Engineering Haptic Devices

“This page left intentionally blank.”

Dr.-Ing. Thorsten A. Kern
Editor

Engineering Haptic Devices

A Beginner’s Guide for Engineers

123
Editor
Dr.-Ing. Thorsten A. Kern
TU Darmstadt
Inst. Elektromechanische
Konstruktionen
Merckstr. 25
64283 Darmstadt
Germany

ISBN 978-3-540-88247-3 e-ISBN 978-3-540-88248-0

DOI 10.1007/978-3-540-88248-0
Springer Dordrecht Heidelberg London New York

Library of Congress Control Number: 2009930680

c Springer-Verlag Berlin Heidelberg 2009

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication
or parts thereof is permitted only under the provisions of the German Copyright Law of September 9,
1965, in its current version, and permission for use must always be obtained from Springer. Violations
are liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.

Cover design: WMXDesign GmbH

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

 The idea for this book was born in 2003. Originally intended as an addition to
my dissertation, it was thought of filling a gap I had noticed: The regrettably small
number of comprehensive recapitulating publications on haptics available for, e.g,. a
technically interested person, confronted with the task of designing a haptic device
for the first time. In 2004, besides a considerable number of conference proceed-
ings, journals or PhD-theses, no document was available giving a summary of the
major findings of this challenging subject. Thanks are due to the support provided
by colleagues, especially by Prof. Dr.-Ing. Dr.-med Ronald Blechschmidt-Trapp and
Dr.-Ing. Christoph Doerrer, enabling me to realize my plan in the following years.
While writing my dissertation, I learnt that the areas to be covered by such a book
would have to be much more extensive than originally expected. Nevertheless my
mentor and supervisor Prof. Dr.-Ing. habil. Roland Werthschützky, to whom I have
a special debt, encouraged me to pursue and finish this project during the time as a
Post-Doc. It was funded by the DFG (KE1456/1-1) with special regard to the con-
solidation of our design methodology for haptic devices. Due to this funding the
financial basis of this task was guaranteed.
The structuring of the themes made clear that the book would be considerably im-
proved by contributions made by specialists of several areas. Thus began the multi-
author project you are holding in your hands. Not only the authors explicitly named
contributed to the book, but also former and current colleagues could be won over
to a committed co-operation during the project, like Dr.-Ing. Markus Jungmann or
Ingmar Stöhr, to name and thank at least two of them.
The first inquiry at Springer, our preferred publisher, resulted in an impressively
positive reaction and consequently in a constructive co-operation with Dr. Christoph
Baumann at any time. We owe to him the project’s final configuration concerning
language versions and date of publication. Let me add a word of gratitude to the
many people involved in proofreading the book and removing major errors, all ahead
of them Anika Kohlstedt, Sebastian Kassner, as well as to the numerous helpers of
both family and friends.
For the English version of this book a professional language institute was originally
hired which, however, was unable to provide the translation in time. Therefore it had
to be done by the authors themselves. Most of the proofreading was done by Ursula
and Jürgen Förnges both of whom I want to thank explicitly.
Without the help of all the people mentioned above and of many other unnamed
supporters, this book would never have reached its present degree of completeness.
I owe a great debt to all of them. A very special thanks is due to the Institut für Elek-
tromechanische Konstruktionen at Technische Universität Darmstadt, Germany, for
the extraordinary support given to me and many of my co-authors in our research
and scientific work.

Thorsten A. Kern
“This page left intentionally blank.”
Preface

The term “haptics” unlike the terms “optics” or “acoustics” is not so well-known
to the majority of people, not even to those who buy and use products related to
haptics. The words “haptics” and “haptic” refer to everything concerning the sense
of touch. “Haptics” is everything and everything is “haptic”, because it does not
only describe pure mechanical interaction, but also includes thermal- and pain- (no-
ciception) perception. The sense of touch makes it possible for humans and other
living beings to perceive the “borders of their physical being”, i.e. to identify where
their own body begins and where it ends. With regard to this aspect, the sense of
touch is much more efficient than the sense of vision, as well in resolution as in the
covered dihedral angle, e.g.: In the heat of a basketball match a light touch on our
back immediately makes us aware of an attacking player we do not see. We notice
the intensity of contact, the direction of the movement by a shear on our skin or a
breeze moving our body hairs - all this is perceived without catching a glimpse of
the opponent.
“Haptic systems” are divided into two classes1 . There are the time-invariant systems
(the keys of my keyboard), which generate a more or less unchanging haptic effect
whether being pressed today or in a year’s time. Structures like surfaces, e.g. the
wooden surface of my table, are also part of this group. These haptically interesting
surfaces are often named “haptic textures”. Furthermore, there are active, reconfig-
urable systems, which change their haptic properties partly or totally dependent on
a preselection - e.g. from a menu. Finally, there are combinations and hybrid forms
of systems, which are presented and explained in the corresponding chapters. The
focus of this book is on the technological design criteria for active, reconfigurable
systems, providing a haptic coupling of user and object in a mainly mechanical un-
1 In engineering there are three terms which are often used but do not have a definite meaning:

System, Device and Component. Systems are - depending on the task of the designer - either a
device or a component. A motor is a component of a car, but for the developer of the motor it
is a device, which is assembled from components (spark-plug, cocks, knocking-sensor). It can be
helpful when reading a technological text to replace each term with the word “thing”. Although this
suggestion is not completely serious, it surprisingly increases the comprehensibility of technical
texts.

vii
viii Preface

derstanding. Thermal and nociceptive perceptions are mentioned according to their

significance but are not seriously discussed. This is also the case with regard to pas-
sive haptic systems.
The fact that you have bought this book suggests that you are interested in haptics.
You might have already tried to sketch a technical system meant to fool haptic per-
ception. And this attempt may have been more or less successful, e.g. concerning
your choice of the actuators. Maybe, you are just planning a project as part of your
studies or as a commercial product aimed at improving a certain manual control or
at introducing a new control concept. Approaches of this kind are quite frequent.
Many of the first active haptic systems were used in airplanes, to make aware of
critical situations by a vibrating control handle. Nowadays, the most wide-spread
active haptic system surely is the vibration of a mobile-phone. It enables its user to
notice the reception of a message without visual or auditory contact, whereby even
the type of the message - SMS or phone call - is coded in this buzzing haptic signal.
More complex haptic systems can be found in automotive technology, as e.g. recon-
figurable haptic control knobs. They are typically located in the center of the control
console and are usually part of complex luxury limousines. Today, multidimensional
haptic interaction is no longer limited to navigation- or modeling purposes of pro-
fessional users, but has also found its way into interaction during computer gaming.
Maybe, you are a member of the popular group of doctors and surgeons actively
using haptics in medical technology. There has been a continuous increase of, the
complexity of the tools for minimally-invasive surgery - longitudinal instruments
with a limited degree of freedom to inspect and manipulate human tissue through
small artificial or natural openings in the human body. This automatically results
in the loss of the direct contact between surgeon and the manipulated tissue. For
decades, the wish to improve the haptic feedback during such kinds of applications
and/or the realization of training methods for minimally-invasive surgery has been
a high motivation for researchers in haptic device design, however without a satis-
factory commercial breakthrough, yet significant improvements in telemanipulation
and simulation have been achieved.
Despite of or even because of the great variety of projects in industry and research
working with haptic systems, the common understanding of “haptics” and the terms
directly referring to it, like “kinaesthetic” and “tactile” are by no means as unam-
biguous and indisputable as they should be. In this book, we, the authors, intend to
offer you a help to act more safely in the area of designing haptic devices. This book
will begin with the presentation of the terminology and its usage according to what
we regard as appropriate. Then it will provide a deeper understanding of haptics and
a simplified engineering description, and will finally lead to concrete instructions
and recommendations for the design of technologically complex haptic systems.
Besides the intention to generate real hardware design, there is another reason for
dealing with haptic device design: A continuing ambition to extend one’s knowl-
edge of haptic perception. This discipline, named “psychophysics” is an “unsharp”,
non deterministic science formulating hypotheses and systematically checking them
with the help of experiments and observations. These experiments are paramount to
any progress. Consequently, special attention has to be paid to their quality and the
Preface ix

parameters observed. As a by-product of this important science of haptic research

a plurality of devices and technical systems have been built. In fact psychophysics
uses expertise in many different disciplines to solve its problems. As a consequence,
important and creative engineers and scientists like Prof. H ONG TAN and Prof. V IN -
CENT H AYWARD have not only been designing high fidelity and very efficient haptic
devices, but are also heavily involved in the research on psychophysical parameters.
Psychophysics with emphasis on haptic questions is a very dynamic science. Every
year, an uncounted number of results and experiments are published at congresses
and in journals. Lately, M ARTIN G RUNWALD [79] has published a notable sum-
mary of the latest state of knowledge. The book you are holding in your hands
does not claim to keep up with every detail of this psychophysical progress. How-
ever, it tries to include as many of its findings as possible into the design of haptic
devices. This book has been written by and is addressed to engineers of all the dis-
ciplines mentioned before: Design-engineers representing mechanical engineering,
hardware-near electrical engineering, control-engineering, software-engineering or
as a synergy of expertise in all disciplines of mechatronics.
As said before, the haptic sense is doubtlessly gaining in importance. This can be
concluded from the great number of scientific publications on this subject and from
the fact that all relevant distal senses like the senses of sight and hearing have already
been provided with synthetic information in almost perfect quality in every-day life.
“Perfect quality” may have different meanings depending on the actual context. A
realistic rendering of a sensual experience can be an important requirement. The
resolution of a 3D-monitor has to be below the resolution capability of the human
eye in color dynamics and spatial distances between the picture elements (pixels).
Sounds have to be traceable in space and must not interfere with artifacts of the
storage- or transmission medium. In different circumstances attracting attention can
be another “perfect quality”. Typically, warning signals in the dashboard of a car
are visual examples, so are acoustic signals in the cockpit of an airplane. Another
demand on “perfect quality” can be the simultaneous requirement of high discrim-
inabilty and large range - just think of navigational signals for ships. Both areas -
optics and acoustics - have been subject to intense research for decades and have
been provided with numerous intelligent device designs. In many cases the borders
of the human capability of perception of the information provided have been reached
or even crossed nowadays. At this point it is obvious to make use of another human
sense to transmit information. Another motivation is the true-to-life simulation of
virtual environments. After visual and auditory presentation having reached a high
quality, the focus is directed to the haptic sense as being the next important one.
Only this sense enables us to experience our physical borders and the synergy of
interaction and perception.
Further areas of haptic research are telepresence and telemanipulation systems. In
these cases, an intuitive and immediate feedback is a prerequisite for a safe han-
dling of e.g. dangerous and / or valuable materials. There are reasons enough and
to spare for dealing with the design of haptic devices which are demanded by the
market. However, experts are rare and the access to this subject is difficult. The de-
sign of haptic devices demands interdisciplinary knowledge which should include
x Preface

the basics of the properties of haptic perception and its dynamic-dependence on am-
plitude and frequency. Furthermore an overview of technological solutions, like the
designs of actuators, kinematics or complete systems including software-solutions
and the interfaces to simulations and virtual reality systems may be extremely help-
ful. For designing virtual reality systems it is also necessary to know the concepts of
haptic renderings to enhance communication between soft- and hardware engineers.

The authors of this book regard their task as being fulfilled as soon as this book
helps to fascinate more design-engineers by the development of haptic devices, thus
speeding up the creation of more and better haptic systems available on the market.

Darmstadt, February 2009 Thorsten A. Kern

Preface xi

Thorsten A. Kern received his Doctor in electrical en-

gineering in 2006 at the Technische Universität Darm-
stadt for the design of a haptic assistive system for
cardio-catheterizations. He is working as a group leader
at Continental Corporation, and is responsible for the
development and production of actuators for visual
and haptic applications. Before he was working in
the area of medical simulation technology and VR-
applications using haptics at PolyDimensions GmbH.
His research focused on engineering methods to quan-
tify the haptic impression of actuators and devices.
He is associated lecturer for biomedical engineering
([email protected]).

Marc Matysek graduated in electrical engineering at

the Technische Universität Darmstadt in 2003. There,
he is working as a PhD student at the Institut für X?Y
Elektromechanische Konstruktionen since 2003. His re-
search focuses on the technology of dielectric elastomer
actuators and their use in applications as tactile displays
([email protected]).

Oliver Meckel received his degree for mechanical en-

gineering in 2002 at Technische Universität Darmstadt.
He is working as group leader of the technical depart-
ment at the Wittenstein motion control GmbH. He is re-
sponsible for the development, simulation, testing, and
technical product support of electrical gear motors and X?Y
actuator systems. Before he was working at the Institute
for Flight Systems and Control at Technische Univer-
sität Darmstadt. His research was focused on adaptive
control systems and the development of Unmanned Au-
tonomous Vehicles ([email protected]).
xii Preface

Jacqueline Rausch received her diploma in electri-

cal engineering and information technology in 2005
at the Technische Universität Darmstadt. There, she
is currently working as a research associate at the
Institute for Electromechanical Design. Her main X?Y
topic is the development of miniaturized piezoresis-
tive multi-axis force/torque sensors to be integrated
at the tip of minimally invasive surgical manipulators
([email protected]).

