Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
DOCTORAL THESIS
Virtual Holonomic Constraintsfrom academic to industrial applications
DANIEL ORTÍZ MORALES
UNIVERSITETSSERVICE
Profil & CopyshopÖppettider:Måndag - fredag 10-16
Tel. 786 52 00 alt 070-640 52 01Universumhuset
Robotics and Control Lab
Department of Applied Physics and Electronics
UMEÅ UNIVERSITY, SWEDEN
Umeå, 2015
Daniel Ortíz Morales
Department of Applied Physics and Electronics
Umeå University, Sweden
SE-90748 UMEÅ
Author e-mail:[email protected]
Thesis submitted for the degree of Doctor of Philosophy (Ph.D.)
in Applied Physics and Electronics with specialization in Control Engineering.
Akademisk avhandling för avläggande av teknologie doktorsexamen (Tekn. Dr.)
i tillämpad elektronik med inriktning mot reglerteknik.
Typeset in LATEX by Daniel Ortíz Morales
Copyright c© Daniel Ortíz Morales, 2015.
ISSN:1654-5419:7
ISBN:978-91-7601-196-6
E-version available at http://umu.diva-portal.org
Printed by Print & Media, Umeå University, Umeå, January 2015
Virtual Holonomic Constraintsfrom academic to industrial applications
Daniel Ortíz Morales
Robotics and Control Lab, Umeå University, Sweden
ABSTRACT
Whether it is a car, a mobile phone, or a computer, we are noticing how automation
and production with robots plays an important role in the industry of our modern world.
We find it in factories, manufacturing products, automotivecruise control, construction
equipment, autopilot on airplanes, and countless other industrial applications.
Automation technology can vary greatly depending on the field of application. On
one end, we have systems that are operated by the user and relyfully on human ability.
Examples of these are heavy-mobile equipment, remote controlled systems, helicopters,
and many more. On the other end, we have autonomous systems that are able to make
algorithmic decisions independently of the user.
Society has always envisioned robots with the full capabilities of humans. However,
we should envision applications that will help us increase productivity and improve our
quality of life through human-robot collaboration. The questions we should be asking are:
“What tasks should be automated?”, and “How can we combine thebest of both humans
and automation?”. This thinking leads to the idea of developing systems with some level
of autonomy, where the intelligence is shared between the user and the system. Reason-
ably, the computerized intelligence and decision making would be designed according to
mathematical algorithms and control rules.
This thesis considers these topics and shows the importanceof fundamental mathe-
matics and control design to develop automated systems thatcan execute desired tasks.
All of this work is based on some of the most modern concepts inthe subjects of robotics
and control, which are synthesized by a method known as the Virtual Holonomic Con-
straints Approach. This method has been useful to tackle some of the most complex prob-
lems of nonlinear control, and has enabled the possibility to approach challenging aca-
demic and industrial problems. This thesis shows concepts of system modeling, control
design, motion analysis, motion planning, and many other interesting subjects, which can
be treated effectively through analytical methods. The useof mathematical approaches al-
lows performing computer simulations that also lead to direct practical implementations.
ii
Sammanfattning
Oavsett om det är en bil, en mobiltelefon eller en dator, märker vi hur produktionen med
robotar påverkar industrin i vår moderna värld. Det förekommer i fabriker, i anläggnings-
maskiner, som farthållare för våra fordon och som autopiloti kommersiell flygtrafik, för
att nämna några exempel.
Automation kan variera mycket beroende på användningsområdet. I ena änden har vi
standardsystem som kontrolleras av användaren och förlitar sig helt på människans för-
måga. Exempel på dessa är anläggningsmaskiner, R/C-system, helikoptrar och mycket
annat. I andra änden har vi autonoma system som kan fatta beslut oberoende av använ-
daren.
Samhället har alltid föreställt sig robotar med den fulla kapaciteten hos människor.
Vi bör dock istället föreställa oss tillämpningar som hjälper oss att öka produktiviteten
och förbättra vår livskvalitet genom människa-robot-samarbete. De frågor vi bör ställa
är: “Vilka uppgifter bör automatiseras?”, och “Hur kan vi kombinera det bästa från både
människor och automation?”. Detta leder till tanken att skapa maskiner med viss grad av
autonomi, där intelligens definieras av en uppsättning matematiska algoritmer och kon-
trollregler.
Denna avhandling behandlar användandet av matematiska modeller och reglerdesign,
så att ett automatiserat system kan utföra förutbestämda arbetsuppgifter. Allt detta arbete
har baserats på ett flertal moderna koncept i ämnena roboteknik och reglerteknik som syn-
tetiserats genom metoden ”virtuella holonoma bivillkor” (virtual holonomic constraints).
Denna metod har varit användbar för att ta itu med några av de mest komplexa problemen
inom olinjär reglering, och har gjort det möjligt att närma sig industriella problem som
annars är svåra att angripa. Genom avhandlingen visas begreppen systemmodellering,
reglerdesign, rörelseanalys, rörelseplanering, etc., som kan behandlas effektivt genom
analysmetoder. Detta tillåter utförandet av datorsimuleringar som också leder till prak-
tiska tillämpningar.
iii
iv
v
To my parents Raquel and Daniel and my sister Raquel, with gratitude, appreciation, and
love. They have given me their love, unequivocal support throughout.
In memory of Joshua David Flanders
October 10, 1980 – April 14, 2011
vi
Preface
“A new scientific truth is usually not
propagated in such a way that
opponents become convinced and
discard their previous views. No, the
adversaries eventually die off, and the
upcoming generation is familiarized
anew with the truth.”
Max Planck
Parts of the contributions presented in this thesis have previously been accepted to
conferences or submitted to journals. A list of these publications is given below.
⊲ Daniel Ortíz Morales, P. La Hera and S. Ur Rehman. “Generating Periodic
Motions for the Butterfly Robot,” inProc. IEEE International Conference on
Intelligent Robots and Systems, pp. 2527–2532, Tokyo, Japan, 3–7 Novem-
ber 2013. c© 2013 IEEE.
⊲ Daniel Ortíz Morales and Pedro X. La Hera, “Design of energy efficient
walking gaits for a three-link planar biped walker with two unactuated de-
grees of freedom,” inProc. IEEE International Conference on Robotics and
Automation, pp. 148–153, Saint Paul, Minnesota, USA, 14–18 May 2012.
c© 2012 IEEE.
⊲ Pedro X. La Hera, andDaniel Ortíz Morales “Non-linear dynamics mod-
elling description for simulating the behavior of forestrycranes,”Interna-
tional Journal of Modelling, Identification and Control, Vol. 21 No. 2
pp.125–138, 2014, DOI: 10.1504/IJMIC.2014.060006,c© 2014 Inderscience
Publishers.
http://inderscience.metapress.com/content/XX51264172M1U31W
vii
viii Preface
⊲ Pedro X. La Hera, andDaniel Ortíz Morales “Designing and testing control
systems for forestry cranes,”(Submitted Manuscript).
⊲ Daniel Ortíz Morales, Simon Westerberg, Pedro X. La Hera, Uwe Mettin,
Leonid B. Freidovich, and Anton S. Shiriaev “Increasing thelevel of au-
tomation in the forestry logging process with crane trajectory planning and
control,” in Journal of Field Robotics, Vol. 31 Issue 3, pp. 343–363, DOI:
10.1002/rob.21496,c© 2014 Wiley Blackwell.
⊲ Daniel Ortíz Morales, Pedro X. La Hera, Simon Westerberg, Leonid B. Frei-
dovich, and Anton S. Shiriaev “Path-constrained motion analysis. An al-
gorithm to understand human performance on hydraulic manipulators,”Ac-
cepted as a regular paper in the IEEE Transactions on Human-Machine Sys-
tems.
List of Publications
A list of all contributions published by the author is given below.
1. U. Mettin, P. La Hera, D. Ortíz Morales, A. Shiriaev, L. Freidovich, and S. Wester-
berg.Trajectory planning and time-independent motion control for a kinemat-
ically redundant hydraulic manipulator . In Proc. 14th International Conference
on Advanced Robotics, pp. 1–6, Munich, Germany, June 2009.
2. D. Ortíz Morales, P. La Hera, U. Mettin, L. Freidovich, A. Shiriaev, and S. West-
erberg. Steps in trajectory planning and controller design for a hydraulically
driven crane with limited sensing. In Proc. International Conference on Intelli-
gent Robots and Systems, pp. 3836–3841, Taipei, Taiwan, October 2010.
3. D. Ortíz Morales, S. Westerberg, P. La Hera, U. Mettin, L. Freidovich, and A. Shiri-
aev. Open-loop control experiments on driver assistance for crane forestry
machines. In Proc. International Conference on Robotics and Automation, pp.
1797–1802, Shanghai, China, May 2011.
4. D. Ortíz Morales and P. La Hera.Design of energy efficient walking gaits for
a three-link planar biped walker with two unactuated degrees of freedom. In
Proc. International Conference on Robotics and Automation, pp. 148–153, Saint
Paul, Minnesota, USA, May 2012.
ix
5. D. Ortíz Morales and P. La Hera.Design of stable walking gaits for biped robots
with several underactuated degrees of freedom.. In Proc. Dynamic Walking
Conference, pp. 1–2, Pensacola, Florida, USA, May 2012.
6. P. La Hera and D. Ortíz Morales.Modeling dynamics of an electro-hydraulic
servo actuated manipulator: A case study of a forestry forwarder crane. In
Proc. World Automation Congress, pp. 1–6, Puerto. Vallarta, Mexico, June 2012.
7. P. La Hera, B. Ur Rehman and D. Ortíz Morales.Electro-hydraulically actuated
forestry manipulator: Modeling and Identification . In Proc. International Con-
ference on Intelligent Robots and Systems, pp. 3399–3404, Vilamoura, Algarve,
Portugal, October 2012.
8. D. Ortíz Morales, P. La Hera and S. Ur Rehman.Generating periodic motions
for the butterfly robot . In Proc. International Conference on Intelligent Robots
and Systems, pp. 2527–2532, Tokyo, Japan, November 2013.
9. P. La Hera and D. Ortíz Morales.Non-linear dynamics modelling descrip-
tion for simulating the behavior of forestry cranes. International Journal of
Modelling, Identification and Control., Vol. 21 No. 2 pp.125–138, 2014, DOI:
10.1504/IJMIC.2014.060006.
10. D. Ortíz Morales, S. Westerberg, P. La Hera, U. Mettin, L.Freidovich, and A. Shiri-
aev. Increasing the level of automation in the forestry logging process with
crane trajectory planning and control. Journal of Field Robotics, Vol. 31, Issue
3, pp. 343–363, 2014, DOI: 10.1002/rob.21496.
11. P. La Hera and D. Ortíz Morales.Designing and testing control systems for
forestry cranes. Preliminary reviewed:IEEE Transactions Journal on Control
Systems Technology.
12. D. Ortíz Morales, P. La Hera, S. Westerberg, L. Freidovich, and A. Shiriaev.Path-
constrained motion analysis. An algorithm to understand human performance
on hydraulic manipulators. Accepted as a regular paper:IEEE Transactions
Journal on Human–Machine Systems.
x Preface
Acknowledgments
“You have to give yourself credit, not
too much because that would be
bragging.”
Frank McCourt
First and foremost, I would like to thank God for providing meguidance and wisdom
during my time in earth. I thank my parents for all their love and support.
There is a large amount of people without whom this thesis would have not been
as successful, and to whom I am greatly indebted. I would liketo express my special
appreciation and gratitude to my supervisor Professor Anton Shiriaev for encouraging me
during my research time and for allowing me to grow as a researcher. I would also like
to thank my co-supervisor Dr. Leonid Freidovich for all the patience, support and time
he spent helping me understanding even the most simple concepts. Your personal and
professional advice have been priceless.
I am very grateful with all the staff of TFE, and with all the people from the Robotics
Control lab who became my colleagues and friends, and contributed greatly in my devel-
opment. My special thanks and admiration to Drs. Pedro La Hera, Uwe Mettin and Simon
Westerberg with whom I had uncountable discussions, received several corrections, and
developed several projects. I would also like to thank my good friend and office mate Mr.
Szabolcs Fodor, it was really enjoyable to share office with you.
And last but not least, I would like to thank to my country Mexico, to Sweden that
has become my new home, my sister, all my friends, relatives,teachers, partner, and all
others who externally influenced my work in many ways. I am really sorry that I cannot
mention you all.
