IET Electrical Systems in Transportation

Special Issue: Selected Papers from the 2017 and 2018 IEEE Vehicle
Power and Propulsion Conference

Brake force distributions optimised with ISSN 2042-9738

Received on 28th February 2019
Revised 4th May 2019
regard to energy recovery for electric vehicles Accepted on 18th June 2019
E-First on 20th August 2019
with single front-wheel drive or rear-wheel doi: 10.1049/iet-est.2019.0039
www.ietdl.org

drive
Philipp Spichartz1 , Constantinos Sourkounis1
1Ruhr University Bochum, Institute for Power Systems Technology and Power Mechatronics, Universitaetsstrasse 150, 44801 Bochum,

Germany
E-mail: [email protected]

Abstract: Energy recovery by means of braking with the electric machine helps to extend the cruising range of electric vehicles.
By optimising the recuperation system, the total energy efficiency is increased. In general, using a usual brake force distribution
the capability for recuperation is not equal for every type of drive train. Usually, four-wheel drives are advantageous in
comparison to drive trains with only one driven axle. Regarding the single drives, generally, a front-wheel drive provides a higher
capability for recuperation than a rear-wheel drive. This is mainly due to the dynamic axle load shift, which occurs during
decelerations. In order to increase the energy recovery of electric vehicles with single drive, an adaptive brake force distribution
is presented and compared with other brake force distributions in this study. It uses the knowledge of the currently available
adhesion coefficient between tyre and road. The positive effect on the recuperation without impairing the braking performance
on a straight road is demonstrated based on simulations. In curves, the lateral forces additionally need to be considered to
guarantee the stability of the vehicle, especially relating to electric vehicles having a rear-wheel drive.

1 Introduction that these criteria have a higher significance than the recuperation
[1, 5].
Today, the bigger part of the electric vehicles (EVs) still has a drive Electric braking and its influence on the vehicle stability have
train topology with only one electric machine, which drives the often been examined without respecting the recuperated energy.
wheels either of the front axle or of the rear axle via a differential Electric machines offer a much faster torque response than a
gear (see Fig. 1). In [2–4], measurement results of EVs with such a hydraulic braking system [6]. Consequentially, a lot of publications
drive train have been presented, which have shown that during have focused on the use of the electric drive train for stability and
decelerations a remarkable part of the kinetic energy is not brake control systems. A non-linear model predictive slip control,
recuperated but converted into heat by the mechanical friction which operates as traction control in case of acceleration and as
brakes. A central cause for these results is the brake force anti-lock braking system (ABS) in case of deceleration, has been
distribution (BFD) between the two axles. Only the brake force at presented in [7]. In [8], the cooperation of electric and mechanical
the drive axle can be used to recover energy. The brake force at the braking has been used to optimise the ABS of an EV with four-
second axle is solely generated by mechanical brakes. In principle, wheel drive. In [9], an ABS has been developed, which uses the
a modification of the BFD aiming for better recuperation results is generated voltages of the brushless in-wheel machines to determine
possible. However, such a modification can influence the braking the rotational speed of the wheels. An electronic stability control
performance and the stability of the vehicle. It has to be ensured using the four in-wheel machines of an all-wheel drive has been
analysed in [10]. In [11], the effect of a novel ABS approach for a
four-wheel drive on the recuperated energy has been examined.
However, there has been a concentration on high values of brake
accelerations, while this paper also takes small and middle
accelerations into account.
In many publications, there has been a concentration on the
cooperation between the regenerative and the mechanical brakes.
Usually, the BFD has not been regarded or it has been assumed to
be constant. Often, a fuzzy logic control has been applied for the
cooperation, like in [12–14] for an EV with front-wheel drive and
in [15] for a four-wheel drive. In [16], the cooperation of a
centralised friction brake and the electric machines at the four
wheels has based on a model predictive control. In [17, 18],
implementations of cooperative braking systems have been
presented. In this paper, it is assumed that such a cooperative
system with fast and precise blending of the electric and
mechanical brake torques is implemented.
In [19], an optimisation algorithm considering variable BFDs
aiming for a short braking distance has been presented. The
recuperation has not been taken into account in that publication. In
Fig. 1 Typical drive train topologies with one central electric machine [20, 21], several BFDs have been analysed with regard to the best
(based on [1]) recuperation result for an EV with front-wheel drive. These BFDs
(a) Front-wheel drive, (b) Rear-wheel drive have based on a limitation, which is set in an agreement of the UN

IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195 186
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
m is the vehicle mass, g is the gravitational acceleration, h is the
vertical position of the centre of gravity and l is the wheelbase. lr
and l f are the distances between the centre of gravity and the rear
and the front axle, respectively. z is called brake intensity in this
paper. It describes the negative acceleration of the vehicle in
relation to the gravitational acceleration

a
z= − , a<0 (3)
g

The ideal BFD (see Fig. 2) considers the dynamic axle load shift
[12, 29]