Alexander Rettig received his diploma in mathematics

at the Technical Universität Darmstadt in 1998. He is
working as team leader software development for Poly-
Dimensons GmbH in the area of medical graphics and
simulation applications. Before he was research asso-
ciate at the Fraunhofer Institute for Graphics Darmstadt,
where he focused on the integration of haptics into vir-
tual reality systems and development of virtual reality
applications ([email protected]).
Photo: Lichtbildatelier Darmstadt

Andreas Röse received his engineering degree in 2002

in electromechanical engineering. He is working as a
research associate at the Institute of Electromechan-
ical Design at the Technische Universität Darmstadt.
His research work is focussed on parallel-kinematic
mechanisms for minimally invasive surgical telema-
nipulators including numerical calculation techniques,
mechanical design and performance evaluation meth-
ods. From 2002 to 2005 he was with Fresenis Med-
ical Care and developed hemodialysis machines with
emphasis on safety systems and control architecture
([email protected]).
Preface xiii

Stephanie Sindlinger b. Klages received her diploma

in electrical engineering and information technology
at the Technische Universität Darmstadt in 2004. She
is working as a research associate at the Institut für
Elektromechanische Konstruktionen. There she devel-
ops a haptic device for cardiologic catheterisations. Her
expertises are the development of piezoelectric ultra-
sonic actuators and the design of the user-interface for
integrating the system in the operational environment
([email protected]).
“This page left intentionally blank.”
Contents

Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi

Part I Some Basics of Haptics

1 Motivation and Application of Haptic Systems . . . . . . . . . . . . . . . . . . . . . 5

1.1 The Meaning of Haptics from a Philosophical and Social Viewpoint 5
1.1.1 Haptics as a Physical Being’s Boundary . . . . . . . . . . . . . . . . . 5
1.1.2 Formation of the Sense of Touch . . . . . . . . . . . . . . . . . . . . . . . 8
1.1.3 Special Aspects of the Design Process . . . . . . . . . . . . . . . . . . . 11
1.2 The Significance of Haptics in Everyday Professional Life . . . . . . . . 13
1.2.1 The Sense of Touch in Everyday Medical Life . . . . . . . . . . . . 13
1.2.2 The Sense of Touch in the Cockpit . . . . . . . . . . . . . . . . . . . . . . 14
1.2.3 The Sense of Touch at the Desk . . . . . . . . . . . . . . . . . . . . . . . . 16
1.2.4 The Sense of Touch in Music . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.1 Scientific Disciplines as Part of Haptic Research . . . . . . . . . . . . . . . . 19
2.2 Terms and Terminology Used for the Description of Haptic Systems 21
2.2.1 Basic Concepts of Haptics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.2 Definition of Haptic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.3 Parameters of Haptic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.4 Characterization of Haptic Object Properties . . . . . . . . . . . . . 28
2.2.5 Technical examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3 Biological Basics of Haptic Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.1 The Sense of Touch and its Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2 Haptic Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2.1 Psychophysical Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.2.2 Frequency Dependency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.2.3 Characteristics of Haptic Interaction . . . . . . . . . . . . . . . . . . . . 50
3.3 Conclusions from the Biology of Haptics . . . . . . . . . . . . . . . . . . . . . . . 54

xv
xvi Contents

3.3.1 Stiffnesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.3.2 One Kilohertz - Significance for the Mechanical Design? . . . 56

4 Modeling the User . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

4.1 Mapping of Frequency Ranges onto the User’s Mechanical Model . . 59
4.2 A Model of the User as a Mechanical Load . . . . . . . . . . . . . . . . . . . . . 63
4.2.1 Types of Grasps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2.2 Measurement Setup and Equipment . . . . . . . . . . . . . . . . . . . . . 66
4.2.3 Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.2.4 Modeling Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3 Modelling Haptic Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.4 Application Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4.1 Kinaesthetic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4.2 Tactile System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.4.3 Examples in Time and Frequency Domain . . . . . . . . . . . . . . . 87
4.5 Summarizing Remarks on the Application of the Method . . . . . . . . . 93

5 Internal Structure of Haptic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

5.1 System Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.1.1 Open-Loop Impedance Controlled . . . . . . . . . . . . . . . . . . . . . . 96
5.1.2 Closed-Loop Impedance Controlled . . . . . . . . . . . . . . . . . . . . . 97
5.1.3 Open-Loop Admittance Controlled . . . . . . . . . . . . . . . . . . . . . 98
5.1.4 Closed-Loop Admittance Controlled Devices . . . . . . . . . . . . . 100
5.1.5 Qualitative Comparison of the Internal Structures of
Haptic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.2 Abstraction of the System Components . . . . . . . . . . . . . . . . . . . . . . . . 104
5.2.1 Basic System Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.2.2 Analysis of System Components . . . . . . . . . . . . . . . . . . . . . . . 105

Part II Designing Haptic Systems

6 Identification of Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

6.1 The Right Questions to Ask . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.1.1 Interaction as a Classification Criterion . . . . . . . . . . . . . . . . . . 117
6.1.2 Cluster “Kinaesthetic” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.1.3 Cluster “Surface-Tactile” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.1.4 Cluster “Vibro-Tactile” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
6.1.5 Cluster “Vibro-Directional” . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.6 Cluster “Omni-Dimensional” . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.7 Cluster “always” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.2 Experiments Together With the Customer . . . . . . . . . . . . . . . . . . . . . . 122
6.3 Requirement Specifications of a Haptic System . . . . . . . . . . . . . . . . . 124
6.4 Order of Technical Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Contents xvii

7 Control of Haptic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.2 System Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
7.2.1 Linear SISO-System Description . . . . . . . . . . . . . . . . . . . . . . . 132
7.2.2 Linear State Space Description . . . . . . . . . . . . . . . . . . . . . . . . . 136
7.2.3 Nonlinear System Description . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.3 System Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
7.3.1 Analysis of Linear System Stability . . . . . . . . . . . . . . . . . . . . . 141
7.3.2 Analysis of Nonlinear System Stability . . . . . . . . . . . . . . . . . . 145
7.3.3 Passivity in Dynamic Systems . . . . . . . . . . . . . . . . . . . . . . . . . 150
7.4 Control Design for Haptic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.4.1 Structuring of the Control Design . . . . . . . . . . . . . . . . . . . . . . . 152
7.4.2 Requirement Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
7.4.3 Control Law Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

8 Kinematic Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

8.1 Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
8.1.1 Mechanisms and their Classification . . . . . . . . . . . . . . . . . . . . 166
8.1.2 Calculating Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
8.1.3 Transfer Characteristics and Jacobian Matrix . . . . . . . . . . . . . 170
8.1.4 Optimizing the Transfer Characteristics . . . . . . . . . . . . . . . . . . 174
8.2 Serial Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
8.2.1 Topological Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
8.2.2 Calculation of the Kinematic Problems . . . . . . . . . . . . . . . . . . 177
8.2.3 Example of a Serial Mechanism . . . . . . . . . . . . . . . . . . . . . . . . 179
8.3 Parallel Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
8.3.1 Topological Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
8.3.2 Calculation of the Kinematic Transfer Functions . . . . . . . . . . 185
8.3.3 Examples of a Parallel Mechanism . . . . . . . . . . . . . . . . . . . . . . 186
8.4 The Complete Process of Kinematic Design . . . . . . . . . . . . . . . . . . . . 188

9 Actuator Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

9.1 General Facts about Actuator Design . . . . . . . . . . . . . . . . . . . . . . . . . . 192
9.1.1 Overview of Actuator Principles . . . . . . . . . . . . . . . . . . . . . . . . 192
9.1.2 Actuator Selection Aid Based on its Dynamics . . . . . . . . . . . . 196
9.1.3 Gears . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
9.2 Electrodynamic Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
9.2.1 The Electrodynamic Effect and its Influencing Variables . . . 199
9.2.2 Actual Actuator Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
9.2.3 Actuator Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
9.2.4 Examples for Electrodynamic Actuators in Haptic Devices . 224
9.2.5 Conclusion about the Design of Electrodynamic Actuators . . 227
9.3 Electromagnetic Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
9.3.1 Magnetic Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
xviii Contents

9.3.2 Design of Magnetic Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

9.3.3 Examples for Electromagnetic Actuators . . . . . . . . . . . . . . . . . 236
9.3.4 Magnetic Actuators in Haptic Devices . . . . . . . . . . . . . . . . . . . 239
9.3.5 Conclusion about the Design of Magnetic Actuators . . . . . . . 241
9.4 Piezoelectric Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
9.4.1 The Piezoelectric Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
9.4.2 Designs and Properties of Piezoelectric Actuators . . . . . . . . . 247
9.4.3 Design of Piezoelectric Actuators for Haptic Systems . . . . . . 258
9.4.4 Procedure for the Design of Piezoelectric Actuators . . . . . . . 258
9.4.5 Piezoelectric Actuators in Haptic Systems . . . . . . . . . . . . . . . 266
9.5 Electrostatic Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
9.5.1 Definition of the Electric Field . . . . . . . . . . . . . . . . . . . . . . . . . 277
9.5.2 Designs of Capacitive Actuators with Air-Gap . . . . . . . . . . . . 279
9.5.3 Dielectric Elastomer Actuators . . . . . . . . . . . . . . . . . . . . . . . . . 285
9.5.4 Designs of Dielectric Elastomer Actuators . . . . . . . . . . . . . . . 289
9.5.5 Electro-Rheological Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
9.6 Special Designs of Haptic Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
9.6.1 Haptic-Kinaesthetic Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
9.6.2 Haptic-Tactile Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

10 Force Sensor Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

10.1 Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
10.1.1 Topology of the Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
10.1.2 Contact Situation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
10.1.3 Mechanical Properties of Measuring Objects . . . . . . . . . . . . . 315
10.1.4 Texture of Measuring Objects . . . . . . . . . . . . . . . . . . . . . . . . . . 317
10.1.5 Selection of Design Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
10.2 Sensing Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
10.2.1 Basics of Elasto-mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
10.2.2 Resistive Strain Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . 325
10.2.3 Piezoresistive Silicon Sensoren . . . . . . . . . . . . . . . . . . . . . . . . . 329
10.2.4 Further Resistive Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
10.2.5 Capacitive Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
10.2.6 Optic Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
10.2.7 Piezoelectric Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
10.2.8 Exotic Ones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
10.3 Selection of a Suitable Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352

11 Application of Positioning Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357

11.1 Basic Principles of Position Measurement . . . . . . . . . . . . . . . . . . . . . . 357
11.1.1 Incremental Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
11.1.2 Absolute Measurement Principle . . . . . . . . . . . . . . . . . . . . . . . 358
11.2 Requirements in the Context of Haptics . . . . . . . . . . . . . . . . . . . . . . . . 359
11.3 Optical Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
11.4 Magnetic Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
Contents xix

11.5 Other Displacement Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365

11.6 Electronics for Absolute Positions Sensors . . . . . . . . . . . . . . . . . . . . . 367
11.7 Acceleration and Velocity Measurement . . . . . . . . . . . . . . . . . . . . . . . 368
11.7.1 Integration and Differentiation of Signals . . . . . . . . . . . . . . . . 369
11.7.2 Induction as a Velocity Measure . . . . . . . . . . . . . . . . . . . . . . . . 370
11.7.3 Force Sensors as Acceleration Sensors . . . . . . . . . . . . . . . . . . 371
11.8 Conclusion on Position Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . 371

12 Interface Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

12.1 Border Frequency of the Transmission Chain . . . . . . . . . . . . . . . . . . . 374
12.1.1 Bandwidth in a Telemanipulation System . . . . . . . . . . . . . . . . 374
12.1.2 Bandwidth in a Simulator-System . . . . . . . . . . . . . . . . . . . . . . 374
12.1.3 Data Rates and Latencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
12.2 Concepts for Bandwidth Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
12.2.1 Analysis of the Required Dynamics . . . . . . . . . . . . . . . . . . . . . 377
12.2.2 Local Haptic Model in the Controller . . . . . . . . . . . . . . . . . . . 377
12.2.3 Event-Based Haptics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
12.2.4 Movement Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
12.2.5 Compensation of Extreme Dead Times . . . . . . . . . . . . . . . . . . 379
12.2.6 Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
12.3 Technical Standard Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
12.4 Final Remarks about Interface Technology . . . . . . . . . . . . . . . . . . . . . 385

13 Software Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387

13.1 Overview About the Subject “Virtual Reality” . . . . . . . . . . . . . . . . . . 388
13.2 Design and Architecture of VR-Systems . . . . . . . . . . . . . . . . . . . . . . . 391
13.2.1 Hardware Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
13.2.2 Device Integration and Device Abstraction . . . . . . . . . . . . . . . 392
13.2.3 Software Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
13.2.4 Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396
13.2.5 Subsystems for Rendering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
13.2.6 Decoupling of the Haptic Renderer from other Sense
Modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
13.2.7 Haptic Interaction Metaphors . . . . . . . . . . . . . . . . . . . . . . . . . . 403
13.3 Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
13.3.1 Virtual Wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
13.3.2 “Penalty” Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
13.3.3 Constraint-based Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
13.3.4 6 DoF-interaction: Voxmap-PointShell-Algorithm . . . . . . . . . 415
13.4 Collision Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
13.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429