Daniel Ortíz Morales
xi
xii Acknowledgments
Contents
Abstract i
Preface vii
Acknowledgments xi
I Comprehensive Summary of Contributions 1
1 Introduction 1
1.1 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Summary of contributions 5
2.1 Butterfly Robot (Paper I) . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Three-link Planar Biped Walker (Paper II) . . . . . . . . . . . .. . . 9
2.3 Hydraulic manipulators (Papers III-VI) . . . . . . . . . . . . .. . . . 16
3 Concluding remarks 25
II Papers 29
I Generating Periodic Motions for the Butterfly Robot. 33
I.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
I.2 Model of the Butterfly Robot . . . . . . . . . . . . . . . . . . . . . . 35
I.2.1 Reaction Force . . . . . . . . . . . . . . . . . . . . . . . . 37
I.3 Trajectory Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
I.4 Examples of constraint functions . . . . . . . . . . . . . . . . . . .. 40
xiii
xiv Contents
I.5 Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
I.5.1 Dynamics along the target orbit and transverse coordinates . 43
I.5.2 Partial feedback linearization . . . . . . . . . . . . . . . . . 44
I.5.3 Integral of the reduced dynamics . . . . . . . . . . . . . . . 44
I.5.4 Coordinates measuring the distance . . . . . . . . . . . . . . 45
I.5.5 Transverse Linearization . . . . . . . . . . . . . . . . . . . 45
I.6 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
I.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
I.7.1 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 48
II Design of energy efficient walking gaits for a three-link planar biped walker
with two unactuated degrees of freedom. 53
II.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
II.2 Model of the biped walker . . . . . . . . . . . . . . . . . . . . . . . 55
II.2.1 Swing Phase Model . . . . . . . . . . . . . . . . . . . . . . 55
II.2.2 Impact Model . . . . . . . . . . . . . . . . . . . . . . . . . 56
II.3 Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . 57
II.4 Gait Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
II.4.1 Re-parametrization of the trajectories . . . . . . . . . .. . . 58
II.4.2 Equations for constraint functions . . . . . . . . . . . . . .59
II.4.3 Defining a function forφ2 . . . . . . . . . . . . . . . . . . . 61
II.4.4 Optimization procedure to find periodic gaits . . . . . .. . 62
II.4.5 Example of a Gait . . . . . . . . . . . . . . . . . . . . . . . 63
II.5 Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
II.5.1 Dynamics along the target orbit and transverse coordinates . 64
II.5.2 Partial feedback linearization . . . . . . . . . . . . . . . . .65
II.5.3 Integral of the reduced dynamics . . . . . . . . . . . . . . . 66
II.5.4 Coordinates measuring the distance . . . . . . . . . . . . . .66
II.5.5 Hybrid Transverse Linearization . . . . . . . . . . . . . . . 67
II.5.6 Feedback controller design . . . . . . . . . . . . . . . . . . 67
II.5.7 Control Design Simulations . . . . . . . . . . . . . . . . . . 68
II.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
III Non-linear dynamics modelling description for simulat ing the behavior of
forestry cranes. 73
III.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Contents xv
III.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
III.3 Modelling the system dynamics . . . . . . . . . . . . . . . . . . . .76
III.3.1 Kinematic modeling . . . . . . . . . . . . . . . . . . . . . . 76
III.3.2 Modelling rigid-body dynamics . . . . . . . . . . . . . . . . 78
III.3.3 Modelling friction forces . . . . . . . . . . . . . . . . . . . 79
III.3.4 Combining the models of dynamics and friction forces . . . 80
III.3.5 Mapping cylinder forces to joint torques . . . . . . . . .. . 81
III.3.6 Modelling dynamics of the hydraulic system . . . . . . .. . 84
III.4 Calibration and estimation of unknown model parameters . . . . . . . 88
III.4.1 Parameter estimation procedure for the model (III.13) . . . . 88
III.4.2 Parameter estimation procedure for the hydraulic model (III.38) 91
III.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
IV Designing and testing control systems for forestry cranes. 99
IV.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
IV.1.1 Overview of forestry machines . . . . . . . . . . . . . . . . 100
IV.1.2 Overview of control of forestry cranes . . . . . . . . . . . .101
IV.1.3 Overview of Model-Based Design as a tool to go from re-
search to product development. . . . . . . . . . . . . . . . . 102
IV.1.4 Paper organization . . . . . . . . . . . . . . . . . . . . . . . 103
IV.2 System modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
IV.2.1 Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
IV.2.2 Friction forces . . . . . . . . . . . . . . . . . . . . . . . . . 106
IV.2.3 Relation between hydraulic force and mechanical torque . . 106
IV.2.4 Hydraulic cylinder dynamics . . . . . . . . . . . . . . . . . 107
IV.2.5 Electro-hydraulic valve . . . . . . . . . . . . . . . . . . . . 108
IV.2.6 Hydraulic force as a function of the control input . . .. . . 110
IV.2.7 Mechanical motion as a function of the control input .. . . 111
IV.3 Nonlinear control design . . . . . . . . . . . . . . . . . . . . . . . . 111
IV.3.1 First-order sliding mode control . . . . . . . . . . . . . . . 112
IV.3.2 Second-order sliding mode control . . . . . . . . . . . . . . 113
IV.4 Controller implementation . . . . . . . . . . . . . . . . . . . . . . .114
IV.4.1 Model calibration . . . . . . . . . . . . . . . . . . . . . . . 115
IV.4.2 Reducing the sliding-mode chattering . . . . . . . . . . . .122
IV.4.3 Mathematical considerations . . . . . . . . . . . . . . . . . 123
IV.5 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 123
xvi Contents
IV.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
V Increasing the level of automation in the forestry loggingprocess with crane
trajectory planning and control. 133
V.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
V.1.1 Literature Review . . . . . . . . . . . . . . . . . . . . . . . 136
V.1.2 General assumptions . . . . . . . . . . . . . . . . . . . . . 137
V.2 Model of the manipulator . . . . . . . . . . . . . . . . . . . . . . . . 139
V.3 Motion planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
V.3.1 Path Planning . . . . . . . . . . . . . . . . . . . . . . . . . 142
V.3.2 Path-Constrained Trajectory Planning . . . . . . . . . . . .144
V.3.3 Time-Efficient Trajectory . . . . . . . . . . . . . . . . . . . 145
V.4 Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
V.4.1 Dead-band compensation . . . . . . . . . . . . . . . . . . . 147
V.4.2 Velocity estimation . . . . . . . . . . . . . . . . . . . . . . 148
V.4.3 Individual joint PID controller . . . . . . . . . . . . . . . . 149
V.4.4 Hydraulic control . . . . . . . . . . . . . . . . . . . . . . . 150
V.4.5 Open-loop control . . . . . . . . . . . . . . . . . . . . . . . 150
V.5 Human-machine interface . . . . . . . . . . . . . . . . . . . . . . . . 152
V.6 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 153
V.6.1 Laboratory Crane . . . . . . . . . . . . . . . . . . . . . . . 154
V.6.2 Commercial Crane . . . . . . . . . . . . . . . . . . . . . . 155
V.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
V.8 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
V.8.1 Installation of Encoders . . . . . . . . . . . . . . . . . . . . 159
V.8.2 Pressure Sensors . . . . . . . . . . . . . . . . . . . . . . . . 159
V.8.3 Chair Signals . . . . . . . . . . . . . . . . . . . . . . . . . 159
V.8.4 Installing Prototyping Hardware into the Cabin . . . . .. . 160
VI Path-constrained motion analysis. An algorithm to understand human per-
formance on hydraulic manipulators. 167
VI.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
VI.1.1 Literature review related to methods applied for analyzing
performance of forestry operators. . . . . . . . . . . . . . . 170
VI.2 Experimental platform. . . . . . . . . . . . . . . . . . . . . . . . . . 171
VI.3 Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Contents xvii
VI.4 Finding a nominal motion. . . . . . . . . . . . . . . . . . . . . . . . 176
VI.5 Analyzing crane patterns generated by human commands .. . . . . . 182
VI.5.1 Path-constrained motion analysis algorithm . . . . . .. . . 183
VI.5.2 Analysis based on recorded data . . . . . . . . . . . . . . . 185
VI.6 Defining machine settings and reassigning velocity profiles . . . . . . 186
VI.6.1 Suggesting a procedure for customizing machine settings . . 187
VI.6.2 Reassigning the velocity profile. . . . . . . . . . . . . . . . 189
VI.7 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
VI.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
VI.9 Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
VI.9.1 Forward Kinematics . . . . . . . . . . . . . . . . . . . . . . 193
VI.9.2 Computation of inverse kinematics . . . . . . . . . . . . . . 194
VI.10 Data processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
VI.10.1 Extracting the time intervals of trajectories . . . .. . . . . . 195
VI.10.2 Removing offsets in the vectors of time . . . . . . . . . . .195
VI.10.3 Computing the boom-tip Cartesian path . . . . . . . . . . .195
VI.10.4 Spline approximation . . . . . . . . . . . . . . . . . . . . . 195
VI.10.5 Estimating derivatives . . . . . . . . . . . . . . . . . . . . . 196
VI.10.6 Resizing the data set . . . . . . . . . . . . . . . . . . . . . . 196
Bibliography 197
xviii Contents
Part I
Comprehensive Summary of
Contributions
1 Introduction
“Most of the fundamental ideas of
science are essentially simple, and
may, as a rule, be expressed in a
language comprehensible to everyone.”
Albert Einstein
Humans have evolved through history from using rudimentarytools to the use of
highly sophisticated robots. Systems of these kind are making a considerable impact on
many aspects of modern life. From industrial manufacturing, healthcare, entertainment,
human assistance, domestic help, exploration of other planets, and more, we can see how
robots occupy an important place in our modern world (Fig. 1.1).
Although robotics and automation technologies are quite advanced today, they are
still very limited. In most cases robots are designed to accomplish a very specific set of
monotonous tasks in constrained environments to avoid accidents with human workers.
In other cases, robots are designed with some level of autonomy to operate in outdoor
environments through a mixture of human commands and semi-autonomous functions.
However, it is difficult to find robotic systems with full autonomy, resembling the decision
making ability of biological systems found in nature. Therefore, the subject of robotics is
still at its infancy, but we are starting to see all the possible benefits that it can bring to a
variety of industries.
The development of any type of robotic system involves threeimportant aspects:1)
the design of the robot itself, which consists of the mechanics, actuators, electronic hard-
ware, etc.,2) the control algorithms that will ensure that the performance of the robot
goes according to what is expected, and3) planning the desired tasks. Because these are
important subjects in this thesis, some preliminary concepts are provided below.
Thanks to the improvement of processing ability of today’s computers, there are var-
ious aspects of design that engineers are able to manage through computer–aided tools
1
2 1. Introduction
(a) Industrial robot, from ABB (b) PARO, from AIST (c) Groundbot, from Rotundus
(d) HAL-5 exoskeleton, from Cy-berdyne
(e) HRP-4, robot from AIST (f) Robonaut, from NASA
Fig. 1.1: Examples of different types of robots
and simulation technology. This allows to design cost effective solutions and verify their
functionality without the need of having to build a system first. Designing systems in
this form often leads to the need for formulating dynamical models based on physics, ob-
served data, or both, to generate mathematical models that can be conveniently solved by
computer–based numerical methods. Today, there are a lot ofways in which we use this
form of numerical solutions. For example, to design simulators that help to train people,
and also to study the design of different systems such as shuttles, cars, or machines, to
mention some.
In the particular field of control systems, model–based computer–aided design tools
are used for two reasons: to design feedback control algorithms and study their perfor-
mance through simulation tests. Applying this approach in robotics is useful, because this
allows to test different control algorithms and easily transition to real-time implementa-
tion. Understanding the performance of these algorithms isimportant for understanding
1.1. Outline 3
the system’s behavior when executing desired processes, tasks, or motions.
Motion is a behavior that can be planned through a variety of formal methods. To
prescribe the way in which a system will move, be it a vehicle,airplane, legged robot,
manipulator, etc., three important details are needed:
1. the possible path to take,
2. the internal and external constraints such as velocity limitations, obstacles, etc.,
3. the desired requirement; for example, the interest of performing movement in the
least amount of time or with the minimum energy, etc.
To effectively manage the problem of planning some motion, we need to consider the
knowledge of what is really required, so that we can select a particular design that falls
into some specified requirements.
The following thesis presents work that covers many of theseconcepts. In essence,
it presents analytical concepts that can be applied in the subjects of motion/trajectory
planning, motion/trajectory analysis, and control design. Much of this literature is based
on mathematical models that are derived from first principlelaws of physics, as well
as topics that are studied in nonlinear control. Examples ofthese are Euler-Lagrange,
Newtons’ laws, and many more.
One of the characteristics of this work is that it does not concentrate on one example,
but instead, it covers a variety of applications. They fall in the categories of under-actuated
mechanical systems, legged locomotion, industrial machines, and analysis of human abil-
ities. What is peculiar about this work is the method that is used to accomplish all of these
studies. Today, it is known as the Virtual Holonomic Constraints Approach (VHC) [1–3]
and is a procedure that my research group has been developingfor a long time. As it will
be seen, this method can be applied in a variety of challenging research disciplines, as
well as industrial applications.
