Int
ern
at
i
onal
Journ
al of
P
ower E
le
ctr
on
i
cs a
n
d
Drive
S
ystem
s
(
IJ
PEDS
)
Vo
l.
12
,
No.
2
,
Jun
2021
,
pp.
650
~
661
IS
S
N:
20
88
-
8694
,
DOI: 10
.11
591/
ij
peds
.
v12.i
2
.
pp650
-
661
650
Journ
al h
om
e
page
:
http:
//
ij
pe
ds
.i
aescore.c
om
Improve
d DTC s
trateg
y of a
n elect
ric ve
hicle with
f
our
in
-
wh
ee
l
s inducti
on moto
r driv
e 4WDEV
usi
ng fuzzy lo
gic
contr
ol
Na
ir
N
ou
ri
a,
Ga
sb
aoui Br
ahi
m, G
ha
z
oua
ni Abdel
kader
, Ben
ou
d
jafer
Ch
eri
f
Depa
rtment
o
f
T
ec
hnology
,
Univ
ersit
y
of
T
ahr
i
Mohamm
ed
,
Sm
art
Grids
&
R
en
ewa
ble E
n
erg
ie
s
La
bor
at
ory
,
Bec
har
,
Alg
eria
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
N
ov
6
, 2
02
0
Re
vised
Dec
31
, 2
02
0
Accepte
d
Fe
b
28
, 202
1
In
thi
s
pap
er,
we
wil
l
study
a
f
our
-
whee
l
drive
elec
tri
c
veh
icle
(4W
DEV
)
with
two
cont
rol
strategie
s:
Conv
ent
ion
al
Dir
ec
t
Torque
Con
trol
(
CDTC)
and
DTC
base
d
on
f
uzz
y
log
ic
(DTF
C).
Our
over
a
ll
i
dea
in
thi
s
work
is
to
show
tha
t
the
4
WDE
V
equi
pped
with
f
our
induc
t
ion
m
otors
provid
ing
t
he
driv
e
of
the
dr
ivi
ng
wh
eels
cont
ro
ll
ed
by
the
d
irect
fu
zz
y
torque
cont
rol
e
nsures
good
stabi
lity
of
the
4WDEV
in
the
diffe
ren
t
topo
lo
gie
s
of
the
roa
d
,
bends
and
slopes,
and
inc
re
ase
s
th
e
r
ange
of
the
e
lectr
i
c
v
ehicle.
Num
erica
l
s
im
ulations
were
p
er
form
ed
on
an
e
lectr
i
c
veh
ic
l
e
powered
by
four
15k
W
induc
t
ion
mot
ors
in
te
gr
a
te
d
int
o
the
whee
ls
usin
g
the
MA
TLA
B/Sim
uli
nk
envi
ronm
ent
,
w
her
e
the
ref
e
ren
ce
spe
eds
of
eac
h
whee
l
(front
a
nd
rea
r)
ar
e
obta
in
ed
using
a
n
E
lectr
oni
c
Spe
ed
Dif
fer
en
ti
a
l
(
ESD).
Thi
s
ca
n
eve
ntu
al
ly
ca
use
it
to
sync
hroniz
e
th
e
whe
el
spe
eds
in
any
cur
ve
.
The
spe
ed
of
each
whee
l
is
cont
ro
ll
ed
by
two
typ
es
of
PI
and
F
LC
cont
ro
ll
ers
to
im
prov
e
stabi
lity
and
sp
ee
d
response
(
i
n
te
r
ms
of
set
point
tracki
ng,
disturba
nc
e
rej
e
ct
ion
and
cli
mb
time)
.
Simu
la
ti
on
resul
ts
show
tha
t
the
pro
posed
FLC
cont
rol
stra
te
gy
red
u
ce
s
torque,
f
lux
and
sta
to
r
cur
r
ent
ripp
le.
Wh
il
e
the
4WDEV
ran
ge
was
im
prove
d
th
roughout
the
dri
ving
cycle
and
b
at
t
ery
power
consumpt
ion
wa
s re
duce
d
.
Ke
yw
or
ds:
4W ele
ct
ric
ve
hicle
Direct t
orq
ue c
on
t
ro
l
Direct t
orq
ue f
uzzy co
ntr
ol
Ind
uction m
otor
Lit
hiu
m
-
i
on
ba
tt
ery
This
is an
open
acc
ess arti
cl
e
un
der
the
CC BY
-
SA
li
ce
nse
.
Corres
pond
in
g
Aut
h
or
:
Nair
N
ouria
Dep
a
rteme
nt of Tec
hnolog
y
Tahr
i
Mo
ham
med U
niv
e
rsity
B. P 417
rou
te
Kendsa
, Uni
ve
rsity
of Tah
ri
M
oha
mme
d,
B
echar
, Algeria
Emai
l:
nouri
a0
479@g
mail
.com
1.
INTROD
U
CTION
To
day,
in
t
he
automoti
ve
sec
tor,
ma
nufactu
rer
s
a
re
movin
g
to
wards
imp
rovin
g
inte
rn
al
com
busti
on
eng
i
nes
a
nd
hy
br
i
dizat
ion
with
el
ect
ric
mo
t
ors
to
mi
nimize
CO2
e
missi
on
s.
A
more
am
bi
ti
ou
s
al
te
rn
at
i
ve
is
to
do
without
the
inter
nal
c
ombust
ion
e
ng
ine,
a
nd
t
her
e
f
or
e
so
-
cal
le
d
zero
e
missi
on
pro
pu
lsi
on
[
1],
[2].
In
ge
ner
al
,
t
he
mo
st
c
om
m
only
us
e
d
el
ect
ric
act
uato
rs
i
n
the
majo
rity
of
i
ndus
tria
l
app
li
cat
io
ns
ar
e
bu
il
t
arou
nd
t
he
in
duct
ion
m
oto
r
[
3],
[
4].
T
he
in
du
ct
io
n
mo
t
or
in
pa
rtic
ular
i
s
chara
ct
erized
by
it
s
rob
us
t
ness,
reli
abili
ty,
lo
w
cost
a
nd
do
es
no
t
re
qu
ire
re
gula
r
mainte
na
nc
e.
H
oweve
r,
it
s
dy
namic
behavio
rs
a
re
of
te
n
very
com
plex,
beca
us
e
it
's
m
od
el
i
ng
re
su
lt
s
i
n
a
hi
gh
l
y
c
ouple
d
nonlinea
r
m
ulti
var
ia
te
s
ys
t
em
of
eq
uatio
ns
[
5]
,
[6].
Additi
on
al
ly,
some
of
it
s
sta
te
var
ia
bles,
includi
ng
flo
w
s,
can
no
t
be
m
easur
e
d.
