
syst
ems.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 6, June 20
14: 4679 – 4
684
4680
2. Supportin
g
Platform o
f
Machine Ol
fac
t
ion
A machine
ol
faction
syste
m
con
s
ist
s
of
an
array of cross-sen
s
itive
ga
s se
nsors and an
approp
riate p
a
ttern re
cog
n
i
tion method, to automat
ica
lly detect and
discriminate
variou
s odo
rs,
and u
s
e
r
int
e
rface (UI) t
o
sh
ow
processed
dat
a.
Signal an
d d
a
ta proce
s
si
ng with va
ri
ous
algorith
m
s are key elem
e
n
ts in
a
ma
chin
e olfa
cti
on
system
a
s
well
as in
mo
st sen
s
i
n
g
instru
ment
s,
for exam
ple
e-no
se.
Sup
porting
platfo
rm of
ma
chi
ne olfa
ction
i
s
d
e
si
gne
d
with
multilayere
d
stru
cture an
d
databa
se
system,
and i
t
s fun
c
tion b
e
com
e
s mo
re po
we
rful a
nd
informatio
nal.
2.1. The Stru
cture o
f
Sup
porting Platform of Mach
ine Olfac
t
ion
In Figure 1, the platform
has
5 logi
c l
a
yers
that in
clud
e devi
c
e
layer (data
acq
u
irin
g
layer), I/O
bu
s laye
r (data
tran
smitting
layer),
dat
a p
r
ocessin
g
lay
e
r, alg
o
rithm
layer a
nd u
s
er
interface l
a
yer. Th
e
bottom laye
r
ref
e
rs to va
riou
s d
e
vice
s. T
he
sen
s
o
r
a
rray, a
ppli
c
at
ion
templates, control
i
n
stru
ments and a
ssi
st
in
st
rum
ents a
r
e
app
lied to mo
nitor intellig
ent
data
acq
u
isitio
n systems. T
he
se
con
d
laye
r is
com
m
uni
cation
mod
u
l
e
s
ado
pting
PCI, UART,
and
variou
s
kind
s of I/O interfa
c
e. In the
third layer,
data
acq
u
ire
d
in t
he bottom
la
yer an
d
sent
to
serve
r
s or workstation
s
via the comm
unica
tion layer is saved
in databa
se
s and ca
n be
pro
c
e
s
sed
b
y
algo
rithms
the fou
r
th la
yer. In al
go
rithm layer,
it discu
s
se
s t
he m
e
thod
s
and
mean
s of d
a
t
a pre
p
ro
ce
ssing
and
an
alyzing
with
prep
ro
ce
ss
a
l
gorithm and
expert syste
m
s.
User interfa
c
e layer can
pro
c
e
ss u
s
e
r
dat
a and user requ
est
s
, and manag
e use
r
s, dat
a,
databa
se, alg
o
rithm
s
, and
appli
c
ation
system
s.
Figure 1. The
Logic Stru
ctu
r
e of the Platform
In Figure 2, the physi
cal structu
r
e is sh
own.
Hig
h
pe
rforma
nce e-noses that h
a
ve the
comm
uni
cati
on functio
n
, system on
chips (S
OCs)
and intellige
n
t
sensors ha
ve acqui
red f
i
eld
gas
data a
n
d
tran
smitted t
o
wo
rkstatio
n
or g
a
teway. Usi
ng
UART,
PCI bu
s, WLAN, an
d ev
en
wirel
e
ss
sen
s
or
network [10], the workstation
receives the field
data. The di
stributed
syste
m
save
s the dat
a and continu
e
s to se
nd to databa
se
s in
serve
r
.
Figure 2. The Physical Structu
r
e of the Platform
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A Method for
De
signi
ng Co
m
p
lex Machi
ne Olfactio
n System
(Hua
Fang
)
4681
2.2. Data
bas
e
Sy
stem
Datab
a
se system is vitally importa
nt for the su
ppo
rting
platform. In logic
stru
cture
in fig.
3, databa
se
system
save
s kind
s of
dat
a that inclu
d
e
s
a
c
qui
red
d
a
ta from field
via bottom layer
and commu
nication laye
r, gas
sen
o
r array data, gas
informatio
n, application system informati
on,
algorith
m
s, e
x
pert system
s, use
r
inform
ation, and sy
stem files.
