Content uploaded by Michael Ruderman
Author content
All content in this area was uploaded by Michael Ruderman on Mar 10, 2017
Content may be subject to copyright.
IEEJ International Workshop on Sensing, Actuation, Motion Control, and Optimization
Examples of actuator uncertainties in environmental systems of
mechatronic systems (SoMS)
Michael Ruderman∗a) Non-member, Hiroshi Fujimoto∗∗ Senior Member
Shota Yamada∗∗ Student Member, Valentin Ivanov∗∗∗ Non-member
The proposed concept of systems of mechatronic systems (SoMS) focuses on a methodology aimed for robust con-
trol, state estimation, and disturbance compensation in highly dynamic environmental mechatronic systems. Three
interfacing topics – “mechatronic chassis systems of electric vehicles”, “mechatronic-based grid-interconnection cir-
cuitry”, and “offshore mechatronics” – have been identified as comprising a series of research challenges in relation
to the state and parameter estimation, disturbance observation and attenuation, and robust control design. This paper
aims to highlight several examples of actuator uncertainties in environmental SoMS, in particular those used in off-
shore mechatronics and electric vehicles, and to promote the research activities of the collaborative project CLOVER
initiated within European Unions Horizon 2020 framework programme under Marie Sklodowska-Curie actions.
Keywords: Systems of mechatronic systems, motion control, uncertainties, actuators, state estimation, compensators, dead-zone,
backlash, offshore mechatronics, hydraulic cylinders, electric vehicle, in-wheel motor technology
1. Introduction
The environmental systems of mechatronic systems can
be seen as a broad class of integrated and interconnected
environment-friendly technologies, performing in a close
proximity with surrounding eco-system and that under crite-
ria of environmental impact minimization and better energy
efficiency. An obvious example of environmental mecha-
tronic applications lies within the chain “renewable energy
production - smart grids - electric vehicles”. This chain has
all the required characteristics to be considered as a system
of systems (SoS). To those belong the operational and man-
agerial independence; geographical distribution; emergent
behavior; evolutionary development; heterogeneity of con-
stituent systems. In view of the system design and implemen-
tation, with related involvement of various technologies, such
SoS includes a substantial number of mechatronic compo-
nents which lends itself to the concept of “System of Mecha-
tronic Systems” (SoMS). The current state of research in
SoMS exhibits the following significant gap: state-of-the-art
of mechatronic, information, and communication technolo-
gies allows one to design and implement the ground vehicles
(1), grids, and renewable energy production systems. These
can have an extraordinary high dynamic performance but re-
quire an advancement of the relevant control approaches, so
as to ensure operating eco-friendly and energy-efficient and,
at the same time, under highly-dynamic and uncertain exter-
nal conditions.
The operation of environmental SoMS is characterized by
variety of uncertain factors and disturbances. Both can be
a) Correspondence to: michael.ruderman@uia.no
∗University of Agder, 4879, Norway
∗∗ The University of Tokyo, 277-8561, Japan
∗∗∗ Technical University Ilmenau, 98693, Germany
of steady-state, short-term and long-term nature, therefore
occurring and acting on different time scales of the system
dynamics. The development of robust and reliable methods
for detection and estimation of such uncertainties and dis-
turbances constitutes the key part for their attenuation, cor-
respondingly rejection. To be emphasized is that the distur-
bance estimation and attenuation appear simultaneously, dur-
ing the operation of a controlled SoMS.
Taking reference to the formulated intensions and start-up
activities of the collaborative project CLOVER, this paper ad-
dresses several examples of the actuator uncertainties in envi-
ronmental SoMS. In particularly, the examples are associated
with offshore mechatronics, as one of the key technological
fields for operation and maintenance of the offshore renew-
able energy production, and electric cars which pave a way
of the future mobility and transportation systems.