Dif
fere
nt
dr
i
ve
te
ch
ni
qu
es
for
in
duct
ion
machine
s
hav
e
bee
n
int
rod
uc
ed
to
pr
ov
i
de
var
ia
ble
f
reque
ncy
s
peed
co
nt
ro
l.
Most
of
th
em
are
base
d
on
rig
or
ou
s
m
at
hemati
cal
f
ormal
isms
.
Among
al
l
t
he
c
on
tr
ol
meth
od
s
,
DTC
or
direct
to
r
qu
e
c
ontrol
is
consi
der
e
d pa
rtic
ularly
i
nteres
ti
ng
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N: 20
88
-
8
694
Impr
oved D
TC strategy
o
f
an
el
ect
ric
vehicl
e wi
th fo
ur
in
-
w
heels in
duct
ion m
oto
r
… (
N
air No
ur
ia
)
651
Taka
hash
i's
t
he
ory
is
to
s
pec
ific
al
ly
evaluat
e
the
co
ntr
ol
pulse
s
us
e
d
for
vo
lt
age
re
ver
si
ng
s
witc
hes
to
mainta
in
el
e
ct
ro
ma
gnet
ic
to
r
que
an
d
sta
tor
flo
w
within
two
pre
def
ine
d
de
pe
nbro
c
k
hy
ste
resis
ba
nds
[
7],
[8].
S
uc
h
a
n
a
pp
li
cat
io
n
of
t
his
te
ch
nique
a
ll
ow
s
decou
pling
of
to
rque
a
nd
flo
w
c
on
t
rol
without
the
ne
ed
f
or
pu
lse
widt
h
m
odulati
on
(PW
M
)
or
co
ordi
na
te
transfo
rmat
ion
.
Se
ver
al
st
ud
ie
s
a
re
sti
ll
unde
rw
a
y
to
imp
rove
the main
classi
c draw
bac
ks
of D
TC.
Amo
ng
these
dr
a
wb
ac
ks
a
re tor
qu
e
r
i
pp
le
s
and stat
or f
l
ux [8],
[9].
In
this
a
rtic
le
,
we
mainl
y
de
scribe
the
im
plementat
io
n
of
a
r
obust
an
d
ef
fici
ent
co
nt
ro
l
la
w
the
DTF
C,
w
hich
sta
nd
s
f
or
dire
ct
fu
zz
y
to
r
qu
e
co
ntro
l.
F
uzzy
log
ic
is
a
f
uzz
y
li
nguisti
c
ap
proac
h
use
d
by
a
typ
e
conditi
on
(S
i
-
The
n)
ba
sed
on
t
he
imi
ta
ti
on
of
a
ppr
oximat
e
qual
it
at
ive
as
pects
of
huma
n
reas
on
i
ng
[10].
And
to
ap
ply
it
in
our
work,
we
will
study
a
tr
act
ion
s
ys
te
m
of
a
n
el
ect
ric
veh
ic
le
(
4WD
EV).
T
he
4W
DEV
is
equ
i
pp
e
d
with f
our
a
sync
hronou
s
m
oto
r
s
en
s
ur
i
ng
t
he
dri
ve
of
t
he
dr
ivi
ng w
heels
c
ontr
olled
by
a d
irect
f
uzz
y
tor
qu
e
c
on
tr
ol.
T
he
pro
posed
co
ntr
ol
la
w
e
ns
ures
good
st
abili
ty
of
the
4WD
EV
in
different
r
oad
t
opologie
s
,
curves
a
nd
sl
opes
an
d
inc
rea
ses
the
auto
no
my
of
the
el
ec
tric
veh
ic
le
.
T
he
seco
nd
met
hod
is
intr
oduc
ed
to
rep
la
ce
the
tor
qu
e
hy
ste
resis
con
t
ro
ll
ers
,
flo
w
rate
c
on
t
ro
ll
ers,
an
d
switc
h
ta
ble
us
e
d
i
n
t
he
C
DTC
with
f
uzz
y
log
ic
c
on
tr
ollers.
T
he
mai
n
obje
ct
ive
of
the
DTF
C
met
hod
is
to
imp
rove
t
he
dy
namic
pe
rformance
of
e
le
ct
ric
veh
ic
le
s a
nd to
r
e
du
ce
torq
ue a
nd f
l
ow r
ip
ple
s.
2.
DESCRIPTI
ON OF THE
4
-
WHEE
L D
RIV
E
EL
ECT
RIC V
E
HIC
L
E
In
Fi
gure
1,
th
e
4W
DE
V
dr
i
ve
t
rain
sho
ws
that
t
he
pow
e
r
st
ru
ct
ur
e
of
this
dr
i
ve
t
rain
consi
sts
of
four
i
nductio
n
mo
to
rs
bu
il
t
in
to
the
w
heels
dr
i
ven
by
f
our
three
-
phase
in
ver
te
r
s,
the
pr
i
ncipal
po
wer
s
ource
for
the
veh
ic
le
bein
g
the
li
thi
um
-
io
n
(li
-
io
n)
ba
tt
ery.
It
is
connecte
d
to
t
he
DC
bus
via
a
two
-
wa
y
D
C
-
DC
(Buc
k
-
B
oost
conve
rter)
c
onve
rter.
A
G
hez
ouani
[8]
the
four
i
nductio
n
m
otors
are
dr
i
ve
n
by
the
DC
bus
via
a
DC
-
AC
co
nve
rter.
T
he
c
ontr
ol
meth
od
us
e
d
f
or
each
en
gi
ne
is
DT
FC
f
uzzy
tor
que
c
ontr
ol.
T
he
g
oal
of
this
appr
oach
is
to
en
han
ce
the
conve
ntion
al
direct
to
rque
con
t
ro
l
DTF
C
strat
eg
y.
T
he
ro
ll
in
g
e
ng
i
ne
s
ar
e
powe
red
by
a
n
el
ect
ronic
dif
f
eren
ti
al
.
T
his
s
ys
te
m
us
es
t
he
thr
ottl
e
posit
ion
an
d
the
w
he
el
ang
le
,
sp
ec
ifie
d
as
inputs
by the
r
otati
on
of the
w
heel
.
Figure
1.
Ge
ne
ral str
uctur
e
of
the E
V4W
D
f
our
-
w
heel
dr
i
ve e
le
ct
ric v
ehicl
e
stu
died
3.
LOAD B
AL
A
NC
E
OF T
HE FO
UR
-
WHE
EL
D
RI
VE E
LE
CTRIC
V
EHICLE
As
s
how
n
in
Fi
gure 2
,
t
he
tota
l
force
Ft
ot
re
quire
d
to move
t
he
el
ect
ric
ve
hi
cl
e
forw
a
rd
is
the
s
um
o
f
the
dif
fer
e
nt
c
ompone
nts
res
ulti
ng
f
r
om
the
bala
nce
of
the
mec
ha
nical
f
orces
ap
plied
to
the
veh
ic
le
[
8],
[
9].
