TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 14, No. 3, June 20
15, pp. 493 ~ 4
9
9
DOI: 10.115
9
1
/telkomni
ka.
v
14i3.789
5
493
Re
cei
v
ed Fe
brua
ry 4, 201
5; Revi
se
d April 26 20
15; Acce
pted Ma
y 15, 201
5
A Study
on TPMS Pre-warning Threshold Algorithm
Based on Multi-sensor Data Fusion
Wang G
a
ng
1
* Zhao Jiy
i
n
2
1
School of Mec
han
ical En
gi
ne
erin
g, Baich
e
n
g
Normal C
o
l
l
e
ge, Baich
e
n
g
, 137
00
0
2
School of Co
mmunicati
on E
ngi
neer
in
g, Jili
n Univ
ersit
y
, C
han
gch
un, 13
0
025
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: haoh
eh
e53
0
@
16
3.com
A
b
st
r
a
ct
In order to improve the pr
ec
is
ion of the tir
e
pressur
e
m
o
ni
toring system
, the Bayes
m
e
thod is
app
lie
d to establis
h its mathe
m
atic
al
mod
e
l of mult
i-se
n
s
or infor
m
ati
o
n fusion. T
h
e
temp
erature
a
n
d
pressur
e
in th
e tire, w
h
ich are the co
mp
l
e
mentary
i
n
for
m
ati
on, are i
n
tegrat
ed i
n
the mo
de
l throu
gh
ana
ly
z
i
n
g
the
mec
h
a
n
is
m of
tire burst
gen
e
r
ated by
te
mp
erature
an
d pr
essure. T
h
ro
ug
h the te
mperat
ur
e
compe
n
satio
n
of tire
burst
pr
essure
thres
h
o
l
d v
a
lu
e, the
fa
lse
alar
m a
nd f
a
lse
ne
gativ
e
a
r
e av
oid
e
d
to t
h
e
hilt. T
he ex
pe
rimental r
e
su
lts show
that compar
ed w
i
th the trad
ition
a
l T
P
MS, the accuracy
of t
h
e
m
e
asuring r
e
s
u
lts of this model is
improv
ed and thus the system’
s
m
o
nito
ring ability is
improved so t
hat
the traffic safety is guara
n
tee
d
.
Ke
y
w
ords
:
te
mp
eratur
e, pre
ssure, mu
lti-se
nsor data fu
si
o
n
, false alar
m,
false ne
gativ
e, T
P
MS
Copy
right
©
2015 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
Tire i
s
a
n
imp
o
rtant
safety
comp
one
nt of
automo
b
ile. I
f
the tire i
s
o
p
e
rated
un
der
heavy
load, for a lo
ng time or u
nder
high p
r
essure, it
may blow out or leak an
d even it may ca
use
traffic accid
e
n
ts, ca
su
altie
s
an
d prope
rty losse
s
[1]. The tire p
r
e
s
sure
monito
ring
system
can
monitor the ti
re pressu
re a
nd tempe
r
atu
r
e in re
al
time and give a
n
alarm
whe
n
the tire is over-
pre
s
sured or unde
r-pre
s
su
red and
the
t
e
mpe
r
ature
i
s
too
hig
h
so
as to avoid
tire b
u
rst [2].
At
pre
s
ent, the tire pressu
re
monitori
ng sy
stem appli
e
s
the singl
e se
nso
r
to colle
ct the tire pressure
and temp
era
t
ure from th
e sam
e
dire
ction, leadi
n
g
to low p
r
eci
s
ion of m
easurin
g re
sult,
misinfo
r
matio
n
and fal
s
e
negative. Ba
sed o
n
the
above p
r
oble
m
s, the TP
MS pre-wa
rn
ing
threshold al
g
o
rithm ba
se
d on multi-sen
s
or dat
a fu
sion
is put forward in this pap
e
r
.