1.1 Outline
The thesis is organized as follows: The first part is formed bythree chapters containing
an introduction, a summary of contributions, and concluding remarks. The second part is
composed by six papers organized in the following way:
4 1. Introduction
Papers I and II
These two papers study the problems of trajectory planning and control design for under-
actuated systems. In both cases, mathematical modeling, trajectory generation and control
design are presented. The system in paper I is a case of under-actuation degree one, while
the system in paper II is a more challenging example of under-actuation degree two.
Papers III – VI
These papers study an industrial application targeting hydraulic forestry cranes. In paper
III a mathematical model of such machines is derived. Paper IV suggests a framework for
designing non-linear motion controllers that have been verified and tested experimentally.
Paper V presents some insights into the automation problemsof these systems. Finally,
paper VI develops a procedure to analyze the performance of machine operators through
concepts of motion planning techniques.
2 Summary of contributions
“If someday they say of me that in my
work I have contributed something to
the welfare and happiness of my fellow
man, I shall be satisfied.”
George Westinghouse
This chapter contains a brief summary of the contributions presented in each of the
papers from this thesis
2.1 Paper I. Generating Periodic Motions for the Butter-
fly Robot
Many systems in nature are inherently under-actuated, because they have fewer actuators
than the number of degrees of freedom (DOF). Yet, they are able to produce complex
motions by smartly coordinating their movements (see e.g. Fig. 2.1). This characteristic
has inspired the design of different systems used in a variety of applications. Airplanes,
vehicles, vessels, submarines, and satellites, are a few ofthese examples.
In robotics, under-actuation is interesting because it allows designing systems with
fewer actuators than needed. When the mechanical design is planned carefully, this helps
to reduce the weight produced by heavy motors and gears. However, reducing the amount
of actuators also increases the difficulty of controlling the system. This poses different
scientific challenges to the problems of designing controllers for motion control and plan-
ning desired motion.
To study problems of this kind, several researchers in the subject of nonlinear control
have suggested to use simplified versions of complex under-actuated mechanical struc-
tures. Classical benchmark examples are based on pendulum systems among which, the
5
6 2. Summary of contributions
(a) Jugglers (b) Seagull
(c) Jumping diver (d) Seller
Fig. 2.1: Examples of under-actuation in daily life1.
Furuta pendulum [4, 5], the inertia wheel pendulum [6, 7], and the Acrobot [8, 9] are
quite popular. These systems have been studied to analyze challenging control problems
among which, stabilizing the system at an unstable equilibrium or performing a swinging
up maneuver are the most well known [5, 8, 10]. Such studies have helped to elaborate
many fundamental concepts to approach more complex problems, as for example, those
encounter in biologically inspired legged and flying machines.
In recent years, there has been an increasing interest to understand the problem of
dynamic object manipulation, also known as non-prehensilemanipulation. This is some-
thing we experience quite often when we see, for example, a juggler on the streets per-
forming contact juggling, or when we see a waitress carryingglasses on a tray with a
single hand. In fact, maneuvers of this kind require great coordination, a lot of practice,
and they are difficult even for the most skilled.
To study the fundamental concepts of this problem, a group ofresearchers from North-
western University proposed some simplified under-actuated systems. In particular, Prof.
1Sources:jugglers,http://www.scottbot.net/HIAL/?page_id=177seagull,http://andrewchen.co/2013/01/14/confessions-of-a-startup-seagull/diver,http://www.fmn.org.mx/w/ESP/galerias/clavados/clavados/2627seller, http://scotteagan.blogspot.se/2013/09/how-your-conflict-is-solved-will-make.html
2.1. Butterfly Robot (Paper I) 7
Lynch introduced the butterfly robot [11], which is a system with two DOF and one ac-
tuator. It is formed by coupling two plates that form a track where a ball moves freely.
These plates have the shape of a butterfly, hence the name (seeFig. 2.2). This setup is
under-actuated because the coupling is actuated by an electric motor, but the ball has free
motion as a consequence of the movement of the plates and gravity.
Fig. 2.2: Butterfly robot.
Some of the specific scientific challenges that can be analyzed with the butterfly robot
are:
1. Stabilizing the ball around different equilibrium points according to the butterfly
shape.
2. Swinging the ball from its resting position to one of theseequilibrium.
3. Making the ball follow a predefined periodic or finite-timemotion while preserving
its contact with the plates.
4. Catching the ball thrown to the plates.
5. Throwing the ball in planned ways.
6. Catching and throwing the ball.
7. Making the ball bounce ata) a point of the butterfly shape,b) in between different
points, or more radicallyc) among different butterfly systems.
8 2. Summary of contributions
8. Doing all of the above with more than one ball.
Although this list is not exhaustive, most of these challenges still remain, because
the problem of having an under-actuated link and non-holonomic mechanics makes them
quite challenging in both theory and practice. Most of thesechallenges can be treated as
the essential problem of controlling motion under constrained contact forces, and limited
control inputs.
To the best of our knowledge, two approaches have been proposed for this particular
setup:a) an optimization-based motion planning under proportionalderivative (PD) feed-
back control [11], andb) passivity based control [12, 13]. The former shows a peculiar
solution done to rotate the plates180 degrees and roll a disk to the other side without los-
ing contact (a disk is used instead of a ball). The latter considers the swinging-up problem
with a subsequent stabilization of the ball.
The article I am presenting for this example intends to propose a solution to the third
problem, which consists of designing periodic movement that resembles a permanent
juggling motion. However, readers will realize that a variation of similar method can
also be used to perform any desired finite movement. To understand the problem of
controlling motion, it can be imagined that to move the ball in some desired way we have
to smartly maneuver the plates to produce such a movement. This involves a combination
of planning how the butterfly is going to move to produce a natural, graceful motion of
the ball and the way such behavior is kept permanently stablethrough feedback control.
Although the material in this article is new, it comes from theory and concepts that
have been an ongoing development within my research group [1, 14] and have shown to
be quite successful in both theory and practice [15–18]. On the basis of these concepts,
we suggest to select VHC to specify the way in which the butterfly will move, and find
analytically all possible trajectories that are feasible for the ball to follow while keeping
contact with the plates. This solution has a number of advantages in comparison to other
approaches, because, most of all, it is analytical. This fundamental characteristic makes
it simple to apply and derive explicit solutions, rather than running extensive numerical
optimization search.
In summary, this article presents the following:a) a complete system modeling in-
cluding the calculation of reaction forces,b) a step-by-step procedure for selecting VHC
that allow planning feasible ball trajectories, andc) an exemplified approach for control
design of mechanical systems with under-actuation degree one.
The results presented are based on simulation studies underMatlab/Simulink. To this
end, we present two types of periodic motions that depend on the placement of the but-
2.2. Three-link Planar Biped Walker (Paper II) 9
terfly plates and are depicted in Fig. 2.3 and Fig. 2.4. These results show how the ball
exponentially converges to the desired cycle from different starting positions. This proce-
dure can be extended to design more sophisticated motions, e.g. perpetual revolutions of
the ball or trajectories to move the ball from the horizontalposition to the other end, etc.
(a) (b) (c)
Fig. 2.3: Butterfly juggling around its horizontal position
(a) (b) (c)
Fig. 2.4: Butterfly juggling around its vertical position
2.2 Paper II. Design of energy efficient walking gaits for
a three-link planar biped walker with two unactuated
degrees of freedom
In the particular field of biped robotics, or legged locomotion in general, there exists a
substantial amount of work devoted to understand the way biologically inspired walking
or running machines can be designed and controlled. Unlike traditional robotics, this is
also a research area that has grown aside from the standard field and today incorporates
elements of different subjects that include bio-mechanics, motion capturing, bio-inspired
mechanical engineering, prosthetic design, just to mention a few. Because of this reason
certain terminology can be quite different to what is known in traditional robotics or con-
trol system engineering. For instance, while it is often common to call motion/trajectory
10 2. Summary of contributions
planning to the task of specifying the way a given system willmove, for research in legged
locomotion, this is often referred to as gait synthesis. Theterm gait expresses a behavior
that mathematically corresponds to a limit cycle or periodic trajectory. It can be designed
empirically or by observing data recorded with motion capture systems. Likewise, unlike
formal control system terminology, some authors use the term stability to refer to dynamic
balance during motion, and the actual analysis of stabilitymight require more concepts
than that of Lyapunov’s stability theorems. In the text below, some of this terminology
will be adopted, and will also be read in the article supporting this section.
As pointed out, different methods exist to approach the problem of legged locomotion.
Today, they have been roughly classified into two categoriesand these are described in
literature as static and dynamic walking.
In the case of static walking, a conventional approach consists on designing robots
with nearly rigid links, motors and gearboxes. Applying concepts of traditional feedback
control, it is possible to make the robot’s body to stay standing upright keeping dynamic
balance. To design such controllers, we need to make the assumption that the robot’s
body can be reduced to a3D pendulum for simplifying the modeling complexity [19].
Applying high-gain feedback control that produces high joint torques, the controller can-
cels out the natural dynamics of the machine and the links canstrictly follow some desired
trajectories. This can result in behaviors that are similarto walking, but they can be un-
natural compared to their biological counterpart. This approach is widely known as the
zero-moment point (ZMP) method and it should be noted that despite the limitation that
the motion has to be carefully synthesized to preserve the ZMP condition, this method has
remained for a long time the only procedure for biped gait synthesis and control. Today,
this concept has been involved in diverse applications related to numerous anthropomor-
phic locomotion mechanisms of different degrees of complexity, such as ASIMO from
Honda [20], who resembles an astronaut walking with a space suit (see Fig. 2.6(a)). In
setups like ASIMO, the motions are carefully planned, very smooth, but highly unnatural.
Today, most of the biped mechanism joints are powered and directly controlled except
for the contact between the foot and the ground, which can be considered as an additional
passive DOF. The foot cannot be controlled directly, but indirectly by ensuring the ap-
propriate dynamics of the mechanism above the foot. Therefore, the trend of dynamic
walking is more recent and has been inspired on the understanding of these basic charac-
teristics of biped locomotion: (i) the possibility of rotation of the overall system about one
foot, which is equivalent to the appearance of an unpowered (passive) DOF, (ii) gait re-
peatability (symmetry), and (iii) natural interchangeability of single- and double-support
2.2. Three-link Planar Biped Walker (Paper II) 11
phases. Based on these concepts, in 1990 Tad McGeer [21] showed how a planar mech-
anism with two legs could dynamically walk stably down a shallow slope with no other
energy input or control than gravity alone2. This system acts like two coupled pendu-
lums, where the stance leg acts like an inverted pendulum andthe swing leg acts like a
free pendulum attached to the stance leg at the hip. Given sufficient mass at the hip, the
system will have a stable limit cycle, which is a nominal trajectory that repeats itself and
will return to this trajectory even if perturbed slightly. An extension of the two-segment
passive walker is to include knees (see Figs. 2.5 and 2.6(b)), which results in human-like
walking behavior3.
swing leg
pendulum
stance leg
inverted
pendulum
push-o�
trunk
velocity body
weight
support
collision
Dynamic Walking
Fig. 2.5: Passive dynamic walking4.
Mechanisms that are able to walk dynamically without the need of control input are
highly efficient in terms of energy consumption. Therefore,today this interesting concept
is studied deeply and classical research examples are the rimless-wheel [22], the compass-
biped [23], the kneed-biped [24], and different version of the compass-biped with upper
torso [25]. Apart of energy efficiency, the phenomena that attracts researchers to this
idea is the human-like walking behavior produced by these systems5. Most importantly,
all of these characteristics are the result of purely natural dynamics, which motivates the
importance of how the mechanisms and actuation of legged machines should be designed
to make more graceful and energy efficient legged machines.
From the theoretical aspects, authors of [21, 26–29] have laid concrete, fundamental
concepts to understand modeling assumptions, gait synthesis, and simulation examples of
2A video describing the ideas behind passive dynamic walking by the man who first discovered the idea: TadMcGeer.https://www.youtube.com/watch?v=WOPED7I5Lac
3A video showing this setup can be seen inhttps://www.youtube.com/watch?v=CK8IFEGmiKY4Source:
http://dyros.snu.ac.kr/concept-of-passive-dynamic-walking-robot/5A video comparison in between the gait of a passive kneed-walker and a person can be seen inhttp:
//drei.mech.nitech.ac.jp/~sano/contents/biped/Movie_W04.mpg
12 2. Summary of contributions
passive walking behaviors. This has been greatly important, because this has opened up
the possibility to expand the domain of passive walkers and make systems walk on flat
surfaces and through rough terrain by letting them move by their own natural dynamics.