FB, r lr 2 l FB, f FB, f lr

= + ⋅ − − (4)
FG 2⋅h h FG FG 2⋅h
Fig. 2 Ideal BFD (solid line), curves of constant brake intensity z (dashed
lines), f- and r-lines: limits concerning locking wheels for various adhesion FB, r and FB, f are the brake forces at the rear axle and the front axle,
coefficients μ (dotted lines) and ECE specification (dot-and-dashed line) respectively, and FG is the weight of the vehicle.
(based on [1]) Using the ideal BFD, the front axle and the rear axle do not lock
before each other. When wheels become locked, the ability of the
Economic Commission for Europe (ECE) [22]. A solution based tyres to perform cornering forces is very limited. Consequently,
on this limitation is also implemented in the present paper and steering is restricted or even not possible with locked front wheels.
serves as a reference for the novel solution. In [23–25], the BFD Locked rear wheels might lead to an instable behaviour of the
for EVs with four-wheel drives has been optimised in terms of vehicle, especially for brake applications in curves. In addition, the
maximum energy recovery. In several of the mentioned papers, stopping distance usually rises, since the maximum adhesion for
negative effects caused by locking wheels have not been regarded. longitudinal motion decreases when the wheels become locked.
Publications dealing with an optimised BFD for EVs with rear- Moreover, the electric machine cannot recover energy as long as
wheel drive are rare [1]. the driving wheels are locked.
Hancock and Assadian [26] has addressed the risk of degrading Lines with constant use of the adhesion, which are often called
the stability of an EV, when a series regenerative braking system f-lines and r-lines, are also shown in Fig. 2 for several values of the
generates brake force at the rear axle only. The influence of the adhesion coefficient between tyre and road μ [23, 29]
position of the centre of gravity on the recuperation and the
stability for EVs with front- and rear-wheel drives has been FB, r FB, f l − μ ⋅ h lr
presented in [27]. In [28], EVs with four-wheel drive and front- = ⋅ − (5)
FG FG μ⋅h h
wheel drive have been compared with regard to recuperation and
the likelihood of locking wheels, but only for a single deceleration. FB, r FB, f μ⋅h μ ⋅ lf
Moreover, in the two latter publications only the so-called ideal = − ⋅ + (6)
FG FG l + μ ⋅ h l + μ ⋅ h
BFD has been considered.
In the present paper, an adaptive BFD is presented and
When μ equals the effective maximum adhesion coefficient μmax,
discussed, which uses the estimated adhesion coefficient of tyre
and road. It shall help to maximise the recuperation of EVs with the f- and r-lines represent the limits for locking front wheels and
front- or rear-wheel drive considering also stability issues [1]. In rear wheels, respectively. In Fig. 2, the limits are exemplarily
Section 2, a summary of the necessary theoretical foundations with highlighted for μmax = 0.6. When the brake forces only vary within
regard to the BFD is given before the idea of the adaptive solution the grey coloured area, there is no risk of locked wheels [1].
is explained in Section 3. A simulation model for the examination The UN ECE prescribes in [22] a minimum value for the
is presented in Section 4. In Section 5, the performance of the braking force at the rear axle. The objective of that regulation is
adaptive BFD is analysed based on simulation results. The that locking of the front wheels does not occur at low values of
recuperation results of this solution are compared with the results brake intensity [20]
of two other well-known BFDs. Moreover, the influence on the
stability of the vehicle is examined by simulating brake z ≥ 0, 1 + 0, 85 ⋅ (μmax − 0, 2) (7)
applications in curves. Requirements for an experimental
validation are discussed in Section 6. A final conclusion is given in FB, f z + 0, 07 lr + z ⋅ h
= ⋅ (8)
Section 7. FG 0, 85 l

2 Theoretical foundations FB, r FB, f

=z− (9)
FG FG
A primary objective with regard to safe decelerations is to
guarantee stability of the vehicle and a satisfying stopping distance.
One basic procedure is to generate one part of the total brake force This specification is illustrated in Fig. 2, too. A general rule is, to
FB at the front wheels and the other part at the rear wheels. install a BFD, which runs between the ideal BFD and the ECE
specification.
Principles with regard to a convenient BFD are described in this
The previous explanations were focussed on the longitudinal
section.
movement. A dynamic load shift also occurs between the left and
Generally, a deceleration leads to a dynamic axle load shift. The
the right wheels, when lateral forces act on the vehicle.
effective normal force at the front axle FZ , f increases while the
Implementing a stability control this should be respected.
normal force at the rear axle FZ , r decreases [19, 20]

lr h
3 Approaches for an increase of energy recovery
FZ , f = m ⋅ g ⋅ +z⋅ (1)
l l For EVs with only one drive axle, the ideal BFD is not productive
in terms of the recuperation. The effect is worse for EVs with rear-
lf h wheel drive, as the bigger part of the brake force is distributed to
FZ , r = m ⋅ g ⋅ −z⋅ (2) the front wheels and generated by mechanical brakes. Generally,
l l

IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195 187
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 3 BFD according to the ECE specification (based on [1])
Fig. 5 Adaptive BFD for front-wheel drives in consideration of the
characteristic curve of the electric machine; here illustrated for an effective
maximum adhesion coefficient μmax = 0.6 (based on [1])

Fig. 4 Adaptive BFD for front-wheel drives; here illustrated for an

effective maximum adhesion coefficient μmax = 0.6 (based on [1])