Figure 3. The
Databa
se System and
Dat
aba
se Sub
s
ystem
s
Users, re
sea
r
ch
ers and
admin
i
s
trato
r
s ca
n updat
e their information, uplo
ad files,
manag
e
sen
s
ors a
nd
set
a
rray
s
, mat
c
h
some
kind
of
algo
rithm to
the inte
rrel
a
ted a
ppli
c
atio
n.
The uplo
ade
d file is use
d
to prep
ro
ce
ssing al
go
rithm
and expe
rt sy
stem
with
C langu
age
scri
pt,
execute file,
ASP file an
d so on. Many upl
oaded pattern recogni
tion met
hods have been
introdu
ce
d in
sup
portin
g
p
l
atform ma
chi
ne olfact
io
n for the inte
rrel
ated appli
c
ati
on. Even, other
use
r
s can d
o
wnl
oad file
s of the
sa
me ap
plicati
on, co
mpa
r
e
sen
s
o
r
a
r
ra
y accuracy
and
efficien
cy, an
d ret
u
rn
ne
w
algorith
m
file
with the
de
scription to
serv
er. Th
e d
e
scription p
r
ovide
s
to more us
ers to refer.
I
n
co
nt
ra
st
t
o
e-
no
se,
t
h
e
olf
a
ct
ory
sy
st
em
can
dete
c
t and
interpre
t the info
rmati
on from
volatile sen
s
o
r
s
with a hi
gh
degree of
se
nsitivity,
sele
ctivity and st
ability. Theref
ore, to conq
u
e
r
dif
f
i
cult
ies in
machi
ne olf
a
ct
ion,
r
e
s
e
a
r
ch
er
s st
art to pay more attentions to
research
od
or
pro
c
e
ssi
ng
model
s for testing an
d verifying dat
a
from sen
s
or database subsy
s
tem. So the
databa
se
sy
stem of su
pportin
g
plat
form of
ma
chin
e olfa
ction provide
s
much
data
for
resea
r
chers
whi
c
h mine
machi
ne olfa
ction data,
create algo
rith
ms, and rel
e
ase ne
w data
to th
e
serve
r
data
b
a
s
e sy
stem.
2.3. Web Us
er Inter
f
ace
Bas
e
d on Web Serv
ice a
nd Work
flo
w
MIS based
o
n
web fa
ce
s
use
r
s,
re
sea
r
che
r
s
and a
d
m
inistrators, and p
r
ovide
s
interface
to access da
tabase. The
method of d
e
sig
n
ing
the
MIS that suits the dem
an
d of the mod
e
rn
internet an
d wirel
e
ss net
work, ha
s be
co
me more a
n
d
more impo
rt
ant and po
pul
ar. Accordi
n
g
to
the workflo
w
requi
rem
ent in dynamic en
vironm
e
n
t and advantage
s of present d
e
sig
n
method
of
MIS, the technologi
es of workflo
w
a
nd
web
se
rvice,
the platform
use
s
flexible
MIS to service all
of users, re
se
arche
r
s a
nd a
d
minist
rators.
3. Simulation Template
The tem
p
late
ca
n
re
ceives filed d
a
ta of
det
e
c
ted
od
ou
r
s
or
s
i
mu
latio
n
d
a
t
a or
re
c
o
r
der
data. The sy
stem has be
en
evaluated in
ext
ensive sim
u
lation
s and t
he sh
own re
sults.
3.1. Data
Ac
quisition an
d Sensor Ar
r
a
y
Template
Accuracy, performan
ce, and stability requirem
ents of t
he gas sensors have
increased in
indu
strial produ
ction, en
vironme
n
tal monito
ri
ng, medical ch
e
cki
ng and o
t
her are
a
s.
The
resea
r
ching
and devel
op
ment of the gas
sen
s
o
r
s have be
e
n
more
and
more imp
o
rtant,
inco
nsta
nt an
d compli
cate
d. In Figure 4
,
sensor
set has 24
sen
s
o
r
s that can divi
de in four a
r
rays
each with 6 sensors, three
sen
s
o
r
array
s
ea
ch with 8
sen
s
o
r
s, two
array
s
ea
ch
with 12 sen
s
ors,
two sen
s
o
r
a
rray
s
with 6 sen
s
o
r
s a
nd 18 sen
s
o
r
s, and one a
r
ra
y with 24 se
nso
r
s. Th
e senor
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 6, June 20
14: 4679 – 4
684
4682
array templat
e
that can b
e
divided
in
multic
ha
nnel
to get field
odor data, a
pplie
s to ma
ny
indu
strial e
n
vironm
ents.