To accomplish this, recent designs combine active and passive joints, which result in
under-actuated systems. Some examples are Flame6, the Rabbit robot [30], MABEL
robot [31], MARLO [32] and most recently, the line of robots from Boston Dynamics
[33]. Some of these examples are impressive machines that are able to walk dynamically
even without feet.
(a) ASIMO, image courtesy Honda[20]
(b) Passive walker, image courtesyNagoya Institute of Technology
(c) Petman Robot, image courtesyof Boston Dynamics [33]
Fig. 2.6:3 Examples of different types of walking robots7.
Conceptually, the trend of robots that are designed with passive dynamics fall in the
class of under-actuated mechanical systems, because not all DOF are directly controllable.
As presented in [21, 34], dynamics of these systems can be modeled applying concepts
of classical mechanics, which particularly involve Newton’s laws, Euler-Lagrange, and
impact events. This combination yields discontinuous dynamics, because it is considered
that the robot’s foot impacting the ground at the end of a stepcan be treated as an impulse
effect (see Fig. 2.5). In the case of running, this can also incorporate a short flying time,
6A video of the dutch robot Flame can be seenhttps://www.youtube.com/watch?v=7JU_zQkVOiI
7Sources:ASIMO, http://asimo.honda.com/ASIMO_DCTM/News/images/highres/ASIMO_one_hand_up.jpgBlue biped,http://drei.mech.nitech.ac.jp/~fujimoto/sano/movie/BlueBiped.jpg
2.2. Three-link Planar Biped Walker (Paper II) 13
which happens when both feet are off the ground. This form of dynamical systems falls
in the class of hybrid dynamical systems, which present various scientific challenges:
• performing gait synthesis given under-actuated hybrid dynamics, which is related
to the problem of trajectory planning and consists of designing stable hybrid-limit
cycles, either for fully passive systems, or under-actuated ones,
• designing robust controllers that can ensure the walking stability given perturba-
tions and unmodeled dynamics,
• analyzing stability of walking behaviors, which consist onverifying the stability of
a gait either with passive or active control,
• walking on uneven terrain by adapting the gait according to obstacles, which could
also consists on modifying the control law parameters adaptively,
• repeating the same for running.
Although this list is not exhaustive, it reflects some of the concepts studied in dynamic
walking. Today, most of these items are still challenging topics and proof of this has been
publicly observed during the latest DARPA ROBOTICS CHALLENGE in December of
2013 [35]. In this event, some of the most advanced humanoid robots were presented by
highly ranked institutions that include MIT, CMU, KAIST, NASA and many more. All of
these robots showed the difficulties found in today’s technology to perform simple human
tasks such as walking on uneven terrains8. Many of them lacked the ability to finish the
walking assignments and most of them could perform this taskonly by being remotely
controlled. However, there exists a number of robots that did not participate in this event,
as it is the case of a family of hydraulically-actuated legged robots, such as BigDog,
LS3, WildCat, Petman (see Fig. 2.6(c)) and ATLAS9, all belonging to Boston Dynamics
[33]. In video examples shown by this company, both BigDog10 and LS311 show the
outstanding ability to dynamically balance and walk in unplanned natural rough terrain
environments, which is also true for the biped version ATLAS12. While these robots are
8A video of the challenge related to walking in uneven terraincan be seen inhttps://www.youtube.com/watch?v=RgG1GCgShZQ
9while ATLAS was provided as a platform for the DARPA challenge, it was never operated by its designerBoston Dynamics
10A video showing BigDog’s abilities can be seen inhttps://www.youtube.com/watch?v=W1czBcnX1Ww
11A video showing LS3 abilities can be seen inhttps://www.youtube.com/watch?v=hNUeSUXOc-w
12A video of ATLAS walking on uneven terrain can be seen inhttps://www.youtube.com/watch?v=WYKgHa8hH1k
14 2. Summary of contributions
clearly impressive machines and raise the bar of what is achievable today, many research
questions remain unanswered, since details of neither the design, nor gait synthesis, nor
their control aspects are available to the research community at large.
In contrast, another family of robots pioneered by Prof. Jessy Grizzle, such as Rab-
bit [30], MABLE [31], and MARLO [32], have shown how fundamental analytical math-
ematical concepts in the subject of VHC can be used to achievehighly dynamic walking
and running behaviors exploiting under-actuated dynamics13. The work done with these
machines has been reported not only in research articles, but books [36], Master and PHD
theses, all belonging to public domain. In this literature we can see the advantages of
VHC for being a method that offers systematic and nearly analytical procedures for gait
synthesis and feedback control design. This is particularly true for the case of legged
mechanisms with one passive link, such as point-feet, wherethis method can be followed
as a recipe during development.
In relation to VHC, the work of my supervisor and research group has contributed
substantially to the development of the theory behind this approach. Years of research
have culminated on a number of publications, and one of them stands out for being a
generalization of the VHCs approach for under-actuated dynamical systems with impulse
effects [14, 37, 38]. This theory can be directly applied in many applications, including
the subject of legged locomotion.
In most cases, the work of [1] and [30] can be essentially treated similarly. However,
they present fundamental differences related to the way controllers are designed, and the
way stability is analyzed. Informally, for control design the work of [30] applies a proce-
dure of standard feedback linearization with PD control, and the stability of the resulting
closed-loop system is analyzed based on Poincaré maps. Thisis decisive for demonstrat-
ing exponential orbital stability of a limit cycle. On the other hand, the work of [1] is
based on concepts of transverse dynamics [14], which is a more concise mathematical
approach to analyze stability of limit cycles (gaits), as well as for estimating the region
of attraction and possible deviations from nominal trajectories in response to small para-
metric perturbations. Control systems designed accordingto transverse dynamics have
superior performance and robustness, as they are able to guarantee global exponential
orbital stability.
To provide an insight of how VHC can be used in biped locomotion for systems with
many passive links, the work presented in this article considers a reduced planar model of
a human body, i.e. a torso and two legs, Fig. 2.7. This system has three characteristics:1)
13Video examples of these machines walking and running can be seen inhttps://www.youtube.com/user/DynamicLegLocomotion
2.2. Three-link Planar Biped Walker (Paper II) 15
point feet,2) one actuator in between the legs, and3) a passive torso maintained upright
by springs. Hence, the dynamics of this model considers a case of under-actuation de-
gree two, and the resulting mechanical characteristics resemble the working principle of
the robot MARLO [32]. One of the main challenges behind this system consists on for-
mulating a gait synthesis algorithm (trajectory generation), and the control design having
two un-actuated links. This example has been originally inspired on the work of [27,28],
where the authors attempt to find passive walking gaits. Unlike that case, our attempt is to
make this mechanism walk on flat surfaces. To this end, an initial solution was proposed
earlier by my research group in the work of [39]. My purpose isto present an alternative
solution, which I believe makes the VHC approach easier to apply.
Fig. 2.7: Schematics of the biped in the sagittal plane and level
From the theoretical aspect, this example presents some of the following fundamental
challenges:
1. Finding passive gaits while walking down slopes.
2. Walking on flat ground.
3. Walking on uneven terrain by adapting the gait according to the obstacles.
4. Repeating the same for running.
This article presents solutions to some of these challengesand particularly:
16 2. Summary of contributions
• The way gait synthesis can be formulated to apply VHC in under-actuated mechan-
ical systems with more than one passive links. To this end, wealso incorporate
concepts of optimization to involve energy efficiency.
• The way concepts of transverse dynamics can be applied, including the procedure
we use for deriving it.
• The way concepts of linearization of transverse dynamics can be used for designing
robust feedback controllers.
• The way simulation methods (see Fig. 2.8) can be used to analyze performance
under modeling uncertainty, external disturbances, etc.
(a) (b) (c) (d) (e)
Fig. 2.8:3-Link planar walker biped simulation
2.3 Papers III-VI. Hydraulic manipulators
In our daily life we find hydraulic systems all around us. Theyare so fundamental to our
world, but we hardly notice it. Nevertheless, hydraulic technology provides the muscles
to lift, steer, build and mine resources that power countless industries. Whether these
systems are used for transportation, construction, heavy duty, robotics, process industry,
or entertainment, our industrial world depends heavily on hydraulics.
The greatest example of brute strength that hydraulic technology can deliver is found
in the heavy-duty machine industry. Over the years the excavators, cranes and trucks
that have helped building our modern world have grown exponentially in size, achieving
unmatched efficiency empowered by hydraulics. Several of these machines are equipped
2.3. Hydraulic manipulators (Papers III-VI) 17
with hydraulic manipulators that work as large scale arms tohandle heavy loads and can
be seen in excavators, loaders, and cranes, just to mention afew.
Cylinders dominate the market in the hydraulic industry, because they represent a
robust component for converting the hydraulic energy into workable energy, allowing to
make well controlled back-and-forth motions. In heavy-duty machines, they are used as
the main actuator for suspension control and to move different DOF.
Most heavy-duty machines are formed by three main components: the cabin, the ve-
hicle and the manipulator. The cabin is highly important because it is used to comply with
ISO safety regulations, which involve protecting the operator from the working environ-
ment’s dangers. Therefore, most of the technology used for controlling these machines is
embedded in the cabin, and includes seats, computers, embedded processing units, key-
boards, joysticks, levers –all of which result in a complex human-machine interaction.
Levers or joysticks are quite relevant to this thesis, because they are the primary tools
used to directly command proportional valves, either mechanically or electrically. These
valves are fundamental for the control of the manipulators,because they regulate how the
hydraulic oil flows to produce the movement of the actuators.Operating machines in this
mode is a standard in many industries, and despite the difficulties associated with it, this
has prevail as the primary approach for decades.
When we look at different hydraulic manipulators, we can recognize standard designs
that are found in industrial robotics. For instance, it is common to see designs based
on parallel or serial kinematics where each DOF is controlled by one hydraulic actua-
tor. Despite differences in the mechanical design concepts, most machines are operated
according to the same principle, since each DOF has to be manually controlled either
by a joystick or a lever. Thus, the complexity of controllingthe system becomes obvi-
ous, as the operator needs to coordinate each DOF to positionthe end-effector in desired
locations. Since these manipulators are often designed with more than four DOF, this op-
eration involves a high redundancy problem. To this we have to include the simultaneous
problem of coordinating the vehicle and end-effector’s tools, which all result in the need
for advanced motor skills, multitasking, and concentration, that are highly demanding
both mentally and physically.
When we involve open-loop manual control, the operator is given the full responsibil-
ity for providing control inputs that cause the machine to complete a given task optimally;
For example, in minimum time or with the least energy consumption. However, the op-
timal control solutions of even very simple nonlinear systems are not intuitive and may
be beyond the capabilities of the human operator responsible for providing the control
18 2. Summary of contributions
input. As a result, machines controlled manually through open-loop commands are being
operated sub-optimally.
Sensing technology is the fundamental component needed fordesigning feedback
control that can be used for making autonomous decisions to relieve some of the oper-
ators’ tasks. However, in today’s hydraulic machinery the absence of sensors restricts the
possibility to introduce computerized support, as there isno indication of how the ma-
chine is operating. Because of this reason, even computerized decision-making tools are
missing, which restricts the possibility to provide working suggestions to the operator.
Consequently, estimating the efficiency of different processes is difficult, which results
in the absence of fundamental understanding that can be usedfor improving optimality,
training, and working efficiency.
Companies involved in training on virtual reality simulators offer solutions that allow
operators to attempt learning optimal ways to work with machines. This is often done to
reduce the environmental consequences of learning with real machines. However, to this
day, there are no clear methods of training or working with machines that are based on
fundamental understanding of motion/task optimization. This is a subject that is highly
developed in the robotics community and can be of great benefit to the hydraulic machine
industry. To this we should also add all the advantages of introducing all the advanced
motion control technologies and autonomous robotic functionalities that we find in many
other industries.
There are different driving forces that motivate the development in this thesis. Among
others, I see the need for developing smart machines to satisfy the increasing demand of
natural resources and infrastructure as one of the most important. One of the reasons is
the world’s exponentially increasing population that willchallenge the supply of many
fundamental elements such as housing, heating, electricity, and food, which are problems
that we are already facing today. These concerns are shared by many political entities. As
an example, I present below two paragraphs that have been extracted from the European
commission’s website [40]:
“The European Commission has today adopted a strategy to shift the European economy
towards greater and more sustainable use of renewable resources. With the world popu-
lation increasing by 40% by 2030, approaching9 billion by 2050 and natural resources
finite, Europe needs renewable biological resources for secure and healthy food and feed,
as well as for materials, energy, and other products.”
“Europe needs to make the transition to a post-petroleum economy. Greater use of renew-
able resources is no longer just an option, it is a necessity.We must drive the transition
2.3. Hydraulic manipulators (Papers III-VI) 19
from a fossil-based to a bio-based society with research andinnovation as the motor. This
is good for our environment, our food and energy security, and for Europe’s competitive-
ness for the future.”