the share of the effective brake force at the drive axle needs to be Fig. 6 Adaptive BFD for rear-wheel drives; here illustrated for an
increased to raise the capability for energy recovery. effective maximum adhesion coefficient μmax = 0.6 (based on [1])
A solution for front-wheel drives is examined in [30]. The BFD
is geared to the ECE specification and only for emergency brakings Partly, the adaptive BFD does not comply with the ECE
a transition to the ideal BFD is stipulated via the f-line of μ = 0.8 regulations. Therefore, in addition to the influence on the
(see Fig. 3). Problems on roads with low adhesion are expectable, recuperated energy, the impact on stability and safety is examined
which is analysed in Section 5. with the following simulations.
An adaptive BFD is a novel approach, which depends on a
continuous estimation of the current adhesion coefficient. Solutions 4 Simulation model
for that estimation are presented in [23, 31–33]. The principle of
the adaptive BFD is illustrated for a front-wheel drive and an A model of an EV is implemented for the simulation-based
estimated adhesion coefficient μest = 0.6 in Fig. 4. examinations. It is possible to choose a front-wheel drive, a rear-
As long as the f-line of μ = μest − 0.1 is not reached (sector A to wheel drive as well as a four-wheel drive, which serves as a
reference. The four-wheel drive has a maximum power of 42.5 kW
B), the total desired brake force is distributed to the front axle.
at each axle while the single drives have a maximum power of 85
With increasing brake intensity, the BFD approaches to the curve
kW at the drive axle. A second difference is the position of the
of the ideal BFD via a quadratic function (sector B to C). The
centre of gravity. The essential parameters of the model are
curves meet in point C, which is the intersection of the f-line and
summarised in Table 1.
the r-line belonging to the estimated adhesion coefficient and
A two-track model is the basis for the modelling of the vehicle
which is also the point of the highest possible brake intensity.
dynamics (see Fig. 7). Longitudinal as well as lateral movements
Using the adaptive solution, the wheels of the front and the rear
can be simulated. Motion of the vehicle body in vertical direction
axle should not lock before each other. The transition to the ideal
is neglected. The longitudinal motion and the lateral motion of the
BFD is not performed via the f-line of μest but via the quadratic
vehicle are described by the following equations [34]:
function to ensure robustness. Otherwise, a slight error of the
estimated adhesion coefficient could cause an unnecessary locking
m ⋅ (v̇X − ψ̇ ⋅ vY ) = FX , W f l + FX , W f r ⋅ cos δ
of the front wheels [1].
Furthermore, the solution respects the characteristics of the − FY , W f l + FY , W f r ⋅ sin δ (10)
electric drive. When the torque or power limits of the machine are
exceeded, the missing brake force is performed at the rear wheels +FX , Wrl + FX , Wrr − F A
at first (see Fig. 5). Thereby, the BFD gets nearer to the ideal curve
but does not lead to a loss of recovered energy [1]. m ⋅ (v̇Y + ψ̇ ⋅ vX ) = FX , W f l + FX , W f r ⋅ sin δ
The procedure can also be applied to rear-wheel drives to
optimise the energy recovery. Here, more brake force is distributed + FY , W f l + FY , W f r ⋅ cos δ (11)
to the rear wheels if possible (see Fig. 6). In doing so, a general +FY , Wrl + FY , Wrr
rule with regard to stability needs to be considered, which says that
the rear wheels should not become locked first [1].
vX and vY are the longitudinal and the lateral velocity of the vehicle
and ψ̇ is the yaw rate. FX , Wi j and FY , Wi j are the longitudinal and the

188 IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1 Main parameters of the modelled vehicle
Parameter Value
max. power of the electric machine(s) Pmax 85 kW
max. torque of the electric machine(s) T max 270 Nm
gear ratio iG 9.8
vehicle mass m 1510 kg
drag plane (frontal area) A 2.189 m2
drag coefficient cw 0.29
vertical position of the centre of gravity h 0.52 m
vertical position of the roll axis hRA 0.33 m
vertical position of the pitch axis hPA 0.38 m
wheel base l 2.631 m
distance between front axle and centre of gravity
lf
front-wheel drive and four-wheel drive 1.1 m
rear-wheel drive 1.25 m
front and rear wheel track b f , br 1.543 m, 1.512 m
moment of inertia around the roll axis J X 482.7 kg m2
moment of inertia around the pitch axis JY 2394.3 kg m2
moment of inertia around the yaw axis JZ 2585.6 kg m2
Fig. 7 Two-track model (based on [34])
(a) Top view, (b) Side view tyre radius rR 0.316 m
moment of inertia of a wheel JR 1.0 kg m2
lateral forces at the four wheels (index ij – fl: front left, fr: front
right, rl: rear left, rr: rear right) in the relative wheel coordinate
system. δ is the steer angle and F A is the air resistance.
which starts with an initial velocity v0 and brakes with a constant
The yawing of the vehicle is a crucial factor for the examination brake intensity z until it stops. The two values vary in the ranges
of the stability [34] (see (12)) . JZ , J X and JY are the moments of 0.05 ≤ z ≤ 0.8 and 10 km/h ≤ v0≤100 km/h. Only decelerations
inertia around the yaw axis, roll axis and pitch axis, respectively. ϕ̇ conducted while moving straight ahead (steer angle δ = 0) are
is the roll rate and θ̇ is the pitch rate. Pitch and rolling motion of examined. A road with wet surface and μmax = 0.6 is regarded. On
the vehicle body, which influence the normal forces at the wheels, a dry road with higher adhesion coefficient, the different effects of
are modelled based on [34], too. the adaptive BFD and the ECE-based BFD could not be
A modified Dugoff model [35], which allows the examination highlighted that well. An ABS function is not implemented in the
of situations with combined longitudinal and lateral slip, is chosen model [1].
for the tyre model. It describes the longitudinal and lateral forces at
each wheel as functions of the current maximum adhesion
5.1.1 Four-wheel drive: First of all, the recuperated energy is
coefficient μmax, the longitudinal wheel slip λ, the tyre slip angle α
examined for an EV with four-wheel drive using the ideal BFD to
and the wheel load FZ . Definitions and equations for the three latter have a reference (see Fig. 8). Generally, the recuperated energy
quantities can be found in literature, e.g. [34, 36, 37]. increases with a rising initial velocity, due to the higher kinetic
The desired value of brake intensity and the initial velocity are energy of the vehicle at the beginning of the deceleration.
the input parameters of the simulation model. An implemented However, there is not a constant proportionality between the
torque blending system for all wheels is assumed. Electric braking recuperated energy and the initial velocity, because of load-
is preferred, i.e. the mechanical friction brakes are not activated dependent efficiencies and limitations in the drive train [38].
before the required braking torque or power at the particular drive At first, the recuperation result also increases with rising brake
wheel reaches the maximum available torque or power delivered intensity, which is mainly due to the efficiency of the electric
by the electric drive. The recuperated energy is measured at the machines, which tends to increase for a rising load. In addition, the
electrical terminals of the electric machine. Therefore, a battery duration of the brake applications and thus also the term, in which
model and the consideration of power consumption caused by the driving resistance is effective, shorten. For the maximum initial
auxiliary units are not necessary. The efficiency of the electric velocity v0, max = 100 km/h the maximum value of energy is
machine(s) is dependent on the effective working point. This is recuperated with the brake intensity z = 0.21. For a further increase
modelled by means of an efficiency map [1]. of the brake intensity the recuperated energy declines, since the
maximum power of the machines is partially reached and the
5 Simulation results mechanical brakes are applied. Due to the chosen BFD and the
5.1 Recuperated energy equal dimensioning of the machines at the front axle and the rear
axle, the limits of the front machine are exceeded for smaller brake
The performance of the three presented BFDs (ideal, ECE based, intensities than the limits of the rear machine [38]. There is also a
adaptive) is examined regarding the energy recovery of an EV, small impact on the recuperated energy due to the increasing slip.