Table
1
cl
arifie
s the c
oncepts
us
e
d
(
2) to
(8).
F
tot
=
F
rou
l
+
F
a
e
ro
+
F
S
lop
e
+
F
acc
=
F
R
+
F
acc
(1)
−
F
rou
l
is
t
he
r
olli
ng
r
esi
sta
nce
force
relat
ed
to
t
he
ro
ll
in
g
c
oe
ff
ic
ie
nt
of
the
wh
e
el
s
(
C
rr
)
.
The
ro
ll
in
g
resist
ance
forc
e is:
F
roul
≈
g
M
S
c
oote
r
C
rr
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
12
, N
o.
2
,
J
une
2021
:
650
–
661
652
−
F
a
e
ro
is
t
he
aer
odyn
amic
re
sist
anc
e
f
orce,
pr
oport
ion
al
t
o
t
he
ai
r
de
ns
it
y,
to
t
he
s
qu
a
re
of
the
wi
nd
sp
ee
d,
to
the
fron
ta
l
sect
io
n
of
the
ve
hicle
and
to
it
s
ai
r
pe
netrati
on
c
oeff
ic
ie
nt
(
C
px
)
.
Its
e
xpr
ession
is give
n b
y
(
2).
F
a
e
ro
=
1
2
ρ
S
f
C
px
(
V
v
e
h
−
V
wind
)
2
(2)
Fr
onta
l
sect
io
n
of
t
he
veh
ic
le
and
to
it
s
a
ir
pe
netrati
on
coeffic
ie
nt
(
C
px
)
.
Its
ex
pr
essi
on
is
giv
e
n
by (3).
F
a
e
ro
=
1
2
ρ
S
f
C
px
(
V
v
e
h
−
V
wind
)
2
(3)
−
F
pe
n
t
e
is
the r
esi
sta
nc
e
f
or
ce
of
t
he
slop
e
to b
e
cl
imbe
d.
In
the
c
ase
w
her
e
the
elec
tric
veh
ic
le
w
ou
l
d
hav
e
to
cl
imb
a
co
rn
e
r
slo
pe
(
α
p
)
as
s
how
n
in
Fi
gure
2,
the
re
is
an
a
dd
it
io
nal
force
pro
portio
nal
to
the total
mas
s
of the
ve
hicle
that is a
pp
li
ed
to
it
s fo
rw
a
rd
m
otion t
his
force
is g
i
ven by:
F
pente
=
g
M
v
e
h
.
sin
(
α
p
)
(4)
F
a
c
c
is t
he d
yn
a
mic t
erm fo
r
t
he
ac
cel
erati
on
or
de
cel
erati
on
of t
he
el
ect
ric
veh
i
cl
e.
F
a
c
c
=
M
v
e
h
d
V
veh
dt
=
M
v
e
h
γ
(5)
−
Finall
y,
t
he
t
otal eff
or
t
of the
veh
ic
le
's
for
wa
rd r
esi
sta
nce
is
wor
t
h.
F
tot
=
g
M
v
e
h
C
rr
+
1
2
ρ
S
f
C
px
(
V
v
e
h
−
V
win
d
)
2
+
g
M
v
e
h
.
sin
(
α
p
)
+
M
v
e
h
γ
(6)
−
The w
heel resi
sta
nce torq
ue
C
r
is relat
ed
t
o
t
he resi
sta
nce
forc
e by
(7).
C
r
=
F
tot
.
R
ω
(7)
−
The
a
ngular
ve
locit
y
ω
_(r
-
i)
(rad/s
) of
eac
h dr
i
ven wheel
is
r
el
at
ed
t
o
the
veh
ic
le
s
pee
d by (
8).
ω
r
−
i
=
V
veh
2
R
ω
(8)
Figure
2
.
F
or
ce
s ex
e
rted o
n
th
e f
our
-
w
heel dri
ve
el
ect
ric v
ehicl
e
Table
1.
A
pp
ea
ran
ce
prope
rtie
s of acce
pted
man
us
cri
pts
ℎ
Total
m
ass
of
th
e
v
eh
icle
ℎ
.
2
Veh
icle
in
ertie
ℎ
.
−
1
Veh
icle
sp
eed
=
0
.
−
1
W
in
d
sp
eed
=
9
.
81
.
−
2
Acceler
atio
n
of
Gr
av
ity
Air
p
en
etration
co
eff
icien
t
2
Fron
t
sectio
n
of
th
e
v
eh
icle
.
3
Air
v
o
lu
m
e
Den
sity
W
h
eel
radiu
s
Rig
h
t
Rear
W
h
eel
Lef
t
Re
ar
W
h
eel
4.
THE
IN
-
WH
EE
L E
LE
CTRIC DRI
VE I
M MO
DEL
Fo
r
the
el
ab
or
at
io
n
of
c
ontr
ol
strat
egie
s,
it
is
nece
ssary
to
fi
nd
a
c
ompr
om
i
se
bet
ween
the
c
omplexit
y
an
d
the
acc
uracy
of
t
he
modeli
ng
a
nd
si
nc
e
the
ob
je
ct
iv
e
of
the
prese
nt
wor
k
is
the
di
rect
tor
qu
e
c
ontrol
base
d
on
f
uz
zy
lo
gic
(
DTFC
)
base
d
on
t
he
kn
ow
le
dg
e
of
th
e
am
plit
ud
e
a
nd
posit
ion
of
the
sta
to
r
flu
x
[11],
[12
]
the
com
plete
m
odel
of
the
mac
hi
ne
in
the
Par
k
re
fer
e
nce
fr
a
me
li
nked
t
o
the
sta
to
r
ref
e
ren
ce
fram
e (
α
-
β
) (9) to
(12).
̇
=
+
(9)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N: 20
88
-
8
694
Impr
oved D
TC strategy
o
f
an
el
ect
ric
vehicl
e wi
th fo
ur
in
-
w
heels in
duct
ion m
oto
r
… (
N
air No
ur
ia
)
653
su
c
h
as:
=
[
]
,
=
[
]
(10)
[
−
−
−
0
0
0
0
0
0
]
;
=
[
1
0
0
1
1
0
0
1
]
(11)
with:
=
1
−
2
,
=
,
=
,
=
Ω
,
=
1
,
=
−
1
(
1
+
1
)
(12)
5.