Many machine manufacturers are well aware of these challenges, but they are also
aware of the difficulties found with the introduction of semi-automated technologies to
the dominating working generation. Despite this, many companies have been investing
on research for more than a decade in topics related to control systems, robotics, and au-
tomation. Today, as we are approaching a younger generationthat is prone to technology,
it is becoming increasingly important to introduce new products into the market to keep
younger customers interested, ease the learning process, and improve efficiency. Proof
of this is the recently released technologies that involve machines with embedded sen-
sors, computer controlled cranes, voice-activated control, (semi)autonomous navigation,
advanced Human-Machine Interfaces, Head-up displays, automatic data analysis, etc, see
Fig. 2.9. Many of these technologies are found in industriesrelated to mining, agriculture,
oil platforms, just to mention a few. However, they are missing in many other important
areas, as it is the case of the forestry industry.
(a) Loader with analog sensors (b) Excavator with wireless sensors
Fig. 2.9: Recent embedded sensing technology for hydraulicmachines14.
In terms of research and development, a lot of interesting research results in the in-
troduction of automation in several hydraulic machines which can be found in excava-
tors [41–47], loaders [48–50], handlers for construction of montage buildings [51], and
forestry cranes [52–56].
14Sources:Loader,http://www.ivtinternational.com/news.php?NewsID=52748Excavator,http://g-nestle.de/en/product/topcon-x-22/
20 2. Summary of contributions
The work found in this thesis adds some further insights intothe development of au-
tomation technology for forestry machines, and in particular, forestry cranes (see Fig. 2.10).
These are hydraulic machines that are used in a variety of assignments related to wood,
and forest biomaterial extraction, which are important renewable natural resources used
to produce several products, such as housing, electricity,heating, and more. Although
forestry machinery is quite similar to most heavy-duty machines, it present some impor-
tant differences:
• The suspension system and moving vehicle are designed to access highly dynamic
and rough terrain, which is often found in deep forest, and mountains. Because of
this reason, even military tank manufacturing companies are involved in the pro-
duction of forestry machine bogies.
• Cranes are often designed with at least one prismatic link toregulate the crane’s
length. This is used to avoid hitting the cabin, and reach a larger radius of operation
without needing to move the vehicle.
• The hydraulic circuits applied to these machines are not standard, and each manu-
facturer has a different way to treat the design. In many cases, this information is
private and represents a difficulty for performing research.
• Some forestry machines, particularly the harvesters, are recognized as some of the
most difficult machines to operate, and are highly comparable to helicopters.
The concept of designing autonomous services for the forestindustry has been on-
going since the1980s. For example, laser pointer assisted motion control of a forestry
crane was implemented and tested in [57]. Similarly, principles of interactive robotics for
forestry cranes were presented in [52]. Nevertheless, these premature solutions made no
impact for commercialization, since they were, at that time, too costly to be produced.
Today, the situation is different. Embedded processing units and sensing devices have
become more mature and accessible in terms of size and price,allowing to consider com-
plex solutions for motion control of forestry cranes [54]. Additionally, social factors, such
as urbanization, long learning period, and the increasing difficulty to attract or recruit new
operators, are making automation increasingly relevant. Therefore, forestry companies
have been recently embracing the overall benefits that robotics and automation can bring
to this industry [58–62].
Certainly, the scenario of full automation of these machines is unrealistic today, but
considering semi-autonomous computerized solutions to simplify the operation it is clearly
2.3. Hydraulic manipulators (Papers III-VI) 21
Fig. 2.10: Forwarder machine used in the forest industry forlogging operations.
interesting and feasible soon. However, experience has shown to me that transporting
methods found in the robotics and control literature into the forestry industry is easier
said that done, because there exists various theoretical and practical challenges that are
still unresolved. For instance, if we name the three fundamental subjects of system mod-
eling, identification and control design, we see that it is difficult to generalize a method
and apply it to an entire family of machines. Some of the reasons are the following:
1. Each degree of freedom in a forestry machine is controlledaccording to a differ-
ent hydraulic configuration. Thus, modeling the hydraulic systems can be quite
challenging.
2. When designing the hydraulic circuits it is thought that some links of the manipu-
lators are used for motion control, as well as for suspensionpurposes when lifting
heavy loads. This introduces complex flexibility phenomena, which reduce the
possibility of performing precise motion.
3. The load carried by a forestry crane can overpass its overall weight, which intro-
duces huge variations in the system parameters.
4. Although mechanically these systems might look alike, the design of the hydraulic
22 2. Summary of contributions
circuitry changes greatly from machine to machine.
5. The changes in temperature and weather conditions can alter the dynamic perfor-
mance of the cranes. One example of this is related to the changes in friction forces
during winter and summer.
Apart of these problems, we can also add some of the practicalones found in today’s
commercial machines:
1. Low processing capacity to implement control algorithms, since current machines
simply use a power amplifier to convert the joystick commandsinto control inputs
for the valves.
2. Absence of cranes with sensors, which are fundamental to design feedback con-
trollers.
3. Absence of visual technology to perform semi-autonomousnavigation.
4. The need of reliable embedded codes that can be easily integrated into the main
system without complete understanding of the programmers.
5. Absence of global position sensing needed to locate the machine in the forest.
In relation to automation and robotic motion control, thereare a couple of ideas that
are interesting to this industry, because they can facilitate the work, ease the learning
process, and increase efficiency. Some of these ideas are thefollowing:
1. To control the crane boom-tip directly instead of controlling the individual links,
which can reduce the amount of joystick inputs needed and work with the machine
more intuitively.
2. Performing autonomous motions for lifting logs, which can be used in machines
used for log-collection.
3. Performing autonomous motions for dragging and cutting trees, which can be used
in machines used for harvesting.
4. Teaching autonomous motions to the machine, such that both crane and vehicle are
able to repeat these motions.
5. Learning from experienced operators, such that similar data is used for training
beginners.
2.3. Hydraulic manipulators (Papers III-VI) 23
6. Navigation through pre-selected paths.
7. Automated working planning.
Although some of these scenarios can partly be solved today,by introducing sensors
and better processing capacity to machines, several of themrequire further development
in other areas. For instance:
1. It is difficult to find computer vision solutions that can easily transition to use in
forestry.
2. GPS-like technology for the forestry brand does not existtoday.
3. Autonomous navigation planning requires information about forest terrain, location
of trees, wet-surfaces, water streams, just to mention a few.
4. It is difficult to place sensors, since some machine parts are prone to collisions and
external damage, provoked by the working conditions.
5. To interact with a robotic machine the Human-Machine Interface needs to be up-
dated in order to cope with a diverse class of new functionalities.
6. Autonomous data analysis that is performed either offlineor in real-time, such that
the operator is able to receive feedback about his own work with suggestions on how
to improve it. This can also be used to make autonomous adaptation of parameters
without the need of human intervention.
7. Autonomous data analysis, that is performed either offline or in real-time, such that
simulators can provide feedback to the user about his own work, with suggestions
about how to improve it.
All the lists presented above are just examples of research areas derived from the
analysis of forestry automation. This shows the complexityof the subject and the rich
amount of solutions that can be derived from it, which have direct applications in other
areas. The solutions found in the papers of this thesis, related to this topic, cover a wide
variety of these problems and they can briefly be categorizedaccording to the following
subjects:
1. Modeling, in which I describe the mathematics needed to model forestry cranes,
presented inPapers III and IV .
24 2. Summary of contributions
2. Control Design, in which I present some of the results found by implementingboth
linear and nonlinear controllers. This is found inPapers IV and V.
3. Motion and trajectory planning, in which I present both theory and practice for
implementing ideas of semi-autonomous motions. As it will be inferred from these
articles, the main approach used for this task is based on VHC, which is an effi-
cient method for designing optimal movement. This approachhas its roots in the
development done by my research group, and has some relations with the method
widely known aspath constrained trajectory planningused extensively in robotics.
The mathematical foundations of this approach can be seen inPaper V, and some
of the applications will be seen inPapers V and VI.
4. Understanding the working methods and operators efficiency, in which I con-
sider one of the applications of our trajectory planning method for understanding
how operators work, and how this could be improved. To the best of my knowl-
edge, this is the first article attempting to understand how forestry operations could
be analyzed through motion data. This information is highlyrelevant to different
industries, and can be applied in a variety of forms, including machine learning,
online trajectory planning, autonomous adaptive control parameter configurations,
and design of simulators that provide feedback information, just to mention a few.
All of this is explained in detail inPaper VI.
3 Concluding remarks
“I am turned into a sort of machine for
observing facts and grinding out
conclusions.”
Charles Darwin
I have been working with the VHC approach for quite some time.In my opinion, this
approach is undoubtly a valuable tool for motion planning and control of different type
of mechanical systems, and in particular, under-actuated ones. In addition, I have found
applications outside the domain of under-actuated systemswhere this approach is very
useful to design optimal motion performance requirements.This is always needed in the
subject of robotics, since systems cannot be designed with infinite capabilities. Examples
of these are the motion planning and analysis of mechanical manipulators that are used in
a large portion of industry.
Although VHC can work as a recipe for a variety of problems, I see that there are quite
many challenges ahead that need further research. Among these challenges, the following
are of great importance:
1. Real-time adaptation. For instance, in the case of leggedlocomotion, it is unrealis-
tic to believe that robots would only walk on flat surfaces with small disturbances.
In reality, we would expect robots to walk in a variety of terrains with the need to
adapt their gaits according to obstacles and terrain. Therefore, motion planning and
controller design should be able to adapt according to theseunknown situations. In
current VHC implementations, trajectories are planned by solving differential equa-
tions that describe the reduced dynamics of the system, and transverse linearization
is used for controller design by solving a time-variant Riccati equation. Perform-
ing these operations in real-time is unfeasible today, and requires great advances in
computational numerical methods.
25
26 3. Concluding remarks
2. Collaboration with industry is quite challenging for researchers, since industry has
a different expectation of experimental results than thoseresearchers have. Often
industrial engineers are unfamiliar with modern nonlinearcontrol methods and it
is difficult for them to judge their advantages over simpler approaches. This is
particularly true for those industries that are recently adopting feedback control
system technologies. Therefore, simpler methods for practical implementation of
VHC should be applied when initiating this form of cooperation.
Despite these challenges, the rich mathematical concepts of VHC make this method
quite friendly. This is particularly true when a person has become familiar with this
approach. Nevertheless, there are cases that could challenge most approaches, including
VHC. For instance, Fig. 3.1 presents some problems derived from the butterfly robot,
that I believe some readers will find interesting. These sketches address challenges with
under-actuation degree higher than one, that could perhapsbe unsolvable1
(a) (b) (c)
Fig. 3.1: Sketches of butterfly-like challenges.
In a similar way, I consider that studying the problem of legged locomotion in the
sagittal plane is quite important to understand walking characteristics. However, if we
intend to design robots able to operate naturally, studyingthe problem in all dimensions
is highly relevant. Nevertheless, we often believe that machines are supposed to look like
some biological system, such as bipedal creatures. I believe that designing robotic legged
systems that are able to move safely and highly efficient according to their own dynamics
1Since there is no website for the work in this thesis, I will make a couple of things available to the readers.For example, the Solidworks models of the sketches presented here have been tested with 3D printers and lasercutting techniques, and they can be requested through email.Additionally, the algorithms presented can alsobe requested, and they have been tested in low-cost hardware, such as Arduino UNO, MEGA, and Due. Thecomputer vision software needed for implementation can also beavailable under request. All of the softwarehas been implemented with MATLAB/Simulink, including the solution for the planar biped model.
27
is more relevant, even when their motion behavior is unnatural. This can be achieved by
designing mechanisms with variable topologies. Some pioneering examples of this kind
are the12-TET robot from NASA, and the Virginia Tech’s IMPASS, to mention a few.
After working with the problem of automation in heavy-duty machines, I find many
open questions that need further research. Some of these arelisted below.
1. When we have a good understanding of forestry operations and the way experienced
operators handle the machine, it is feasible to design a new family of simulators able
to prepare beginners more efficiently. This can be done through methods presented
in this thesis and, in particular, applying the algorithms proposed in paper VI. These
type of simulators could also be used for several other purposes, e.g.a) to design
scientific methods to teach operators,b) to involve gamification with the aim of
challenging operators and improve their working performance. One example is the
standard self-competition against a simulated ghost representing the optimal way
of operation (see Fig. 3.2).
2. Cabins, at the moment are filled with buttons and joysticksand it seems inefficient
to add more buttons or joysticks, since this can complicate the work. If automa-
tion is incorporated in heavy-duty equipment, new human-machine interfaces will
be needed. This will be important to select among different automated tasks and
functions.