Workstat
ion
re
ceives field
ga
s sen
s
or ar
ra
ys tran
smittin
g
data, an
d the
function
s of
workstation a
r
e sho
w
n in fig. 5.The
se
n
s
or
array tem
p
late ca
n sim
u
late field da
ta
and hi
story d
a
ta in databa
se for u
s
e
r
up
dating ne
w al
gorithm
s an
d expert sy
ste
m
s.
Figure 4. The
Simulation Template an
d Applicatio
n
Figure 5. The
Function
s of
Wo
rkstation
s
3.2. Application and Tem
p
late of
Appl
ication
The d
a
ta pro
c
e
ss
of ma
ch
ine olfa
ction t
he st
ru
cture i
s
di
scusse
d i
n
the ab
ove section
s
.
The
sign
al m
odel
of sen
s
o
r
a
r
rays
asso
ciated
with
specifi
c
o
dor classe
s i
s
p
r
e
s
ente
d
in
Fig
u
re
4. The sp
ecif
ic cla
s
se
s acquire
data of
sen
s
o
r
array
template, an
d these
cla
s
ses an
d sp
ecif
ic
appli
c
ation
s
are
con
c
eive
d as an inte
rrelated w
hole.
Industri
a
l saf
e
prod
uctio
n
adopt
s co
mm
on
gas sen
s
o
r
s
as
TGS38
7
0
is u
s
e
d
to
det
ect
CH4 i
n
co
al min
e
s. A
n
d
ha
rmful
and
dang
ero
u
s g
a
s
sen
s
o
r
s
of MQ se
rie
s
a
r
e u
s
ed to
e
n
vironm
ent
monitori
ng. T
e
mplate of a
pplication uni
fies
indu
strial safe pro
d
u
c
tion
, environme
n
t moni
torin
g
, food pro
c
essing a
nd
storin
g, medi
cal
che
c
king
an
d
medi
cin
e
d
e
v
eloping,
and
even
spe
c
ial
and
u
n
known od
or an
alyzing
a
s
well
as
spi
c
e t
e
st
in
g.
3.3. Verification
The supporti
ng platform
of machi
ne
olfact
ion
was utilized
to realize monit
o
r to the
environ
menta
l
system
s
with simul
a
tio
n
template.
The
kind
ap
plicatio
n wa
s de
sig
ned
b
y
Kingview
sof
t
ware, SQ
L
databa
se a
n
d
web
se
rv
ice de
sign too
l
set incl
udin
g
asp.net. The
Kingview is
a kind of ind
u
strial
config
urati
on
software for d
e
vel
oping
supe
rvi
s
ory control and
data acquisiti
on system. P
r
inci
pal comp
onent analy
s
i
s
(PCA
) is wi
dely used fo
r dimensi
onali
t
y
redu
ction
in
p
a
ttern
re
cog
n
i
t
ion an
d h
a
s
been
ap
plied
in many
linea
r a
r
ea
s
su
cce
ssfully [1
1]. So
the Verification sy
stem a
dopts th
e alg
o
rithm to
re
cogni
ze p
o
llut
ed ai
r. The li
near algo
rith
m is
visual to
air testing
an
d e
a
sy to
reali
z
e
.
Figur
e 6
sh
ows p
a
rts of
gas monito
rin
g
re
sult
s, ba
sed
on the
su
ppo
rting pl
atform
of ma
chin
e
olfaction.
T
h
e
monito
ring
system run in
the workstati
on
that marked
No.3 in Figu
re 7 and de
cri
bed in Tabl
e 1.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A Method for
De
signi
ng Co
m
p
lex Machi
ne Olfactio
n System
(Hua
Fang
)
4683
The supp
orti
ng platform
o
f
machin
e olfaction p
r
e
s
e
n
t
ed an ambi
e
n
t real-tim
e a
i
r quality
monitori
ng
sy
stem that detected 1
2
ki
nd
s of normal g
a
se
s in air in
cludi
ng O
2
, H
2
, CO
2
, CO,
an
so on.