2. Framework of collaborative project CLOVER
The following examples relates to the collaborative project
CLOVER within the Horizon 2020 framework of Marie
Sklodowska-Curie Actions (MSCA) established by the Eu-
ropean Commission. The project is funded through the Re-
search and Innovation StaffExchanges (RISE) scheme be-
tween universities and industrial organizations from Ger-
many, Austria, Belgium, Norway, UK, Mexico, and Japan.
Global approach to SoMS in the CLOVER project is based
on a consecutive implementation of development activities
on three methodological levels as shown in Fig. 1 – Con-
trol Engineering; Mechatronic Systems; Testing Technology.
The related control engineering problems allows formulating
a set of key objectives and that Research (R) objectives and
Innovation and technological (IT) objectives.
R1 Benchmarking tools for comparative analysis of differ-
c
2016 The Institute of Electrical Engineers of Japan. 1
Preprint of the manuscript accepted to IEEJ International
Workshop on Sensing, Actuation, Motion Control, and
Optimization (SAMCON2017)
Examples of actuator uncertainties in environmental systems of mechatronic systems (SoMS) (Michael Ruderman et al.)
Fig. 1. Research and innovation areas related to the
CLOVER project objectives
ent control and estimation technique applied to mecha-
tronic systems;
R2 Methodological approach for switching between vari-
ous control strategies under criteria of environmental im-
pact minimization and better energy efficiency;
R3 Advanced methods for observers and disturbance rejec-
tion /attenuation as applied to highly dynamic mecha-
tronic systems;
IT1 Development and real-time hardware-in-the-loop vali-
dation of plug-in EV dynamics controller with optimized
performance by criteria of energy efficiency, energy har-
vesting and system safety;
IT2 Development and real-time hardware-in-the-loop vali-
dation of robust controllers for mechatronic systems op-
erating for and on the offshore wind-park platforms, as
smart grid components, and service vessels;
IT3 Advancement of open development platform aimed at
model-based design of SoMS.
The methodology and implementation of the CLOVER
project is characterized by a high grade of interdisciplinar-
ity in the research activities as visualized in Fig. 2. Here
the involved areas of knowledge and their contribution to the
project topics are represented around the core of highly dy-
namic environmental mechatronic systems. The multidisci-
plinary character of the CLOVER project aims at providing
synergies from different competencies and close intersectoral
collaboration between the project partners with their specific
expertise and research focus.
3. Actuators in offshore mechatronics
Hydraulic linear-stroke cylinders and hydraulic rotary mo-
tors have been, for decades, and yet still remain by far the
mostly used actuators in the offshore mechatronics and re-
lated maritime applications (2). This is among others due to
a high power density and, consequently, high force-to-mass
ratio of hydraulic drives, their robustness when operating in
harsh and open-air environments, and required safety against
the short-circuits and humidity which, otherwise, appear as
critical for the electrically actuated drive systems. Hydraulic
actuators exhibit, on the other hand, high level of nonlinear-
ities and heightened uncertainties in view of the nominal pa-
rameters and varying operation conditions.
The dead-zone and backlash type nonlinearities, which
pose general challenges for the motion control systems, are
in addition weakly known and subject to uncertainties when
Fig. 2. Inter- and multidisciplinary of CLOVER project
dealing with hydraulic drives and mechanisms. The dead-
zones (3) (4) appear mainly due to the closed center spool in
directional control valves and therefore nearer to the con-
trolled system input. The backlash (5)∼(8) type nonlinearities
in the couplings and gearing of hydraulic motors connected
to winches are rather located in the thick of the drive-trains
and therefore more challenging for detection and proper es-
timation. Furthermore, the open-loop transfer functions of
the valve-controlled hydraulic motors and correspondingly
hydro-mechanical winch systems (9) exhibit the uncertain and
state-varying gain characteristics associated with hydraulic
volume displacement, approximated valve constant, not ide-
ally compensated differential pressure over the valve edge,
and others. The wire elasticities and varying damping of the
wire rolled on the drum constitute additional uncertainties in
the open-loop gain of the winch systems to be controlled.