CONVE
NTI
ONAL DT
C
F
OR ONE
IN
-
WHEE
L IN
D
UC
TI
ON M
O
TOR
DRIVE
In
the
mid
-
ei
ghti
es,
I
s
uggest
ed
a
M
et
hod
f
or
the
Direct
T
orq
ue
Co
ntro
l
of
Ind
uction
Moto
rs
(
DTC)
in
the
li
te
rature
of
Ta
ka
hash
i
T
[13],
[14],
Nog
uch
i
a
nd
Deerbr
ook.
T
he
DTC
the
ory
is
base
d
on
a
direc
t
determi
nation
of
the
puls
es
us
ed
for
the
volt
age
in
ver
te
r
s
wi
tc
hes.
This
i
s
done
to
mainta
in
the
el
ect
romag
netic
tor
qu
e
an
d
the
sta
t
or
flo
w
in
t
wo
hyste
resis
ba
nds.
Such
a
pp
li
cat
ion
ens
ur
es
t
hat
to
rque
and fl
ux contr
ol
are disco
nn
e
c
te
d.
The
vo
lt
age
i
nverter
al
lo
ws
f
or
7
locat
io
ns
i
n
the
ph
as
e
pla
ne,
wh
ic
h
co
rresp
onds
to
t
he
8
seq
ue
nces
of
the
vo
lt
ag
e
vecto
r
at
the
in
ver
te
r
outp
ut
[
15],
[
16]
.
The
blo
c
k
dia
gr
am
in
Fi
gure
3
s
hows
a
s
ynopti
c
D
T
C
diag
ram
us
e
d
i
n
a
th
ree
-
wh
ee
le
d
el
ect
ric
sc
oote
r
i
nductive
mo
to
r.
T
he
flu
x
cal
c
ulati
on
c
an
be
e
sti
mate
d
from
the
sta
tor
c
urre
nt
an
d
volt
age
measu
reme
nts
of
t
he
in
du
ct
io
n
mac
hin
e
[17
].
Ta
ble
2
s
ho
ws
the
DTC
c
on
t
ro
l
truth t
able.
φ
s
α
=
∫
(
v
s
α
−
R
s
i
s
α
)
t
0
dt
(13)
φ
s
β
=
∫
(
v
s
β
−
R
s
i
s
β
)
t
0
dt
(14)
The
sta
to
r flu
x mo
du
le
is
.
φ
s
=
√
φ
s
α
2
+
φ
s
β
2
(15)
The
fiel
d
in
wh
ic
h
t
he
vect
or
is
locat
ed
i
s
dete
rmin
e
d
f
rom
the
c
omp
on
e
nts
an
d
th
e
an
gle
betwee
n
the
r
e
po
sit
ory (α
-
β)
and the
vect
or
[18].
θ
s
=
a
rctg
(
φ
s
β
φ
s
α
)
(16)
Wh
e
n
t
he
tw
o flu
x
c
ompone
nt
s ar
e
reache
d,
the ele
ct
romag
netic
torq
ue
c
a
n be calc
ulate
d b
y [19].
T
em
=
3
2
p
[
φ
s
α
i
s
β
−
φ
s
β
i
s
α
]
(17)
Table
2.
DTC
con
t
ro
l t
r
ut
h
ta
ble
Secto
r
1
2
3
4
5
6
∆
=
1
∆
=
1
V
2
V
3
V
4
V
5
V
6
V
1
∆
=
0
V
7
V
0
V
7
V
0
V
7
V
0
∆
=
−
1
V
6
V
1
V
2
V
3
V
4
V
5
∆
=
0
∆
=
1
V
3
V
4
V
5
V
6
V
2
V
1
∆
=
0
V
0
V
7
V
0
V
7
V
0
V
∆
=
−
1
V
5
V
6
V
1
V
2
V
3
V
4
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
12
, N
o.
2
,
J
une
2021
:
650
–
661
654
Figure
3.
Co
nventional
DTC
for i
nductio
n m
otor
dr
ive
in
t
he
wheel
s
us
ed
in
the
4W
DE
V
6.
DIRECT
T
ORQUE FU
ZZY C
ONTROL
OF TH
E ASYN
CHRON
OUS M
ACHINE
The
tra
diti
on
al
DTC
co
ntr
ol
prov
i
des
ra
pi
d
an
d
preci
se
respo
ns
e
to
el
ect
ro
ma
gnet
ic
tor
qu
e
a
nd
sta
tor
fl
ux.
T
he
gr
eat
est
dr
a
wb
ac
k
of
t
his
powe
r,
howe
ve
r
is
the
la
r
ge
tor
qu
e
,
sta
tor
f
lux
a
nd
cu
rr
e
nt
ripp
le
du
e
to
the
us
e
of
hyste
resis
c
omparat
or
s
[
20
].
This
sect
io
n
pro
po
ses
t
o
boos
t
t
he
e
ff
ic
ie
nc
y
of
t
he
tra
diti
on
a
l
CDTC
Co
ntr
ol
,
Direct
To
rque
F
uzz
y
C
ontrol
(
DT
FC).
T
his
met
hod
pro
po
ses
to
re
place
the
hyst
eresis
com
par
at
or
s
a
nd
the
sel
ect
io
n
ta
ble
with
a
con
t
ro
ll
er
ba
se
d
on
a
f
uzz
y
i
nf
e
ren
ce
s
ys
te
m
[
21]
,
[22
].
F
igure
4
sh
ows
t
he
sc
he
mati
c
diag
ram
of
t
he
DTF
C
c
on
t
ro
l
of
an
i
nductio
n
mo
t
or
integrate
d
in
t
he
w
heels
of
the
fou
r
-
wh
eel
dr
i
ve
el
ect
ric
veh
ic
le
.
The
ob
ta
ine
d
tor
qu
e
(e
Tem
)
and
flu
x
(
)
error
s
a
s
well
as
the
an
gle
ar
e
require
d
by
t
he
f
uzzy
in
fere
nc
e
sy
ste
m
to
e
va
luate
the
re
fere
nce
volt
age
ve
ct
or
to
dri
ve
the
t
orq
ue
a
nd
f
lux
to
their
desire
d v
al
ues.
Figure
4
.
Sc
he
mati
c d
ia
gram
of the
direct
fuzzy c
on
t
ro
l
(DTFC)
of the
MI inte
gr
at
e
d
in
the
wh
eel
s
of th
e EV
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N: 20
88
-
8
694
Impr
oved D
TC strategy
o
f
an
el
ect
ric
vehicl
e wi
th fo
ur
in
-
w
heels in
duct
ion m
oto
r
… (
N
air No
ur
ia
)
655
Figure
5
disp
l
ays
t
he
me
mbe
rsh
i
p
functi
ons
for
t
he
f
uz
zy
in
fer
e
nce
method
i
nput
and
outp
ut
var
ia
bles.