3. I have not covered anything related to the problem of computer vision or navigation.
However the use of cameras, radars, Lidar systems, can help to automate grasping
tasks, as well as autonomous navigation. These are basic technologies required to
automate large portions of the process and to reduce furtherthe necessity of high
human experience.
4. Systematic methods for designing controllers that can autonomously adapt to dif-
ferent types of cranes.
5. Mathematical models describing dynamics of the completeforestry machines, i.e.
cabin, manipulator and vehicle. As these systems are coupled, each component has
a strong effect on the remaining parts.
6. Considering that mathematical optimization is a mature subject in different areas,
designing cranes can be done differently today. Optimization algorithms could lead
to designs that are naturally efficient, unlike the empirical designs found in most
industries. This could lead to machines that instead of one crane are equipped with
28 3. Concluding remarks
a pair of cranes that work through autonomous decisions and managed by advanced
human-machine interfaces.
7. Human-Machine collaboration, and Machine-Machine collaboration is a topic that
will be needed to automate the work among machines.
Fig. 3.2: Gamer competing against his ghost that performed the fastest lap2.
2Source:https://www.lfs.net/attachment/3715
Bibliography
Bibliography
[1] A. Shiriaev, J. Perram, and C. Canudas-de-Wit, “Constructive tool for orbital stabilization
of underactuated nonlinear systems: Virtual constraints approach,”IEEE Transactions on
Automatic Control, vol. 50, no. 8, pp. 1164–1176, 2005.
[2] U. Mettin, Principles for planning and analyzing motions of underactuated mechanical sys-
tems and redundant manipulators. PhD thesis, Umeå Universitet, Institutionen för Tillämpad
Fysik och Elektronik, 2009.
[3] P. La Hera,Underactuated mechanical systems: Contributions to trajectory planning,anal-
ysis, and control. PhD thesis, Umeå Universitet, Institutionen för Tillämpad Fysik och Elek-
tronik, 2011.
[4] K. Furuta, M. Yamakita, and S. Kobayashi, “Swing-up control of inverted pendulum us-
ing pseudo-state feedback,”Proceedings of the Institution of Mechanical Engineers, Part I:
Journal of Systems and Control Engineering, vol. 206, no. 4, pp. 263–269, 1992.
[5] I. Fantoni and R. Lozano, “Stabilization of the furuta pendulum around its homoclinic orbit,”
International Journal of Control, vol. 75, no. 6, pp. 390–398, 2002.
[6] M. Spong, P. Corke, and R. Lozano, “Nonlinear control of the reaction wheel pendulum,”
Automatica, vol. 37, no. 11, pp. 1845–1851, 2001.
[7] D. J. Block, K. J. Åström, and M. W. Spong,The Reaction Wheel Pendulum. Morgan &
Claypool Publishers, 2007.
[8] M. W. Spong, “The swing up control problem for the acrobot,”Control Systems, IEEE,
vol. 15, no. 1, pp. 49–55, 1995.
[9] M. W. Spong and D. J. Block, “The pendubot: A mechatronic systemfor control research
and education,” inDecision and Control, 1995., Proceedings of the 34th IEEE Conference
on, vol. 1, pp. 555–556, IEEE, 1995.
[10] K. J. Åström and K. Furuta, “Swinging up a pendulum by energy control,” Automatica,
vol. 36, no. 2, pp. 287–295, 2000.
199
200 Bibliography
[11] K. Lynch, N. Shiroma, H. Arai, and K. Taine, “The roles of shape and motion in dynamic
manipulation: The butterfly example,” inProceedings of IEEE International Conference on
Robotics and Automation (ICRA), 1998.
[12] M. Cefalo, L. Lanari, and G. Oriolo, “Energy-based control ofthe butterfly robot,” inPro-
ceedings of the 3rd IFAC workshop on Periodic Control Systems, 2006.
[13] M. Cefalo, Controllo energy-based dei sistemi meccanici sottoattuati con applicazione al
sistema Butterfly. PhD thesis, Universita Degli Studi di Roma "La Sapienza", 2005.
[14] A. Shiriaev and L. Freidovich, “Transverse linearization for impulsive mechanical systems
with one passive link,”IEEE Trans. on Automatic Control, vol. 54, no. 12, pp. 2882–2888,
2009.
[15] L. Freidovich, A. Robertsoon, A. Shiriaev, and R. Johansson,“Stable periodic motions of
inertia wheel pendulum via virtual holonomic constraints,” inProc. European Control Con-
ference, (Kos, Greece), pp. 3771–3776, 2007.
[16] L. Freidovich, A. Robertsson, A. Shiriaev, and R. Johansson,“Periodic motions of the Pen-
dubot via virtual holonomic constraints: Theory and experiments,”Automatica, vol. 44,
no. 3, pp. 785–791, 2008.
[17] P. L. Hera, L. Freidovich, A. Shiriaev, and U. Mettin, “New approach for swinging up the
furuta pendulum: Theory and experiments,”Mechatronics, vol. 19, no. 8, pp. 1240–1250,
2009.
[18] L. B. Freidovich, P. La Hera, U. Mettin, A. Robertsson, A. S. Shiriaev, and R. Johansson,
“Shaping stable periodic motions of inertia wheel pendulum: theory and experiment,”Asian
journal of control, vol. 11, no. 5, pp. 548–556, 2009.
[19] M. Vukobratovic and B. Borovac, “Zero-moment pointUthirty five years of its life,”Interna-
tional Journal of Humanoid Robotics, vol. 1, no. 01, pp. 157–173, 2004.
[20] Honda, “ASIMO Robot.”http://world.honda.com/ASIMO/index.html, Ac-
cessed 2007.
[21] T. McGeer, “Passive dynamic walking,”The International Journal of Robotics Research,
vol. 9, no. 2, pp. 62–82, 1990.
[22] M. J. Coleman, A. Chatterjee, and A. Ruina, “Motions of a rimless spoked wheel: a simple
three-dimensional system with impacts,”Dynamics and Stability of Systems, vol. 12, no. 3,
pp. 139–159, 1997.
[23] A. Goswami, B. Espiau, and A. Keramane, “Limit cycles in a passive compass gait biped and
passivity-mimicking control laws,”Autonomous Robots, vol. 4, no. 3, pp. 273–286, 1997.
[24] F. Asano, M. Yamakita, N. Kamamichi, and Z.-W. Luo, “A novel gait generation for biped
walking robots based on mechanical energy constraint,”Robotics and Automation, IEEE
Transactions on, vol. 20, no. 3, pp. 565–573, 2004.
Bibliography 201
[25] J. Grizzle, G. Abba, and F. Plestan, “Asymptotically stable walking for biped robots: Anal-
ysis via systems with impulse effects,”IEEE Transactions on Automatic Control, vol. 46,
no. 1, pp. 51–64, 2001.
[26] S. Collins, M. Wisse, and A. Ruina, “A three-dimensional passive-dynamic walking robot
with two legs and knees,”The International Journal of Robotics Research, vol. 20, no. 7,
pp. 607–615, 2001.
[27] M. Wisse, A. L. Schwab, and F. C. van der Helm, “Passive dynamic walking model with
upper body,”Robotica, vol. 22, no. 06, pp. 681–688, 2004.
[28] S. Collins and A. Ruina, “A bipedal walking robot with efficient and human-like gait,”
in Proceedings of the 2005 IEEE International Conference on Robotics and Automation,
(Barcelona, Spain), Apr. 18-22 2005.
[29] A. Kuo, “Stabilization of lateral motion in passive dynamic walking,”The International
Journal of Robotics Research, vol. 18, no. 9, pp. 917–930, 1999.
[30] C. Chevallereau, G. Abba, Y. Aoustin, F. Plestan, E. Westervelt, C. Canudas-de-Wit, and
J. Grizzle, “RABBIT: A testbed for advanced control theory,”IEEE Control Systems Maga-
zine, vol. 23, no. 5, pp. 57–79, 2003.
[31] J. Grizzle, J. Hurst, B. Morris, H.-W. Park, and K. Sreenath, “Mabel, a new robotic bipedal
walker and runner,” inAmerican Control Conference, 2009. ACC’09., pp. 2030–2036, IEEE,
2009.
[32] B. Buss, A. Ramezani, K. Hamed, B. Griffin, K. Galloway, and J.Grizzle, “Preliminary
walking experiments with underactuated 3d bipedal robot marlo,” inIntelligent Robots and
Systems (IROS 2014), 2014 IEEE/RSJ International Conference on, pp. 2529–2536, Sept
2014.
[33] Boston Dynamics.http://www.bostondynamics.com, Accessed 2014.
[34] Y. Hurmuzlu, F. Génot, and B. Brogliato, “Modeling, stability and control of biped robotsUa
general framework,”Automatica, vol. 40, no. 10, pp. 1647–1664, 2004.
[35] E. Ackerman, “Darpa robotics challenge trials: What you should (and shouldnŠt) expect to
see,”IEEE Spectrum, vol. 19, 2013.
[36] E. Westervelt, J. Grizzle, C. Chevallereau, J. Choi, and B. Morris, Feedback Control of Dy-
namic Bipedal Robot Locomotion. CRC Press, Taylor and Francis Group, 2007.
[37] L. B. Freidovich and A. S. Shiriaev, “Transverse linearization for mechanical systems with
passive links, impulse effects, and friction forces,” inDecision and Control, 2009 held jointly
with the 2009 28th Chinese Control Conference. CDC/CCC 2009. Proceedings of the 48th
IEEE Conference on, pp. 6490–6495, IEEE, 2009.
[38] A. Shiriaev, L. Freidovich, and S. Gusev, “Transverse linearization for controlled mechanical
systems with several passive degrees of freedom,”Automatic Control, IEEE Transactions on,
vol. 55, no. 4, pp. 893–906, 2010.
202 Bibliography
[39] P. La Hera, A. Shiriaev, L. Freidovich, U. Mettin, and S. Gusev, “Stable walking gaits for
a three-link planar biped robot with one actuator,”Robotics, IEEE Transactions on, vol. 29,
pp. 589–601, June 2013.
[40] European Comission , “Research and innovation.”http://ec.europa.eu/
research/bioeconomy/news-events/news/20120213_en.htm, Accessed
2014.
[41] L. Płonecki, W. Trampczynski, and J. Cendrowicz, “A concept of digital control system
to assist the operator of hydraulic excavators,”Automation in construction, vol. 7, no. 5,
pp. 401–411, 1998.
[42] A. Stentz, J. Bares, S. Singh, and P. Rowe, “A robotic excavatorfor autonomous truck load-
ing,” Autonomous Robots, vol. 7, no. 2, pp. 175–186, 1999.
[43] Q. Ha, Q. Nguyen, D. Rye, and H. Durrant-Whyte, “Impedancecontrol of a hydraulically
actuated robotic excavator,”Automation in Construction, vol. 9, no. 5, pp. 421–435, 2000.
[44] M. Haga, W. Hiroshi, and K. Fujishima, “Digging control system forhydraulic excavator,”
Mechatronics, vol. 11, no. 6, pp. 665–676, 2001.
[45] H. Araya and M. Kagoshima, “Semi-automatic control system for hydraulic shovel,”Au-
tomation in construction, vol. 10, no. 4, pp. 477–486, 2001.
[46] M. D. Elton, A. R. Enes, and W. J. Book, “A virtual reality operatorinterface station with
hydraulic hardware-in-the-loop simulation for prototyping excavator control systems,” inAd-
vanced Intelligent Mechatronics, 2009. AIM 2009. IEEE/ASME International Conference on,
pp. 250–255, IEEE, 2009.
[47] G. J. Maeda, S. P. Singh, and D. C. Rye, “Improving operational space control of heavy
manipulators via open-loop compensation,” inIntelligent Robots and Systems (IROS), 2011
IEEE/RSJ International Conference on, pp. 725–731, IEEE, 2011.
[48] E. Spencer, K. Chipperfield, F. Wagner, and A. Kelkar, “Modeling and control of a wheel
loader with a human-in-the-loop assessment using virtual reality,”Journal of dynamic sys-
tems, measurement, and control, vol. 127, p. 415, 2005.
[49] S. Mukhopadhyay, G. Gupta, and J. Howarth, “Embedded microcontroller and sensors based
front end loader control system,” inInstrumentation and Measurement Technology Confer-
ence Proceedings, 2008. IMTC 2008. IEEE, pp. 1509–1514, IEEE, 2008.
[50] H. Almqvist,Automatic bucket fill. PhD thesis, Linköping, 2009.
[51] J. Cinkelj, R. Kamnik, P.Cepon, M. Mihelj, and M. Munih, “Closed-loop control of hy-
draulic telescopic handler,”Automation in Construction, vol. 19, no. 7, pp. 954–963, 2010.
