You are on page 1of 11

Annual Reviews in Control 28 (2004) 207217

www.elsevier.com/locate/arcontrol

Combined active steering and traction for mechatronic bogie


vehicles with independently rotating wheels
J. Pereza,*, Jesus M. Busturiab,1, T.X. Meic,2, J. Vinolasd
a

Department of Engineering Technology, Manchester Metropolitan University, Manchester, M1 5GD, UK


b
GAMESA Parque Tecnologico de Alava, 01510 Minano Menor, Spain
c
School of Electronic and Electrical Engineering, University of Leeds, Leeds, LS2 9JT, UK
d
CEIT, Paseo Manuel de Lardizabal, 15 - 20018 San Sebastian, Spain
Received 20 December 2003; received in revised form 26 January 2004; accepted 17 February 2004

Abstract
The benefits of substituting the conventional solid wheelset by independently rotating wheels (IRW) in actively steered trailed railway
vehicles have already been presented in many works. If a traction system is included in this kind of vehicle it will interact strongly with the
active steering system. This paper investigates how traction and active steering systems can be combined in an IRW bogie to produce a good
curving performance. The combined steering and traction control concepts developed are implemented in simulation models and validated
against a conventional passive vehicle.
# 2004 Elsevier Ltd. All rights reserved.
Keywords: Railways; Active control; Active steering; Traction control

1. Introduction
The ability of railway vehicles to follow the track path is
based on the properties of the solid wheelset. The conic
shape of the wheels makes their rolling radius dependent on
the lateral position with respect to the track. As the two
wheels are constrained to have the same rotating speed, the
yaw movement of the wheelsets is coupled to their lateral
displacements, thus generating a self-centering mechanism.
Unfortunately, this movement is not stable. The conventional solution to this problem is to join wheelset pairs
through a bogie and a relatively stiff horizontal suspension,
which constrains the relative movements of the two
wheelsets. This mechanism can stabilize the movement of
the wheelset effectively, but prevents it from adapting to the
shape of the track on curves, producing unbalanced wheel to
rail lateral forces and high wear. Therefore, in the case of
* Corresponding author. Tel.: +44 161 247 6620; fax: +44 161 247 1633.
E-mail address: j.perez@mmu.ac.uk (J. Perez), jmbusturia@gamesa.
es (J.M. Busturia), t.x.mei@ee.leeds.ac.uk (T.X. Mei).
1
Tel.: +34 945 185 823; fax: 34 945 185 667.
2
Tel.: +44 113 343 2066; fax: +44 113 343 2032.
1367-5788/$ see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.arcontrol.2004.02.004

conventional vehicles, the design requirements for preserving stability at high speeds and for getting a good curving
performance are contradictory.
The use of new mechatronic technologies can help to
improve those and other features of railway vehicles. The
integrated design of mechanics and control electronics
provides the designer with more freedom and flexibility than
conventional design methods based on the exclusive use of
mechanics, allowing achieving simultaneously different
goals that have been traditionally considered contradictory.
This philosophy has already been applied in practice to
systems such as active tilting and suspension (Goodall,
1997).
Methods to steer the wheelsets by means of active
controls have been already explored by some researchers
(Goodall & Kortum, 2000). This new approach allows for
the optimization of both curving performance and stability at
the same time. Furthermore, it can change the fundamental
design of railway vehicles by eliminating the bogies and
other mechanical elements, leading to the Mechatronic Train
Concept (Ellis & Goodall, 1999).
Previous works have shown the potential of those
techniques (Goodall, 1998). A range of specific studies of