2. Tire Burs
t Mecha
n
ism
2.1. Influenc
e of Inflation
Pressure on
Tire Burs
t
Whe
n
the tire pre
s
su
re i
s
lowe
r than t
he st
a
nda
rd value
du
ring the
driving proce
s
s
at
high
spe
ed, the be
ndin
g
d
e
formatio
n of
the tire
side
wall may
be i
n
crea
sed, the
tire tempe
r
at
ure
may rise
sha
r
ply and it may delaminat
e and the tire stre
ngth a
nd loadi
ng capa
city may be
impaired, resulting in l
o
o
s
i
ng an
d
ruptu
r
e of cord
thre
ad on
the i
n
n
e
r
wall
of tire
and ai
r l
eaka
g
e
or tire b
u
rst.
Whe
n
the
tire
pre
s
su
re i
s
h
i
gher than
th
e sta
nda
rd va
lue, the
co
nta
c
t a
r
ea
bet
we
en
tire and
roa
d
surfa
c
e
de
creases
and th
e pre
s
su
re
b
o
rne
by the
central a
r
ea
of the tire tre
ad
increa
se
s so that the abra
s
ion b
e
come
s seve
re a
n
d
bottom of the tread p
a
ttern cra
c
ks; si
n
c
e
the tire
rigidi
ty incre
a
ses
at this m
o
m
ent an
d d
o
e
s
n
o
t have
the fun
c
tion
of bufferin
g
,
the
dynamic loa
d
between ti
re
and
ro
ad
su
rface i
s
i
n
crea
sed
an
d thu
s
the a
u
tomob
ile sm
oothn
e
ss
become
s
wo
rse
and the
handli
ng qua
lity reduce
s
[3, 4]. From Figure 1, we can
see t
h
e
influen
ce of tire pressu
re o
n
tire perfo
rm
ance.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 493 – 49
9
494
Figure 1. Influence of Tire
Pressu
re on
Tire Perfo
r
ma
nce
2.2. Influenc
e of Temper
ature o
n
Tire
Burst
Duri
ng the d
r
iving pro
c
e
s
s, the tire tem
perature m
a
y rise sharply due to the heat
dissipatin
g, le
ading to
rubb
er a
nd
co
rd t
h
rea
d
st
re
ngt
h re
du
ction
s
.
Whe
n
the
tire
tempe
r
ature
is
risin
g
fr
om
0
℃
to 100
℃
,
the cord thre
ad
stre
ngth
o
f
nylon tire m
a
y re
du
ce
ab
out 20%
an
d
its
rubb
er
stre
ng
th may redu
ce about 5
0
%; when the
ti
re tempe
r
atu
r
e is
highe
r than the
critical
temperature (within
1
0
0
℃
belon
gs to
n
o
rmal tem
p
e
r
ature, bet
we
en 10
0 an
d 1
2
1
℃
be
lo
ng
s to
critical tem
p
e
r
ature, an
d a
bove 1
2
1
℃
b
e
long
s to
da
n
gero
u
s temp
eratu
r
e),
the
rubb
er an
d
cord
thread
stre
ng
th may redu
ce gre
a
tly, therefore, t
he tire temperature rise may exert a treme
n
d
ous
influen
ce
on i
t
s servi
c
e life
[5]. Figu
re
2
sho
w
s th
e
inf
l
uen
ce
s of te
mperature
an
d spee
d o
n
ti
re
perfo
rman
ce durin
g
driving
.
Figure 2. Influen
ce
s of Tempe
r
ature and Sp
eed on Ti
re Perform
a
n
c
e
3. Multi-sen
s
o
r Data Fusi
on Model
Even though
certai
n o
r
se
veral sen
s
o
r
s in a certain
plane
domai
n
fail whe
n
u
s
i
ng the
multi-sen
s
or
data fusio
n
techn
o
logy to measure t
he pre
s
sure and
temperatu
r
e
data of the tire in
a limited tim
e
, the accurate results
can
still be obta
ined through the inform
ation
offered by
other
non-fail
ure sensors. Th
e
multi-se
n
s
o
r
data fu
sion
of pressu
re
and te
mpe
r
ature i
n
the
tire
inclu
d
e
s
failure data reje
cti
on and valid
data optimiza
t
ion and fusi
o
n
[6].