Th
e
sy
stem
con
s
iste
d
of several distrib
u
ted m
onitorin
g
stations
that com
m
unicated ea
ch
other. The S
O
C ma
rked
No.4 was e
q
u
ippe
d with
gaseou
s an
d
meteorol
ogi
cal
sen
s
ors a
n
d
correl
ative p
a
ram
e
ter
se
nso
r
s
a
s
well
as d
a
ta lo
gging
and
wirele
ss/
UART
comm
uni
cat
i
on
capabilities.
The server
m
a
rked No.1
collected real t
i
me dat
a from the
station and converted it
into, and deli
v
ered to users throu
gh a
web p
o
rtal
an
d a mobile a
pplication as
the moblie ph
one
marked No.6. The serve
r
comp
uter co
uld be very
far away from
stations an
d
it could be also
flexibly assi
g
ned. Th
e
server comp
uter su
ppo
rt
ed
several th
e re
mote me
asurement u
n
its
with
flexibly assig
ned fun
c
tion
s. Therefo
r
e, t
he
mea
s
u
r
e
m
ent re
sult
s
and g
a
s
cha
nging trend
s
wa
s
widely spre
ad
via Internet access in re
al time.
Figure 6. The
Kingview Mo
nitoring Syste
m
and Ten
d
e
n
cy Ch
arts of
Three O
dou
rs
Figure 7. An Applicatio
n System
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 6, June 20
14: 4679 – 4
684
4684
Table 1. The
Comp
one
nt Parts of the Ve
rification Syst
em
Number
Description
Function
1
server
The server is the
integration of th
e
database and
w
e
b service. IIS (V5.1 for fo
rmer site
and V7.0)
runs o
n
it and users UI
is an inner web site.
2
e-nose
The device assembles 12 gas sensors.
It can acquire gas conc
entr
a
tions via inner
expe
rt s
y
stem, s
h
o
w
test results and send all data
to
w
o
rkstation via RS232.
3
w
o
rkstation
In the distributed
supporting platfo
rm, t
he
w
o
rkstation is a gate
w
a
y
t
hat receives field
data and tr
ansmits to server. The
Kingview
runs on
it. Fig. 7 sho
w
s monitoring
softw
a
r
e based o
n
the Kingview
.
4
SOCs
The kind of devices assembles 6
unfixed
sensors.
Parts of T
G
S series and MQ serie
s
can match its port.
5
sensors
The sensors via SOC A/D acquir
e
fi
eld odor data,
and transmit to
workstation via
RS232/USB.
6
terminal
The mobile phon
e as user termin
al accesses data
base, and bro
w
s
e
the
w
eb site.
7
probe
The assist instrument can go dee
p into
various kin
d
s of containers, and imbibe gases
w
h
en the cont
roller turns on t
he p
u
mp.
4. Conclusio
n
This
pap
er
p
u
rpo
s
e
s
a n
o
vel app
ro
ach to
con
s
tru
c
t supp
orting
platform
of
machi
n
e
olfaction,
and
appli
e
s this
method i
n
in
d
u
strial
env
i
r
o
n
ments.
The
r
efore, th
e
cla
ssifie
d
d
a
ta h
a
ve
compli
cate
d stru
ctures a
nd
the simu
lation
te
mpla
te may be
more
useful
for data
with
compli
cate
d
and vari
ou
s stru
ctures.
Main
Co
ncl
u
sions
d
r
a
w
n f
r
om
the
Pape
r: 1) Th
e g
a
s
con
c
e
n
tration
wa
s
simul
a
ted by
usi
n
g
pre
p
rocess
of the templ
a
te of
sen
s
o
r
a
rray; 2
)
We
focu
sed on t
e
ch
nolo
g
ies
of suppo
rting
platform of
machi
ne olfa
ction that pro
v
ided sign
al, data
pro
c
e
ssi
ng solution
s and
appli
c
ation
s
; 3) Expandi
ng
supp
orting p
l
atform of ma
chin
e olfactio
n
benefited
bui
lding the
da
ta wa
reh
o
u
s
e and
on
-lin
e analytical
pro
c
e
ssi
ng t
o
solve the
s
e
probl
em
s in the future.