208

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

active systems for active steering have already been


presented by different authors (Aknin, Ayasse, & Devallez,
1995; Gretzschel & Bose, 1999; Mei & Goodall, 1999). In
Pe rez, Busturia, and Goodall (2002) a completely conventional vehicle mounted with solid wheelsets was actively
steered, showing that significant performance improvements
can be obtained in curves while preserving stability.
Nevertheless, the performance of those strategies was
limited by the fact that the solid wheelset relates the good
performance properties to a particular lateral position
respect to the track, which leads to flange contact in sharp
curves. Once the natural self-guidance properties of the solid
wheelset have been substituted by active control, it is
possible to use independently rotating wheels (IRW) instead
of solid wheelsets (Mei & Goodall, 1999). Pe rez, Mauer, and
Busturia (2001) shows that strategies based on independently rotating wheels can avoid flange contact in a vehicle
with bogies by centering the wheelset on the track.
The studies mentioned above did not account for the
existence of a traction system in the vehicle. This can be an
important issue, as the industry tends nowadays to produce
trains with a distributed traction, delivering higher and
better-controlled tractive effort. The supply of separated
traction to each of the wheels of a wheelset with
independently rotating wheels can generate a yaw torque
to the whole wheelset that can interact with the steering
action. Therefore, the design and control of the traction and
the steering system cannot be done separately, and combined
strategies must be sought. This work intends to show some
principles that can be applied to the combined design of
active steering/traction systems and to verify its behaviour
through complex simulation models.
For the study a fully conventional configuration of
railway vehicle with bogies has been chosen. A plan-view
diagram of the vehicle is shown in Fig. 1. The vehicle
includes a secondary suspension between the bogies and the
vehicle body, which is designed with comfort criteria, and
the primary suspension between the wheelsets and the bogie,
designed with stability and curving criteria. Actuators are
added at the tip of each wheelset to provide the active
steering action. Each solid-axle wheelset is replaced by a

pair of wheels attached to independent motors. There is a


connection between the two motors so that the yaw
movement can be transmitted to the whole wheelset directly,
but the rotating axle is cut and each of the wheels is
independently driven by its own motor. This paper considers
the problem of combining the differential traction of the
motors mounted on the same wheelset with the steering
action provided by the actuators.
The rest of the paper is organized in the following
manner: First of all, some strategies for the active steering of
railway vehicles that have already been validated with
trailed vehicles are reviewed. Then general concepts on how
traction should be provided to interact adequately with the
active steering strategies given are presented, producing a
range of combined strategies. After that some specific issues
on the control of motors are discussed. Finally, the strategies
proposed are validated through simulation using complex
models of vehicle, control systems motors and actuators.

2. Strategies for active steering of trailed railway


vehicles
Even though the objective of this paper is not to study the
active steering system itself, it is necessary to introduce the
basic strategies that will be used in the rest of the work.
From the broad range of strategies presented in the literature
for the active steering of railway vehicles, only two will be
used here, one for vehicles with solid wheelsets and the other
for vehicles with independently rotating wheels. Though, it
is likely that the concepts of traction/steering combination
are extensible to most of the strategies presented in other
works.
2.1. Vehicles with solid wheelsets: control of the yaw
moment transmitted from wheelset to bogie
This steering strategy was first proposed by Shen and
Goodall (1997). It is well known that a single wheelset
without primary suspension is able to orient itself on a curve
and to run in pure rolling condition, but that this movement

Fig. 1. Plan-view diagram of the vehicle.

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

209

Fig. 2. Schematic representation of the yaw moment control strategy.

is unstable at the hunting frequency (generally between 2


and 8 Hz, depending on the speed, and various other
parameters of the vehicle). The passive primary suspension
introduces a damping by constraining the movement of the
wheelset, but it deteriorates its self-orienting properties. The
solution proposed consists of controlling the yaw moment
transmitted from the wheelset to the bogie to be zero at low
frequencies, in order to recover the properties of the free
wheelset. The frequency range controlled is limited so that
the passive suspension is still working at the hunting
frequency, thus stabilizing the vehicle. A simple schematic
representation of a control system of this type is represented
in Fig. 2.
The application of this strategy has shown to be relatively
simple but yet effective. Their two main limitations are that
the bandwidth is inherently limited, ensuring that the active
guidance does not interact with the passive stabilization, and
that there is a tendency to run in flange contact in sharp
curves, which is a feature of most of the active steering
strategies for solid wheelsets.
2.2. Vehicles with IRW wheelsets: control of the differential
rotational speed
Due to the shape of the wheel profile, the lateral
displacement of the wheelset respect to the track modifies
the relative rotational speed of the two wheels in an IRW
wheelset. If the layout of the track is known, the speed
difference can be controlled to keep the vehicle running on
the center of the track. The expression of the desired value
for the difference of the rotational speeds is developed in
Pe rez et al. (2001):
vl  vr obj

2y
r
r0

(1)

where r is the track curvature, is the semi distance of


rolling radii, r0 is the wheelset nominal radius and y is the
vehicle speed.
The strategy is illustrated in the schematic representation
of Fig. 3.
The performance of those strategies is similar to the
strategy for solid wheelsets presented before, with the

Fig. 3. Schematic representation of the differential rotational speed control


strategy.

advantage that, as for IRW the pure rolling does not depend
on the lateral position relative to the track, flange contact can
be avoided.