14 Final Remarks on the Design of Haptic Systems . . . . . . . . . . . . . . . . . . . 431

xx Contents

Part III Appendix

15 URLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

16 Mechanical Impedances and Admittances for Translatory and

Rotatory Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443

17 Details about Gyrators and Transformers . . . . . . . . . . . . . . . . . . . . . . . . . 445

18 Impedance Values of Grasps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Symbols

Symbol Description Unit

m
a acceleration s2
a stimulus-constant
a displacement (Denavit-Hartenberg-parameter) m
a vector, summarizing actuator displacement
and angles ai (chapter 8)
δa virtual displacement of drives (chapter.8) m
A area m2
ACoil cross section of a winding m2
AConductor surface of a coiling m2
ACore area of the conductive copper area m2
AIron cross-section of the iron core m2
AG cross-section of the airgap of an electromag- 2
m
netic actuator
An cross-section of the element n of a magnetic 2
m
circuit
Δ A j , j=1,2,3 surface element m2
A(ω ) amplitude response (chapter 7) dB
A matrix of a linear system of equations
α roll-angle (rotation around x-axis) degree
αc border-angle (chapter 8) degree
αi rotation (Denavit-Hartenberg-parameter)
degree
(chapter 8)
αV K coefficient of thermal expansion K−1
b0 transfer factor for differential movements

xxi
xxii Symbols

Symbol Description Unit

BIron V·s
magnetic flux density in the iron core T= m2
BG magnetic flux density of an electrodynamic ac-
T
tuator
Bn magnetic flux density in the element n of a
T
magnetic circuit
Br remanence flux density T
B magnetic flux density T
B matrix of a linear system of equations
BHmax J
energy density m3
β pitch angle (rotation around the y-axis) degree
c perceptional constant
ci jlm m2
elastic constants N
C,CQ A·s
capacity F= V
C0 photoelastic coefficient of a Bragg-lattice
Cb coupling capacity (at mechanical full-stop) F
Cd N
field control constant of an ERF V2
Cs N
material-specific field constant of an ERF V2
C matrix of a linear system of equations
ΔC
C0 capacity change
circ circumference of a conductive coil m
d N
damping/friction m·s
di j,k V
piezoelectric charge constant m
di displacement (Denavit-Hartenberg-parameter) m
D dielectric displacement/ electrical displace- C
ment density m2

D (transmission-)matrix of a linear system of

equations
DL difference limen/ amplitude resolution
DOF degrees-of-freedom
ΔD position-discrete resolution
δ phase difference (chapter 10)
e A·s
piezoelectric voltage coefficient m2
ei directional unit vector
E N
E-modulus m2
E electrical field strength (with dielectric mate- V
rial) m
Symbols xxiii

Symbol Description Unit

E0 electrical field strength (without dielectric ma- V

terial) m

Ere f reference field strength, with Cs of an ERF be- V

ing given m
E V
electrical field m
ε permittivity (ε = ε0 · εr ) A·s
V·m
ε relative dielectric constant at constant mechan- A·s
ical tension V·m
ε remaining error (chapter 7)
ε0 electrical field constant (ε0 = 8, 854 · C
10−12 V·m
C
) V·m
εr relative permittivity (εr = EE0 )
η basic-viscosity m2
s
f frequency Hz
f0 , fR resonance-frequency Hz
fb border-frequency Hz
ftot sum of all joint degress-of-freedom of a mech-
anism
fi , i . . . g degree-of-freedom of the ith joint in a mecha-
nism
fid sum of all identical bindings in a mechanism
fink dynamics of the detection of all increments for
Hz
positioning measurement
f (.) static non-linearity
F bearing-/movement-DOF of a mechanism
Fa force in direction of a border a N
FA force acting upon the point A N
Fb force in direction of a border b N
Fe unit force on the mass m of a system N
FF counter-force N
Fin input-force N
Fl force in direction of l N
FN normative force N
Fnorm normal force N
Fw transformer input force N
Fx force in x-direction N
Fy force in y-direction N
xxiv Symbols

Symbol Description Unit

Fz,el force in an electrical field in z-direction N
Fη velocity dependent viscosity force N
Fτ field dependent strain term N
Fξ force at position ξ N
F force N
FD device related force N
Fm force on mass m N
F noise disturbing force N
F out output/generated force N
F force N
Fa vector of all force/torques on the driven joints N
FF counter-force N
FLorentz L ORENTZ-force N
Fx vector of all forces/torques on the tool-center-
N
point
Fx force for displacement in x-direction N
Fy force for displacement in y-direction N
Fz,el force in the electrical field in z-direction N
ΔF force-resolution N
Δ Fi , i = 1, 2, 3 replacement-forces N
φ magnetic flux Wb=V · s
φ (ω ) phase plot degree
Φ phase-angle rad
Φ phase-displacement rad
Φ stimulus
g number of joints in a mechanism (chapter 2)
g V·m
piezoelectric constant N
G transfer function
GCD transfer function of a controller
GDn , n ∈ ℵ device-related transfer function
GED transfer function of a driver (transformation of
a force signal in energy)
GFF feedforward-transfer function
GFIP transfer function for the transformation of per-
ception of mechanical oscillations
Symbols xxv

Symbol Description Unit

GFSense transfer function of a force sensor
GHn , n ∈ ℵ user-related transfer function
GK transfer function of a compensator
GM transfer function of a measurement unit
GR transfer function of a controller
G(s) transfer function in L APLACE domain
GS transfer function of the control/the actuator
GSZ noise transfer function
γ yaw angle (rotation around the z-axis) degrees
γ̇ shear-rate s−1
h height m
H A
magnetic field strength m
Hc A
coercitive field strength m
HFe A
magnet. field strength of a coil with iron core m
HG magnet. field strength in the air gap of a iron A
core m

Hn magnet. field strength in the element n of a A

magnetic circuit m
i current A
iL current through and inductance L A
iSource source current A
iw transform input current A
I moment of inertia m4
Ib current at the input of an operational amplifier A
j A
current density m2
J= ∂x Jacobi-matrix with actuator DOF a and DOFs
∂a of the tool center points x
JND Just-Noticeable-Difference
geometrical design dependent constant of
k m·s
ERFs
k fill-factor of a coil (≥ 1)
k gradient in the P OPOV-criterion
k coupling-factor or k-factor
k chain-number in a mechanism
k N
stiffness (mechanical) m
kL N
stiffness in idle mode m
xxvi Symbols

Symbol Description Unit

K user reaction
Kkrit critical amplification
KR controller amplification
K̃ perceptional space
ΔK time dependency of perception curve dB
K force-perception
K amplification matrix
l length m
lConductor length of a conductor m
lFe length of a coil with iron core/of a magnetic
m
circuit
lMag length of a rare-earth magnet m
ln length of the element n of the magnetic circuit m
L V·s
inductivity H= A
LM Successiveness Limen s
L amplification matrix
λ mechanical control ratio
λ wavelength m
λ0 original wavelength at a defined position in the
m
spectrum
λb Bragg-wavelength m
Λ Bragg-lattice period m
m mass kg
Mα Input torque in axial direction of the drive Nm
Mb bending moment Nm
M torque Nm
M0 torque-source
μ movability of a charge-carrier m2
V·s
μ frictional coefficient
μ mean value
μ permeability (μ = μ0 · μr ) V·s
A·m
μ0 magnetic field constant H
m
μr relative permeability
n, N number ∈ ℵ
n N
elasticity m
Symbols xxvii

Symbol Description Unit

n0 , ni refraction index
nK elasticity of the coupling of a mechanical N
short-circuit m
δ n̄e f f ektiv medium value of index modulation
Δ nRK elasticity of the rotary coupling of a mechani- N
cal short-circuit m
NConductor number of coilings
NA numerical aperture
ω rotational velocity s−1
ω, Ω angular velocity rad
s
p, P N
pressure m2
P C
polarization m2
Pel electrical power loss W = V·A
pel , Pel A·V·s
electrostatic pressure m3
PLoss power loss W
Pmech mechanical power W
PSource source power W
Pη N
presssure loss due to viscosity m2
Pτ N
field dependent pressure loss m2
π piezoresistive coefficient m2
N
πl piezoresistive coefficient in longitudinal direc- m2
tion N

πq piezoresistive coefficient in transversal direc- m2

tion N
Ψ perception amplitude
q, Q electrical charge C = A·s
qi , i ∈ ℵ driven joint i
q vector of a driven joint qi
r distance m
r radius m
ri , i ∈ ℵ active resistors Ω= V
A
R electrical resistance Ω
R0 electrical basic resistance Ω
Rcoil coiling resistance Ω
Ri , i ∈ ℵ reference resistance Ω
RL conductor resistance Ω
xxviii Symbols

Symbol Description Unit

Rm A
magnetic resistance/reluctance V·s
RMag A
magnetic resistance of a rare-earth magnet V·s
RmFe A
magnetic resistance of an iron circuit V·s
RmG magnetic resistance of an air gap within a mag- A
netic circuit V·s
RSense sensing resistor of a PWM Ω
Rspez f . specific resistance Ω
dR
R0 relative resistance change
Δ Rinch position resolution given in dots-per-inch DPI
Δ Rmm position resolution given in millimeter mm
ρ small number ≥ 0
ρ density kg
m3
ρ specific resistance/conductivity Ω ·m
s m2
elasticity coefficient at a constant field strength N
s measure for the level of modulation in an opti-
cal fibre
s sum of the passive links of a mechanism
S external supply rate
S mechanical tension/deformation
S constraints in a mechanism
Si , i ∈ ℵ ith switch
Sx transversal-tension m
Sz longitudinal compression m
σ conductivity S
m = A
V·m
σ standard deviation
t time/point in time s
tr transmition ratio of a gear
T N
mechanical tension m2
T time constant s
Tan rise time (100% of the target value are reached
s
for the first time)
TD rate time s
Tl mechanical tension in longitudinal direction mN2
Tmax point in time for xd,max s
TN reset time constant s
Symbols xxix

Symbol Description Unit

Tq N
mechanical tension in transversal direction m2
Tt dead time s
T transparency
T N
tension m2
T special case of the homogeneous transforma-
tion matrix
Tε transient time until ε s
TCP tool center point
τ shear force N
τ time constant of the step response of an elec-
s
trical transmission system (τ = RL )
τERF yield stress of an electrorehological fluid N
m2
τF,d dynamic yielding point N
m2
τF,s static yielding point N
m2
τMRF yield stress of a magnetorheological fluid N
m2
τy yield stress N
m2
θ rotational angle (Denavit-Hartenberg-
degree
parameter)
Θ magnetomotive force A
Θ mass inertia kg · m2
Θc accepted angle degree
ϑ temperature K
u voltage V
uC voltage at a capacitor V
uRi , i ∈ ℵ voltage at a resistor Ri V
uSource source voltage V
u multidimensional input value of a linear sys-
tem
U electrical voltage V
Ub supply voltage V
Uin input voltage V
Uind induced voltage V
Upull−in voltage at which the pull-in happens V
USense current-proportional voltage at RSense V
va m
velocity at point A s
ve m
input velocity of a system s
xxx Symbols

Symbol Description Unit

vexplor m
velocity of a movement s
vin m
input velocity s
vm m
velocity of a mass m s
vmax m
maximum velocity s
v m
velocity s
v0 m
velocity of a linear movement s
vD m
movement velocity of a device s
vH m
user-related velocity s
vind m
velocity of an induced movement s
vout m
ouput velocity/ generated velocity s
vspo m
velocity of a spontanious movement s
vZ m
velocity at a mechanical impedance s
V magnetic tension A
V volume m3
VERF volume of a electrorheological fluid m3
VMRF volume of a magnetorheological fluid m3
V (x) scalar nonlinear positive definite stoarge func-
tion of system states x
V pre-filter matrix
V̇ m3
volume flow s
ΔV volume-element m3
w general value for in- and ouput values
w wave-amplitude of the stator m
Wel kg·m2
electrical work/energy J= s2
Wel,ERF electrical energy of an electrorheological fluid J
Wel,MRF electrical energy of a magnetorheological fluid J
Wmag magnetic energy J
Wmech mechanical energy J
δ Wα virtual work of a propulsion cylinder J
δ Wx virtual work of an actuator J
x distance m
xd (t) control deviation of a closed-loop control cir-
cuit
xd,max maximum overshoot of a closed-loop control
circuit
Symbols xxxi

Symbol Description Unit

x displacement/translation m
x = (x, y, z) cartesian coordinates
x inner states of a linear system
x vector summarizing translations and angles x j
of the control handle
δx virtual displacement of the tool-center-point m
Δx position resolution m
Δx position of a point in three-dimensional space
X transformation constant
ξ displacement m
ξG air-gap length in a magnetic circuit m
dξ spatial displacement
y control value
y multidimensional output value of a linear sys-
tem
Y gyratoric transformation constant
Y m
admittance N·s
YH m
user-related admittance N·s
zi disturbance variable
z(t) disturbance variable in a closed-loop control
circuit
Z N·s
impedance (usual mechanic) m
ZD N·s
display-/interface-impedance m
ZD N·s
impedance of the device m
ZH N·s
user-impedance m
Z in N·s
impedance as input value m
Z max N·s
maximum impedance/ full load m
Z min N·s
minimum impedance in idle movement m
Z out N·s
impedance as output value m
Z rot M N·s
rotatory impedance of a motor (= α ) m
Z transl translatory impedance of a motor (= Fv ) N·s
m
Z width impedance width, measure of performance N·s
(Z width = Z max − Z min ) m
Structure of this Book

This book subdivides into two major parts. Part I “Some Basics of Haptics” presents
the sense of touch- and movement from different scientific and popular perspectives.
It defines the specific area, in which context “haptics” will be used throughout this
book. Starting from the definition of terms, an introduction into the biological basics
of haptic perception is given, to gain some insight into the specialties of this spa-
tially distributed sensing organ, and sharpens the reader’s own senses for generating
technical illusions for haptics. Afterward some basic structures and classes of haptic
systems are introduced, and first requirements on the technical design are derived.
The first part ends with a discussion about the possibilities to quantify haptic per-
ception illustrated with several examples.

