World Electric Vehicle Journal Vol.

6 - ISSN 2032-6653 - © 2013 WEVA Page Page 0095

EVS27
Barcelona, Spain, November 17-20, 2013

Advanced VDC simulations of In-wheel electric vehicle

using Carsim and Simulink
Hyunuk Ha1, Jongmoo Kim1, Shiuk Chung1, Jangmyung Lee2
1
Electric motor research center, KERI(Korea Electrotechnology Research Institute), sungjudong, sungsan-gu,
Changwon-si, Gyeongsangnam-do, Korea, Republic of,
[email protected], [email protected], [email protected]
2
Dept. Of Electronic engineering, Pusan National University, jangjeon-dong, geumjung-gu,
Pusan, Korea, Republic of,
[email protected]

Abstract
Conventional engine based vehicles inevitably have complicated structures due to lots of elements. Because
of this characteristic, research, development, and also marketing are mainly conducted by the conglomerate,
like GM, BMW, Honda, and Hyundai motors, etc. But environmental pollution and fuel exhaustion by the
stated vehicle increase the necessity of EV(Electric Vehicle) as a representative of green car. First of all,
the structure of EV is relatively simple, and energy transmission ratio of electric motor is more efficient
than the engine based power train system. In addition to these, inwheel EV can estimate exact wheel torque,
which is the most fundamental information for slip control, such as VDC(Vehicle dynamics
control)/TCS(Traction control system)/ABS(Anti-lock brake system). Various kinds of expectable
situations during EV’s navigation have been simulated through the coordination between ‘Carsim’ and
‘Simulink’.

Keywords: In-wheel EV, Direct torque control, VDC(Vehicle dynamics control), slip control

this structure guarantees high stability, smooth

1 Introduction cornering and ABS-like system through optimal
Recently, most of car-manufacturers have torque distribution to each wheel. But it’s vulnerable
concentrated on EV by the social needs to the green for external disturbances and impulsive shock which
vehicles. First of all, EV can significantly decrease contains high frequency, and it requires precise
fuel consumption which causes serious speed control for straight driving. In this paper,
environmental problems. In addition to this, EV is
various kinds of expectable situations during the EV’s
expected to realize demand oriented production
system because of its relatively simple structure. navigation have been simulated through the
Direct torque transmission structured EVs consist collaboration between ‘ Carsim ’ and ‘ Simulink ’ .
of in-wheel type and in-axis type. Both kinds of
Carsim’s fundamental features are highly focused at
structures have their own pros and cons, such as in-
wheel EV could be controlled wheel by wheel, so conventional power-train system, engine torque
converter-transmission-differential gear-wheel. In

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 1
World Electric Vehicle Journal Vol. 6 - ISSN 2032-6653 - © 2013 WEVA Page Page 0096

order to simulate in-wheel EV, make practical vehicle. Slip control of in-wheel EV is mainly
motor modeling in Simulink, and insert it into the conducted by Hori lab, the university of Tokyo[5].
Hori lab realized TCS by applying MFC(Model
in-wheel EV’s power train instead of engine in the
Following Control) in 2001[6], and recently
conventional power train system. Detailed proposed MTTE(Maximum transmissible torque
simulation environment is shown at figure 2 in estimation)[7] to enhance robustness as well as
section 2. This paper consists of four sections, reliability.
including introduction. Section 2 describe proposed
VDC algorithm of in-wheel EV and important 2.2 Proposed VDC algorithm for in-
simulation environments, section 3 illustrates wheel EV
simulation results for most representative four cases, Figure 2. illustrates overall block diagram of
section 4 finally concludes and presents some proposed VDC algorithm.
future works.

2 VDC algorithm of in-wheel EV

2.1 Conventional In-wheel EV
researches

In-wheel EV is mainly researched by the

traditional motor manufacturer, especially
Protean[1], Michelin[2], AISIN[3], NTN[4], etc.
Figure 2 : Proposed VDC algorithm for in-wheel EV

Torque controller calculates wheel torque based on

error of slip ratio, yaw rate, and driver ’ s
accelerator pedal information. Vehicle model is
replaced by Carsim block, slip ratio calculator
attains forward wheel speed as estimator for
vehicle speed, and rear wheel speed. The ratio
between vehicle speed and wheel speed is defined
as vehicle ’ s slip ratio, and difference between
desired slip ratio and estimated vehicle’s slip ratio
is controller input along vehicle’s longitudinal axis.
On the other hand, the between desired yaw rate
and vehicle’s yaw rate is another controller input
along lateral axis. Detailed derivation of desired
slip ratio and yaw rate is described in section 2.2.1
and 2.2.2, respectively.
Figure 1 : Newly developed KERI’s in-wheel motor
2.2.1. Desired slip ratio
Fixed value, generally between 0.15~0.2, is widely
The most significant technologies for the in- utilized in the former researches[8]. Figure 3.
wheel EV is high torque density and robustness illustrates relationship between friction coefficient,
against external disturbances, and precise mu, and vehicle’s slip ratio, lambda by Pacejka’s
speed/torque control ability. Traditionally, the magic formula[9]. Most of the cases, except for
ratio between sprung mass and un-sprung mass cobblestone environment, road ’ s adhesion
affects vehicle’s stability and driving comfort, so coefficient is maximized when vehicle’s slip ratio
heavy weighted wheels diminish these two is around 0.15. So, conventional researches set
important vehicle characteristics. Robustness
desired slip ratio as fixed value, 0.15. But vehicle’s
against external disturbances is directly related to
maintenance problems and the life cycle of adhesion coefficient is varying by the vehicle ’ s