Ackn
o
w
l
e
dg
ements
This work wa
s finan
cially suppo
rted by
Nation
al
Nat
u
ral
Sci
e
n
c
e
Found
ation
of
China
(Grant No. 6110
318
0), '085'
p
r
oje
c
ts
of
Shan
gh
ai
Educatio
na
l Committee (405ZK1
241
2
7
),
Shanghai Excellent
Young
Teachers Program
(ZZyyy12044,
ZZyy
y
12045), Tal
ent Devel
o
pment
Special
F
und
ed
Proje
c
ts o
f
Shanghai In
stitut
e of Technolo
g
y (YJ2
012-16,
YJ2
012-17, YJ2
0
13-
37).
Referen
ces
[1]
Gardner JW
, Bartlett PN. A brief histor
y
o
f
electronic n
o
s
es.
Sensors and Actuators
B: Chemical.
199
4; 18(1-
3): 210-
211.
[2]
Marta RL, Pe
d
r
o MM, José R
,
López PA. R
e
mote o
p
ti
cal s
ensor for r
eal-ti
m
e resid
u
a
l
sal
t
monitori
n
g
on roa
d
surfac
es.
Sensors a
n
d
Actuators B: Che
m
ic
al
. 20
1
4
; 191(2): 3
71-
376.
[3]
Rajes
h
, T
a
rush
ee A, D
e
ve
ndr
a K. Rec
ent
pr
ogress
in t
he
d
e
vel
opme
n
t of
nan
o-structure
d
co
nducti
ng
pol
ym
ers/nan
o
c
ompos
ites fo
r sensor ap
pl
ic
ations.
Se
nsors
and Actuators
B: Chemical.
200
9;
136(
2):
275-
286.
[4]
Nisar A, Afzulp
urkar N, Maha
i
s
avari
y
a B, T
uant
ran
ont A. MEMS-base
d
mi
cropum
ps in dr
ug de
liver
y
and b
i
ome
d
ic
al
appl
icatio
ns.
Sensors a
nd Act
uators B: Che
m
ic
al.
20
08; 13
0(2): 917-
94
2.
[5]
Bani
ukevic J, Kirl
y
t
e J, Ram
anav
icius A, Rama
navic
ien
e
A. Applic
ati
on
of ori
ented
and ra
nd
o
m
antib
od
y immo
biliz
atio
n meth
ods in imm
u
n
o
s
ensor d
e
si
gn.
Sensors a
nd A
c
tuators B: Ch
emical.
20
13
;
189(
12): 21
7-2
23.
[6]
Herrasti Z
,
Et
xabe
E, Mitxel
e
na JM, Ga
bil
o
ndo
MP
, Martínez F
M
. Dev
e
l
opme
n
t an
d i
n
t
egrati
on
of a
n
electroc
hemic
a
l
s
y
stem in a
LOC devic
e for DNA detecti
on.
Sensors a
nd Actuators
B: Chemical.
201
3; 189(
12): 66-7
0
.
[7]
Lee J
H
. Gas s
ensors
usin
g h
i
erarch
ica
l
an
d
hol
lo
w
o
x
i
de nan
ostructures
.
Sensors and
Actuators
B:
Chemic
al. 20
0
9
; 140(1): 3
19-
336.
[8]
Kumar A, Kim
H, Hanck
e
GP. Enviro
nment
al
Monitor
i
ng
S
ystems: A Revi
e
w
.
Sens
ors J
ournal, IEEE.
201
3; 13(4): 13
29-1
339.
[9]
Lee D
D
, Lee D
S
. Environme
n
t
gas sensors.
IEEE Sensors J
ournal.
20
01; 1
56(1): 21
4-2
1
5
.
[10]
Kermit M,
T
o
mic O.
Indepe
nd
ent compo
n
e
n
t anal
ys
is app
li
ed on g
a
s sens
or arra
y
m
eas
u
r
ement data
.
Sensors Journal, IEEE.
2003;
3(2): 218-2
28.
[11]
Sunn
y, Mishr
a
VN, D
w
i
v
e
d
i R
,
Das RR. Cla
ssifi
cation of Gases/Odors Us
ing Dy
namic Responses of
T
h
ick F
ilm Gas Sensor Arra
y.
Sensors Journal, IEEE.
2013;
13(12): 49
24-
493
0.
Evaluation Warning : The document was created with Spire.PDF for Python.