Part II “Designing Haptic Systems” deals with the actual technical design pro-
cess in all facets relevant for haptic systems. Following a chapter on requirement
identification, methods for control engineering support the identification of interde-
pendencies between the components within a system. A chapter on the engineering
of kinematics, with an emphasis on the frequently - in the context of haptic devices
- used parallel kinematics, and their calculation builds the basis for the next steps in
the design process. The most compelling chapter of part II covers actuator design.
All important actuation principles are introduced and discussed for their applica-
tion in haptic devices. Common actuation principles are detailed and described up
to a point, where any engineer should be able to start with own designs. Within the
chapter about force-sensors technology especially relevant for telemanipulation and
closed-loop haptic devices is discussed. The next chapter adds information about
positioning sensors and realizable resolutions. A presentation of relevant interfaces
closes the hardware design part. The subsequent chapter about software design in-
troduces terminology from a neighboring discipline, which is very relevant for any
device engineer - especially in the context of simulator and virtual-reality environ-
ments. The part ends with some finalizing remarks about the design of haptic sys-
tems.
 Part I
Some Basics of Haptics
“This page left intentionally blank.”
 3

The design of haptic devices aims at deceiving the human sense of movement
and touch. In the first part of the book, a general view of this task, starting with
the terminology and biology of haptics is given. Additionally, the application of
psychophysical insights into technological problems is discussed.

• Chapter 1: The motivation for the design of haptic systems is presented. The
significance of haptics in a social context and in professional life is given in
order to emphasize the developer’s responsibility when fooling this sense.
• Chapter 2: A specific terminology is introduced and the understanding of various
disciplines participating in haptic science and research is detailed.
• Chapter 3: A basic understanding of the biological receptors responsible for
haptic perception is conveyed. This enables the reader to assess the parameters
of the technological influence and enhances his sensibility for the complexity of
the haptic sense.
• Chapter 4: A technically motivated model of the human user as mechanical load
- as a consequence of the technological influence on haptic perception - is given
(section 4.2) for the system, and a frequency dependent parameter (section 5) for
its perception. Together this results in a model of the human-machine interaction
and the perception of dynamic mechanical, physical magnitudes.
• Chapter 5: The most important classes of haptic devices based on this model are
introduced.

In the second part of the book the focus switches from the haptic sense to the
technical design of haptic devices.
“This page left intentionally blank.”
Chapter 1
Motivation and Application of Haptic Systems

T HORSTEN A. K ERN

1.1 The Meaning of Haptics from a Philosophical and Social

Viewpoint

Haptics describes the sense of touch and movement. An engineer tends to describe
haptics in terms of forces, elongations, frequencies, mechanical tensions and shear-
forces. This of course makes sense and is important for the technical design process.
However haptics is more than that. Haptic perceptions range from minor interactions
in everyday life, e.g., drinking from a glass or writing this text, to a means of so-
cial communication, e.g. shaking hands or giving someone a pat on the shoulder,
and very personal and private interpersonal experiences. This chapter deals with
the spectrum and influence haptics has on the human being beyond technological
descriptions. It is also a hint for the development engineer, to be responsible and
conscious when considering the capabilities to fool the haptic sense.

1.1.1 Haptics as a Physical Being’s Boundary

Haptics is derived from the Greek term “haptios” and describes “something which
can be touched”. In fact the consciousness about and understanding of the haptic
sense has changed many times in the history of humanity. A RISTOTELES puts the
sense of touch in the last place when naming the five senses:
1. sight
2. hearing
3. smell
4. taste
5. touch

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_1, 5

c Springer-Verlag Berlin Heidelberg 2009
6 1 Motivation and Application of Haptic Systems

Nevertheless he attests this sense a high importance concerning its indispensabil-

ity [3]:

The social estimation of the sense of touch experienced all imaginable phases.
Frequently it was afflicted with the blemish of squalor, as lust is transmitted by it:

“Sight differs from touch by its virginity, such as hearing differs from smell and
taste: and in the same way their lust-sensation differs [289].”

It was called the sense of excess [78] . In a general subdivision between lower
and higher senses, touch was almost constantly ranged within the lower class. In
western civilization the church once stigmatized this sense as forbidden due to the
pleasure which can be gained by it. However in the 18th century the public opinion
changed and K ANT is cited with the following statement [126]:
“This sense is the only one with an immediate exterior perception; due to this it is
the most important and the most teaching one, but also the roughest. Without this
sensing organ we would be able to grasp our physical shape, whose perception the
other two first class senses (sight and hearing) have to be refered to, to generate
some knowledge from experience.”

K ANT thus emphasizes the central function of the sense of touch. It is capable of
teaching the spatial perception of our environment. Only touch enables us to feel and
classify impressions collected with the help of other senses, put them into context
and understand spatial concepts. Although stereoscopic vision and hearing develop
early, the first-time interpretation of what we see and hear, requires the connection
between both impressions perceived independently and information about distances
between objects. This can only be provided by a sense, which can bridge the space
between a being and an object. Such a sense is the sense of touch. The skin, being a
part of this sense, covers a human’s complete surface and defines his or her physical
boundary, the physical being.

Wearing glasses is another fascinating example of the effect of the relationship be-
tween distance and perception. Short- sightedness requires glasses that demagnify
the picture of the environment on the retina due to the distance between eyeball and
lenses. Shortsighted people have a different view of size, e.g. concerning their own
body height, dependent on whether they wear glasses or contact lenses. At every
change between both optical aids the perception of their body has to adapt. Depen-
dent on a person’s kind of defective vision this is a consciously perceivable process.
It can be performed within seconds by using the well known references of one’s own
arms which touch things or one’s legs which walk.
1.1 The Meaning of Haptics from a Philosophical and Social Viewpoint 7

Especially in the 20th century art deals with the sense of touch and plays with
its meaning. Drastically the furry-cup (fig. 1.1) makes you aware of the significance
of haptic texture for the perception of surfaces and surface structures. Whereas the
general form of the cup remains visible and recognizable, the originally plane ce-
ramic surface is covered by fur. “Fighting the mud” (fig. 1.2) remembers you that
not only hands and fingers are relevant for haptic perception, but that the whole body
surface is able to touch and feel. In 1968 the “Pad- and Touch-Cinema” (fig. 1.3)
allowed visitors to touch VALIE E XPORT’s naked skin for 12 seconds through a box
being covered by a curtain all the time. According to the artist this was the only
valid approach to experience sexuality without the aspect of voyeurism [70]. These
are just a few examples of how art and artists played with the various aspects of
haptic perception.

Fig. 1.1 M ERET O PPENHEIM: furry-cup 1936 [70, 186].

Also with virtual worlds and surroundings, haptic interaction has characteristics
of artistry. This is repeatedly demonstrated by expositions during the Worldhap-
tics Conferences. At the same time, Prof. I SHII from MIT Media Laboratory or the
Graduate School of Systems and the Information Engineering group of the Univer-
sity of Tsukuba of Prof. I WATA demonstrate startling exhibits (fig. 1.4) of “tangible
user interfaces” (TUI). These interfaces couple visual displays with haptically re-
configurable objects to provide intuitive human-machine interfaces. There is much
more to find when the senses are sharpened to search for it.
The sense of touch can be a lot of things, e.g. a limitation of the physical being,
which helps to assess distances and calibrate other senses like vision, as well as a
means of social communication and a mediator of very personal experiences .Ad-
ditionally it is - like all the other senses - a target of art which makes us aware of
the importance of haptic experiences by fooling, distorting and emphasizing them.
. Besides these facets of the haptic sense, its function and its dynamic mechanical
8 1 Motivation and Application of Haptic Systems

Fig. 1.2 K AZUO S HIRAGA: Doro ni idomu (Fighting the mud) 1955 [70][217].

Fig. 1.3 VALIE E XPORT: Pad- and Touch-Cinema 1968 [70].

properties are also very impressive. Haptic perception in all its aspects is presented
in the following section.

1.1.2 Formation of the Sense of Touch

As shown in the prior section, the sense of touch has numerous functions. The
knowledge of these functions enables the engineer to formulate demands on the
technical system. It is helpful to consider the whole range of purposes the haptic
sense serves. However, at this point we do not yet choose an approach by measuring
its characteristics, but observe the properties of objects discriminated by it.
1.1 The Meaning of Haptics from a Philosophical and Social Viewpoint 9

Fig. 1.4 Example for “Tangible Bits”, with different data streams accessible by opening bottles.
In this case single instrumental voices are combined to a trio [109] .

The sense of touch is not only specialized on the perception of the physical
boundaries of the body, as said before, but also on the analysis of surface prop-
erties. Human beings and their predecessors had to be able to discriminate e.g. the
structure of fruits and leaves by touch, in order to identify their ripeness or whether
they were eatable or not, like e.g. a furry berry among smooth ones. The haptic
sense enables us to identify a potentially harming structure, like e.g. a spiny seed,
and to be careful when touching it,in order to obtain its content despite its dangerous
needles. For this reason, the sense of touch has been optimized for the perception
and discrimination of surface properties like e.g. roughness. Surface properties may
range from smooth ceramic like or lacquered surfaces with structural widths in the
area of low μ m, to somewhat structured surfaces like coated tables and rough sur-
faces like coarsely woven cord textiles with mesh apertures in the range of several
millimeters. Humans have developed a very typical way how to interact with the-
ses surfaces enabling them to draw conclusions based on the underlying perception
mechanism. A human being moves his or her finger along the surface (fig. 1.5),
allowing shear forces to be coupled to the skin. The level of the shear forces is de-
pendent on the quality of the frictional coupling between the object surface and the
skin. It is a summary of the tangential elasticity of the skin depending on the normal
pre-load resulting from the touch Fnorm and the velocity Fexplr of the movement and
the quality of the coupling factor μ .
Everyone who has ever designed a technical frictional coupling mechanism
knows that without additional structures or adhesive materials viscous friction be-
tween two surfaces can hardly reach a factor of μr = 0.1. Nevertheless nature, in
order to be able to couple shearing force more efficiently into the skin, has “in-
10 1 Motivation and Application of Haptic Systems

m
Fnorm

vexplor

Fig. 1.5 Illustration for the interaction of movements, normal forces on the finger pad and frictional
coupling.

vented” a special structure at the most important body-part for touching and explo-
ration: the fingerprint. The epidermal ridges couple shearing forces efficiently to the
skin, as by the bars a bending moment is transmitted into its upper layers. Addition-
ally these bars allow form closures within structural widths of similar size, which
means nothing else but canting between the object handled and the hand’s skin. At
first glance this is a surprising function of this structure. When one looks again, it
just reminds you of the fact that nature does not introduce any structure without
a deeper purpose. Two practical facts result from this knowledge: First of all the
understanding of shear-forces’ coupling to the skin has come into focus of current
research [65] and has resulted in an improvement of the design process of tactile
devices. Secondly, this knowledge can be applied to improve the measuring accu-
racy of commercial force sensors by building ridge-like structures [275]. Additional
details of the biological basics of tactile perception are given in chapter 3.
Consequently the sense of touch, as said before, has been developed for the dis-
crimination of surface structures. Although the skin may be our most sensitive or-
gan, it is still not the only haptically relevant one. Additional receptors are located
within muscles and joints, which enable us to get an impression of acting forces.
Anyone who has ever lifted a four pound weight (e.g. a well filled pitcher) with
an outstretched arm in a horizontal position, will have little recollection of the tac-
tile surface properties of the handle. The much more impressive experience of such
an experiment is the tensing up of the muscles, their slowly increasing fatigue and
the resulting change in the lifting angles of the joints. This is called “kinaesthetic
perception”. Whereas tactile perception describes forces (≈ 5 mN..5 N) and elonga-
tions between skin and object which are low in amplitudes (≈ 1 μ m..1 mm) and high
in frequencies (≈ 10 Hz..1000 Hz), kinaesthetic perception happens within muscles
and joints at higher forces but with lower dynamics (≈ static..10 Hz). This enables
the human being and every other biological system with a firm supportive structure
- may it be bones or shells of chitin - to perform coordinated movements and tar-
geted interactions with its environment. While tactile perception generates similar
impressions during passive (e.g. a relative movement between a static finger-tip and
a moving surface) and active (e.g. a relative movement between a static surface and
a moving finger-tip) movement, kinaesthetic perception is more complex and influ-
1.1 The Meaning of Haptics from a Philosophical and Social Viewpoint 11

enced by additional factors. The human being is able to change deliberately his or
her mechanical properties. A handshake of the same person can be firm and rigid,
but it may be also loose and amicable. The coupling between muscles, joint position
and perception enables us to consciously influence the kinaesthetics of ourselves,
and to influence the intensity of our kinaesthetic perception in one and the same
situation. This makes us capable of blocking a blow with the same hands we use to
rock a baby to sleep. It gives us the ability to touch a structure carefully before we
grasp it firmly. The borders between action and reaction, active and passive become
blurred in the perspective of kinaesthetics. The awareness of this fact is important
for the requirements on systems with closed-loop control, which are important for
the design of haptic devices (chapter 5). At the same time this adaptability of the
human being and the connected ambiguity of the system’s borders are a significant
challenge for the design of a technical device.