3.1. Failure Data Rejecti
on Metho
d
Based on
Co
mpatibilit
y
M
a
trix
Whe
n
usi
n
g
several p
r
essure temp
eratu
r
e sen
s
ors to me
a
s
ure pressu
re and
temperature i
n
the tire, let
the mea
s
u
r
e
d
data
of the
ith and jth
sensors be Xi
and Xj an
d th
ey
sho
u
ld
com
p
l
y
with
Gau
s
sian di
stri
butio
n. In o
r
de
r to
reflect
the
de
viation bet
we
en Xi
and
Xj,
the
confidence di
stance me
asure (probabilit
y metrics)
dij is introduced.
(
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A Study on T
P
MS Pre-warning Th
re
shol
d Al
gorithm
Base
d on Multi
-
se
nsor… (Wang Ga
ng)
495
σ
i is the mea
n
squ
a
re e
r
ro
r of the measured d
a
ta of the ith sen
s
o
r
; the smaller t
h
e value
of dij, the
clo
s
er the
mea
s
ured
value
s
of the it
h
an
d
jth sen
s
o
r
s,
or
otherwi
s
e,
the d
e
viation
is
big, therefo
r
e
,
dij is called t
he fusi
o
n
deg
ree of the ith and jth se
nso
r
s.
If measu
r
ing
the sam
e
ind
e
x param
eter thr
oug
h m sensors, the confiden
ce di
stance
measure
)
,
2
,
1
,
(
m
j
i
d
ij
forms
a matrix
D
m
:
(2)
Whe
n
u
s
ing
several sen
s
ors to
mea
s
u
r
e a p
a
ra
met
e
r from
different dire
ction
s
, give the
threshold val
ue
of
ij
d
based
on the experi
ence or the
t
e
st
re
sult
s a
n
d
supp
os
e
Comp
ose rel
a
tion matrix
m
R
by
ij
r
:
(3)
If rij=1, it is
consi
dered tha
t
the com
pati
b
ilit
y of the ith and jth
sen
s
ors i
s
b
ad o
r
they do
not support
each other. If
ri
j =0,
it is considered that
the comp
atibility of the ith
and
jth sensors is
good o
r
they sup
port ea
ch
other. If a senso
r
ca
n
only be supp
orte
d by few sen
s
ors, the data of
this se
nsor
should b
e
failu
re data
whi
c
h
shoul
d be
rej
e
cted. If the readin
g
numb
e
r of this
sen
s
or
is invalid in a
long pe
riod, the se
nsor sh
ould be
che
c
ked.
3.2. Multi-se
nsor Data Fu
sion Metho
d
Based o
n
Ba
y
es Estimation
Defini
tion 1
:
(Baye
s
e
s
ti
mation): l
e
t the p
a
ra
mete
r Q i
n
the
a
ggre
g
ate
dist
ribution
func
tion F(x, Q) be
random variable, and for the
dec
is
ion func
tion d(x1...x
n), if B(d*)=
min(B
(
d))
is
obtained through any dec
i
s
i
on func
tion d *(x1
...x
n), then d* is c
a
lled as
Bayes
es
timation
value of parameter Q and B(d) is
c
a
lled as
the
Bayes
risk
of dec
i
s
i
on func
tion d(x1...xn) [7].
Theorem 1
:
if quadri
c
expression i
s
take
n for the loss
function:
2
1
,
xn
x
d
Q
d
Q
Then the Bay
e
s e
s
timation
value of para
m
eter Q is:
dQ
x
x
Q
QP
x
x
Q
E
x
x
d
n
n
n
1
1
1
|
|
)
(
Hen
c
e, i
n
o
r
d
e
r to
evaluat
e the Baye
s
esti
mation
of
Q, the p
r
ob
a
b
ility density
curve
P
(Qlx1...xn) s
h
ould be evaluated firs
tly.