Below, the weakly known dead-zone and varying open-
loop gain are addressed in more details as relevant examples
of the actuator uncertainties in offshore mechatronic systems.
3.1 Weakly known nonlinearities The dead-zone in
the valve-controlled hydraulic cylinders and motors mani-
fests itself as input nonlinearity, for which modeling and
identification either the pure static characteristics or those
coupled with internal valve dynamics can be pursued. Due
to the lack of full-order dynamic modeling and state mea-
surements in the complex chain of energy conversion of
the valves, i.e. electric-magnetic-mechanic-hydraulic, the
dead-zone nonlinearity appears as often uncertain and state-
dependent. The internal spool dynamics with implication on
the valve orifices are strongly dependent on the operation
conditions, like hydraulic pressure and flow, which in turn
can appear as dynamic quantities with uncertainties arising
out of elements of hydraulic circuits and mechanical loads.
The experimental setup of hydraulic cylinder to be used for
dead-zone analysis and estimation is shown in Fig. 3. The
extended measurements of the system include the coil cur-
rent in electromagnetic solenoid and spool displacement of
the directional control valve (DCV), on the one hand, and the
2
Examples of actuator uncertainties in environmental systems of mechatronic systems (SoMS) (Michael Ruderman et al.)
Fig. 3. Experimental setup of valve-controlled hy-
draulic cylinder with pressure and displacement measurement
Fig. 4. Experimental setup of two-inertia system
piston stroke and pressure in both chambers of the hydraulic
cylinder, on the other hand. Additional measurements of the
oil flow within hydraulic loop and temperature on the surface
of components in hydraulic circuits, like for example valves,
connectors, or cylinder self, provide further operation states
which appear significant for analysis and evaluation.
Furthermore, the methods for detection and estimation of
the dead-zone and backlash type nonlinearities can be inves-
tigated and evaluated on a two-inertia system with exchange-
able couplings and both-side encoders which provide an ex-
act reference value of incorporated nonlinearities (10). The ex-
perimental setup consists of two surface permanent magnet
synchronous motors with 20 bits high resolution encoders as
shown in Fig. 4. To evaluate the robustness of control meth-
ods against model uncertainties, flexible joints with various
torsional rigidity and inertia weights can be equipped. By
replacing a flexible coupling with a gear coupling, backlash
can be added and removed easily. The backlash width can be
directly evaluated by using two high resolution encoders.
3.2 Varying open-loop gain The hydro-mechanical
systems, like for example winches actuated by hydraulic mo-
tors (9), are often modeled through linearizing the continuity
equations of hydraulic flow and approximating the hydraulic
motor volumetric displacement, valve gain, and other opera-
tion states by the constant values. During the operation, the
assumed constants are subject to uncertainties, therefore re-
sulting in a varying open-loop gain – differing from that de-
termined by the measured frequency response functions at
the stage of system identification. Furthermore, the varying
drum radius, which depends on the state of the rolled-on wire,
and structural and contact damping of the winch bring addi-
tional uncertainties into the open-loop gain of system under
control. The weakly known leakage coefficient and total vol-
ume of hydraulic pipes and motor displacement equally fall
into the scope of varying gain characteristics when consid-
ering the open-loop transfer function of the actuated hydro-
mechanical winch system. For the quite similar reasons as
above, another types of hydraulic actuators like linear-stroke
cylinders, but also any other complex drive-trains irrespec-
tive of hydraulic or electric actuation principles, are subject
to uncertain and condition-based variations of the open-loop
gain characteristics. For designing robust and efficient feed-
forward and feedback control systems the methods for esti-
mating and coping with varying system gains are required.
4. In-wheel motor control of electric vehicle
Electric vehicles can demonstrate higher motion perfor-
mance than internal combustion engine vehicles thanks to the
following three advantages: 1) torque responses of electric
motors are more than hundred times faster as that of inter-
nal combustion engines; 2) distributed arrangement of mul-
tiple motors makes torque vectoring possible; 3) motors can
measure the output torque from their current sensors. Wheel
slip control is required to achieve high vehicle motion per-
formance. In electric vehicles, wheel slip control with high
control bandwidth can be realized thanks to the three afore-
mentioned advantages (11).