1.1.3 Special Aspects of the Design Process

The design of any technical system always includes a long chain of compromises.
The achievement of the engineer lies in the selection of those compromises which
ensure that existing requirements are still fulfilled. Often these compromises are fi-
nancially motivated - a product should be inexpensive during the production process
without losing performance. Concerning these demands, an optimization of systems
with interfaces to other purely technical systems is often elegantly possible. The
technical systems are quite exactly known as to their characteristics and a technical
design can anticipate these characteristics with a certain security margin. Thus the
interpretation of a sensor capturing the rotation of a wheel, e.g. a speedometer, is
a relatively clear task. The necessary speeds are known, and disturbance variables
like temperature areas as well as humidity. can be identified Alternatively they can
be measured with high exactness. It is also relatively easy to identify the require-
ment of measuring a two-dimensional movement of a human operated device on a
level surface - e.g. a computer mouse. The temperature range of the appliance is
known; the disturbance variables are limited to the optical measurement path and
the mechanical surface state and can easily be investigated. Only the speed is not
given as precisely as by a technical system. It results from the consideration about
the maximum speeds a human hand can reach. Here uncertainties become evident,
soon. Although the dynamics of human movement can be measured - technically,
a high variance between different people will be observed. This variance also con-
cerns the technical requirements of any object used by humans, and may it only be
the physical dimensions of tables and chairs. Dealing with such variances, matching
measuring methods and statistical analysis methods have entered anthropometric
modeling up to ergonomic design of work-places [153] as well as ergonomic stan-
dardization ISO norms 9241/DIN 33 402 1 . The science of anthropometrics applies

1 Ergonomics is the science of the conformity of human work with natural law.
12 1 Motivation and Application of Haptic Systems

to static (lengths, dimensions) and dynamic (speeds) cases. As a matter of fact: Ev-
ery human’s-applicable characteristic value is affected by such a wide variance that
with the information of ergonomic or anthropometric data only a proportional esti-
mation can be made. These estimations are called percentiles (fig. 1.6). A percentile
is a percentage of the totality of the data subject to analysis (e.g. European female
children between 10 and 15 years) and, depending on the context, encloses all peo-
ple who exceed or are below the percentage.

Fig. 1.6 Anthropometric design for sitting and standing work places considering the 5% and the
95% percentile according to DIN 33406.

With regard to the description of body masses and dynamics the description via
percentiles introduced above is well established, as it fits quite well the natural vari-
ance of people. With regard to the description of senses and their performances aver-
age values are more common, e.g. when using a threshold 2 . Thresholds themselves
are a key parameter in finding physical values to quantify human performance. De-
rived from such values the technical system’s requirements like amplitude, ampli-
tude change or dynamics can be employed for deceiving a human sense and for
generating a “realistic” or “sufficient” haptic impression. The choice of words al-
ready shows that requirements seldom comprise a concrete verifiable measurement.
They mostly represent a well-known structure, so that a group of people - or just
the superior or the board of directors - is content with its haptic impression. For the
design engineer this is an unsatisfying benchmark. Alternatives will be discussed to
a large extent. in the course of this book and especially in chapter 6.

2 The use of mean values for the description of the performance of sense organs is doubtful as it
would be surprising that our senses variability are met best with a Gaussian distribution. Never-
theless it is found in almost all publications. At least reporting variance should be provided when
talking about mean values to improve the quality of the results for interpretation. A concept com-
parable to the approach of percentiles can hardly be found in the relevant literature. It is likely that
the number of experiments and available data is not yet sufficient to use this concept.
1.2 The Significance of Haptics in Everyday Professional Life 13

1.2 The Significance of Haptics in Everyday Professional Life

The importance of haptics for professional life differs dependent on the profes-
sion considered. In handcraft or manual trades the word ‘hand’ already implies the
relevance of haptics for performing these jobs. No bricklayer, carpenter, butcher,
plumber or barber would be able to do his job, if the sense of touch did not give
then important information about the material they work on. May that be the hair
they hold between their fingers, the humidity of the wall (as a change of heat trans-
mission), the cable core within the insulation, the difference between tendons and
muscles, the graining of pine and beech trees, the consistency of mortar. Even with
today’s state-of-the-art technology the involvement of man increases with the re-
quired complexity and carefulness of a manual work. With this involvement and
the use of sense of touch the tools usually become less complex. Whereas during
archaeological excavation a first layer of earth is removed with an excavator, when
approaching a hidden structure a shovel will be used or maybe a spatula or for preci-
sion work a brush or even the bare hands. However even in handcraft jobs machines
of increased flexibility made people turn away from the workpiece and its haptic
properties. Today master craftsmen criticize apprentices either for not having any-
more a sensation for materials and their properties or for lacking the information
-based technological know-how for the control of machines. By optimizing the in-
terface between manual work and machine-programming, engineers try to overcome
this gap. But in other areas of professional life, not only in jobs carrying the word
“manual” in their name, the loss of the sense of touch for everyday work has already
taken place.

1.2.1 The Sense of Touch in Everyday Medical Life

In many medical disciplines high manual skills are required. The capabilities of the
sense of touch are necessary for diagnostics and therapy, be it for the identification
of skin diseases, the diagnosis of joints, and the palpation of inner organs from the
outside or via natural openings; or for a direct surgical application like the trans-
plantation of a heart, the sawing of the cranium or the punctuation of the spinal
cord. The sense of touch transmits a plurality of information about texture, elastic-
ity and temperature to the medical professional - information which would either
be inaccessible or not so easily accessible in other ways. Nevertheless, in certain
situations it is necessary to substitute the sense of touch in diagnosis and therapy.
Via magnetic resonance imaging e.g. tendons and menisci of the knee can be visu-
alized. Thus a demanding manual examination of the joints’ movement range is not
necessary; especially as performing the procedure and interpreting the haptically
felt data requires experience and still leaves room for misinterpretation. Addition-
ally the results of a manual investigation are harder to explain to the patient than
the distinctiveness of a real image. However, when comparing the expenses of both
14 1 Motivation and Application of Haptic Systems

diagnostic procedures, the precedence should be given to the haptic diagnosis. A

compromise can be seen in devices like the “Wristalyzer” [77]. This device either
puts varying loads on a moving joint - the wrist - or actively moves it, while dy-
namically measuring the angle vs. displacement curves. Additionally it acquires a
complete electro-myography of the muscles. Besides for diagnosis, devices of this
kind are already planned for therapy. By actively generating forces and torques,
they can be used for the training of all joints of our extremities, of the cervical spine
and of the pelvis. Considering all these factors, there seems to be a tendency for the
mechanization in diagnostics and therapy. In orthopedic areas there is, however, still
some room to discuss its necessity, whereas in surgery there is an urgent need for
mechanization which, however, leads to a loss of haptic impressions. After surgical
interventions like e.g. an appendectomy, the wish for small wounds and scars for
medical and cosmetic reasons has therefore led to the design of laparoscopic instru-
ments (fig. 1.7). Simply by their length, mass and stiffness they also resemble a filter
for the haptic information. This decoupling between patient and surgeon has found
its temporary climax in the DaVinci system (fig. 1.8) - a laparoscopic telemanipula-
tion system without force feedback. This loss of the sense of touch during surgical
(or any other internistic) interventions is obvious and regrettable. As a result numer-
ous research projects were and still are focusing on an adequate substitute for the
direct haptic interaction by alternative technologies [73] or improved instruments
with integrated force-feedback [209] (fig. 1.9).

Fig. 1.7 Rigid laparoscopic instrument by Karl Storz.

1.2.2 The Sense of Touch in the Cockpit

Besides the aim of getting information which is already mechanically available

(elasticity, surface structure, etc.), there is the necessity to provide artificially gener-
ated tactile data in addition to overloaded visual or auditory senses in information-
1.2 The Significance of Haptics in Everyday Professional Life 15

Surgeon at control station Surgical roboter system

DaVinci

Fig. 1.8 Surgical telemanipulator DaVinci

R
by Intuitive Surgical, installation in Munich.

Control-
unit

15 mm

Actuator housing
with ultrasonic
actuators

Parallel-kinematic intrument tip

Fig. 1.9 Functional muster of a hand-held laparoscopic telemanipulator with increased number of
degrees of freedom at the instrument’s tip, such as a prepared intracorporal force measurement
with haptic feedback on the control unit [209].
16 1 Motivation and Application of Haptic Systems

loaded working places. Such working places can be found in control stations where
the human has to make time critical and responsible decisions, e.g. within a jet, air-
plane or at the steering wheel of a common car. The designers of a cockpit typically
choose between visual, acoustic and haptic transmission paths. Even the choice of
a scroll-wheel with hard stops instead of a pure incremental sensor is influenced by
the knowledge that a selection within a certain range can be much faster done if
the limits of this range are explicitly given [14]. Control knobs like the i-drive in
a BMW allow a reconfiguration of its haptic properties during operation. Warning
signals are already given via vibrating motors or so called “tactons”. Especially in
the military area a complex spatial orientation based on vibrating clothing (fig. 1.10)
for marines and flight-personnel is subject to actual research [267, 115], whereas
active sidesticks in military and civil airplanes and vibrating braking assistance or
in-lane guidance in cars are already established.

Fig. 1.10 West equipped with vibrators for the spatial coding of positioning and bearing data
(TNO, Netherlands) [267].

1.2.3 The Sense of Touch at the Desk

There is hardly any other job where the sense of touch has lost so much of its sig-
nificance than in the office. Just a few decades ago the use of paper, pens in a large
variety, rulers, folders and files was a joyful source of haptic information for the
sense of touch. Today the haptic interface to an office working place is defined by
a keyboard and a mouse. Due to this extreme focus on a single type of haptic in-
terface for a variety of things, the ergonomics of a keyboard is of extraordinarily
high importance. Besides the switching characteristics of the key itself, the surface
structure and the tactile markers on the letters F and J (fig. 1.11) and the size of
the key are necessary and considerable design criteria. ISO 9241-400 defines clear
1.2 The Significance of Haptics in Everyday Professional Life 17

decision paths for both, the designer and the buyer of keyboards. Nevertheless it is
beyond doubt that major ergonomic improvements are not done by the optimization
of keyboard and mouse, but by improvements of office software ergonomics. Con-
trary to many cases where the term “interface” refers only to the graphical interface,
R ASKIN’s “The Humane Interface” [203] is a decided and enjoyable collection of
software with unergonomic graphical interfaces offering methods and design crite-
ria for their improvement.

Fig. 1.11 Keyboard of the author with tactile markers on the letters J and F.

1.2.4 The Sense of Touch in Music

If regarded from an abstract standpoint, haptic sense and acoustic perception have
multifarious parallels. Both are sensitive to the perception of mechanical oscillations
and cover a comparable frequency range. Thereby the haptic sense rather perceives
frequencies covering two decades below 1 kHz, whereas the acoustic sense rather
perceives frequencies up to two decades above 100 Hz. Music quite often makes
use of these parallels which may be used to perceive the oscillations of the string of
a valuable violin or harp; or to touch the soft vibration of a wind instrument giving
a low A. They are even to be found in studio technology. Devices like the “But-
tKicker” (fig. 1.12) from The Guitammer Company are electrodynamic actuators
which are used as tactile feedback devices during concerts. They transmit the lower
frequency range to the drummer giving the rhythm of the band without drowning his
own instrument. Additionally the acoustic pressure for the musicians is reduced, as
they may not necessarily want to be exposed to the same loudness as their excited
audience. These kinds of actuators are also suitable for e.g. the couch in a home
18 1 Motivation and Application of Haptic Systems

cinema or chairs in front of gaming PCs to increase the perception of bass-intense

effects. Here again, the tactile effect is of similar intensity as the perception of a
bass impulse, connected with the advantage that little acoustic pressure is emanated
resulting in almost no disturbing noise for people around.

Fig. 1.12 Electrodynamic actuator “ButtKicker” for generating low-frequency oscillations on a

drum-stool.
Chapter 2
Terminology

T HORSTEN A. K ERN

In the introduction a number of terminologies originating from the context of

haptic science and device design has already been used. In this chapter a systematic
introduction into the area of designing haptic devices begins. The following sections
explain the scientific and industrial disciplines participating in the research and de-
velopment of haptic devices. Afterward terms and their definitions are introduced
and illustrated by examples how to characterize haptic systems based on some con-
crete technical devices.

2.1 Scientific Disciplines as Part of Haptic Research

In haptic science there are three groups of interest (fig. 2.1) with quite fluent bor-
ders in between: Scientists working within the area of “haptic perception” proceed
according to strictly deductive scientific principles: Resulting from an observation a
hypothesis is derived. For this hypothesis an experiment is designed by testing the
point of the hypothesis by the exclusion of other varying parameters. As a result the
hypothesis is veri- or falsified leading to a new and improved hypothesis.