3. Strategies for combined active steering and traction


The strategies presented up to now have been formulated
and verified for the case that the mechanical relationship
between the behaviour of the two wheels is completely
defined by the wheelset configuration (i.e. solid or IRW).
When the separated traction is provided to the wheels it is
necessary to define the mechanical relationship between
the traction torques transmitted to the two wheels of a
wheelset so that the active steering strategy formulated is
still valid, or even improved by the action of the motors. Two
different concepts are presented, one to be used with solid
wheelset steering strategies and the other for IRW steering
strategies.
3.1. Differential speed control of the traction motors
(solid wheelset-like strategy)
It can be intuitively stated that, if the two wheels of the
same wheelset are forced by the motors to run with the same
rotational speed, the IRW wheelset will behave as if it was a
solid wheelset from the guidance and stability points of
view. In this case, solid wheelset strategies can be used for
the steering system of an IRW wheelset with independent
motors. The behaviour expected using this technique is, in
the best case, the same as the one obtained with the trailed
vehicle with solid wheelsets. Improvements can be obtained
if the command for the differential speed control of the
wheels is made dependent of the radius of the track, making
the outer wheel run faster than the inner wheel so that flange
contact is avoided. The difference between the rotational
speeds necessary to make the wheelset run by the center of
the track is again given by Eq. (1)

210

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

Fig. 4. Combined traction control/solid wheelset active steering strategy.

Fig. 5. Combined traction control/IRW active steering strategy.

Supposing an ideal perfect control of the differential


speed the dynamic behaviour is the same as the one of the
solid wheelset, but a lateral offset has been added so that the
wheelset always runs by the center of the track. The scheme
is shown in Fig. 4.

in the present work, but it is considered an attractive option


for future development.

3.2. Differential torque control of the traction motors


(IRW-like strategy)

Modern electrical railway traction systems are very


complex. Asynchronous induction motors are often used to
replace dc traction motors for a number of practical reasons,
but they are much more difficult to control and significantly
more complex to model mathematically. However, advances
in high power switching devices as well as motor control
methods (such as PWM and vector control techniques) have
enabled an induction motor to behave very similarly to a
separately excited dc motor in the range of frequencies of
interest for this application (below 15 Hz). Taking advantage
of this fact, all the complexity of the AC traction motor and
its associated electronics and control is substituted in this
work by a dc motor model with the separate excitation
control of the torque and flux producing currents, under the
reasonable expectation that no significant difference will be
introduced in the dynamic behaviour of the wheelset.
The vector control of asynchronous motors enables the
control of flux and torque separately. In the control of the dc
motor, the equivalent inputs are the field and armature
voltages, respectively (Vf, Va). In this case, the control of
those variables is performed in the following way: The
desired motor behaviour is the sum of two different goals,
which are the traction and steering commands. The
command coming from the steering control will be much
smaller than the traction component, but the required
bandwidth will be higher. Therefore, only the control of Va is
used for steering because of its faster dynamics. The traction

The behaviour of the IRW wheelset is determined by the


fact that almost no longitudinal contact force is produced
because of the freedom of the wheels to rotate at different
speeds. From the point of view of steering, this means that no
yaw moment is introduced to the wheelset by the contact
forces. This feature is obtained not only in absence of
longitudinal contact forces, but also whenever the contact
forces are equal in the two wheels of the same wheelset.
Therefore, if the torque provided to both wheels is controlled
to be equal, the system theoretically works in a very similar
way to an IRW trailed vehicle from steering point of view.
This allows combining active steering strategies for IRW
with differential torque control of the motors on the same
wheelset. This combined control is schematically presented
in Fig. 5.
In this case no performance improvement is expected
when compared to the trailed vehicle, as the flange contact
was already avoided by the original steering strategy.
In the two-combined control strategies presented here,
the control systems for the steering actuators and the motors
are implemented using separated control loops. A different
approach that could be considered would be the definition of
an integrated traction/steering control using MIMO control
strategies such as H2 and H1. This approach is not treated

4. Control of traction motors

J. Perez et al. / Annual Reviews in Control 28 (2004) 207217

211

4.1. Single motor modelling


The behaviour of each motor is given by the equations of
armature and field winding:
(2)
Va La Ia Ra Ia Ke If v
Vf Lf If Rf If