Tra
pe
zoidal
a
nd
t
riangular
ass
oci
at
ion
f
un
ct
io
ns
ha
ve
bee
n
sel
ect
ed.
The
in
put
of
the
to
rqu
e
er
ror
consi
sts
of
3
f
uzzy
set
s
N
(
ne
gative)
,
Z
(ze
ro)
a
nd
P
(
posi
ti
ve)
.
T
w
o
f
uz
zy
set
s
wer
e
c
on
si
der
e
d
for
t
he
fl
ow
error
mem
ber
s
hip
functi
ons,
N
(
ne
gative)
a
nd
P
(
posit
ive)
[23].
T
he
sta
tor
flu
x
an
gle
c
an
b
e
de
fine
d
by
si
x
li
ng
uisti
c
va
riables
(
1
⟶
6
),
t
o
ha
ve
a
fi
ne
a
dj
us
tm
ent.
The
i
nf
e
re
ntial
de
vice
outp
ut
var
ia
ble
is
di
vid
e
d
into
ei
ght
in
di
vid
ualto
ns,
t
wo
nu
ll
volt
ages
(
V0
a
nd
V7)
an
d
six
nu
ll
volt
ages.
T
he
outp
ut
var
ia
bl
e
membe
rs
hip
f
unct
ions
are
s
hown
in
Fig
ur
e
5.
T
he
diff
e
ren
t
possible
c
omb
inati
on
s
of
3
f
uzzy
set
s
for
t
orq
ue
error, 2
fuzzy
s
et
s
f
or
fl
ux
er
r
or
an
d
six
sect
or
s
f
or
sta
to
r
fl
ux
a
ng
le
f
orm 3
6
r
ules
in
the
b
asi
s o
f
the
in
f
eren
ce
sy
ste
m.
Figure
5.
The
membe
rs
hip
f
unct
ions f
or the
input a
nd
outp
ut v
a
riables
of
the fuzz
y
in
fere
nce s
ys
te
m
The
r
ule
base
is
base
d
on
a
s
ta
tor
flu
x
diag
ram
(
α
-
β
)
i
n
t
he
plane
.
For
exam
ple,
i
f
t
he
an
gle
θs
of
the
sta
to
r
flu
x
li
es
in
the
valu
e
of
θ2
if
on
e
wan
ts
to
slo
wly
decr
ea
se
t
he
tor
qu
e
an
d
qui
ckly
incr
ease
t
he
flo
w
then
the
vect
or
V
1
is
the
m
ost
su
it
able
al
te
r
native.
T
he
sa
me
rati
on
al
e
is
us
e
d
to
co
ns
tr
uct
the
r
ule
ba
se
f
or
the
fl
uid
di
rec
t
tor
que
c
on
t
r
ol
in
Ta
ble
3.
The
la
w
s
a
re
the
flui
d
c
ontr
ol
i
nf
e
rio
r
e
ngine.
T
he
y
e
xpress
in
a relat
ion bet
w
een eleme
ntar
y flu
ffy pr
oposa
ls or co
njuncti
on
s
of
fun
dam
ental
prop
os
al
s
[
21],
[
23].
:
ℎ
(18)
With
,
an
d
de
are
the
li
ng
uisti
c
var
ia
bles
of
t
he
fl
ux
e
rror,
the
to
r
qu
e
error
a
nd
the
s
ta
tor
fl
ux
a
ngle
,
resp
ect
ivel
y.
is
the
ou
t
pu
t l
i
nguisti
c v
aria
ble
and
is r
ule
numb
e
r
i.
Table
3.
Fu
zz
y l
og
ic
switc
hi
ng
ru
le
s
e
φ
s
e
Tem
θ
1
θ
2
θ
3
θ
4
θ
5
θ
6
N
N
V
5
V
6
V
1
V
2
V
2
V
4
P
V
0
V
7
V
0
V
7
V
0
V
7
Z
V
6
V
4
V
5
V
6
V
1
V
2
P
N
V
3
V
1
V
3
V
3
V
4
V
5
P
V
7
V
0
V
7
V
0
V
7
V
0
α
i
=
min
(
μ
A
i
(
e
φ
s
)
,
μ
B
i
(
e
Te
m
)
,
μ
C
i
(
θ
s
)
)
(
19)
By fuzz
y reaso
ning, Ma
md
a
ni
's minim
um p
r
ocess (
20).
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
12
, N
o.
2
,
J
une
2021
:
650
–
661
656
μ
V
i
(
v
)
=
min
(
α
i
,
μ
V
i
(
v
)
)
(20)
(
)
,
(
)
,
(
)
and
(v)
(
)
design
at
in
g
re
spe
ct
ively
the
de
gr
ees
of
mem
ber
s
hip
of
,
,
and
t
to
t
he
f
uzzy
sets
,
,
an
d
.
In
our
ca
se, t
he
outp
ut is const
it
uted
by a
set
of
sin
gletons,
we wil
l app
l
y
the
M
A
X met
hod (
21).
The
value
co
rr
e
spo
nd
i
ng
to
(
)
sho
uld
the
n
be
c
onve
rted
t
o
a
volt
age
vect
or.
I
n
the
pro
pose
d
fu
zz
y
con
t
ro
ll
er
for
defuzzifi
cat
io
n
the
meth
od
of
the
center
of
gr
a
vity
was
us
ed.
Fi
gure
6
shows
t
he
c
har
ac
te
risti
c
su
r
face
of
t
he
pro
po
se
d
f
uzz
y
co
ntr
oller,
it
e
xpresses
the
v
a
riat
ion
s
of
the
actual
va
lue
of
the
co
ntr
oller
o
ut
pu
t
as a fu
nction o
f
the
in
pu
ts
w
he
n
the
lat
te
r
ar
e trave
rsing t
he
sp
eec
h u
niv
e
rse.
μ
V
ou
t
(
v
)
=
ma
x
i
=
1
36
(
μ
V
i
(
v
)
)
(21)
Figure
6.
Cha
r
act
erist
ic
su
r
fa
ce o
f
the
fuzz
y sel
ect
ion
ta
ble
−
Desig
n
of fuzz
y
lo
gic s
pee
d
c
on
t
ro
ll
er
The
cl
assic
PI
con
t
ro
ll
er
has
been
use
d
in
s
peed
c
on
tr
ol
for
the
majo
rity
of
dif
fere
nt
i
nduction
m
otor
con
t
ro
l
st
rateg
ie
s.