[52] A. Halme, A. Visala, M. Paakkunainen, J. Joensuu, P. Forsman, and T. Torvikoski, “Model
based on-site description of robotized processings,”IFAC Man-Machine Systems, 1988.
Bibliography 203
[53] A. A. Goldenberg, J. Wiercienski, P. Kuzan, C. Szymczyk, R. G. Fenton, and B. Shaver, “A
remote manipulator for forestry operation,”Robotics and Automation, IEEE Transactions
on, vol. 11, no. 2, pp. 185–197, 1995.
[54] P. Viljamaa, H. Koivo, and A. Peltomaa, “Adaptive feed control of a forest harvester, icom
2003 int,” inConf. Mechatron., Loughborough University, 2003.
[55] E. Papadopoulos, B. Mu, and R. Frenette, “On modeling, identification, and control of a
heavy-duty electrohydraulic harvester manipulator,”Mechatronics, IEEE/ASME Transac-
tions on, vol. 8, no. 2, pp. 178–187, 2003.
[56] P. La Hera, U. Mettin, I. Manchester, and A. Shiriaev, “Identification and control of a hy-
draulic forestry crane,” inProceedings of the 17th IFAC World Congress, (Seoul, Korea),
pp. 2306–2311, July 6-11 2008.
[57] M. Manninen, A. Halme, and R. Myllyla, “An aimable laser time-of-flight range finder for
rapid interactive scene description,” inProceedings of the 7th Annual Conference of British
Robot Association, 1984.
[58] A. Halme and M. Vainio, “Forestry robotics-why, what and when,” in Autonomous Robotic
Systems, pp. 149–162, Springer, 1998.
[59] U. Hallonborg, “Unmanned forestry machines can be competitive,” Skogforsk Results, no. 4,
2003.
[60] M. Brander, D. Eriksson, and B. Löfgren, “Automation of knuckleboom work can increase
productivity.” Skogforsk Results, No 4, 2004.
[61] J. Billingsley, A. Visala, and M. Dunn,Springer Handbook of Robotics, ch. Robotics in
Agriculture and Forestry, pp. 1065–1077. Berlin Heidelberg: Springer-Verlag, 2008.
[62] A. Jongebreur and L. Speelman, “Future trends in agricultural engineering,”NJAS wagenin-
gen journal of life sciences, vol. 45, no. 1, pp. 3–14, 1997.
[63] C. Canudas-de-Wit, B. Espiau, and C. Urrea, “Orbital stabilizationof an underactuated three-
link planar robot,” inProceedings of the 15th IFAC Word Congress, (Barcelona, Spain), Jul.
21-26 2002.
[64] L. Freidovich, A. Shiriaev, F. Gomez-Estern, F. Gordillo, and J.Aracil, “Modification via av-
eraging of partial-energy-shaping control for orbital stabilization: cart-pendulum example,”
Int. J. Control, vol. 82, no. 9, pp. 1582–1590, 2009.
[65] L. Consolini and M. Maggiore, “Virtual holonomic constraints for euler-lagrange systems,”
in Symposium on Nonlinear Control Systems (NOLCOS), 2010.
[66] CNRS – French National Center for Scientific Research, “Biped robot RABBIT.” URL:
http://robot-rabbit.lag.ensieg.inpg.fr/English/, accessed on 2007.
204 Bibliography
[67] C. Nielsen, C. Fulford, and M. Maggiore, “Path following using transverse feedback lin-
earization: Application to a maglev positioning system,”Automatica, vol. 46, no. 3, pp. 585–
590, 2010.
[68] M. Spong, S. Hutchinson, and M. Vidyasagar,Robot Modeling and Control. New Jersey:
John Wiley and Sons, 2006.
[69] L. Freidovich, A. Shiriaev, and I. Manchester, “Stability analysis and control design for an
underactuated walking robot via computation of a transverse linearization,” in Proceedings
of the 17th IFAC World Congress, (Seoul, Korea), Jul. 6-11 2008.
[70] J.-Y. Kim, I.-W. Park, and J.-H. Oh, “Experimental realizationof dynamic walking of the
biped humanoid robot KHR-2 using zero moment point feedback and inertial measurement,”
Journal of Advanced Robotics, vol. 20, no. 6, pp. 707–736, 2006.
[71] M. Spong, J. Holm, and D. Lee, “Passivity-based control of bipedal locomotion,”IEEE
Robotics and Automation Magazine, vol. 14, no. 2, pp. 30–40, 2007.
[72] K. Sreenath, H. Park, I. Poulakakis, and J. Grizzle, “A compliant hybrid zero dynamics
controller for stable, efficient and fast bipedal walking on mabel,”The International Journal
of Robotics Research, vol. 30, no. 9, pp. 1170–1193, 2011.
[73] B. Morris and J. Grizzle, “Hybrid invariant manifolds in systems withimpulse effects with
application to periodic locomotion in bipedal robots,”IEEE Transactions on Automatic Con-
trol, vol. 54, no. 8, pp. 1751–1764, 2009.
[74] L. Freidovich and A. Shiriaev, “Transverse linearization for mechanical systems with passive
links, impulse effects, and friction forces,” inProc. of the 48th IEEE Conference on Decision
and Control and 28th Chinese Control Conference, (Shanghai, China), pp. 6490–6495, 2009.
[75] P. X. L. Hera, A. S. Shiriaev, L. B. Freidovich, and U. Mettin, “Gait synthesis for a three-link
planar biped walker with one actuator,” inProc. of the 2010 IEEE International Conference
on Robotics and Automation, (Anchorage, Alaska, USA), pp. 1715–1720, 2010.
[76] C. Chevallereau, G. Bessonnet, G. Abba, and Y. Aoustin,Bipedal Robots: Modeling, Design
and Walking Synthesis. New Jersey: John Wiley and Sons, 2009.
[77] Y. Hurmuzlu and D. Marghitu, “Rigid body collisions of planar kinematic chains with multi-
ple contact points,”The International Journal of Robotics Research, vol. 13, no. 1, pp. 82–92,
1994.
[78] A. L. Schwab and M.Wisse, “Basin of attraction of the simplest walking model,” in Pro-
ceedings International Conference on Noise and Vibration, (Pittsburgh, Pennsylvania), Sept.
9-12 2001.
[79] L. Freidovich, U. Mettin, A. Shiriaev, and M. Spong, “A passive 2-DOF walker: Hunting for
gaits using virtual holonomic constraints,”IEEE Trans. on Robotics, vol. 25, no. 5, pp. 1202–
1208, 2009.
Bibliography 205
[80] Oryx Simulations.http://www.oryx.se/, Accessed 2013.
[81] Algoryx , “Multiphysics and 3D simulations.”http://www.algoryx.se, Accessed
2007.
[82] UMIT Research Lab.http://www.org.umu.se/umit/, Online since 2011.
[83] B. Löfgren, “Automation of forest machines–an important piecein precision forestry,” in
Precision Forestry in plantations, semi-natural and natural forests. Proceedings of the Inter-
national Precision Forestry Symposium, Stellenbosch, pp. 135–148, 2006.
[84] U. Mettin and P. M. La Hera, “Modelling and control design for a hydraulic forestry crane,”
Research Results, Department of Applied Physics and Electronics, Umeå University, 2005.
[85] S. Westerberg, I. Manchester, U. Mettin, P. La Hera, and A. Shiriaev, “Virtual environ-
ment teleoperation of a hydraulic forestry crane,” inProceedings of the 2008 IEEE Inter-
national Conference on Robotics and Automation, (Pasadena, USA), pp. 4049–4054, May
19-23 2008.
[86] Cranab, “Cranab presents an entirely new generation of cranes.” http://www.cranab.
se/static/en/206/, Accessed on 2013.
[87] J. Sprinkle, J. M. Eklund, H. Gonzalez, E. I. Grøtli, B. Upcroft, A. Makarenko, W. Uther,
M. Moser, R. Fitch, H. Durrant-Whyte,et al., “Model-based design: a report from the
trenches of the darpa urban challenge,”Software and Systems Modeling, vol. 8, no. 4,
pp. 551–566, 2009.
[88] Mathworks.http://www.mathworks.com/, Accessed 2014.
[89] V. Arnold, Mathematical Methods of Classical Mechanics (Graduate Texts in Mathematics).
New York: Springer, 2nd ed., 1989.
[90] J. J. Craig,Introduction to Robotics: Mechanics and Control. USA: Prentice Hall, 2003.
[91] L. N. Hand and J. D. Finch,Analytical mechanics. USA: Cambridge University Press, 1998.
[92] H. Olsson, K. Åström, C. Canudas de Wit, M. Gäfvert, and P. Lischinsky, “Friction models
and friction compensation,”European Journal of Control, vol. 9, no. 4, pp. 176–195, 1998.
[93] N. D. Manring,Hydraulic Control Systems. New York, USA: John Wiley & Sons, first ed.,
2005.
[94] M. Jelali and A. Kroll, Hydraulic Servo-Systems: Modeling, Identification and Control.
Great Britain: Springer, second ed., 2004.
[95] O. Egeland and J. T. Gravdahl,Modeling and Simulation for Automatic Control. Trondheim,
Norway: Marine Cybernetics, 2002.
[96] L. Ljung, System Identification: Theory for the User. Upper Saddle River: Prentice Hall,
1999.
206 Bibliography
[97] P. Belanger, P. Dobrovolny, A. Helmy, and X. Zhang, “Estimationof angular velocity and
acceleration from shaft-encoder measurements,”The International Journal of Robotics Re-
search, vol. 17, no. 11, pp. 1225–1233, 1998.
[98] John Deere , “Simulators.”http://www.deere.co.uk/wps/dcom/en_GB/
products/equipment/simulators/simulators.page, Accessed 2013.
[99] Interaction Design.http://www.interactiondesign.se/terry, Online since
2010.
[100] M. Linjama,The Modelling and Actuator Space Control of Flexibile Hydraulic Cranes. PhD
thesis, Tampere University of Technology, 1998.
[101] M. Münzer,Resolved Motion Control of Mobile Hydraulic Cranes. PhD thesis, Aalborg
University, 2002.
[102] U. Mettin, P. La Hera, D. Ortiz Morales, A. Shiriaev, L. Freidovich, and S. Westerberg,
“Trajectory planning and time-independent motion control for a kinematically redundant
hydraulic manipulator,” inAdvanced Robotics, 2009. ICAR 2009. International Conference
on Robotics, pp. 1–6, IEEE, 2009.
[103] U. Mettin, S. Westerberg, A. Shiriaev, and P. La Hera, “Analysisof human-operated motions
and trajectory replanning for kinematically redundant manipulators,” inIntelligent Robots
and Systems, 2009. IROS 2009. IEEE/RSJ International Conference on, pp. 795–800, IEEE,
2009.
[104] D. Ortiz Morales, P. La Hera, U. Mettin, L. B. Freidovich, A. S. Shiriaev, and S. Westerberg,
“Steps in trajectory planning and controller design for a hydraulically driven crane with
limited sensing,” inIntelligent Robots and Systems (IROS), 2010 IEEE/RSJ International
Conference on, pp. 3836–3841, IEEE, 2010.
[105] D. Ortiz Morales, S. Westerberg, P. La Hera, U. Mettin, L. B. Freidovich, and A. S. Shiri-
aev, “Open-loop control experiments on driver assistance for crane forestry machines,” in
Robotics and Automation (ICRA), 2011 IEEE International Conference on, pp. 1797–1802,
IEEE, 2011.
[106] A. Hansson and M. Servin, “Semi-autonomous shared controlof large-scale manipulator
arms,”Control Engineering Practice, vol. 18, no. 9, pp. 1069–1076, 2010.
[107] HIAB, “Hiab xs 622 product brochure.”http://www.atlaspolar.com/pdfs/PB-
622-EN-WW_L.pdf, Accessed on 2012.
[108] John Deere, “Smooth boom control.”http://www.deere.co.uk/wps/dcom/en_
GB/industry/forestry/learn_more/sbc.page, Accessed on 2012.
[109] John Deere first with crane tip control on forest machines.http://www.
internationalforestindustries.com, Online since 2013.
Bibliography 207
[110] P. M. La Hera, U. Mettin, S. Westerberg, and A. S. Shiriaev, “Modeling and control of hy-
draulic rotary actuators used in forestry cranes,” inRobotics and Automation, 2009. ICRA’09.
IEEE International Conference on, pp. 1315–1320, IEEE, 2009.
[111] R. Osypiuk, T. Kroger, B. Finkemeyer, and F. M. Wahl, “A two-loop implicit force/position
control structure, based on a simple linear model: Theory and experiment,” in Robotics
and Automation, 2006. ICRA 2006. Proceedings 2006 IEEE International Conference on,
pp. 2232–2237, IEEE, 2006.