Fig. 6. Curve of the operation limits of the motor (solid) and limit used for
traction purposes (dashed).

is provided controlling both Va and Vf, but the maximum


value of the armature voltage considered for traction control
will be slightly smaller than the maximum VaMax available
(xVaMax ; x < 1), in order to account for the steering demand
with the remaining voltage (Fig. 6). Thus, the steering
control will be performed by simply adding a small value to
the Va considered for traction.
Figs. 7 and 8 show the schematic diagram for the
combined traction/steering control of the motors in one
wheelset for the IRW wheelset-like and solid wheelset-like
strategies, respectively. The traction command would be
directly given by the driver or by an automatic train
operation system, while the differential torque command
would be produced by the steering control (IRW like
strategy), as shown in Fig. 7. If solid wheelset strategies are
used for steering a differential speed control is necessary,
such as the one presented in Fig. 8.
As this research is directed to the study of steering,
and the mechanical models of vehicle used do not account
for the existence of traction forces, the traction control
loop is not implemented, but an approximate value of
the If required to run at the desired speed is calculated in
order to reproduce at some extent the conditions of the
operation of the motor. This is explained further in the next
section.

(3)

where Va and Vf are the armature and field winding voltage,


Ia and If are the armature and field winding current, v is the
rotational speed of the motor, La and Lb are the equivalent
inductance of armature and field winding, Ra and Rf are the
equivalent resistance of armature and field winding and Ke is
the EMF constant.
The torque generated by the motor is given by the
equation:
TM Ke If Ia

(4)

In order to linearise Eq. (4), it is assumed that the speed of


the vehicle is constant, or changes slowly compared to the
action of the differential traction control system, which is
generally true. It is supposed that the traction torque
necessary to circulate at a given speed is approximately
constant, and so is the value of the field current If.
Accounting for the gear ratio N the steady-state traction
torque provided to the wheel is expressed as a non-linear
relationship of the speed and the field current If:
TWheel

Ke
xVamax If  Ke vIf2
NRa

(5)

This expression allows calculating the value of If for each


speed if the value of the traction torque TWheel is known, as
shown in Table 1. This is given by an estimated value of the
running resistance, which is obtained using an empirical
formula commonly used in the railway industry (Profillidis,
2000).
In the subsequent analyses, the value of If and the
circulating speed are considered constant for each case.
Under this assumption, the equations of the motor become

Fig. 7. Differential speed control scheme for the motors in a wheelset.

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

212

Fig. 8. Differential torque control scheme for the motors in a wheelset.

linear, being Va the only input. Defining the constant value


KT, which must be calculated for every speed:
KT Ke If

(6)

Substituting KT in (4) The expression of the torque


provided by the motor becomes linear:
TM KT Ia

(7)

Being N speed reduction factor of the gearbox, the


equivalent speed and torque after the gearbox (vE, TE) are
calculated as:
TM TE N
(8)
vE
(9)
vM
N
The model is completed by including the dynamics of the
rotating parts:
JM v M CM vM TM  NTOUT

(10)

Being JM and CM the inertia and damping of the rotating


elements referred to the motor, and TOUT the torque supplied
to the wheel. Finally, the elasticity of the transmission
elements is accounted for through the following expression:
TOUT KOUT uE  uOUT COUT vE  vOUT

(11)

Table 1
Winding current for different circulating speeds
Speed (km/h)

Traction torque per wheel (N m)

If (normalized)

108
140
176
216
265

108
140
176
216
265

108
140
176
216
265

Being KOUT and COUT the stiffness and damping of the


transmission referred to the output, and vOUT the speed of
the wheel.Operating on Eqs. (2), (3), (7), (10) and (11), it is
easy to show that the output torque provided to each of the
wheels can be expressed in Laplace transform form as:
TOUT s Gu sVa Gv svOUT

(12)

Expression (12) shows a dependence of the torque


provided by the motor on both the rotating speed and the
armature voltage provided. The expressions of the transfer
functions Gu(s) and Gv(s) can be written as:
Gu s

AsBs
Cs Bs

(13)

Gv s

CsBs
Cs Bs

(14)

where
KT =N
La s Ra


KOUT
COUT
Bs
s


JM
CM KT
As
s

Cs
N2
N2
N
As

(15)
(16)
(17)