To realize advanced motion control by precisely control-
ling traction force, control methods using wheel angular ve-
locity and acceleration have been proposed (12). Angular ve-
locity and acceleration are usually calculated by angles of
resolvers equipped on motors. Wheels are connected to mo-
tors with drive shafts and reduction gears, which introduce
model uncertainties such as low resonance modes, backlash,
and friction (13) . It appears as not realistic to estimate load-
side (wheel-side) information from the motor-side resolvers
with gears’ nonlinearities (14) and unknown environmental dis-
turbances. Therefore, the research group at the University of
Tokyo has developed an in-wheel motor electric vehicle with
load-side encoders. Fig. 5 shows an electric vehicle setup
and Fig. 6 shows an in-wheel motor unit with a load-side en-
coder. As shown in Fig. 6, there is an output shaft between
the load and the load-side encoder through the hollow mo-
tor, and the motor-side encoder and the load-side encoder are
equipped side by side. The load-side encoder enables to ob-
tain precisely the wheel-side information without influences
of the gears’ nonlinearities.
Precise slip ratio estimation based on a high-resolution en-
coder has been already evaluated experimentally with on the
setup (15). High-resolution encoders can reduce quantization
errors of the angular velocity and acceleration which are cal-
culated by the angle information. The setup is to be used for
further advanced research, with aim to improve the motion
performance under gears’ nonlinearities and unknown envi-
ronmental disturbances.
3
Examples of actuator uncertainties in environmental systems of mechatronic systems (SoMS) (Michael Ruderman et al.)
Fig. 5. In-wheel motor electric vehicle with load-side
encoders
5HGXFWLRQJHDU
LQWHJUDWHG+8%EHDULQJ
5RWRU
0RWRUVLGH
HQFRGHU
:KHHOVLGH
HQFRGHU
7UDQVIHUVKDIW
8SULJKW
Fig. 6. In-wheel motor unit with load-side encoder
5. Summary
The multidisciplinary concept of highly dynamic environ-
mental applications allows addressing the topical problems of
uncertainties and disturbance rejection /attenuation in com-
plex systems of mechatronic systems (SoMS). The frame-
work of the collaborative project “CLOVER - Robust Con-
trol, State Estimation and Disturbance Compensation for
Highly Dynamic Environmental Mechatronic Systems” has
identified a number of theoretical and technological issues
that requires intensive research and innovation staffexchange
between the networked project partners. The project con-
sortium includes partners with competencies in different do-
mains of SoMS. Industrial participants are Tenneco Automo-
tive (Belgium), AVL List GmbH (Austria), Red Rock Marine
AS (Norway), and Siemens Industry Software (Belgium).
Academic partners are TU Ilmenau and FH Kempten (both
Germany), Universitetet i Agder (Norway), TU Graz (Aus-
tria), University of Leeds (UK), University of Tokyo (Japan),
and Universidad Nacional Autnoma de Mexico (Mexico). To
guarantee a strong focus of the project activities on real-
world problems, the CLOVER concept is based on the R&D
and training in three interfacing topics: “mechatronic chas-
sis systems of electric vehicles”, “mechatronic-based grid-
interconnection circuitry”, and “offshore mechatronics”.
In this paper, the following examples of the actuator un-
certainties in environmental systems of mechatronic systems
have been addressed. For the load and person transfer sys-
tems, which are widely used in offshore mechatronics and
mostly equipped with hydraulic type actuators, the uncer-
tain and state-varying dead-zone nonlinearities and open-
loop gains have been considered. For electric vehicles, which
are required to work under unknown disturbances, the gear’s
nonlinearity in the drivetrain has been addressed with apply-
ing a high-resolution encoder at the load side.