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 2
World Electric Vehicle Journal Vol. 6 - ISSN 2032-6653 - © 2013 WEVA Page Page 0097

rotation elements. In this paper, desired varying

slip ratio is determined as, Vy
Vx tan(δ f ) cos(tan −1 ( ))
Vx (4)
ψ ψ ψ d =
λd = λdefault − 0.05 = 0.15 − 0.05 ⋅ (1) l f + lr
ψ max
 ψ max

2.2.3. Wheel torque calculation algorithm
In this paper, proposed wheel torque is calculated
with the driver’s accelerator pedal information, slip
ratio error, and yaw rate error. Its detailed
numerical formula is defined as,

eλ eψ
τ=
rl T ( P) ⋅ {1 − (1 − k ) }{1 − k } (5)
eλ max eψ max
eλ eψ
τ rr= T ( P) ⋅ {1 − (1 − k ) }{1 + k } (6)
eλ max eψ max

Figure 3 : Tire’s mu-lambda curve Where T(P) is driver’s accelerator pedal, eλ is slip
e
ratio error, ψ is yaw rate error, k is vehicle ’ s
In default slip ratio is set as 0.15 as usual, and rotational factor defined as,
proposed desired slip ratio is varying according
to the vehicle’s normalized yaw rate as rotational
ψ
information. Maximum yaw rate is tentatively k= (7)
selected as 100 rad/s by heuristic method, which ψ max

is maximum yaw rate during slalom driving.

2.2.2. Desired yaw rate 3 Simulation

By the vehicle’s lateral dynamics in [10], yaw rate
is stated as, 3.1 Simulation environment

Vx {tan(δ f ) − tan(δ r )}cos( β ) As mentioned in introduction, carsim ’ s default

ψ = (2)
l f + lr interface is established on engine based powertrain
system. So we have to make a suitable motor model
Where is Vx is vehicle speed, δ f and δ r are front to simulate inwheel EV as fig. 4. shows carsim’ s
and rear wheel’s steering angle, β is vehicle slip, internal setting so as tso in-wheel EV simulation and
its fundamental simulation scheme. T-N curve(Rpm-
and finally l f and lr are distance between front torque curve) is designed using user defined function,
wheel to COM and rear wheel to COM, which is torque command input of motor to Carsim.
respectively. Also from [10], vehicle’s slip angle is Other parameters, like tire radius/ friction constant/
overall length/ weight/ slanted angle of road is set in
derived as,
the Carsim as shown in table 1.

lr tan(δ f ) Table 1. Some important internal Carsim parameters

β = tan −1[ ] (3)
lr + l f Category Value Unit
Front wheel-COM 1103 mm
Front wheel-Rear wheel 2347 mm
In order to derive vehicle’s desired yaw rate, (3) is Vehicle mass 747 kg
Slanted angle 0 deg
substituted into (2). So we can attain vehicle
Tire radius 292 mm
desired yaw rate as in (4), which is a function of Road friction coefficient 0.85 -
steering angle, vehicle’s longitudinal and lateral
speed when vehicle’s COM is fixed.

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 3
World Electric Vehicle Journal Vol. 6 - ISSN 2032-6653 - © 2013 WEVA Page Page 0098

Figure 7 : Slip ratio and slip angle of without VDC/with

VDC
Figure 4 : Internal setting of Carsim for in-wheel EV and
fundamental simulation scheme

We propose five most representative environments

to simulate in-wheel EV, which is split mu braking
test(ISO14512), double lane change(ISO3888),
slalom(S-turn), Circle, and J-turn. Our proposed
VDC algorithm has been demonstrated under stated
five environments by comparing slip ratio, lambda,
Figure 8 : Tire ’ s longitudinal/lateral force of without
slip angle, beta, tire’s longitudinal and lateral force, VDC/with VDC
and yaw rate, etc.

3.2 Simulation result

This section describes simulation results for stated

environment. Each figure has its own short caption
for most distinctive features.