The complete system studied will also include the


dynamics of the vehicle. The relationship between the
dynamics of the motor and the one of the rest of the vehicle is
schematically represented in Fig. 9.
The fact that the dynamics of the vehicle needs to be
simulated using a multibody software, which generally use
forces and torques as inputs and displacements and
velocities as outputs, makes necessary to represent the

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

213

Fig. 9. Schematic representation of the coupling between vehicle and


traction motor dynamics.

motor dynamics with vOUT as input and TOUT as the output in


Eq. (12).
Expression (12) applies for a single wheel, but the
strategies presented in the previous section require the
control of the differential behaviour of wheel pairs.
4.2. Motor pair modelling
The rotational speed of each of the two motors is
considered separately:
vLEFT
OUT

VS
vdiff
r0

vRIGHT
OUT

VS
vdiff
r0

(18)
(19)

Fig. 10. Block diagram of a wheelset with a differential torque controlled


motor pair.

torque provided is represented by the block diagram in


Fig. 10:
From this diagram the expression of the differential
torque is obtained, being a function of the differential torque
command and of the differential rotational speed of the
wheelsets:
Tdiff

Gu sCs
T Command
1 Gu sCs diff

Gv s
vdiff
1 Gu sCs

(25)

LEFT
TOUT
TTRACTION Tdiff

(20)

The differential rotational speed of the wheels vdiff is


related to the differential torque Tdiff through the dynamics
of the complete vehicle. As it would be too complex to
consider the whole dynamic equations of the vehicle here, a
very simplified model of the dynamics of the rotational
speed between the two wheels is used for control design
purposes.

RIGHT
TOUT
TTRACTION  Tdiff

(21)

vdiff Gwset sTdiff

Being VS the running speed of the vehicle. The torque


transmitted to each of the wheels can be written as the sum of
the traction torque, which is the same for both wheels, and a
differential torque:

It is assumed that the differential torque in the two motors


is obtained using a single control signal u:
VaLEFT

xVaMax

VaRIGHT xVaMax  u

(22)
(23)

The expressions for left and right wheels can now be


developed separately from (12). If Eqs. (18)(23) are
substituted and the equations for the two wheels are
subtracted, the following expression for the differential
torque of the wheel pair is obtained:
Tdiff s Gv svdiff Gu su

1
Tdiff
If s f11 r02 =VS

(26)

where If is the rotatory inertia of the wheelset, f11 is the


longitudinal creep coefficient, r0 is the nominal rolling
radius and n is the circulating speed
Including this relationship between differential rotational
speed and differential torque, the block diagram of the
complete design model is obtained (Fig. 11).

(24)

Starting from this expression, the differential speed and


torque control systems are developed in the next two sections.
4.3. Differential torque control of motor pairs
Considering the behaviour of the motor-pair of the same
axle given by expression (24), the control of the differential

Fig. 11. Complete design model for differential torque control of motor
pairs.

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

214

Fig. 12. Block diagram of a wheelset with a differential speed controlled


motor pair.

Fig. 13. Complete design model for differential speed control of motor
pairs.

The transfer function of the system for control design is


obtained from this block diagram:
Tdiff

Gu s
u
1  Gv sGwset s

(27)

This transfer function is used for the derivation of the


traction motor pair differential torque controller C(s),
taking into account both the vehicle and the motor pair
dynamics.
4.4. Differential speed control of motor pairs
Now lets consider a steering control strategy that
requires the control of the differential rotational speed vdiff
instead of the differential torque Tdiff. The diagram of the
controlled system is given in Fig. 12.
The expression of the differential torque supplied by the
motor pair is obtained from the block diagram:
Tdiff Gu sCsvCommand
Gv s  Gu sC
diff
 svdiff

(28)

As in the previous case, this expression along with a


simplified model of the vehicle dynamics are used for
control design (Fig. 13).

The transfer function of the system to be controlled is


easily derived:
vdiff

Gu sGwset s
u
1  Gv sGwset s

(29)

This transfer function is used for the derivation of the


traction motor pair differential speed controller C(s),
taking into account both the vehicle and the motor pair
dynamics.
5. System validation methodology
The complex dynamic behaviour of railway vehicles
cannot be represented by the use of simple linear models
with acceptable accuracy. The simulation software SIMPACK has been used here to generate and simulate complex
models of the mechanical part of the vehicle which has to be
controlled actively including all the complex features of
railways vehicles, such as large displacements and rotations
between bodies, non-linear characteristics of some suspension components and non-linear wheel rail contact geometry
(Schupp, Netter, Mauer, & Gretzschel, 1999).