H
oweve
r,
the
P
I
c
on
t
ro
ll
er
has
no
t
pro
vid
e
d
sat
isfact
ory
perf
or
ma
nc
e
in
case
of
s
udde
n
sp
ee
d
cha
nges
,
load
t
orqu
e
disturba
nces,
low
s
peed
c
on
t
ro
l
due
to
con
ti
nu
ou
s
va
riat
ion
s
in
m
achin
e
par
a
mete
rs
an
d
operati
ng
co
ndit
ion
s
.
T
he
re
fore,
in
orde
r
t
o
overc
om
e
th
ese
dra
wb
ac
ks
,
co
ntr
ollers
ba
sed
o
n
fu
zz
y
lo
gic
a
re
highly
desi
rabl
e.
In
this
wor
k
the
cl
assic
al
PI
c
on
t
ro
ll
er
is
rep
la
ce
d
by
a
rtif
ic
ia
l
intel
li
gen
ce
te
chn
iq
ues
, suc
h
as
Fuzzy
Log
ic
Con
tr
ol
(F
L
C) to
im
pro
ve dri
ve per
forma
nce
[23, 2
4].
The
dev
ia
ti
on
betwee
n
the
ref
e
ren
ce
spe
ed
an
d
t
he
act
ual
sp
ee
d
of
t
he
in
du
ct
i
on
mac
hin
e,
(
)
=
∗
(
)
−
(
)
,
an
d
t
he
var
i
at
ion
of
t
his
de
viati
on
∆
(
)
=
(
)
−
(
−
1
)
,
a
re
us
e
d
as
fu
zz
y
con
t
ro
ll
er
in
put
fu
zz
y
va
riabl
es
of
t
he
sp
ee
d
an
d
t
he
c
ontr
ol
le
r
outp
ut
is
t
he
ref
e
re
nce
el
ect
ro
ma
gnet
ic
tor
qu
e
∗
,
th
e
blo
c
k
dia
gr
a
m
of
w
hich
is
sho
wn
in
Figure
7
th
e
f
uzzifica
ti
on
of
the
f
uzz
y
co
nt
ro
ll
er
in
pu
t
a
nd
ou
t
pu
t
var
ia
bles
is
show
n
i
n
Figure
7.
Eac
h
of
the
thre
e
li
nguisti
c
var
i
ables
is
re
pr
e
s
ented
by
fi
ve
fu
zz
y
su
bse
ts
(
G
N=
Larg
e
Ne
gativ
e,
P
N=Small
Neg
at
ive
,
Z=
Zero,
P
P=Sm
al
l
Po
sit
ive,
GP
=La
rg
e
P
osi
ti
ve)
.
M
ore
ov
e
r,
the
defuzzifi
cat
io
n
ha
s
bee
n
pe
rformed
by
the
center
of
gravi
ty
met
hod
ass
oc
ia
te
d
with
t
he
s
um
-
pro
du
ct
i
nterf
e
ren
ce
meth
od
[
8], [2
5].
(a)
(b)
(c)
Figure
7. F
un
ct
ion
s
of the
FL
C spee
d
c
on
t
rol
le
r
mem
ber
s
for (a
)
the
r
e
fere
nce tor
qu
e
∗
, (b) the
var
ia
ti
on
of the s
pee
d
e
r
ror
∆e
(k)
a
nd (c
)
the
sp
ee
d
e
rror e(
k)
-4
-2
0
2
4
0
0
.
2
0
.
4
0
.
6
0
.
8
1
R
e
f
e
r
e
n
c
e
T
o
r
q
u
e
T
e
m
*
D
e
g
r
e
e
o
f
m
e
m
b
e
r
s
h
i
p
n
Z
p
N
P
-4
-2
0
2
4
0
0
.
2
0
.
4
0
.
6
0
.
8
1
C
h
an
g
e
i
n
e
r
r
o
r
s
p
e
e
d
D
e
g
r
e
e
o
f
m
e
m
b
e
r
s
h
i
p
PN
Z
PP
GN
GP
-4
-2
0
2
4
0
0
.
2
0
.
4
0
.
6
0
.
8
1
S
p
e
e
d
E
r
r
o
r
D
e
g
r
e
e
o
f
m
e
m
b
e
r
s
h
i
p
PN
Z
PP
GN
GP
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N: 20
88
-
8
694
Impr
oved D
TC strategy
o
f
an
el
ect
ric
vehicl
e wi
th fo
ur
in
-
w
heels in
duct
ion m
oto
r
… (
N
air No
ur
ia
)
657
Su
c
h
t
hat,
for
the
ℎ
r
ule:
is
it
s
degree,
is
the
abscissa
of
it
s
center
of
gr
a
vi
ty
a
nd
is
the
su
r
face
of
t
he
ou
t
pu
t
f
uzz
y
s
ub
s
et
.
T
he
r
ule
ba
se
for
deci
di
ng
the
ou
t
pu
t
of
the
in
fer
e
nc
e
s
ys
te
m
co
ns
i
sts
of
2
5
I
f
-
The
n
r
ules
in
this
ca
se
because
th
ere
are
5
fu
z
zy
set
s
in
each
of
th
e
in
pu
ts.
Ta
ble
4
s
hows
re
pr
e
sentin
g
the in
fer
e
nce
r
ule b
a
se.
∆
T
em
∗
=
∑
μ
ci
X
Gi
S
i
25
i
=
1
∑
μ
ci
S
i
25
i
=
1
(22)
Table
4.
Fu
zz
y r
ule
∆
⁄
GN
PN
Z
PP
GP
GN
N
Z
N
N
Z
PN
N
N
N
Z
P
Z
n
N
Z
P
P
PP
n
Z
P
P
P
GP
Z
P
P
P
P
7.
PROP
OSED
SP
EED CY
CLE FO
R T
HE 4WD EL
ECTR
IC V
E
HI
C
LE
We
ha
ve
pro
pose
d
a
relat
ive
ly
short
10s
s
pe
ed
lo
op
to
te
s
t
the
eff
ic
ie
nc
y
of
t
he
D
TFC
direct
f
us
z
y
tor
qu
e
c
on
tr
ol
strat
egy
of
the
4W
DEV
tract
ion
syst
em,
an
d
Fig
ur
e
8
pr
e
sents
t
he
s
pee
d
pro
file
of
the
cycle
.
This
r
ou
te
is
c
har
act
erize
d
by
seve
n
s
ucces
sive
p
hases.
In
the
first
sta
ge,
the
ve
hicle
is
pu
s
he
d
strai
gh
t
at
a
sp
ee
d
of
50K
m/h
i
n
the
sec
ond
sta
ge
.
a
righ
t
tu
rn
is
im
po
s
ed
on
the
ve
hicle
by
a
ste
erin
g
a
ngle
c
omman
d
(
=
25°
)
as
s
how
n
i
n
Figure
9,
i
n
th
e
thir
d
ph
ase
,
the
4W
DEV
runs
on
a
strai
gh
t
r
oad
at
the
sa
me
s
pee
d,
in
the
four
t
h
phas
e
a
le
ft
t
urn
is
imposed
on
th
e
ve
hicle
with
a
ste
eri
ng
a
ng
l
e
co
mma
nd
(
=
−
15°
).