Based o
n
Bayes estimatio
n
theory [8], the optimal fusion p
r
e
s
sure an
d tem
peratu
r
e
data of the four pressu
re temp
e
r
ature sensors are ob
tained:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 493 – 49
9
496
L
k
k
L
k
k
k
T
T
T
1
2
0
2
1
2
0
0
2
1
1
L
k
k
L
k
k
k
P
P
P
1
2
0
2
1
2
0
0
2
1
1
(
4
)
Thereinto :L
≤
4
P
k
: observed
value of the kth pressu
re
sensor;
σ
k
: standard deviation of measured val
ue of the kth pre
s
sure se
n
s
or;
P
0
: mean value of observe
d value of the Lth pre
ssu
re
sen
s
o
r
;
σ
0
: standard deviation of o
b
se
rved
valu
e of the Lth sensor;
T
k
: obse
r
ved
value of the kth temperatu
r
e sen
s
o
r
;
σ
k
: standard deviation of measured val
ue
of the kth tempe
r
ature sensor;
T
0
: mean value of observe
d value of the Lth temperat
ure sen
s
o
r
;
σ
0
: standard deviation of o
b
se
rved valu
e
of the Lth temperature
se
nso
r
.
4. Temperature Comp
en
sation o
f
Tire Burs
t Pre-w
a
r
n
ing Thr
eshold
In order to o
b
tain the sci
entific and a
c
curate th
re
shold value, the cau
s
e
s
th
at may
cha
nge p
r
e
s
sure threshol
d value shou
ld be an
alyzed; in the pa
rt, the main focu
s i
s
on t
h
e
influen
ce of e
n
vironm
ent temperature o
n
tire pre
s
sure.
Let the up
pe
r and lo
we
r li
mits of tire
pressure thresh
old value
be
Pmax and P
m
in, the
pre
s
sure d
a
ta of the
tire
measured i
n
time be
Pr
a
nd me
asured
tempe
r
atu
r
e
data b
e
T
r
; when
the tempe
r
at
ure i
s
Tr, th
e adju
s
ted p
r
essure
value
that need
s t
e
mpe
r
ature compen
satio
n
is
∆
Pr, then the
uppe
r and lo
wer th
re
shold
values of
pre
s
sure afte
r temperature
co
mpen
sation a
r
e
Pmax
＋＋
∆
Pr and Pmin
∆
Pr.
If the system
doe
s n
o
t provide temp
erature
c
o
mp
en
s
a
tion
,
th
e
p
r
e-
wa
r
n
in
g
s
h
ou
ld
be
judge
d as foll
ows:
Whe
n
Pr
＋
≥
Pm
ax or Pr
≤
Pmi
n
and Pmin
＜＜
＋
∆
Pr
Pr
Pmax
∆
Pr, false alarm
may occur
i
f
the system al
arm
s
, when
Pmin
＜＜
＋
Pr
Pmax and Pr
≥
Pmax
＋
∆
Pr or Pr
≤
Pmin
∆
Pr, false neg
ative
may occur if the syste
m
do
es not ala
r
m.