JZ ⋅ ψ̈ − (J X − JY ) ⋅ ϕ̇ ⋅ θ̇ = (FX , W f l ⋅ sin δ + FX , W f r ⋅ sin δ

+FY , W f l ⋅ cos δ + FY , W f r ⋅ cos δ) ⋅ l f
+(FX , W f r ⋅ cos δ − FX , W f l ⋅ cos δ
(12)
bf
+FY , W f l ⋅ sin δ − FY , W f r ⋅ sin δ) ⋅
2
br
+ FX , Wrr − FX , Wrl ⋅ − FY , Wrl + FY , Wrr ⋅ lr
2
IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195 189
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 8 Energy recuperation of a four-wheel drive using the ideal BFD for Fig. 10 Energy recuperation of a front-wheel drive using the ECE-based
several decelerations with varying initial velocities and values of brake BFD for several decelerations with varying initial velocities and values of
intensity (μmax = 0.6) brake intensity (μmax = 0.6) (based on [1])

Fig. 9 Energy recuperation of a front-wheel drive using the ideal BFD for Fig. 11 Energy recuperation of a front-wheel drive using the adaptive
several decelerations with varying initial velocities and values of brake BFD for several decelerations with varying initial velocities and values of
intensity (μmax = 0.6) (based on [1]) brake intensity (μmax = 0.6) (based on [1])

For lower initial velocities, the limits of the machines and, the ideal BFD. For z > 0.44, using the ECE-based BFD, the
therefore, the maximum recuperation results are reached at higher adhesion at the front wheels is not high enough to transmit the
brake intensities. requested force, so that the wheels become locked. As a result, it is
At z = 0.6 the maximum adhesion is effective. Higher desired not possible to recover energy. Using the ideal BFD or the adaptive
brake intensities lead to locked wheels, so that recuperation is not BFD, all wheels get locked not before z > 0.6 = μmax and
possible. recuperation is possible up to that value of brake intensity [1].
The reduction of recuperation is not the only consequence
5.1.2 Front-wheel drive: Applying the ideal BFD in an EV with caused by the earlier locking of the wheels with the ECE-based
front-wheel drive, the recuperated energy is lower in nearly all BFD. It also influences the stopping distance and thus safety. The
cases compared to the recuperated energy with the four-wheel stopping distances and the longitudinal slip of one drive wheel,
drive (see Fig. 9). The mechanical brakes at the rear axle contribute achieved with the three BFDs, are compared in Fig. 12. The initial
brake force in every brake situation and transform a part of the velocity is v0 = 50 km/h and the brake intensity is varied. Using the
kinetic energy in thermal energy. Only for high brake intensities, adaptive BFD, the slip is higher than with the ideal BFD. That
when the power limits of the machines are reached in both cases, lowers the ability to realise lateral forces at the tyre. In the regarded
nearly the same results are achieved. Generally, the maximum case, in which a vehicle brakes while driving straight-ahead, that
recuperation energy is reached for higher brake intensities, as the does not directly cause a disadvantage. Both solutions reach the
whole drive power is installed at the front axle and higher loads are same stopping distances for all regarded cases. In contrast, once the
necessary to reach the limits. wheels lock, the stopping distance achieved with the ECE-based
Using the ECE-based BFD, the energy recovery is increased for BFD is up to 44% longer than with the two other BFDs [1].
low and medium values of brake intensity compared with the Moreover, there is no chance to steer in that situation. Generally, an
results obtained with the ideal BFD (see Fig. 10). Using the ABS would prevent the locking of the wheels. Usually, these
adaptive BFD, better results can be obtained additionally for high systems still work with the mechanical brake system and electric
values of brake intensity (see Fig. 11). braking is stopped in emergency situations, so that the reduced
For brake intensities up to z = 0.2, both BFDs distribute the recuperation result would persist.
total brake force to the front axle. Therefore, they achieve the same
results, which are almost equal to the results with the four-wheel 5.1.3 Rear-wheel drive: The examination is performed for EVs
drive. In the range 0.2 < z ≤ 0.38, the adaptive BFD reaches with rear-wheel drive using the ideal BFD or the adaptive BFD,
slightly higher values than the ECE-based BFD as the latter starts too. First of all, the recuperation performance achieved with the
to brake partly with the mechanical brakes at the rear axle. In the rear-wheel drive (see Figs. 13 and 14) does not come up to the
range 0.38 < z ≤ 0.44 it is the opposite way around, as the working performance achieved with the front-wheel drive (see Figs. 9 and
point of the adaptive BFD passes the described transition towards 11).