T
he
fifth
ph
a
se,
the
ve
hicle
runs
on
a
st
raig
ht
ro
a
d
with
a
sp
ee
d
of
50K
m/h.
I
n
the
sixth
ph
ase
,
th
e
VE4W
D
cl
imbs
a
n
incli
ned
r
oad
with
a
n
a
ng
le
of
10°
(slope
)
with
a
sp
ee
d
of
70K
m/h.
Fi
nally,
t
he
la
st
on
e
(
7)
prese
nt
s
the
decele
rati
on
phase
w
her
e
t
he
s
peed
of
the
ve
hicle
is
30
Km/h.
The
c
onstrai
nts
of
th
e
r
oa
d
a
re
pre
sented
in
Table
5.
Table
5.
T
op
olo
gies
of s
pecifi
ed dri
ving ro
utes
Ph
ase
Tim
e
(
Se
c)
Even
t info
rm
atio
n
Véh
icu
le sp
eed K
m
/h
01
0
s < t
< 1,5
s
Straigh
t r
o
ad
5
0
km/h
02
1
,5s
<
t <
2
,5s
Cu
rved
r
o
ad
SI
DE
r
ig
h
t
5
0
km/h
03
2
,5s
<
t <
4
s
Straigh
t r
o
ad
5
0
km/h
04
4
s < t
< 5s
Cu
rved
road
SI
DE
lef
t
5
0
km/h
05
5
s < t
< 6s
Straigh
t r
o
ad
5
0
km/h
06
6
s < t
< 8s
Climbin
g
slo
p
e 10%
7
0
km/h
07
8
s < t
< 10
s
Straigh
t r
o
ad
3
0
km/h
Figure
8. S
pecify
dri
ving
r
oad topolo
gy
Figure
9. Stee
r
ing
a
ngle
v
a
ria
ti
on
8.
RESUL
TS A
ND DIS
CUS
SION
The
numerical
simulat
io
ns
in
this
sect
io
n
w
ere
do
ne
with
the
MAT
LAB/
Simuli
nk
e
nv
i
r
onment
on
an
el
ect
ric
ve
hi
cl
e
dr
ive
s
ys
te
m
dri
ve
n
by
f
our
15kW
in
duc
ti
on
m
otors,
w
hich
wer
e
i
ntegr
at
e
d
in
the
w
heels
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
12
, N
o.
2
,
J
une
2021
:
650
–
661
658
as
sho
wn
in
Fi
gure
1.
T
he
obje
ct
i
ves
of
the
simulat
ion
car
r
ie
d
out
e
valuat
ed
th
e
ef
fecti
ve
ness
of
the
diff
e
ren
t
con
t
ro
l
strat
e
gi
es
propose
d
(
conve
ntion
al
DTC
an
d
DTFC
)
on
the
dyna
mics
of
t
he
e
le
ct
ric
veh
ic
le
,
and
a
com
par
is
on
w
as
ma
de
betwe
en
t
he
t
wo.
T
hi
s
sy
ste
m
was
simulat
ed
us
i
ng
a
re
f
ere
nce
wh
eel
sp
ee
d
gi
ven
by
the
to
po
l
ogy
s
how
n
in
Fig
ure
8.
The
dy
na
mic
an
d
ae
rod
yn
a
mic
cha
rac
te
risti
cs
of
t
he
ve
hicle
an
d
T
able
6
include
s
in
duc
ti
on
m
otor
paramet
ers.
The
aerod
yn
a
mic
tor
que
is
re
duc
ed
with
D
TF
C
con
t
ro
l
relat
ive
to
CDTC.
56.
6Nm
wit
h
DTF
C
a
nd
57.1Nm
pe
r
CD
TC
(pha
se
6,
see
Fig
ur
e
10)
.
T
his
val
ue
ca
n
be
ex
pl
ai
ned
by
the
la
r
ge
f
ront
al
zon
e
in
t
he
case
of
CD
TC
versus
DT
FC.
It
can
be
seen
that
the
overa
ll
resist
ive
tor
qu
e
is
impro
ved in
D
TFC co
mp
a
re
d t
o
C
DTC (see
Figure
11).
Table
6.
Propo
sed 4
WD el
ect
ric v
e
hicle
and
IM p
a
rameter
s
Para
m
eters
Na
m
e
Sy
m
b
o
l
Valu
e
Para
m
eters
Na
m
e
Sy
m
b
o
l
Valu
e
W
h
eel r
ad
iu
s
R
w
(m)
0
.32
Ro
to
r
Ind
u
ctan
ce
(
)
0
.14
9
Veh
icle
m
ass
M
(kg
)
1300
Ro
to
r
Ind
u
ctan
ce
(
)
0
.14
9
Aerod
y
n
am
ics d
ra
g
coeff
icien
t
C
d
0
.3
Mutu
al I
n
d
u
ctan
ce
(
)
0
.14
1
Veh
icle f
ron
tal ar
e
a
A
f
(m
2
)
2
.60
Stato
r
Res
istan
ce
(
Ω
)
1
.37
Tir
erollin
g
r
esis
t
an
ce
co
eff
icien
t
C
r
0
.01
Ro
to
r
Res
istan
ce
(
Ω
)
1
.1
Air
d
en
sity
Ρ
a
ir
(kg
/
m
2
)
1
.2
Nu
m
b
er
o
f
po
le pairs
2
Gear
co
eff
ici
en
t
k
g
ea
r
5
Moto
r
-
lo
ad
inertia
(
.
2
)
0
.1
W
id
th
of veh
icle
d
ω
(m)
1
.5
Rated
po
wer
(
)
10
Leng
th
of veh
icle
L
ω
(m)
2
.5
Visco
u
s fr
ictio
n
co
eff
icien
t
(
.
.
)
0
.00
0
1
4
Figure
10. Ve
hi
cl
e
Aerody
na
mics t
orq
ue va
riat
ion
with
CDTC a
nd D
T
FC
Figure
11.
Global
ly
veh
ic
le
r
esi
sti
ve
tor
que
The
dri
ve
r
pr
ovides
th
e
ste
ering
a
ngle
of
t
he
fron
t
wh
eel
s;
the
el
ect
ronic
diff
e
re
ntial
is
centere
d
on
the
sp
ee
ds
of
t
he
dri
ving
wh
e
el
s.