The exp
e
rim
ent proved th
at [9] wh
en t
he e
n
vironm
ent tempe
r
at
ure i
s
betwe
en 0
an
d
24
Ԩ
, there i
s
no ne
ed to revise the
inflation pre
s
sure; ho
we
ver, when th
e environ
me
nt
temperature i
s
hig
her th
a
n
24
Ԩ
or lo
wer t
han 0
Ԩ
, the tire pre
s
sure
may chang
e with t
h
e
environ
ment
temperature
and the b
a
s
ic i
n
flat
ion pre
s
sure
sh
ould
be revi
sed. Whe
n
the
environ
ment t
e
mpe
r
ature i
s
hi
ghe
r tha
n
24
Ԩ
, it
s infl
uen
ce
deg
re
e on
tire
pre
s
sure i
s
sho
w
n in
table 1. Whe
n
the environ
m
ent tempe
r
ature i
s
bet
ween
－
1
Ԩ
and
－
40
Ԩ
, the b
a
si
c pressu
re
of
0.025Pa sho
u
ld be ad
ded
once the temperatu
r
e i
s
re
duced by 1
Ԩ
,
starting fro
m
0
Ԩ
.
Table 1. Te
st Data Tabl
e of Tire Pre
s
sure
Incre
m
ent
Rate un
der
Different Environment
T
e
mp
er
a
t
ur
e
Acco
rdi
ng to
United States Standa
rd [10], w
hen th
e tire pre
s
su
re is hig
h
e
r
than 1.2
times
of the
stand
ard
tire
pressu
re
or lower than
85
％
of
the
standard tire
pre
s
sure, ala
r
m
sho
u
ld be
gi
ven and th
e temperature
comp
en
satio
n
of tire bu
rst pressu
re th
reshold val
u
e
is
sho
w
n
in Fi
g
u
re
3.
When
the tem
pera
t
ure
within t
he tire
is hi
gher than
85
℃
(
t
he
no
r
m
a
l
temperature
i
s
b
e
twe
en
6
0
℃
a
nd
70
℃
), the
syste
m
ala
r
ms.
In th
is p
ape
r, the
pre
s
sure valu
e
that is 1.2 ti
mes
of the
standard tire
pre
s
sure
is set as
“up
p
e
r
limit value
of pre
s
sure
pre-
Environment Te
mperatu
r
e
（℃
）
Tire Pressure Inc
r
ement Rat
e
（
%
）
25
～
29
4
30
～
34
6
35
～
39
8
40
～
45
10
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A Study on T
P
MS Pre-warning Th
re
shol
d Al
gorithm
Base
d on Multi
-
se
nsor… (Wang Ga
ng)
497
warning th
re
shold”
and th
e
pre
s
sure val
ue that is
85
％
of the stan
dard ti
re p
r
e
s
sure is
set a
s
“lower limitin
g value of p
r
essu
re p
r
e
-
warning th
re
shol
d”; 85
℃
is set a
s
“t
empe
rature
pre
-
warning th
re
shold value
”
.
Figure 3. Te
mperature
Co
mpen
sation o
f
Pressure Th
reshold Val
u
e
5. Measure
d Data Fusion
Experiment
Bora autom
o
b
ile tire is take
n as an
example in the experim
e
n
t. According
to the
experie
nce of Bora engi
ne
ers, the p
r
e
s
sure of
front whe
e
l sh
ould
be 230KPa
and that of back
whe
e
l should
be 25
0KPa if runni
ng at hi
gh spee
d (f
ull
y
loaded
). If the intermedia
t
e value of 2
4
0
KPa is taken
for the fro
n
t and b
a
ck
wheel
s du
ring
air inflation,
then the u
p
per limit of ti
re
pre
s
sure pre
-
wa
rni
ng thresh
old an
d the lower lim
it of tire pre
s
sure
pre
-
wa
rning; the tire
temperature pre
-
warni
ng
thre
shol
d.
Put the four MENS se
n
s
ors in
a tire unifo
rmly
and mo
nito
r its p
r
e
s
sure and
temperature.
In ord
e
r to e
nhan
ce
the
measurem
ent
accu
ra
cy an
d facilitate
calcul
ation, e
a
c
h
sen
s
o
r
ne
ed
s four data, a
s
shown in table 2.