190 IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 14 Energy recuperation of a rear-wheel drive using the adaptive BFD
for several decelerations with varying initial velocities and values of brake
intensity (μmax = 0.6) (based on [1])

Fig. 12 Comparison of three BFDs for a front-wheel drive braking with

varying values of brake intensity (v0 = 50 km/h, μmax = 0.6)
(a) Stopping distance sB (based on [1]), (b) Longitudinal brake slip λB of the front left
tyre

Fig. 13 Energy recuperation of a rear-wheel drive using the ideal BFD for Fig. 15 Comparison of two BFDs for a rear-wheel drive braking with
several decelerations with varying initial velocities and values of brake varying values of brake intensity (v0 = 50 km/h, μmax = 0.6)
intensity (μmax = 0.6) (based on [1]) (a) Stopping distance sB (based on [1]), (b) Longitudinal brake slip λB of the front left
tyre
This applies to both BFDs. The main reason is the mentioned
dynamic axle load shift [1]. achieves the same results like the EV with front-wheel drive using
The ideal BFD distributes the bigger part to the front axle, the adaptive or the ECE-based BFD (see Fig. 14). At z = 0.22 the
where only mechanical brakes are installed. The power and torque r-line belonging to μ = μmax − 0.1 is already reached, so that the
limits of the electric machine are not exceeded for any brake transition towards the ideal BFD begins and the recuperated energy
application. Two effects nearly compensate each other, so that the strongly decreases for higher brake intensity values.
results for a constant initial velocity and a varying brake intensity Similar to the behaviour of the front-wheel drive, the
are quite equal (see Fig. 13). On the one hand, the efficiency of the longitudinal slip of the drive wheels is higher for the adaptive than
electric machines usually increases with raising load, which should for the ideal solution. However, the stopping distances are not
lead to better results. On the other hand, the relation between the different (see Fig. 15). Furthermore, the wheels of the rear axle do
brake forces at the rear and the front axle decreases with rising not lock first, which could lead to dangerous oversteer [1].
brake intensity. That restricts the potential of recuperation.
Using the adaptive BFD, a significant improvement of the
5.1.4 Comparison based on a driving cycle: Finally, the
recuperation is obtained, especially for low values of brake
performance of the examined drive train topologies and BFDs with
intensity. In comparison with the rear-wheel drive using the ideal
respect to the recuperation of energy during decelerations is tested
BFD, a gain of up to 170% is realised for some simulated
by passing the New European Driving Cycle (NEDC). The NEDC
decelerations [1]. For z < 0.22 the EV with rear-wheel drive

IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195 191
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 2 Recuperated energy for different drive train topologies and BFDs completing the NEDC (based on [38])
Drive train topology BFD Erec, Wh
four-wheel drive ideal 350.4
front-wheel drive ideal 201.6
front-wheel drive ECE 343.8
front-wheel drive adaptive 343.8
rear-wheel drive ideal 151.9
rear-wheel drive adaptive 343.7

One indicator for the stability is the vehicle side slip angle,
which is the angle between the longitudinal axis of the vehicle and
the velocity vector in the centre of gravity

vX
β = arctan (13)
vY

A side slip angle, that stays in the range β < 5° = 0.0873 rad, is a
sign of a stable behaviour [39].
Another indicator is the self-steering response of the vehicle in
curves. According to the definition of Olley, it can be identified by
means of the difference between the slip angle of the front tyres α f
and the slip angle of the rear tyres αr [29]

Δα = α f − αr (14)

A neutral behaviour is obtained for Δα = 0. A positive Δα

signalises a tending to understeer while a negative Δα signalises a
tending to oversteer. Generally, the latter behaviour is assessed to
be more dangerous, because it may lead to skidding and, usually,
drivers with average skills are not able to handle it [29].
The following manoeuvre is used for the examinations. On a
dry road (μmax = 1), the vehicle has a velocity v0 = 61 km/h and
neither the electric machine nor the mechanical brakes generate
any torque. It gets to a curve with the radius rC = 90 m. Therefore,
the steer angle increases up to δ = 1.67° = 0.0292 rad within 1 s, in
accordance with the approximation for small angles

l
δ ≃ tan δ ≃ , [δ] = rad (15)
rC

At the time t = 3 s, the vehicle is decelerated until it stands still.