The
s
peeds
of
the
t
wo
rig
ht
-
hand
dri
ve
wh
eel
s
loc
at
ed
on
th
e
outsi
de
of
the
bend
(
rig
ht
tu
r
n
phase
2)
Swi
tc
h
at
s
pee
ds
gr
eat
er
tha
n
th
e
tw
o
i
ns
ide
le
ft
dri
ve
wh
eel
s
of
t
he
be
nd.
At
the
mo
me
nt
t=
4s
t
he
veh
ic
le
is
in
the
sec
ond
l
eft
tur
n
9
(
ph
a
se
4);
the
sam
e
thin
g
for
t
he
el
ect
ronic
di
fferentia
l
cal
culat
es
the
ref
e
ren
ces
of
t
he
new
s
peed
s
to
tu
r
n
the
w
heels
to
sta
bili
ze
the
ve
hicle
inside
the
le
ft
tur
n.
Table
7
s
hows
the s
peed val
ue
s for eac
h w
he
el
f
or
bo
t
h
t
urns (phases
2 an
d 4).
A
F
uzz
y
L
ogic
Co
ntro
ll
er
(FL
C)
wa
s u
sed
in
place of
an
othe
r
tra
diti
on
al
P
I
ty
pe
t
o
help
i
mpro
ve
t
he
sp
ee
d
res
pons
e
of
the
ve
hicle
.
The
ad
va
ntag
e
o
f
this
c
on
t
r
oller
li
es
in
it
s
robustness
a
ga
inst
sp
ee
d
va
r
ia
ti
on
s
and
f
ollo
ws
th
e
set
point
without
overs
hoot
ing
a
nd
with
good
acc
uracy
.
Fi
gure
12
s
hows
the
sim
ul
at
ion
resu
lt
s
of
th
e li
near
velocit
y o
f
the
4WD veh
ic
le
u
sing
t
he
two
c
ontr
ol stra
te
gies (
D
TC w
it
h
a PI
ty
pe ve
locit
y
con
t
ro
ll
er
a
nd
DTF
C
with
FL
C).
From
the
r
esults
we
can
no
ti
ce
that
t
he
eff
ect
of
the
di
sturbance
s
is
c
le
arly
visible
in
the
li
near
velocit
y
respo
ns
e
of
th
e
ve
hicle
by
usi
ng
the
DTC
strat
egy
(
w
here
the
ve
hicle
is
dri
ve
n
on
a
10%
sl
ope
phase
6
road)
with
a
n
overta
king
of
0.1
5%.
T
he
res
ult
of
the
tw
o
co
ntr
ol
la
ws
can
be
su
m
marized
in
Table
8.
On
t
he
oth
e
r
ha
nd,
t
he
D
TFC
strat
egy
gi
ves
us
a
good
dy
namic
in
te
rms
of
f
ollow
i
ng
the
inst
ru
ct
io
n
(setp
oi
nt)
without
over
sp
ee
ding
i
n
the
sta
ti
on
a
r
y
ca
se
w
it
h
a
lo
w
-
rise
ti
me
an
d
ze
ro
sta
ti
c
error.
The
e
voluti
on
of
t
he
fou
r
e
le
ct
ro
ma
gn
et
ic
tor
que
pro
pulsi
on
en
gin
e
s
(
IM)
of
the
4WD
el
ect
ric
veh
ic
le
is
give
n
in
Fig
ur
e
13
(a)
an
d
Fi
gure
13
(
b)
,
us
in
g
bo
t
h
co
nv
e
ntio
nal
DT
C
and
DT
FC
con
t
rol
strat
egies.
T
he
res
ults
obta
in
ed
il
lustrate
quit
e
cl
early
good
t
orq
ue
res
pons
e
dy
namic
outp
ut
of
the
propose
d
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N: 20
88
-
8
694
Impr
oved D
TC strategy
o
f
an
el
ect
ric
vehicl
e wi
th fo
ur
in
-
w
heels in
duct
ion m
oto
r
… (
N
air No
ur
ia
)
659
DTF
C
c
on
tr
ol.
In
a
dd
it
io
n,
a
sign
ific
a
nt
re
du
ct
io
n
in
t
orq
ue
rip
ple
can
be
seen
,
co
mpa
red
to
c
onve
nt
ion
al
direct to
rque
c
on
t
ro
l
(DTC
).
Table
7.
Value
s of th
e
fo
ur
-
w
heels s
peed in
ph
a
ses
2
to
4
W
h
eel sp
eed
(Km
/h
)
Ph
ase 0
2
Ph
ase 0
4
CDTC
DTFC
CDTC
DTFC
Fron
t lef
t
wh
eel
6
1
,12
6
1
,10
5
3
,54
5
3
,49
Fron
t r
ig
h
t wheel
5
5
,31
5
5
,25
5
1
,85
5
1
,80
Rear le
ft
wh
eel
4
5
,35
4
5
,15
4
8
,35
4
8
,33
Rear r
ig
h
t whee
l
3
9
,52
3
9
,43
4
7
,21
4
7
,19
Table
8.
Per
for
mances
of the
DTC a
nd DTF
C
in the spee
d re
sp
onse
Co
n
trol
Typ
e
Ris
in
g
T
im
e
[Sec]
Ov
er
sh
o
o
t
[%]
Sp
eed Er
ror
[%]
DTC
0
.16
0
.15
0
.01
DTFC
0
.11
0
0
(a)
(b)
Figure
12
. Ve
hi
cl
e V
ariance
of linear
s
peed (
a) a
nd
er
ror rat
e (
b)
at
va
rio
us p
oin
ts
(a)
(b)
Figure
13.
Ele
c
trom
a
gnet
ic
torqu
e
r
es
ponse
de
velo
ped by t
he
four m
otors
usi
ng (
a
)
c
onve
ntion
al
DTC,
(b) DTFC
Figure
14
(a
)
and
Fi
gure
14
(
b)
s
hows
t
he
traj
ect
ory
of
the
sta
to
r
fl
ux
vect
or
in
t
he
plane
(
α
-
β)
relat
ed
to
the
s
ta
tor
f
or
t
he
ri
gh
t
fron
t
(LF)
mo
to
r.
It
can
be
seen
that
the
traj
ect
ory
of
the
en
d
of
t
he
sta
tor
flu
x
in
the
case
of D
TFC
c
on
t
ro
l (Fi
gure 14
(b))
ta
kes
a uni
form
ci
rc
ular
s
hap
e
w
it
h
a
ra
diu
s
eq
ual
t
o
0.97W
b
centere
d
at
the
or
i
gin
w
hich p
resen
ts
a
good
decou
pling o
f
t
he flu
x
f
r
om
t
he
torq
ue.
Evaluation Warning : The document was created with Spire.PDF for Python.