3.2.1. Double lane change(ISO3888)

Figure 9 : Wheel torque, vehicle speed, and yaw rate for
slalom simulation

4 Conclusions
Most EV experts have consentaneously announced
that the most significant bottleneck of EV is battery
problem. Its low energy density, short driving
distance per full charging state, insufficient life cycle,
and heavy weight are slowing EV down not to be
Figure 5 : Slip ratio and slip angle of without VDC/with next generation ’ s transportation. But EV is an
VDC for double lane change
inevitable and huge wave heads for our future. As
emphasize on transportation ’ s conveniences, the
reliability and stability issues should be considered
concurrently. Moreover, in-wheel EV might be a
good and the only solution for specialized EV fields,
such as personal mobility system with distinctive
chassis. This paper suggests in-wheel EV ’ s
prospective forecasts, and stability problems. As a
solution for the suggested issues, efficient VDC
algorithm based on precise wheel torque/speed
Figure 6 : Tire’s longitudinal/lateral force of without
estimation has been proposed. Proposed algorithm
VDC/with VDC
has been demonstrated through some practical
simulations with Carsim and Simulink. As some
3.2.2. Slalom(S-turn)

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 4
World Electric Vehicle Journal Vol. 6 - ISSN 2032-6653 - © 2013 WEVA Page Page 0099

expected future works, 1. Optimal torque Control: Mechanical Engineering Series,

distribution to minimize side-slip angle, 2. ISSN 0941-5122, 3rd edition, 2011.
Substitute motor modeling to real motor to realize
HILS(Hardware In the Loop Simulation) system. Authors
Hyunuk Ha received the B.Sc. and
M.Sc degrees in Electrical
Acknowledgments Engineering from Pusan National
This research was supported by the MOTIE(The University, Busan, Korea, in 2008,
Ministry of Trade, Industry and Energy), Korea, and 2010, respectively, where he
under the Human Resources Development Program is currently pursuing Ph.D degree.
for Special Environment Navigation/Localization His current research interests are
National Robotics Research Center support in-wheel EV and mobile inverted
program supervised by the NIPA(National IT pendulum.
Industry Promotion Agency)." (H1502-13-1001)
Jongmoo Kim born in Korea in 1970, He
received the B.S., M.S., and Ph.D
degrees in electrical engineering from
References Yeungnam University, Kyungbuk,
Korea, in 1992, 1993, and 2009,
[1] http://www.proteanelectric.com/ respectively. Since 1994, he has worked
[2] Espanet, C., Tekin, M.; Bernard, R.; with the Korea Electrotechnology
Research Institute(KERI). His research
Miraoui, A.; Kauffmann, J.-M.: A new interests are high performance motion
structure of an high torque in-wheel motor,: and motor control, control theory, DSP
Electrical Machines and Systems, 2003. application, contactless power supply
ICEMS 2003. Sixth International and electric vehicle.
Conference on, vol. 1, pp. 158-162, 2003.
Shiuk Chung received the B.S, M.S. and
[3] Ryoji Mizutani, Nishikamo-gun, Atsushi Ph.D degrees in mechanical engineering
Torii, Shigetaka Isogai: In-wheel motor from Pusan National University, Busan,
capable of efficiently cooling electric motor S.Korea in 1997, 1999, and 2010,
and reduction gear: United states patents, respectively. From 2002 to 2005, he was
with Samick THK as a researcher. Since
1st ,Apr., 2008. 2005, He has been with Electric Motor
[4] Minoru Suzuki, Koichi Okada, Kayo Sakai, Research Center, Korea
Yusuke Makino: Development of an In- Electrotechnology Research Institute,
Wheel Motor Axle Unit: NTN technical Changwon, S.Korea, as a Seniotr
Researcher. His research interests
review, no, 75, 2007. include the design and analysis of linear
[5] http://hflab.k.u-tokyo.ac.jp/ and rotary direct drive electric machines.
[6] Dejun Yin, Yoichi Hori: A Novel Traction
Control for Electric Vehicle without Chassis
Velocity: Jangmyung Lee received the B.S and
M.S degrees in electronic engineering
[7] Jia Sheng, Dejun Yin, Yoichi Hori, Feng from Seoul National University, Seoul,
rung hu: Electric vehicle traction control: Korea, in 1980 and 1982, respectively,
IEEE trans. On Industry applications, vol. and the Ph. D degree in computer
18, no. 2, pp. 23-31. 2012. engineering from the University of
Southern California (USC), Los Angeles,
[8] Rongrong Wang, Junmin Wang: Tire–road in 1990. Since 1992, he has been a
Professor at Pusan National University,
friction coefficient and tire cornering Busan, Korea. His research interests
stiffness estimation based on longitudinal include intelligent robotics, advanced
tire force difference generation: Control control algorithm, and specialized
engineering practice, vol. 21, no. 1, pp. 65- environment navigation/localization.
75. 2013.
[9] Mohammad Safwan Burhaumudin,
Pakharuddin Mohd Samin, Hishamuddin
Jamaluddin, Roslan Abd Rahman, Syabillah
Sulaiman: Modeling and Validation of
Magic Formula Tire Model: International
Conference on Automotive, Mechanical and
Materials Engineering (ICAMME'2012)
Penang (Malaysia), May 19-20, 2012.
[10] Rajesh Rajamani: Vehicle Dynamics And

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 5