Since
the measure
d
data shoul
d be subje
c
t to
Gau
ssi
an di
stribution, the e
x
pectation E (T) and va
rian
ce D
(T)
sho
u
l
d be:
4
1
4
1
)
(
i
4
1
2
2
4
1
)
(
)
(
T
T
T
D
I
Table 2. Te
st Data Tabl
e of Pressu
re Te
mperature Se
nso
r
(u
nit: kg/cm,
℃
)
Sensor
1
2 3 4
Data1
292.6(85.
4)
289.7(84.
3)
291.5(84.
9)
300.3
（
92.4)
Data2
292.1(86.
1)
289.4(83.
1)
292.2(83.
4)
302.1(91.
2)
Data3
292.8(85.
4)
290.1(84.
5)
291.8(83.
1)
301.2(90.
1)
Data4
292.3(83.
7)
293.2 (86.
2)
291.6 (82.
4)
299.2(89.
7)
Expectation
292.45(85
.15)
290.6(84.
53)
291.77(83
.45)
300.7(90.
85)
Variance
0.0725(0.
78)
0.232(0.3
7
)
0.072(0.8
33)
0.115(0.1
1
)
From Table
2,
we ca
n get
the conf
iden
ce
di
sta
n
ce
matrix
Dmp fo
rme
d
by the
confid
en
ce di
stan
ce mea
s
ure dij of pressure data:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 493 – 49
9
498
Get the
thre
shol
d valu
e
1
.
0
)
4
,
3
,
2
,
1
,
(
j
i
ij
of dij
ba
se
d
on
expe
rien
ce
or test
results, and t
he matrix of relation is:
Then the me
asu
r
ed d
a
ta of the four sensor
s are a
ll valid data, the optimal fusion
numbe
r is L
=
4 and Bayes
optimal fusio
n
data P of the measured p
r
essu
re i
s
:
)
(
84
.
290
1
1
1
2
0
2
1
2
0
0
2
KPa
P
P
P
L
k
k
L
k
k
k
The co
nfiden
ce di
stan
ce
matrix formed
by
the confid
ence dista
n
ce measure dij
of
temperature data
is:
Take th
e thre
shol
d value
1
.
0
)
4
,
3
,
2
,
1
,
(
j
i
ij
of dij, and the matrix of relat
i
on is:
Then the me
asu
r
ed d
a
ta of the four sensor
s are a
ll valid data, the optimal fusion
numbe
r is L
=
3, and Bayes
optimal fusio
n
data T of the measured tempe
r
ature is:
)
(
32
.
85
1
1
1
2
0
2
1
2
0
0
2
C
T
T
T
O
L
k
k
L
k
k
k
After cal
c
ulat
ion, P=29
0.8
4
kPa a
nd
T=85.
32
℃
;
at this time, both pressure a
n
d
temperature
excee
d
the set pre-wa
rnin
g threshold.
The syste
m
sho
u
ld wa
rn
the driver of
tire
burst; if it continues to ri
se,
raise the al
ert le
vel; if it
tends to de
cline,
stop alarmin
g
.
6. Conclusio
n
In this pape
r, the decisi
o
n model on
TPMS pre-warnin
g thre
sh
old ba
sed o
n
multi-
sen
s
o
r
data
fusion i
s
put
forwa
r
d. Co
mpared with
the tradition
al tire pressure mo
nitori
n
g
system, thi
s
model
ha
s th
e adva
n
tage
s of i
n
form
at
ion inte
grity, uniformity, di
versity an
d f
ault
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
A Study on T
P
MS Pre-warning Th
re
shol
d Al
gorithm
Base
d on Multi
-
se
nsor… (Wang Ga
ng)
499
toleran
c
e. Th
e experim
ent
results
sho
w
t
hat the
syste
m
can
achieve the
functio
n
of pre-wa
rnin
g
for tire bu
rst and it ca
n effectively avoid
false
ala
r
m a
nd false
neg
a
t
ive and prev
ent the tire bu
rst
durin
g driving
so that the driver’s lif
e and
prop
erty safet
y
can be gu
aranteed.