The authors intend to introduce corresponding results in
the subsequent publications.
Acknowledgement
This work has received funding from the European
Unions Horizon 2020 research and innovation programme
(H2020-MSCA-RISE-2016) under the Marie Sklodowska-
Curie grant agreement No 734832.
References
( 1 ) V. Ivanov and D. Savitski, “Systematization of integrated motion control of
ground vehicles,” IEEE ACCESS, vol. 3, pp. 2080–2099, 2015.
( 2 ) J. Woodacre, R. Bauer, and R. Irani, “A review of vertical motion heave com-
pensation systems,” Ocean Engineering, vol. 104, pp. 140–154, 2015.
( 3 ) X.-S. Wang, H. Hong, and C.-Y. Su, “Model reference adaptive con-
trol of continuous-time systems with an unknown input dead-zone,” IEE
Proceedings-Control Theory and Applications, vol. 150, no. 3, pp. 261–266,
2003.
(4) C.V
´
azquez, S. Aranovskiy, and L. B. Freidovich, “Input nonlinearity com-
pensation and chattering reduction in a mobile hydraulic forestry crane,”E&
I Elektrotechnik und Informationstechnik, vol. 133, no. 6, pp. 248–252, 2016.
( 5 ) M. Nordin and P.-O. Gutman, “Controlling mechanical systems with back-
lash – a survey,” Automatica, vol. 38, no. 10, pp. 1633–1649, 2002.
( 6 ) R. Merzouki and J. Cadiou, “Estimation of backlash phenomenon in the elec-
tromechanical actuator,” Control Engineering Practice, vol. 13, no. 8, pp.
973–983, 2005.
( 7 ) S. Villwock and M. Pacas, “Time-domain identification method for detecting
mechanical backlash in electrical drives,” IEEE Transactions on Industrial
Electronics, vol. 56, no. 2, pp. 568–573, 2009.
( 8 ) M. Ruderman, F. Hoffmann, and T. Bertram, “Modeling and identification
of elastic robot joints with hysteresis and backlash,” IEEE Transactions on
Industrial Electronics, vol. 56, no. 10, pp. 3840–3847, 2009.
( 9 ) M. B. Kjelland and M. R. Hansen, “Offshore Wind Payload Transfer Using
Flexible Mobile Crane,” Modeling, Identification and Control, vol. 36, no. 1,
pp. 1–9, 2015.
( 10) S. Yamada and H. Fujimoto, “Proposal of high back-drivable control using
load-side encoder and backlash,” in IEEE 42nd Annual Conference of the
IEEE Industrial Electronics Society, 2016, p. pp.
( 11) V. Ivanov, D. Savitski, and B. Shyrokau, “A survey of traction control and
antilock braking systems of full electric vehicles with individually controlled
electric motors,” IEEE Transactions on Vehicular Technology, vol. 64, no. 9,
pp. 3878–3896, 2015.
( 12) K. Maeda, H. Fujimoto, and Y. Hori, “Four-wheel driving-force distribution
method for instantaneous or split slippery roads for electric vehicle,” Au-
tomatika - Journal for Control, Measurement, Electronics, Computing and
Communications, vol. 54, no. 1, pp. 103–113, 2013.
( 13) M. Ruderman and M. Iwasaki, “Observer of nonlinear friction dynamics for
motion control,” IEEE Transactions on Industrial Electronics, vol. 62, no. 9,
pp. 5941–5949, 2015.
( 14 ) M. Ruderman, T. Bertram, and M. Iwasaki, “Modeling, observation, and con-
trol of hysteresis torsion in elastic robot joints,” Mechatronics, vol. 24, no. 5,
pp. 407–415, 2014.
( 15) T. Emmei, H. Fujimoto, Y. Hori, K. Omata, and D. Gunji, “Slip ratio esti-
mation for in-wheel motors with reduction gears using load-side high reso-
lution encoders,” in IEEJ Industry Applications Society Conference, vol. in
Japanese, 2016, p. pp.
4