Research in the area of “haptic perception” is done by two scientific disciplines:

Psychophysics and Neurobiology. Psychophysics deals with the analysis of the im-
pression of physical stimuli - in the case of haptic perception this mainly refers to
oscillations and forces of different spatial orientation. The aim of psychophysics is
to create a model explaining perception. Neurobiology observes biologically mea-
surable connections and analyzes the direct conversion of physical stimuli into neu-
ronal signals and their processing within the brain. Both disciplines complement

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_2, 19

c Springer-Verlag Berlin Heidelberg 2009
20 2 Terminology

product
knowledge
demands Service oriented research

Haptic Industry
Measure
! Standardisation and
normative approach
! Force measurement
! Vibration engineering
Science of basic principles ! Translation, acceleration Functional oriented research
and velocity measurement
Haptic Haptic
Perception Synthesis

Psychophysics Neurobiology Virtual Reality Telemanipulation

! Perception of ! receptor types and ! Hardware Interfaces ! Hardware Interfaces

- maximum and densities - tactile - tactile
minimum forces ! interconnectivity of - kinesthetic - kinesthetic
- absolute thresholds receptors - Information - actuator and
- resolutions and ! nerve fibre transmissions processing kinematic design
dependencies ! place and procedure of - actuator and ! Hardware endeffector
- Dynamic information processing kinematic design - Force measurement
! Interaction ! reactions ! Software Simulation ! Control engineering for
- strategies of - Algorithm closed loop systems
- spontanious reactions - Graphic and haptics ! Telepresence
! ... - Augmented reality ! Communication and data
! Control engineering for delay
time-discrete systems

Fig. 2.1 Overview about the disciplines participating in haptic research.

each other so that the neuronal observation should be able to explain a part of the
psychophysical model and vice versa. These scientific disciplines formulate tech-
nical tasks for the preparation of experiments which are processed by two groups
interested in “haptic synthesis” or “haptic measurement”, respectively
On an alternative track both groups get assignments from industry making them-
selves use of the knowledge gathered by research on haptic perception. These groups
work according to engineering solution strategies An assumption of requirements is
derived from a technical question based on the current state of knowledge. A func-
tional prototype and later a product to fulfill the requirements is designed in a devel-
opmental process accompanied by a continuous tracking of the prior assumptions
and their meaning. Then the product obtained can be used for the analysis of psy-
chophysical questions, or, respectively as a a product of the gaming-, automotive or
aviation industry.
In the case of the generation of haptic impressions for Virtual-Reality (VR) appli-
cations the technical requirements typically ask for tactile, kinaesthetic or combined
feedback systems. In that area the emphasis is on the correct choice of actuators,
control and driver electronics and on the processing and transmission of signals.
Due to the coupling of devices and time-discrete simulation systems a consideration
of discretization-effects and their influence on the haptic quality of the impression
is necessary. In the case of telemanipulation systems technical challenges are com-
2.2 Terms and Terminology Used for the Description of Haptic Systems 21

parable. The main difference lies in the necessary measurement technology for the
acquisition of haptic object properties. Additionally, the control engineering ques-
tions are more complex, as this area typically deals with closed-loop systems with
unknown loads on both ends.

2.2 Terms and Terminology Used for the Description of Haptic

Systems

The definition of the terminology within the context of haptic systems is subject
to the current ISO 9241-910 norm. Many of the definitions used in this book fol-
low the terminology presented there. According to the author’s experience, all these
terminologies have the status of recommendations shared by a large number of re-
searchers associated with the haptic area. However, there is no binding consent for
their usage within the haptic community, so that many actual and future papers differ
from the definitions presented above. The nomenclature mentioned here is based on
prior publications to this material, especially by H AYWARD [90], C OLGATE [176],
H ANNAFORD [85], B URDEA [34], A DAMS [2] and many papers by other authors.

2.2.1 Basic Concepts of Haptics

Haptics means the combined sensation of mechanical, thermal and noci-perception

(fig. 2.2). It is more or less defined by the exclusion of the optical, acoustic, ol-
factory and gustatory perception from the sum of sensory perceptions. As a result
haptics consists of nociceptive, thermoceptive, kinaesthetic and tactile perceptions.
The sense of balance takes an exceptional position as it is not counted among the
five human senses having receptors of their own Yet, it really exists making use of
all other senses’ receptors, especially the haptic ones.

Haptics describes the sensory as well as the motor capabilities within the skin,
joints, muscles and tendons.
Tactile means the mechanical interaction with the skin. Therefore tactile percep-
tion is the sensation of exclusively mechanical interaction. Please note that tactile
perception is not exclusively bound to forces or movements.

Kinaesthetics describes both, actuatory and sensory capabilities of muscles and

joints. It refers to their forces, torques, movements, positions and angles. As a result
any kinaesthetic interaction has a tactile component due to this definition.
22 2 Terminology

Human Senses:

haptic

Smelling Hearing Equilibrum Thermal tactile

reception reception reception

Vision Tasting Nocio Kinaesthetic

reception reception

Fig. 2.2 Distribution of senses.

2.2.2 Definition of Haptic Systems

The technical terminology is listed from the special to the general and illustrated
by block diagrams. The arrows between the components of the block diagrams may
represent different kinds of information depending on the devices they refer to. They
remain unlabeled. Haptic devices are capable of transmitting elongations, forces and
temperature differences and in a few realizations they also stimulate pain receptors.

The terms “system” and “device” and “component” are not defined on an interdis-
ciplinary basis. Dependent on one’s point of view the same object can be e.g. “a
device” for a hardware-designer ,“a system” for the software-engineer, or “just a
component” for another hardware-engineer. These terms are nevertheless part of any
engineering discipline and are used accordingly here but should anyhow be read with
this knowledge in mind.

A haptic device is a system generating an output which can be perceived hapti-

cally. It has (fig. 2.3) at least one output, but not necessarily any input. The tactile
markers on the keys F and J of a keyboard represent information for the position-
ing of the index finger. By these properties the keys are already tactile devices. At
a closer look the key itself shows a haptically notable point of actuation, the hap-
tic click. This information is transmitted in a kinaesthetic and tactile way by the
interaction of the key’s mechanics with the muscles and joints and the force being
transmitted through the skin. Such a key is a haptic device without a changing input
and two outputs.

A user (in the context of haptic systems) is a receiver of haptic information.

A haptic controller describes a component of a haptic system for processing

haptic information flows and improving transmission. Quite pragmatic in the case
of telemanipulation systems these kinds of controllers are frequently either a spring-
2.2 Terms and Terminology Used for the Description of Haptic Systems 23

damper coupling element between end-effector and the operating element or a local
abstraction model of the area of interaction to compensate transmission delays. In
the case of a haptic simulator it is quite frequently a simple LTI-model with a high
in- and output rate. The LTI model itself is then updated on a lower frequency than
the actual speed of the haptic in- and output.

Haptic Haptic
controller device User

Fig. 2.3 Haptic device, user and controller.

Haptic interaction describes the haptic transmission of information. This trans-

mission can be bi- or unidirectional (fig. 2.4). Moreover, specifically tactile (unidi-
rectional) or kinaesthetic (uni- or bidirectional) interaction may happen. A tactile
marker like embossed printing on a bill can communicate tactile information (the
bill’s value) as a result of haptic interaction.

Fig. 2.4 Haptic interaction.

The addressability of haptic systems refers to the subdivision (spatial or tempo-

ral) of an output signal of a device (frequently a force) or of the user (frequently a
position).

The resolution of a haptic system refers to the capability to detect a subdivision

(spatial or temporal) of an input signal. With reference to a device this is in accor-
24 2 Terminology

dance with the measuring accuracy. With respect to the user this corresponds to his
perceptual resolution.

A haptic marker refers to a mark communicating information about the object

carrying the marker by way of a defined code of some kind. Examples are markers
in Braille on bills or road maps. Frequently these markers are just tactile, but there
are also kinaesthetically effective ones marking sidewalks and road crossings for
visually handicapped people.

A haptic display is a haptic device permitting haptic interaction, whereby the

transmitted information is subject to change (fig. 2.5). There are purely tactile as
well as kinaesthetic displays.

A tactor is a haptic purely tactile haptic display generating a dynamic and os-
cillating output. They usually provide a translatory output (e.g. fig. 9.19), but could
also be rotatory (e.g. fig. 2.14).

Fig. 2.5 Haptic display.

A haptic interface is a haptic device permitting a haptic interaction, whereby the

transmitted information is subject to change and a measure of the haptic interaction
is acquired (fig. 2.6). A haptic interface always refers to data and device.

Force-Feedback (FFB) refers to the information transmitted by kinaesthetic in-

teraction (fig. 2.6). It is a term coined by numerous commercial products like FFB-
joysticks, FFB-steering wheels and FFB-mice. Due to its usage in advertising, the
term Force Feedback (FFB) is seldom consistent with the other terminology given
here.

A haptic manipulator is a system interacting mechanically with objects whereby

continuously information about positions in space and forces and torques of the in-
teraction is acquired.
2.2 Terms and Terminology Used for the Description of Haptic Systems 25

Fig. 2.6 Haptic interface.

A telemanipulation system refers to a system enabling a spatially separated hap-

tic interaction with a real physical object. There are purely mechanical telemanipu-
lation systems (fig. 2.7), scaling forces and movements via a lever-cable-system. In
the area of haptic interfaces, there are mainly electromechanic telemanipulation sys-
tems according to figure 2.8 relevant. These systems allow an independent scaling
of forces and positions and an independent closed-loop control of haptic interface
and manipulator.

Fig. 2.7 Mechanical telemanipulator for handling dangerous goods (CRL model L) .

A haptic assistive system is a system adding haptic information to a natural in-

teraction (fig. 2.9). For this purpose object or interaction properties are measured via
a sensor and used to add valuable information in the interaction path. An application
would be a vibrating element indicating the leaving of a lane in a drive assistance
system.

A haptic simulator is a system enabling interaction with a virtual object (fig. 2.10).
It always requires a computer for the calculation of the object’s physical properties.
Haptic simulators and simulations are important incitements for the development of
26 2 Terminology

Fig. 2.8 Scheme of an electric telemanipulator.

Haptic Haptic
Sensor controller device
+ User
+
Object

Fig. 2.9 Haptic assistive system.

haptic devices. They can be found in serious training applications, e.g. for surgeons,
as well as in gaming applications for private use (see also chapter 13).

Fig. 2.10 Haptic simulator.

2.2.3 Parameters of Haptic Systems

In [156] L AWRENCE defines the transparency T as a factor between impedance as

the input source of the haptic interface Z in and the actually felt output impedance
Z out of the device.
2.2 Terms and Terminology Used for the Description of Haptic Systems 27

Z in
T= . (2.1)
Z out
The principle of transparency is mainly a tool for control engineering purposes
analyzing stability and should be within the range ± 3 dB. T may be regarded as
the sole established, frequency dependent, characteristic value of haptic interfaces.
Frequently only the transparency‘s magnitude is considered. A transparency close
to “one” shows that the input impedance is not altered by the technical system. The
user of the haptic device being the end of the transmission chain experiences the
haptic input data in a pristine way. The concept of transparency can be applied to
telemanipulation systems and as well as to haptic simulators .
In [39] C OLGATE describes the impedance width (Z-width) of a haptic system

Z − width = Z max − Z min (2.2)

as the difference between the maximum load Z max and the perceivable friction
and inertia at free space movement Z min . The Z-width describes the potential of de-
vices and enables the comparability between them, after technical changes, e.g. by
the integration of a closed-loop control and a force measurement.

Active haptic devices are systems requiring an external energy source for the
display of haptic information. Usually, these are at least haptic displays. Passive
haptic devices, on the contrary, are systems transmitting haptic information solely
by their shape. This may lead to a false conclusion: A passive system in a control
engineering sense is a system with a negative energy flow at its input, e.g. a system
not emitting energy into the outside world. This concept of passive control is an
important stability criterion which will be discussed in detail in subsection 7.3.3.
For the moment, it should be noted that a passive haptic system is not necessarily
identical with a haptic system designed according to the criterion of passivity1 .

The mechanical impedance Z is the complex coefficient between force F and ve-
locity v respectively torque M and angular velocity Ω . Impedance and its reciprocal
value - the mechanical admittance Y - are used for the mathematical description of
dynamic technical systems. High impedance means that a system is “stiff” or “in-
ert” and “grinds”. Low impedance describes a “light” or “soft” and “sliding” system.
The concept of impedance is applied to haptic systems by way of the terms display-
impedance or interface-impedance Z d . It describes the impedance a system shows
when it is moved at its mechanical output (e.g. its handle).The concept of impedance
cannot be applied only to technical systems, but also to a simplified model of the
user and his mechanical properties. This is described by the term user-impedance
Z H . User-impedance - how stiff a user tends to be - can be influenced at will up
to a certain point. Shaking hands can either be hard or soft depending on its fre-
quency. The mechanical resistance of a handshake is lower at low frequencies and
1 However, in the meaning of this definition, a passive haptic device is indeed a system which is
passive, according to the control-engineering classification Yet not all passive systems meeting this
criterion of control-engineering are necessarily passive haptic devices.
28 2 Terminology

higher at high frequencies resulting simply from the inertia of the hand’s material.
Detailed descriptions of the building of models and the application of the concept
of user-impedance are given in section 4.2. An introduction into calculating with a
complex basis and mechanical systems is given in appendix 16. Understanding com-
plex calculation rules and the mechanical impedances are fundamental to the design
of haptic devices in the context of this book. Therefore it is recommended to update
one’s knowledge by self-studies of the relevant literature of electromechanics [158]
and control-engineering [167].