Ackn
o
w
l
e
dg
ements
Scien
c
e an
d Tech
nolo
g
y Develo
pment
Plan of Jilin provin
ce (201
3011
47); Sci
ence and
Tech
nolo
g
y Prog
ram of Jili
n Educatio
nal
Committee d
u
ring "1
2th Five-Year Pla
n
" in 2013.
Referen
ces
[1]
Jeff Burgess.
T
i
re Pressure
Monitori
ng: A
n
Industr
y
Und
e
r Pres
s
u
re. S
ensors. P
e
terb
orou
gh, N
H
:
200
3; (07): 29-
33.
[2]
W
ang Ga
ng. S
i
mple
Ana
l
ysis
on T
e
chno
log
y
A
ppl
icati
on
of T
i
re Burst Early W
a
r
n
in
g S
y
st
em.
Journ
a
l
of Chan
gch
un Institute of T
e
c
hno
logy (S
ocia
l Scienc
e Editi
on)
. 200
9; 10(3
)
: 109-11
2.
[3]
Cui Sh
engm
in,
Yu Qun
.
Run
n
in
g Performa
nce an
d T
e
st
of Automotive
T
i
re. Beijing: C
h
in
a Machi
n
e
Press. 1995.
[4]
Liu Gui
p
i
ao, Luo Y
u
tao. A
Stud
y
on Influe
nce
of T
i
re Pressure o
n
Automobi
le
Performanc
e.
Guang
do
ng Pu
blic Sec
u
rity Scienc
e an
d T
e
chno
logy.
2
003;
(1): 58-61.
[5]
Z
hou Yu
la
n, Z
hao
Nin
g, Guo
San
x
u
e
. T
he T
i
re Pressure
Monitori
ng S
y
s
t
em Based
on
Multi-Se
nso
r
Information Fusion].
Co
mp
ute
r
Measure
m
ent
& Control.
20
0
5
; 13(3): 24
1-2
43.
[6]
Yan Ho
ng
w
e
i,
Li Hu
iju
an, L
i
Xi
n, Gao Qia
n
g
. Ba
sed
on A
r
tificial Immun
e
Algor
ithm of
Robot Mu
lti-
Sensor S
i
gn
al
Variati
on C
har
acte
ristics of th
e Detecti
on Me
thod.
T
E
LKOMNIKA Indon
esi
an Jo
urna
l
o
f
Electrical E
ngi
neer
ing.
2
013;
11(1
2
): 759
9-7
604.
[7]
Bian Gua
ngro
ng, Li Ho
ngsh
eng, He Ni
ng
h
u
i. Mu
lti-sens
o
r
Data Proces
sing a
nd F
u
si
ng Base
d o
n
Kalma
n
Filter
i
ng.
T
E
LKOMN
I
KA Indo
nesi
a
n Jo
urna
l
of
Electrical
En
gi
neer
ing.
20
12;
11(
3): 12
66-
127
7.
[8]
Xi
an
g
Xin
jia
n. T
e
mperature
Monitori
ng S
y
s
t
em of
Grain
Dep
o
t Based
on Multi-s
ens
o
r
Data F
u
sio
n
.
Chin
ese Jo
urn
a
l of Sc
ientific I
n
strument.
20
0
3
; 24(5): 52
5-5
28.
[9]
Hall
DL, Llinas J.
An intro
ducti
on to
multise
n
s
o
r da
da f
u
sio
n
. Proceedings
of the IEEE. 1997; 85(l):
16-
23.
[10]
W
ang
Qiume
i
, T
an
Bu. HT
ML--T
i
re Monitor:
Re
d
u
cin
g
Caus
e
of
T
i
re Burst. http://
w
ww
.
pcauto.com.cn/
q
c
y
p/
pe
iji
an/1
0
309/2
2
3
28.
Evaluation Warning : The document was created with Spire.PDF for Python.