The steer angle is constant.
At first, the EV with front-wheel drive using the ideal BFD is
analysed. It decelerates with the brake intensity z = 0.6. The slip
angles of the front tyres are higher than the slip angles of the rear
Fig. 16 Brake application (z ≃ 0.6) of an EV with front-wheel drive and tyres during the whole manoeuvre (see Fig. 16), which indicates
ideal BFD in a curve (radius rC = 90 m) on a dry road (μmax = 1.0) that the vehicle slightly tends to understeer in curves, generally.
The reason is that the chosen position of the centre of gravity
(a) Longitudinal velocity of the vehicle vX and steer angle δ, (b) Side slip angle of the
within the vehicle is quite near to the front. However, the vehicle
vehicle β, (c) Tyre slip angles α f l (front left), α f r (front right), αrl (rear left), αrr (rear
shows a stable behaviour. When the brake application starts,
right), (d) Longitudinal brake slip of the tyres λ f l, λ f r, λrl, λrr
dynamic alterations of the values do not occur and the side slip
angle is small. The side slip angle is dependent on the centripetal
consists of 18 brake applications and the maximum brake intensity force, so that the alteration with decreasing velocity is regular.
is zmax = 0.14. In the second case, the EV with front-wheel drive uses the
The results concerning the recuperated energy after one cycle adaptive BFD. The manoeuvre remains the same. Observing the
are presented in Table 2. Due to the quite low decelerations in the longitudinal slip, one can assert that nearly the entire brake force is
NEDC, the EVs with single drive using the adaptive or ECE-based generated at the front wheels (see Fig. 17). As already mentioned
BFDs nearly reach the result of the EV with four-wheel drive. above, the potential for conducting lateral forces at the front wheels
About 70% more energy is recuperated in comparison with the is lower for high values of longitudinal slip. As a consequence,
front-wheel drive using the ideal BFD and even 126% more energy when the vehicle decelerates, at first, the difference between the
than with the rear-wheel drive using the ideal BFD. front and rear slip angles increases. That means, the tending to
understeer is slightly higher than with the ideal BFD. There is also
5.2 Stability of the vehicle a small increase of the negative side slip angle in the first moment
of the deceleration. However, the vehicle is still stable and
The previous examinations were limited to longitudinal dynamics controllable and the stopping distance does not become longer.
and the recuperated energy. The following simulations focus on the Perhaps, negligible steering corrections might be necessary.
behaviour of an EV, which brakes while cornering. It is analysed, The comparison between the two BFDs is conducted for the EV
whether the adaptive BFD affects the stability of the vehicle with rear-wheel drive, too. In this case, the brake intensity is
adversely. z = 0.4. Irrespective of the used BFD, the vehicle with the chosen
parameters has a nearly neutral self-steering response. The front
192 IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 17 Brake application (z ≃ 0.6) of an EV with front-wheel drive and Fig. 18 Brake application (z ≃ 0.4) of an EV with rear-wheel drive and
adaptive BFD in a curve (radius rC = 90 m) on a dry road (μmax = 1.0) ideal BFD in a curve (radius rC = 90 m) on a dry road (μmax = 1.0)
(a) Longitudinal velocity of the vehicle vX and steer angle δ, (b) Side slip angle of the (a) Longitudinal velocity of the vehicle vX and steer angle δ, (b) Side slip angle of the
vehicle β, (c) Tyre slip angles α f l (front left), α f r (front right), αrl (rear left), αrr (rear vehicle β, (c) Tyre slip angles α f l (front left), α f r (front right), αrl (rear left), αrr (rear
right), (d) Longitudinal brake slip of the tyres λ f l, λ f r, λrl, λrr right), (d) Longitudinal brake slip of the tyres λ f l, λ f r, λrl, λrr

and rear slip angles are almost equal, when the vehicle is cornering between tyre and road Fres, W cannot become bigger than the
and no torque is delivered in the period 1 s<t<3 s (see Figs. 18 and product of the effective maximum adhesion coefficient and the
19). In comparison to the EV with front-wheel drive, the centre of normal force at the wheel [29, 36]
gravity is more centrally arranged, which leads to that self-steering
response. Using the ideal BFD this neutral and stable behaviour is Fres, W = FX2 , W + FY2 , W ≤ μmax ⋅ FZ , W (16)
still identifiable during the deceleration process (see Fig. 18).
Dynamic alterations of the side slip angle or of the tyre slip angles Using the resulting total force for the determination of the suitable
are not detectable. BFD, the described transition to the ideal BFD begins at a lower
For the examined case, the behaviour during the brake value of brake intensity, when a lateral force occurs. Consequently,
application is significantly different, when the adaptive BFD is the demanded longitudinal and lateral forces can be provided and
used (see Fig. 19). The curves of the longitudinal slip show that stability is ensured in the possible range.
most of the brake force is generated at the rear axle. A remarkable
increase of the rear slip angles occurs, which leads to oversteer of
the vehicle. The side slip angle reaches the region of critical values. 6 Requirements for an experimental validation
In the here regarded case, the vehicle does not skid. However, The validation of the adaptive BFD concept is quite complex and
many influences might lead to that dangerous event. Examples for demands several preconditions, which are discussed in this section.
these influences are a temporary change of the adhesion factor at The concept is based on several technical features that are not
one or more tyres or an (unmindful) increase of the steer angle. standard in most of the vehicles yet or still in the research stage.
Brake applications with higher values of brake intensity would not These include in particular a BFD which is adjustable during
necessarily lead to a more instable behaviour. The additional brake operation, the estimation of the maximum adhesion coefficient and
force, which is required for an increase of the brake intensity, is a cooperative braking system that allows torque blending, e.g. by
mainly generated at the front axle. Consequently, the adaptive BFD means of an electrohydraulic or electromechanical system for the
approximates the ideal BFD. friction brakes. Moreover, EVs with different drive train
The adaptive BFD can be modified to mitigate or prevent the topologies, equal dimensioning of the drive power and the
negative effects presented and to guarantee the stability of the mentioned features are necessary, when the results of the presented
vehicle when braking in curves. Not only the longitudinal forces at comparison shall be validated.
the drive wheels (for the desired deceleration) have to be In addition, a proving ground that allows for a reproducible,
considered but also the lateral forces caused by the cornering, as low or even alterable adhesion coefficient is required. On a dry
the resulting total forces are decisive. The total transferable force road with a usually high adhesion, the avoidance of locking wheels

IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195 193
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
has to be estimated with high precision and low delay [1]. Second,
the adaptive solution with the presented approach has an influence
on the handling and braking performance, when the vehicle brakes
during cornering. The tending to understeer is increased for EVs
with front-wheel drive while the tending to oversteer rises for EVs
with rear-wheel drive. Especially, the latter might lead to an
instable behaviour. Consequently, in addition to the desired
longitudinal brake intensity and the estimated maximum adhesion,
the adaptive BFD has to consider the current lateral acceleration or
forces caused by the steer angle to ensure a stable and safe braking
behaviour in all situations. Third, today, most of the vehicles have
an electronic stability control system, which can prevent skidding.
However, the BFD, the drive train and the general construction of a
vehicle should be dimensioned in such a way, that the stability
control system has to intervene preferably rarely.