2.2.4 Characterization of Haptic Object Properties

Besides the terminology for haptic systems, there is another group of terms describ-
ing solely haptic objects and their properties:
Haptic texture refers to those object properties, which can exclusively be felt by
touch. The roughness of a surface, the structure of leather, even the haptic markers
already mentioned are haptic textures of the objects they are located on. In some
cases a differentiation is made between tangential and normal textures, whereby
the directional information refers to the skin’s surface. This specific differentiation
is more a result of technical limitations, than of a specialty of tactile perceptions
as tactile displays are frequently unable to generate a feedback covering a two or
three-dimensional movement.

Haptic shape refers to object properties which can mainly be felt kinaestheti-
cally. This can be the shape of a cup held in one’s hand. But it can also be the shape
and geometric design of a table rendered to be touched in a virtual environment.

In fact terms like texture and shape are used analogically to their meaning in
graphical programming and software techniques for 3D objects, where meshes pro-
vide shape and surface-textures give color and fine structures. However, in com-
parison with graphical texture, haptic texture mainly describes three-dimensional
surface properties incorporating properties like adhesion or friction, i.e. a realis-
tic haptic texture is much more complex in its parameters than a typical graphical
texture, even when considering bump-, specular or normal-maps. Therefore numer-
ous haptic surface properties, e.g. specific haptic surface effects are defined and
described from the perspective of a software engineer. These surface effects are
partly derived from physical equivalents of real objects, narrowed down to software-
motivated concepts in order to increase the degree of realism of haptic textures:

• Surface friction describes the viscose (velocity-proportional) friction of a con-

tact point on a surface.
• Surface adhesion Surface adhesion describes a force binding the movement of
a contact point to a surface. This concept allows simulating magnetic or sticking
effects.
2.2 Terms and Terminology Used for the Description of Haptic Systems 29

• Roughness describes an uniform, sinoid structure of a small, defined amplitude

making the movement of a contact point on a surface appears rough.

Tacton refers to a sequence of stimuli adressing the tactile sense. It usually en-
codes an event within the sequence’s pattern. The stimuli vary in intensity and fre-
quency. Both, stimuli and tacton, may even be overlayed with a time-dependent
amplitude modulation, such as fade-in or fade-out.

2.2.5 Technical examples

In the following section several technical examples will demonstrate the usage of the
terminology introduced before and will give further insight into the interdependency
of several terminology groups. The systems presented here should be understood
as illustrating examples, not necessarily as technological milestones or devices of
extraordinary performance.

2.2.5.1 Force-Feedback Interfaces

There are several commercial haptic control units available on the market for the ap-
plication in design, CAD and modeling. One major player on the market is SensAble
with their PHANTOM-series
R and the actually most low-cost product PHANTOM
Omni (fig. 2.11a). The PHANTOM-series can most easily be identified by the free
positioning of a pen-like handle in a three dimensional space. The position and
orientation of this handle is measured in three translational and three rotational de-
grees of freedom. Depending on the model of the series, the tip force can act on the
handle in at least three translational dimensions. The generation of forces is done
via electrodynamic actuators; depending on the model these are either mechani-
cally or electronically commutated. The actuators are located within the device’s
basis and transmit their mechanical energy via levers and Bowden cables on the
corresponding joints. As a result of changing level-lengths the transmission-ratio
of the PHANTOM devices is nonlinear. For the static situation these changes are
compensated within the software driver. The PHANTOM devices are connected to
common PCs. The electrical interface used depends largely on the device’s product
generation and ranges from parallel ports to IDE cards and FireWire connectors.
The PHANTOM devices from SensAble are haptic devices (fig. 2.11c) primarily
addressing the kinaesthetic perception of the whole hand and the arm. As the force
transmission happens via a hand-held pen tactile requirements are automatically rel-
evant for the design too. This bidirectional haptic display is a haptic interface to the
user transmitting force information of a software application in a PC and feeding
back positioning information to her or him.
30 2 Terminology

M0 F0 Fout
F0=M1/l
Q dR m
W1 v0 ZH
v0=W1 l

a) b)

FSignal Fout
Haptic
device User
xSignal xout
c)

Fig. 2.11 Haptic tool with a SensAble PHANTOM Omni(a), R and corresponding mechanical
network of one degree-of-freedom (b), such as block-structure (c).

The network model of one degree of freedom (fig. 2.11b) shows the electronic
commutated electrodynamic motor as an idealized torque source M 0 with inertia of
Θ of the rotor and a rotary damping dR resulting from bearings and links. By the
use of a converter resembling levers the rotary movement is transformed in a linear
movement with a force F 0 and a velocity v0 . An inertia m describes the mass of
the hand-held pen. The portion of the generated force F out is dependent on the ratio
between the sum of all display-impedances against the user impedance Z H .

2.2.5.2 Reconfigurable Keyboard

The reconfigurable keyboard (fig. 2.12a) is made of a number of independent ac-

tuators arranged in a matrix. The actuators are electrodynamic linear motors with
a moving magnet. Each actuator can be controlled individually either as an open-
loop controlled force source or as a positioning actuator by a closed-loop control.
When being used as force source, the primary purpose of the actuator is to follow
a configurable force/displacement curve of a typical key. The application of this re-
configurable keyboard [46] is an alternative to the classical touchscreen - a surface
providing different haptically accessible functions depending on a selection within
a menu. For this purpose single actuators can be combined to larger keys and may
change in size and switching characteristics.
The reconfigurable keyboard is a haptic device (fig. 2.12c) mainly addressing the
kinaesthetic sensation, but has strong tactile properties, too. The user of the device is
the controller of the keyboard, receiving haptic information in form of the changing
shape of keys and their switching characteristics during interaction. The keyboard is
2.2 Terms and Terminology Used for the Description of Haptic Systems 31

at least a haptic display. As it communicates with another unit about the switching
event and the selection, it is also a haptic interface.

F0 Fout

m d
ZH

a) b)

F/x curve FSignal Fout

Haptic
Haptic
device User
controller
keypress- xSignal Key xout
event

c)

Fig. 2.12 Haptic key as a reconfigurable actuator with a single degree-of-freedom (a), and corre-
sponding mechanical network (b), [46] such as block-structure (c).

The network model (fig. 2.12b) of a single key shows the open-loop controlled
force source F 0 of the electrodynamic actuator, the mass of the moving magnet m
and the friction in the linear guides d. Elasticity does not exist, as the design does
not contain any spring. This is in contrast to what could be expected from the typical
designs of electrodynamic speakers and their membranes. The actuator is capable
of generating a force Fout dependent on the ratio between the complex impedance
of the haptic display Z D = s m + d and the user’s impedance Z H .

2.2.5.3 Tactile Pin-Array

Tactile pin-arrays are the archetype of all systems generating spatially coded infor-
mation for the haptic sense. Conceptually they are based on Braille-displays whose
psychophysical impression has been studied comprehensively since the middle of
the 20th century, e.g. by B ÉKÉSY [23]. Many approaches were made ranging from
electromagnetic actuators of dot matrix printers [232] to piezoelectric bending actu-
ators [149] and pneumatic [292], hydraulic [231], electrostatic [290] and thermal [5]
actuators. Tactile pin arrays mainly focus on the skin’s stimulation in normal direc-
tion. Only lately spatially resolved arrays with lateral force generation are receiving
an increased interest [142].
32 2 Terminology

A tactile pin-array with excitation in normal skin direction is a haptic device

(fig. 2.13c) mainly addressing the tactile perception. The user is in continuous hap-
tic interaction with the device and receives haptic information coded in changing
pin heights. A tactile pin array is a haptic display. In contrast to the systems exam-
ined before this time the user’s interaction does not include any user-feedback. As a
result the device is not necessarily a haptic interface2 .

k Fout

v0 vH ZH

a) b)

xSignal xout

Haptic
display User
Fout
c)

Fig. 2.13 Classical tactile display with pins oriented normal to the display surface [276] (a), and
corresponding mechanical network (b), such as its block-structure(c).

In the mechanical network model (fig. 2.13) a tactile pin array corresponds to
a positioning or velocity source v with a mechanical stiffness k in series to it (a
combination of actuator and kinematics). In a stiff design the mechanical admittance
of the technical system is small resulting in the elongation being totally unaffected
by the user’s touch. The system is open-loop position controlled.

2.2.5.4 Vibration-Motor

Vibration-motors are used to direct attention to a certain event. There is a vibration

motor similar to figure 2.14a within every modern mobile phone, made of a rotary
actuator combined with a mass located eccentrically on its axis. Its rotation speed is
controlled by the voltage applied. It typically ranges from 7000 to 12000 rotations
2 There are possibilities for tactile systems measuring the position of the finger and reporting it to

a control unit. But this functionality is not typical of this kind of displays.
2.2 Terms and Terminology Used for the Description of Haptic Systems 33

per minute (117 to 200 Hz). It is possible to encode information into the felt vibra-
tion by varying the control voltage. This is often done with mobile phones in order
to make the ring tone haptically perceptible.
A vibration-motor is a haptic device (fig. 2.14c) addressing tactile perception.
The user is haptically interacting with the device and receives haptic information
in the form of oscillation coded in frequency and amplitude. A vibration-motor is a
pure form of a haptic display, or more precisely a purely tactile display.

k
F0 Fout

m
d ZH

a) b)

xSignal xout

Haptic
display User
Fout
c)

Fig. 2.14 Vibration-motor of a mobile phone (a), and corresponding mechanical network (b), such
as block-structure (c).

With vibration motors the relevant force effect is the centripetal force. Assuming
a rotational speed of ω = 2π 1000060RPM Hz and a moving mass of 0.5 g on a radius
of 2 mm a force amplitude of F = m ω 2 r = 1.1 N is generated, producing a sinoid
force with a peak-to-peak amplitude of 2.2 N. This is an extraordinary value of an
actuator with a length of only 20 mm. Considering the network model (fig. 2.14b)
the vibratory-motor can be regarded as a force-source with sinoid output. It has to
accelerate a housing (e.g. the phone) with a mass m which is coupled to the user
via an elastic material, e.g. clothes. It is important for the function of the device
that the impedance of the branch with spring/damper coupling and user-impedance
Z H is large against the mass m. This guarantees that most of the vibration energy is
directed to the user, thus generating a perception.
“This page left intentionally blank.”
Chapter 3
Biological Basics of Haptic Perception

T HORSTEN A. K ERN

3.1 The Sense of Touch and its Biology

Examples of haptic systems and the importance of the haptic sense have been dis-
cussed in the preceding chapters without actually giving an exact idea of the function
of haptic perception. For the design of haptic systems it is vital to have a basic under-
standing of characteristic biological parameters, as only these will help to identify
relevant technical requirements. This chapter introduces the most important termi-
nology and basics for understanding the neurobiology of haptic perception. Please
note that research on haptic perception is far from being complete. As a consequence
this short presentation of complex biological coherences is a well-founded working
hypothesis which will be extended or confuted by further research. In order to per-
ceive information from our surroundings, man is equipped with five senses: Hear-
ing, Smelling, Tasting, Sight and Touch. The physiology of senses distinguishes five
sensors and sensory-systems [219] differing from this very popular definition. They
allow a classification in a vocabulary lent from a technical approach to describe
things:

• Thermal sensor for registering the change of temperature especially within the
skin,
• Chemical sensors reacting on odorous or gustatory substances,
• Optical sensors reacting on the impact of photons, especially within the cones
and rods in the retina,
• Pain sensors, also named nociceptors, to identify chemical and physical tissue
damage,
• Mechanical sensors for detecting mechanical tensions and strains e.g. within the
skin or muscles.

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_3, 35

c Springer-Verlag Berlin Heidelberg 2009
36 3 Biological Basics of Haptic Perception

The sensory capacity and its importance for haptic perception are valued dif-
ferently. The visual sensors register ≈ 10 Mio. bit/s, the sense of touch ≈ 1 Mio.
bit/s and the acoustic sense ≈ 100 kbit/s [18]. The processing of these sensory data
happens within the cerebral cortex. It is structured in functional brain areas. The pri-
mary motor cortex is the physiological location for processing data from the sense of
touch. A visualization of the distribution of body parts on the primary motor cortex
(fig. 3.1) shows a significant portion being used for fingers and hand.

heaoulder
shm

torso
ar

neck

hip
ellbowrm

leg
forea
wrist

d
han e fin ger
litt g fi

t
foo
rin

m
l
d ge

fo idd toe
thu refin le
n

ey m g fin ia
ital
r

e b er ge gen
nos r
e
face
uppe
r lip
lips

bottom lip
teeth, palatine, jaw
tongue
a vity
es al c
fauc omin
abd

Fig. 3.1 Visualization of the functional brain areas in the motor cortex (somatotrope ordering) [52].