Fig. 14. Scheme of the co-simulation procedure.

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

215

Fig. 17. Attack angles of wheelsets.


Fig. 15. Lateral position of wheelsets.

To carry out simulations of the actively steered vehicle it


is necessary to consider simultaneously the dynamics of the
actuators and motors, the control system and the vehicle.
The SIMAT interface of SIMPACK has been used to cosimulate the model of the vehicle in SIMPACK together with
the rest of the system modelled in SIMULINK through an
appropriate description of the inputs and outputs needed by
each of the sub-models (Fig. 14). In the present case, the cosimulation of the vehicle and the control system allows
obtaining accurate results of the dynamic behaviour of the
active vehicle and performing the final tuning of the
controllers under realistic operation conditions.

6. Results and discussion


Simulations have been carried out in different track
conditions using the co-simulation model presented above.
Five curved tracks with radiuses from 300 to 1800 m have
been used for the assessment. The speed of the vehicle is

Fig. 16. Lateral position of wheelsets.

adjusted for each curve so that a maximum cant deficiency of


300 mm is obtained in all cases, which gives a range of
speeds from 108 to 265 km/h. Fig. 15 shows the lateral
displacement of the wheelset in one of the curves considered
(R = 800 m, V = 176 km/h).
The lateral displacement of the wheelset is controlled in
both active cases to a maximum value of around 1.5 mm,
thus avoiding flange contact. In the constant curve, both
active strategies lead to zero displacements. Therefore, the
goals have been achieved from the point of view of
guidance.
Fig. 16 shows the maximum values of the lateral
displacements in all the curves studied. The passive vehicle
leads to flange contact, while with both active strategies the
maximum lateral displacement is always around 1.5 or
2 mm.
The angle of attack (Fig. 17) is reduced on sharp curves
but not on smoother curves, in which the values are similar.
The lateral contact forces transmitted from the wheelset
to the track are reduced down to values near the minimum
achievable, represented by the dotted line in Fig. 18. The
solid wheelset-like strategy shows a slightly worst performance, possibly due to its lower bandwidth.

Fig. 18. Maximum lateral contact forces.

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217

216

References

Fig. 19. Mean wear indices.

The most impressive reductions of wear are logically


obtained for the smallest curve radiuses (Fig. 19). It is
noticeable that the amount of wear is kept almost constant
regardless of the curvature of the track when active systems
are used. For the smoothest curves wear is still reduced to
the half with the active system, which is still a very good
result.

7. Conclusions
This paper shows that the active steering of IRW vehicles
with distributed traction is possible with an appropriate
design of the controls of the traction motors and the active
steering actuators. The approach followed consists of
reproducing the features of either a solid wheelset or an
IRW wheelset through the control of the amount of
differential traction supplied to the wheels, so that almost
any of the active steering strategies for solid wheelsets or
IRW existing in the literature can be applied. It has been seen
that an appropriate control of the traction motors preserves
or improves the curving performance of the actively steered
vehicle.
Due to the flexibility inherent to the approach presented,
many different strategies for the combined control of
traction and steering can be implemented without significant
changes in the hardware and the mechanical configuration.
This feature can be used, for example, to reconfigure the
system in presence of failures, thus enhancing the fault
tolerance, which is a critical issue for actively steered
vehicles.

Acknowledgements
Authors wish to acknowledge the support of the European
Community for funding the project BE97-4387, which made
this study possible.