8 References
[1] Spichartz, P., Xu, B., Sourkounis, C.: ‘Optimized recuperation strategies for
single front and rear wheel drives’. 2017 IEEE Vehicle Power and Propulsion
Conf. (VPPC), Belfort, Fance, 2017, pp. 1–6
[2] Spichartz, P., Sourkounis, C.: ‘Measurement of braking energy recuperation in
electric vehicles’. 2013 15th European Conf. on Power Electronics and
Applications (EPE), Lille, France, 2013, pp. 1–9
[3] Spichartz, P., Dost, P., Sourkounis, C.: ‘Recuperation strategies for electric
vehicles’. 2014 Ninth Int. Conf. on Ecological Vehicles and Renewable
Energies (EVER), Monte-Carlo, Monaco, 2014, pp. 1–7
[4] Spichartz, P., Dost, P., Sourkounis, C.: ‘Examination and measurement of
range extension using different recuperation modes in electric vehicles’. 2014
IEEE Industry Application Society Annual Meeting, Vancouver, BC, Canada,
2014, pp. 1–9
[5] Jaime, R.-P., Kaestner, F., Erban, A., et al.: ‘Regelalgorithmen fuer
Rekuperation und Traktion bei Elektrofahrzeugen (‘control algorithms for
recuperation and traction of electric vehicles’)’, ATZ elektronik, 2014, 9, (2),
pp. 12–17
[6] Ivanov, V., Savitski, D., Shyrokau, B.: ‘A survey of traction control and
antilock braking systems of full electric vehicles with individually controlled
electric motors’, IEEE Trans. Veh. Technol., 2015, 64, (9), pp. 3878–3896
[7] Yuan, L., Chen, H., Ren, B., et al.: ‘Model predictive slip control for electric
vehicle with four in-wheel motors’. 2015 34th Chinese Control Conf. (CCC),
Hangzhou, China, 2015, pp. 7895–7900
[8] Mutoh, N.: ‘Driving and braking torque distribution methods for front- and
rear-wheel-independent drive-type electric vehicles on roads with low friction
Fig. 19 Brake application (z ≃ 0.4) of an EV with rear-wheel drive and coefficient’, IEEE Trans. Ind. Electron., 2012, 59, (10), pp. 3919–3933
[9] Dadashnialehi, A., Bab-Hadiashar, A., Cao, Z., et al.: ‘Intelligent sensorless
adaptive BFD in a curve (radius rC = 90 m) on a dry road (μmax = 1.0)
antilock braking system for brushless in-wheel electric vehicles’, IEEE Trans.
(a) Longitudinal velocity of the vehicle vX and steer angle δ, (b) Side slip angle of the Ind. Electron., 2015, 62, (3), pp. 1629–1638
vehicle β, (c) Tyre slip angles α f l (front left), α f r (front right), αrl (rear left), αrr (rear [10] Zhai, L., Sun, T., Wang, J.: ‘Electronic stability control based on motor
driving and braking torque distribution for a four in-wheel motor drive
right), (d) Longitudinal brake slip of the tyres λ f l, λ f r, λrl, λrr
electric vehicle’, IEEE Trans. Veh. Technol., 2016, 65, (6), pp. 4726–4739
[11] Xu, G., Xu, K., Zheng, C., et al.: ‘Fully electrified regenerative braking
achieved by the transition to the ideal BFD can hardly be control for deep energy recovery and maintaining safety of electric vehicles’,
demonstrated. Further requirements are to be met by the test track IEEE Trans. Veh. Technol., 2016, 65, (3), pp. 1186–1198
[12] Pan, S., Song, Z., Wang, X.: ‘Electro-mechanical braking force distribution
when cornering is also to be performed. Obviously, a track with strategy for electric vehicle’. 2015 IEEE Int. Conf. on Communication
curves or a sufficiently large area with the mentioned conditions is Problem-Solving (ICCP), Guilin, China, 2015, pp. 387–390
required. If the vehicle skids, this must not lead to a dangerous [13] Xu, G., Li, W., Xu, K., et al.: ‘An intelligent regenerative braking strategy for
situation. electric vehicles’, Energies, 2011, 4, (9), pp. 1461–1477
[14] Peng, D., Zhang, Y., Yin, C.L., et al.: ‘Combined control of a regenerative
braking and antilock braking system for hybrid electric vehicles’, Int. J.
7 Conclusion Autom. Technol., 2008, 9, (6), pp. 749–757
[15] Tao, Y., Xie, X., Zhao, H., et al.: ‘A regenerative braking system for electric
The BFD between front and rear axle has a significant influence on vehicle with four in-wheel motors based on fuzzy control’. 2017 36th Chinese
the capability for recuperation, especially regarding EVs having Control Conf. (CCC), 2017, pp. 4288–4293
[16] Satzger, C., de Castro, R.: ‘Predictive brake control for electric vehicles’,
only one motor-driven axle. Typically, an approximation of the IEEE Trans. Veh. Technol., 2018, 67, (2), pp. 977–990
ideal BFD is desired, which ensures the shortest stopping distance [17] Ko, J., Ko, S., Son, H., et al.: ‘Development of brake system and regenerative
and a stable behaviour of the vehicle during decelerations. braking cooperative control algorithm for automatic-transmission-based
However, the simulation results in this paper show that EVs with hybrid electric vehicles’, IEEE Trans. Veh. Technol., 2015, 64, (2), pp. 431–
440
single drive using this distribution have a restricted potential for [18] Nadeau, J., Micheau, P., Boisvert, M.: ‘Ideal regenerative braking torque in
recuperation particularly with regard to brake applications with low collaboration with hydraulic brake system’. 2017 Twelfth Int. Conf. on
or middle brake intensity. Ecological Vehicles and Renewable Energies (EVER), Monte Carlo, Monaco,
An adaptive BFD was presented in this paper, which can be 2017, pp. 1–5
[19] Xu, J., Zhang, X.: ‘Optimization algorithm for vehicle braking force
implemented for front-wheel drives and rear-wheel drives. distribution of front and rear axles based on brake strength’. 2016 12th World
Simulation results show that it helps to increase the energy Congress on Intelligent Control and Automation (WCICA), Guilin, China,
recovery during decelerations and thus to raise the cruising range. 2016, pp. 3353–3360
It distributes as much brake force as possible to the motor-driven [20] Gao, Y., Chu, L., Ehsani, M.: ‘Design and control principles of hybrid braking
system for EV, HEV and FCV’. 2007 IEEE Vehicle Power and Propulsion
axle as long as the power and torque limits of the drive train are not Conf., Arlington, TX, United States, 2007, pp. 384–391
exceeded and the available adhesion is sufficient. Moving straight [21] Wei, Z., Xu, J., Halim, D.: ‘Study of HEV electro-mechanical hybrid braking
ahead, the wheels become locked at the same value of brake system control strategy’. 2015 IEEE Vehicle Power and Propulsion Conf.
intensity for the adaptive BFD and the ideal BFD. (VPPC), Montreal, QC, Canada, 2015, pp. 1–5
[22] UNECE, Agreement Concerning the Adoption of Uniform Technical
A few issues should be respected. First, the adaptive BFD Prescriptions for Wheeled Vehicles, Equipment and Parts which can be Fitted
requires the knowledge of the currently available adhesion. Thus, it and/or be Used on Wheeled Vehicles and the Conditions for Reciprocal