[112] P. Krus and S. Gunnarsson, “Adaptive control of a hydrauliccrane using on-line identifica-
tion,” in Proc. Third Scandinavian Conf. on Fluid Power, Linkoeping, Sweden, 1993.
[113] M. Sirouspour and S. Salcudean, “Nonlinear control of hydraulic robots,”Robotics and Au-
tomation, IEEE Transactions on, vol. 17, no. 2, pp. 173–182, 2001.
[114] R. V. M. Donath, “Dynamic feedback linearization for electrohydraulically actuated ser-
vosystem,” inProceedings of the 1992 Japan/USA Symposium on Flexible Automation, (San
Francisco, USA), pp. 595–606, July 1992.
[115] J. Mattila and T. Virvalo, “Computed force control of hydraulic manipulators,” in5th Scan-
dinavian Inter. Conf. On Fluid Power, (Sweden), p. 139U154, July 6-11 1997.
[116] G. Sohl and J. Bobrow, “Experiments and simulations on the nonlinear control of a hydraulic
servosystem,”Control Systems Technology, IEEE Transactions on, vol. 7, no. 2, pp. 238–247,
1999.
[117] P. Li and M. Wang, “Passivity based nonlinear control of hydraulic actuators based on
an euler-lagrange formulation,” inASME Proc. Dynamic Systems and Control Conference,
2011.
[118] A. Mohanty and B. Yao, “Indirect adaptive robust control ofhydraulic manipulators with
accurate parameter estimates,”Control Systems Technology, IEEE Transactions on, vol. 19,
no. 3, pp. 567–575, 2011.
[119] H. Zeng and N. Sepehri, “Adaptive backstepping control of hydraulic manipulators with
friction compensation using lugre model,” inAmerican Control Conference, 2006, pp. 6–pp,
IEEE, 2006.
[120] W. Perruquetti and J.-P. Barbot,Sliding mode control in engineering, vol. 11. CRC, 2002.
[121] P. La Hera and D. Ortiz Morales, “Modeling dynamics of an electro-hydraulic servo actuated
manipulator: A case study of a forestry forwarder crane,” inWorld Automation Congress
(WAC), 2012, pp. 1–6, IEEE, 2012.
[122] P. La Hera, B. Ur Rehman, and D. Ortiz Morales, “Electro-hydraulically actuated forestry
manipulator: Modeling and identification,” inIntelligent Robots and Systems (IROS), 2012
IEEE/RSJ International Conference on, pp. 3399–3404, IEEE, 2012.
208 Bibliography
[123] P. La Hera and D. Ortiz Morales, “Non–linear dynamics modelling description for simulat-
ing the behaviour of forestry cranes,”International Journal of Modelling, Identification and
Control, vol. 21, no. 2, pp. 125–138, 2014.
[124] D. Ortiz Morales, S. Westerberg, P. X. La Hera, U. Mettin, L. Freidovich, and A. S. Shiriaev,
“Increasing the level of automation in the forestry logging process with crane trajectory
planning and control,”Journal of Field Robotics, vol. 31, pp. 343–363, May/June 2014.
[125] A. Pourmovahed, N. Beachley, and F. Fronczak, “Modeling of a hydraulic energy regener-
ation system. part i: Analytical treatment,”Journal of dynamic systems, measurement, and
control, vol. 114, no. 1, pp. 155–159, 1992.
[126] V. Utkin, J. Guldner, and J. Shi,Sliding Mode Control in Electro-Mechanical Systems, (Au-
tomation and Control Engineering), vol. 5. CRC Press, 2009.
[127] J.-J. E. Slotine, W. Li,et al., Applied nonlinear control, vol. 199. Prentice-Hall Englewood
Cliffs, NJ, 1991.
[128] L.-W. Chang, “A mimo sliding control with a second order sliding condition,” in ASME
WAM, 1990.
[129] dSPACE.http://www.dspace.com, Accessed 2014.
[130] J. Swevers, C. Ganseman, D. B. Tukel, J. De Schutter, and H.Van Brussel, “Optimal robot
excitation and identification,”Robotics and Automation, IEEE Transactions on, vol. 13,
no. 5, pp. 730–740, 1997.
[131] P. Apkarian and D. Noll, “Nonsmooth h-infinity synthesis,”Automatic Control, IEEE Trans-
actions on, vol. 51, no. 1, pp. 71–86, 2006.
[132] N. Bruinsma and M. Steinbuch, “A fast algorithm to compute the h-infinity norm of a transfer
function matrix,”Systems & Control Letters, vol. 14, no. 4, pp. 287–293, 1990.
[133] P. MacDonald and M. Clow, “”Just one damn machine after another?” technological inno-
vation and the industrialization of tree harvesting systems,”Technology in Society, vol. 21,
no. 3, pp. 323–344, 1999.
[134] S.-A. Axelsson, “The mechanization of logging operations in Sweden and its effect on occu-
pational safety and health,”Journal of Forest Engineering, vol. 9, no. 2, pp. 25–31, 1998.
[135] P. MacDonald and M. Clow, “The industrialization of tree harvestingsystems in the eastern
canadian forest, 1955-1995,”Labour/Le Travail, vol. 58, pp. 145–167, 2006.
[136] J. Bobrow, S. Dubowsky, and J. Gibson, “Time-optimal control of robotic manipulators along
specified paths,”The International Journal of Robotics Research, vol. 4, no. 3, pp. 3–17,
1985.
[137] K. Shin and N. McKay, “Minimum-time control of robotic manipulators with geometric path
constraints,”Automatic Control, IEEE Transactions on, vol. 30, no. 6, pp. 531–541, 1985.
Bibliography 209
[138] F. Pfeiffer and R. Johanni, “A concept for manipulator trajectory planning,” Robotics and
Automation, IEEE Journal of, vol. 3, no. 2, pp. 115–123, 1987.
[139] G. Antonelli, S. Chiaverini, C. Curatella, and A. Marino, “Constrained motion planning for
industrial robots,” inAutomation and Logistics, 2009. ICAL’09. IEEE International Confer-
ence on, pp. 1934–1939, IEEE, 2009.
[140] A. Gasparetto and V. Zanotto, “A technique for time-jerk optimal planning of robot trajecto-
ries,” Robotics and Computer-Integrated Manufacturing, vol. 24, no. 3, pp. 415–426, 2008.
[141] O. Dahl and L. Nielsen, “Torque-limited path following by online trajectory time scaling,”
Robotics and Automation, IEEE Transactions on, vol. 6, no. 5, pp. 554–561, 1990.
[142] S. Macfarlane,On-line smooth trajectory planning for manipulators. PhD thesis, University
of British Columbia, 2001.
[143] P. Hsu, J. Hauser, and S. Sastry, “Dynamic control of redundant manipulators,” inRobotics
and Automation, 1988. Proceedings., 1988 IEEE International Conference on, pp. 183–187
vol.1, 1988.
[144] J.-P. Merlet,Parallel robots. Springer, 2006.
[145] J. Wu, J. Wang, L. Wang, and T. Li, “Dynamics and control of aplanar 3-dof parallel manip-
ulator with actuation redundancy,”Mechanism and Machine Theory, vol. 44, no. 4, pp. 835–
849, 2009.
[146] T. Hellström and O. Ringdahl, “Follow the past: A path-tracking algorithm for autonomous
vehicles,”International journal of vehicle autonomous systems, vol. 4, no. 2, pp. 216–224,
2006.
[147] T. Hellström, T. Johansson, and O. Ringdahl, “Development ofan autonomous forest ma-
chine for path tracking,” inField and Service Robotics, pp. 603–614, Springer, 2006.
[148] O. Ringdahl, O. Lindroos, T. Hellström, D. Bergström, D. Athanassiadis, and T. Nordfjell,
“Path tracking in forest terrain by an autonomous forwarder,”Scandinavian Journal of Forest
Research, vol. 26, no. 4, pp. 350–359, 2011.
[149] O. Ringdahl,Automation in forestry: development of unmanned forwarders. PhD thesis,
LNU, 2011.
[150] Sauer-Danfoss, “Proportional control valves.”http://www.sauer-danfoss.
com/Products/PVGProportionalValves/CartridgeValves/
ProportionalControlValves/index.htm, Accessed on 2014.
[151] J. Kovanen and H. Handroos, “Adaptive open-loop control method for a hydraulically
driven flexible manipulator,” inAdvanced Intelligent Mechatronics, 2001. Proceedings. 2001
IEEE/ASME International Conference on, vol. 2, pp. 942–947, IEEE, 2001.
[152] H. Khalil, Nonlinear systems. New Jersey: Prentice Hall, 3rd ed., 2002.
210 Bibliography
[153] K. Åström and T. Hägglund, “PID Controllers: Theory, Design,and Tuning,”Instrument
Society of America, 1995.
[154] J. Bobrow, B. Martin, G. Sohl, E. Wang, F. Park, and J. Kim, “Optimal robot motions for
physical criteria,”Journal of Robotic systems, vol. 18, no. 12, pp. 785–795, 2001.
[155] SkogsElmia, “Voice operated forestry machines.”http://www.elmia.se/
en/skogselmia/For-Vistors/Fair-news-2011/Voice-operated-
harvester-a-huge-crowd-magnet-/, Accessed 2013.
[156] Optea, “Projecting display systems for forestry machines.”http://www.optea.com/
projects.asp, Accessed 2013.
[157] S. Westerberg and A. Shiriaev, “Virtual environment-based teleoperation of forestry ma-
chines: Designing future interaction methods,”Journal of Human-Robot Interaction, vol. 2,
no. 3, pp. 84–110, 2013.
[158] D. Tiernan, G. Zeleke, P. Owende, C. Kanali, J. Lyons, and S.Ward, “Effect of working
conditions on forwarder productivity in cut-to-length timber harvesting onsensitive forest
sites in ireland,”Biosystems engineering, vol. 87, no. 2, pp. 167–177, 2004.
[159] H. Ovaskainen, J. Uusitalo, and K. Väätäinen, “Characteristics and significance of a harvester
operatorsŠ working technique in thinnings,”International Journal of Forest Engineering,
vol. 15, no. 2, pp. 67–77, 2004.
[160] J. Manner, T. Nordfjell, and O. Lindroos, “Effects of the number of assortments and log
concentration on time consumption for forwarding,”Silva Fennica, vol. 47, no. 4, 2013.
[161] R. Zinkevicius and D. Vitunskas, “Assessment of operator effect on machine performance of
Sponsse buffalo dualT harwarder,”Engineering for Rural Development, no. 23, pp. 73–78,
2013.
[162] T. Nurminen, H. Korpunen, and J. Uusitalo, “Time consumption analysis of the mechanized
cut-to-length harvesting system,”Silva Fennica, vol. 40, no. 2, p. 335, 2006.
[163] D. Lee and K.-W. HAN, “Analysis of human operator’s behavior in high-order dynamic
systems,”IEEE Transactions on Systems, Man, and Cybernetics, vol. 10, pp. 207–213, 1980.
[164] K. Tervo, L. Palmroth, and H. Koivo, “Skill evaluation of humanoperators in partly auto-
mated mobile working machines,”Automation Science and Engineering, IEEE Transactions
on, vol. 7, no. 1, pp. 133–142, 2010.
[165] K. Tervo, L. Palmroth, and A. Putkonen, “A hierarchical fuzzy inference method for skill
evaluation of machine operators,” inAdvanced Intelligent Mechatronics, 2009. AIM 2009.
IEEE/ASME International Conference on, pp. 136–141, IEEE, 2009.
[166] L. Palmroth, K. Tervo, and A. Putkonen, “Intelligent coaching of mobile working machine
operators,” inIntelligent Engineering Systems, 2009. INES 2009. International Conference
on, pp. 149–154, IEEE, 2009.
Bibliography 211
[167] K. M. Newell, Y.-T. Liu, and G. Mayer-Kress, “Time scales in motor learning and develop-
ment.,”Psychological review, vol. 108, no. 1, p. 57, 2001.
[168] Skogforsk, “Fem nyckelmoment for effektiv och bränslesnål skotning.” http:
//www.skogforsk.se/sv/kunskap/db/2013/Fem-viktiga-moment-
for-effektiv-och-branslesnal-skotning/, Accessed on 2013.
[169] S. Westerberg,Semi-Automating Forestry Machines : Motion Planning, System Integration,
and Human-Machine Interaction. PhD thesis, Umeå University, Department of Applied
Physics and Electronics, 2014.
[170] M. Attebrant, J. Winkel, S. Mathiassen, and A. Kjellberg, “Shoulder-arm muscle load and
performance during control operation in forestry machines: Effectsof changing to a new arm
rest, lever and boom control system,”Applied Ergonomics, vol. 28, no. 2, pp. 85–97, 1997.