Aknin, P., Ayasse, J. B., & Devallez, A. (1995). Active steering of railway
wheelsets. 12th IAVSD Symposium.
Ellis, B., & Goodall, R. M. (1999). The mechatronic train: Requirements
and concepts. In Proceedings of World Congress on Railway Research
(pp. 17).
Goodall, R.M. (1997). Active railway suspensions: Implementation status
and technological trends. Vehicle System Dynamics, 28, 87117.
Goodall, R. M., & Kortu m, W. (2000). Mechatronic developments for
railway vehicles of the future. 1st IFAC conference on Mechatronic
Systems.
Goodall, R.M. (1998). Steer-by-wire: A railway technology for the
future. International Congress Railtech98 Tomorrows World (pp.
6776).
Gretzschel, M., & Bose, L. (1999). A mechatronic approach for active
influence on railway vehicle running behaviour. 16th IAVSD Symposium: Dynamics of Vehicles on Roads and Tracks.
Mei, T. X., & Goodall, R. M. (1999). Wheelset control strategies for a 2-axle
railway vehicle. 16th IAVSD Symposium: Dynamics of Vehicles on
Roads and Tracks.
Pe rez, J., Busturia, M., & Goodall, R. M. (2002). Control strategies for
active steering of bogie-based railway vehicles. Control Engineering
Practice, 10(9), 10051012.
Pe rez, J., Mauer, L., & Busturia, J. M. (2001). Design of active steering
systems for bogie-based railway vehicles with independently rotating
wheels. 17th IAVSD Symposium: Dynamics of Vehicles on Roads and
Tracks.
Profillidis, V. A. (2000). Railway Engineering. Asgate Publishing Limited,
ISBN 0-7546-1279-1.
Schupp, G., Netter, H., Mauer, L., & Gretzschel, M. (1999). Multibody
system simulation of railway vehicles with SIMPACK. Vehicle System
Dynamics Supplement, 31, 104118.
Shen, G., & Goodall, R. (1997). Active yaw relaxation for improved bogie
performance. Vehicle System Dynamics, 28, 273289.

J. Perez studied Mechanical Engineering at the University of Navarra,


Spain, where he received his degree in 1995. Since then, he has been
working as a researcher at the department of Applied Mechanics of the
research center CEIT in San Sebastian (Spain). He obtained a Ph.D. in the
area of mechatronics for railway vehicles in 2002. Javier Pe rez has worked
in a variety of research projects for both industry and institutions. His main
areas of expertise are railway vehicle dynamics and mechatronic and control
applications for railway vehicles. He has been involved in industrial R&D
projects for rolling stock manufacturers and rail operators, and also in
European Research projects.
Jesus M. Busturia received a M.Sc. (1982) degree from the University of
the Basque Country and a Ph.D. (1991) from the University of Navarra
(Spain). He spent 18 years working for the Department of Mechanical
Engineering of the CEIT (a research institute in San Sebastia n) in various
positions starting as a Ph.D. student. His last duty was the Head of the
Experimental Dynamics Area. During years 19911992 he made a postdoctoral stage in the University of California at Davis. In 2001 he moved to
GAMESA an industrial company devoted to the design and manufacturing
of wind turbine generators (WTG). He has been working for the Business
Development Department taking part in the acquisition and development of
various companies mainly related with the WTGs gearbox. His research
interest started in the Experimental Modal Analysis area during the first
years and moved to the mechatronics area with special emphasis on railways
active suspensions. Another line of research has been the development of
mathematical models for real time training simulators.
T.X. Mei received a B.Sc. (1982) and a M.Sc. (1985) degree from Shanghai
Tiedao University, China and also a M.Sc. (1991) from Manchester University and a Ph.D. (1994) from Loughborough University, UK. He held an

J. Pe rez et al. / Annual Reviews in Control 28 (2004) 207217


academic position in the department of Electrical Engineering at Shanghai
Tiedao University between 1985 and 1989. He also spent a number of years
working in industry after his Ph.D., and acquired substantial practical
experience through many industrial R&D projects he was involved. Dr
Mei is currently a Senior Lecturer in the School of Electronic and Electrical
Engineering at the University of Leeds, UK. His research is concerned with
a variety of practical applications of advanced control, including motor
drives, traction control, vehicle dynamics, contact mechanics, excitation
control, fault-tolerant systems, system integration and real time implementations. He is a Member of the IEE and a Chartered Engineer.

217

J. Vinolas is at present the Head of the Applied Mechanics Department at


CEIT (Centro de Estudios e Investigaciones Te cnicas) and Associate Professor at the Faculty of Engineering of the University of Navarra. He obtained
his Ph.D. in 1991, the title of the thesis was A New Experimental Methodology for Testing Active and Conventional Active Suspensions. He has been
involved in different research projects related with railway dynamics and
railway noise and vibration. Being rail research topics one of the priorities of
the Department he has close links to Spanish rolling stock manufacturers,
infrastructure owners, component suppliers and other international companies
and institutes through European Research projects and Networks.

You might also like