194 IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195
© The Institution of Engineering and Technology 2019
 20429746, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-est.2019.0039 by INASP - NEPAL, Wiley Online Library on [05/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Recognition of Approvals Granted on the Basis of these Prescriptions – [32] Kaldenbach, M., Spichartz, P., Sourkounis, C.: ‘Assessment of methods for
Addendum 12: Regulation No. 13, 4 August 2011 estimating the maximum coefficient of friction between road and tire’. 2017
[23] Paul, D., Velenis, E., Cao, D., et al.: ‘Optimal mu-estimation-based IEEE Vehicle Power and Propulsion Conf. (VPPC), Belfort, France, 2017, pp.
regenerative braking strategy for an AWD HEV’, IEEE Trans. Transp. 1–6
Electrification, 2017, 3, (1), pp. 249–258 [33] Castro, R.D., Araujo, R.E., Freitas, D.: ‘Real-time estimation of tyre-road
[24] Fujimoto, H., Harada, S.: ‘Model-based range extension control system for friction peak with optimal linear parameterisation’, IET Control Theory
electric vehicles with front and rear driving-braking force distributions’, IEEE Applic., 2012, 6, (14), pp. 2257–2268
Trans. Ind. Electron., 2015, 62, (5), pp. 3245–3254 [34] Isermann, R.: ‘Fahrdynamikregelung’ (Friedr. Vieweg & Sohn Verlag,
[25] Zhang, X., Goehlich, D., Li, J.: ‘Energy-efficient toque allocation design of Wiesbaden, 2006)
traction and regenerative braking for distributed drive electric vehicles’, IEEE [35] Bian, M., Chen, L., Luo, Y., et al.: ‘A dynamic model for tire/road friction
Trans. Veh. Technol., 2018, 67, (1), pp. 285–295 estimation under combined longitudinal/lateral slip situation’. SAE Technical
[26] Hancock, M., Assadian, F.: ‘Impact of regenerative braking on vehicle Paper. SAE International, 2014
stability’. IET Hybrid Vehicle Conf., Coventry, UK, 2006, pp. 173–184 [36] Schramm, D., Hiller, M., Bardini, R.: ‘Modellbildung und Simulation der
[27] Spichartz, P., Sourkounis, C.: ‘Influence of the drive train topology and the Dynamik von Kraftfahrzeugen’ (Springer Vieweg Verlag, Berlin, Heidelberg,
center of mass on the regenerative braking in electric vehicles’. 2017 Twelfth 2018, 3rd edn.)
Int. Conf. on Ecological Vehicles and Renewable Energies (EVER), Monte [37] Wu, J., Wang, Q., Wei, X., et al.: ‘Studies on improving vehicle handling and
Carlo, Monaco, 2017, pp. 1–8 lane keeping performance of closed-loop driver-vehicle system with
[28] Spichartz, P., Bokker, T., Sourkounis, C.: ‘Comparison of electric vehicles integrated chassis control’, Math. Comput. Simul., 2010, 80, (12), pp. 2297–
with single drive and four wheel drive system concerning regenerative 2308
braking’. 2017 Twelfth Int. Conf. on Ecological Vehicles and Renewable [38] Spichartz, P., Sourkounis, C.: ‘Comparison of drive train topologies for
Energies (EVER), Monte Carlo, Monaco, 2017, pp. 1–7 electric vehicles with regard to regenerative braking’. 2019 Fourteenth Int.
[29] Haken, K.-L.: ‘Grundlagen der Kraftfahrzeugtechnik’ (Carl Hanser Verlag, Conf. on Ecological Vehicles and Renewable Energies (EVER), Monte Carlo,
Munich, 2015) Monaco, May 2019. (accepted paper)
[30] Luo, N., Jiang, J., Yu, A.: ‘Research on the control strategy of the [39] Breuer, B., Bill, K.-H.: ‘Bremsenhandbuch’ (Springer Vieweg Verlag,
regenerative braking system’. 2014 Int. Conf. on Mechatronics and Control Wiesbaden, 2017, 5th edn.)
(ICMC), Jinzhou, China, 2014, pp. 2514–2517
[31] Stellet, J.E., Suchaneck, A., Giessler, M., et al.: ‘Fahrbahnreibwertschaetzung
mit optimaler linearer Parametrierung’, at – Automatisierungstechnik, 2014,
62, (8), pp. 570–581

IET Electr. Syst. Transp., 2019, Vol. 9 Iss. 4, pp. 186-195 195
© The Institution of Engineering and Technology 2019