Within the sensorimotor functions the haptic sense has the highest importance.
It consists of a group of mechanical sensors detecting force induced deformations
within tissues in the skin, muscles and joints. As a consequence haptic perception
is the sum of signals from a large number of measurement points distributed among
the human body, consisting of at least 6 types of sensors which can be divided into
two basic groups: Tactile and kinaesthetic sensors (fig. 3.2).
Tactile sensors are located in the outer areas of the skin in exposed positions (e.g.
the fingertips). They react on strains of the skin and are activated either proportion-
ally to the elongation, to the velocity or to the acceleration. The neuro-pyhsiology
distinguishes between four different types of tactile sensors [236, 219]:

• Rapid-adaption or Fast-Adaption (RA or FA-I) Meissner corpuscles - with ve-

locity dependent activation.
3.1 The Sense of Touch and its Biology 37

haptic
receptors

noci tactile kinesthaetic thermal

receptors receptors receptors receptors
- located within the skin - located within muscles and joints
- high dynamic range >100 Hz - low dynamic range < 100 Hz
- reacting on stretch in - reacting on stretch & tension
surrounding tissue

RA PC spindle spindle
SA-I SA-II FA-II stretch tension
FA-I receptors receptors
- rapid/fast - slow adapting - slow adapting - Pacinian corp. - in parallel to - in series to
adapting - Merkel-cell - Ruffini corp. - fast adapting muscle fibres muscle fibres
- Meisner corp. - dx/dt & dx -x - d²x/d²t
- dx/dt - unclear whether
present in
unhaired skin
nuclear nuclear Golgi
bag chain tendon
fibres fibres organ
- small and - force sensing
slender - frequency depend.
- steady
streatching
forces
DBF SBF

- Dynamic Bag - Static Bag

Fibres Fibres
- Velocity dep. - Static forces

Fig. 3.2 Classification of the haptic receptors and their naming, adapted from [236].

• Slow-adapting (SA-I and SA-II) Merkel cells and Ruffini-corpuscles with velocity-
dependent and elongation proportional activation. They show a lower dynamic
response 1 compared to the Meissner-corpuscles.
• Fast-Adaption (FA-II) Pacinian corpuscles with acceleration-proportional acti-
vation.
The distribution of sensors varies within different skin areas (fig. 3.3) and is part
of current research. For example in [193], the existence of Meissner corpuscles has
to be put into question in contrast to established doctrines.
Unlike tactile sensors kinaesthetic sensors are located mainly within muscles,
joints and tendons. They acquire forces affecting whole extremities only. Their dy-
namic requirements are reduced as a result of the mechanical low-pass character-
istics of the extremities (their mass, damping and stiffness). Their requirements on
the relative resolution between the smallest perceivable force vs. the maximum de-

1 Dynamic refers to the width in frequency domain the actuator or sensor is designed for. There is
an important difference to amplitude-dynamic referring to the height of the output from an actuator
or the maximum input value of a sensor.
38 3 Biological Basics of Haptic Perception

Cornea
Epidermis
Merkel- Ruffini-
Corium cells corpuscle

Subcutis
Meissner- Pacini-
corpuscle corpuscle

Fig. 3.3 Distribution of the sensors in the skin layers [46].

tectable force (amplitude-dynamics) can be compared with the group of tactile sen-
sors. Kinaesthetic sensors can be divided in two groups:
• spindle-stretch-receptors Dynamic Bag fibres and Static Bag fibres placed in par-
allel to the muscle fibres.
• spindel-tension-receptors Golgi tendon organ - in serial orientation to the muscle
fibres.

When summarizing all information about biological sensors contributing to hap-

tic perception, it is interesting to see that nature chooses a design to identify forces
and vibrations, which does not significantly differ from the technical solution of
comparable problems. However comparable technical solutions are older than the
biological understanding of the sense of touch. Therefore it seems likely that with
the given physical constraints only solutions optimized in such a manner are ade-
quate.
Besides dealing with theses sensors as the first part of the haptic perception chain,
in the next step it is necessary to consider a model for the neurological processing
of haptic information, in order to get a feeling for the complexity of the system and
outline components relevant for the design of technical haptic systems.
Figure 3.4 shows a simplified understanding of neuronal subsystems participat-
ing in a task like “grasping a glass of water”. The motivation phase starts with thirst
due to e.g. the body salinity being too high, and with the knowledge about the avail-
ability of a glass of water. As a result a decision is taken to “seize the glass of
water”. Within a programming phase this decision results in a definition of move-
ments for single extremities and body parts. As a subcomponent each of them has
a position controller, which controls the movement. Feedback is given by the motor
sensors within the joints but also by visual control from a superordinate control loop.
Subordinate to the visual control, a closed-loop circuit with force feedback exists,
enabling the safe and secure holding of the glass based on a maximum force to be
exerted. As an alternative, a feedback loop can be assumed controlling the grasping
force to avoid a slipping through of the glass.
3.2 Haptic Perception 39

Motivation-phase: Decision-phase: Programming-phase:

Target-value upper arm
Salinity
BB Catch the t1
glass of water Target-value lower arm
Blood-sugar A Thirst
& B & C t2
Target-value finger
t3
…
t n

Glass of water
available
Glass
available in
“Recognition”
positon x

Processing phase movement:

t
1
… Differential-position

t
2
… action-potential
t
3 - 1st position
controller
- 2nd position
controller
- Muscles &
locomoter system
t
n Force-controller
& Transformer

Maximum force - “Force sensors”

Actual position value

“Position-sensors”

Fig. 3.4 Neuronal processing chain with the example of grasping a glass of water:
A, B, C are decision elements of no further detailation; τi delay element.

It is remarkable that the analogy to technical control systems can so easily be

drawn. Decision phase, programming phase and processing phase are accepted ref-
erences for central-neural structures [219]. The interconnection between position
and force-controller is a direct result from dynamic ranges and measurement errors
unique to the components of the close-loop circuit. The position control loop in-
cluding the locomotor system and kinaesthetic sensors shows a dynamic range of
≤ 10 Hz [287, 83]. Additionally angle-positioning and absolute position measure-
ment without a line of sight show large errors (2◦ to 10◦ dependent on the joints
participating [34]). Movements including visual control are much more precise.
The visual perception is able to resolute movements with up to 30 Hz depending
on the illumination level. By the aid of sight a human is able to move to a position
and hold it until immediately before physical contact - which is strictly impossible
with closed eyes. On the other hand tactile sensors show a dynamic range of many
hundred Hertz. This capability combined with high amplitude-dynamics enables
humans to hold even slippery and fragile objects without breaking them.

3.2 Haptic Perception

Knowledge about the performance of haptic perception is essential. for the formu-
lation of requirements as a basis for system design, For each body part there are
different characteristic values, as the haptic sense is not located in a single organ.
Additionally haptic interaction is always bidirectional, which means - especially
in case of kinaesthetic interaction - that haptic interaction can be mapped only by
40 3 Biological Basics of Haptic Perception

considering positions and angles of body parts and forces and torques of mechan-
ical specifications,. Furthermore haptic perception is also greatly dependent on the
dynamic excitation in a broad frequency range. Last but not least the aspect of mul-
timodality2 has to be considered. Haptically inconspicuous keys and buttons are
considered to be of high quality when they are accompanied by a loud click-sound,
compared to silent buttons with identical haptic properties. As a consequence of
the complexity of effects influencing haptic perception every characteristic value
taken from literature has to be seen in the context of the individual test design and
weighted with the accuracy of the experiment’s layout. The characteristic values pre-
sented within this chapter shall be taken as points for orientation only, and should
be modified and even disapproved of by future experiments.

3.2.1 Psychophysical Concepts

In order to be able to understand characteristic values of haptic perception, a basic

knowledge about some relevant psychophysical concepts is necessary. The defini-
tions given here are based on G. A. G ESCHEIDER [66], recommended to any reader
interested in this subject.

3.2.1.1 Threshold and Difference-Limen

Two fundamental concepts for the analysis of thresholds are distinguished in psy-
chophysics. On the one hand there is the measurement of thresholds for differential
perception (thresholds of differential sensitivity). On the other hand there are thresh-
olds for absolute perception (thresholds of absolute sensitivity). All measurement
principles in psychophysics can be categorized according to these two principles.
Additionally the analyzed stimuli differ in their dimensions (e.g. space, time, spec-
tral 3 ).
The absolute threshold (fig. 3.5) of a stimulus describes the value, from which
a stimulus φ begins to become perceivable.
As another characteristic value the stimulus’ change is relevant, creating a just-
noticeable difference (JND) 4 . The stimulus’ change is called difference threshold
or alternatively difference-limen (DL). Consequently the DL means the measure-
ment of a Δ φ being the difference to a stimulus φ0 compared to another stimulus
φ1 . The JNDs are numbered discretely as JND being a member of N. The first JND
is the first DL after the absolute threshold; the second JND is that DL following the
sum of the absolute threshold and the first DL (fig. 3.6). To sum up: the JND is the

2 Multimodality: Cumulated influence of different sensory perceptions affecting the perception of

a single event.
3 of different frequency
4 rarely also named just-noteable-difference
3.2 Haptic Perception 41

stnd
mean
stnd

n1 n2 n3 ... nm

Fig. 3.5 Identification of the absolute threshold of a stimulus φ from m tests .

smallest physiological scale unit of the linearized perception of a physical stimulus

φ.

DL = Δ φ = φ0 − φ1 (3.1)

F JND
JND4
F4
JND3
DL3
F3 JND2
DL2
F2 JND1
F1 DL1
DL0 JND0
F0
physical physiological

Fig. 3.6 Concept of the physical scale of the stimulus φ and the linearized physiological scale of
the discrete JND.

The DL of stimuli is analyzed according to different questions. The methods

applied allow conclusions concerning the neuronal processing of stimuli. A classical
method for the analysis of DL is the presentation of a reference stimulus comparing
it to a second stimulus, which is presented to a subject either in an automated way
or manually controlled by the test-person himself (fig. 3.7).
Besides the aspects just mentioned there are others for doing comparable analy-
sis. On the one hand, there is the aspect of masking with the question: “At which
point will two stimuli dependent on a single parameter be perceived as different?”.
Aspects analyzed frequently for masking a time and spatial dependencies. As a re-
sult the terms of temporal masking and spatial masking have been fixed.
42 3 Biological Basics of Haptic Perception

F DF
DFm
DF2
DF1 DF. .

stnd
mean
stnd

n1 n2 ... nm n1 n2 ... nm
n
=F0 =F1 =DF=>JND

Fig. 3.7 Identification of the DL as a result from the Just-Noticeable-Difference (JND) Δ φ be-
tween the stimuli phi0 and phi1 from m experiments.

example dynamic masking: The perception of a change in frequency of a mechan-

ical oscillation of fixed amplitude shall be analyzed. For this purpose two stimuli
are given at the same time to test a subject. The subject is allowed to change the
frequency of one stimulus until he or she detects two independent stimuli. The mea-
sure of the change in frequency Δ f is the value for DL with respect to the reference
stimulus. Results of such a kind of experiments are not always precise and should
be critically analyzed. For example in case of stimulus locations very near to each
other the above experiment can be easily interpreted in a way that only the maxi-
mum amplitude of a summed up signal has been analyzed and not the DL of a fre-
quency change. To prevent this kind of criticism a careful experiment design should
be done with a series of hypotheses for falsi- and verification. In this case an addi-
tional experiment would be adequate showing a statistically significant difference in
the perception of JND between a summed-up amplitude of two stimuli with identical
frequency compared to a signal with two frequencies.

example temporal masking: A stimulus φ0 with a frequency f is presented for a

long period t. Afterwards stimuli φn , e.g. dependent on frequency, are given. The per-
ception of those stimuli (e.g. with regard to the absolute threshold) varies dependent
on the prior period t. The measure of this variation is the temporal-masking-effect
of a certain masking frequency f.

example spatial masking: Two stimuli φ0 and φ1 , e.g. needles on the skin are given
with a spatial distance d. At a certain distance d both stimuli are perceived indepen-
dently from each other. This is a very specific example of spatial masking frequently
used for measuring the resolution of tactile perception. It has therefore been given
its own term: two-point-threshold
3.2 Haptic Perception 43

Another aspect of analysis is the Successiveness Limen (LM) connected with

the question: “How many stimuli presented consecutively can be perceived?”

Example LM: With the help of a vibratory motor a sequence of stimuli is presented
on a body location. The stimuli vary according to a temporal pattern. The LM is the
temporal pace enabling a correct perception of the sequence.

3.2.1.2 Psychophysical Laws

An important way for presenting DL Δ φ is as a value related to a reference stimulus

φ0 according to the formula
Δφ
=c (3.2)
φ0
In 1834, E.H. W EBER found out, that c is a constant quotient for a specific
perception. In his key experiment he placed weights on the skin and found out, that c
1
is almost 30 . This means that the next higher weight differing from a weight of 200 g
is 30 · 200 g + 200 g = 206.66 g. The value c differs significantly between different
1

stimuli, but the comprehensive coherence according to equation 3.2 (Weber’s law)
seems to be universal for many situations. As a consequence Weber’s law allows
putting different senses and their perception in relation to each other. An exception
is the area of lower stimuli (fig. 3.8a) in the range of absolute thresholds where c
increases significantly.

Fig. 3.8 Identification of the Just-Noticeable-Differenz (JND) Δ φ in between the stimuli phi0 and
phi1 of m experiments.
44 3 Biological Basics of Haptic Perception

A modification of Weber’s law

Δφ
=c (3.3)
φ0 + a
compensates this dependency in the range of absolute thresholds (fig. 3.8b). The
constant a is - identical to c - specific for each sense and as compared to c, quite
low. The physiological reason for a has not finally been determined. A hypothesis
existing assumes it to be a measure for the background noise of the corresponding
receptors.
Some senses, especially the acoustic but also the haptic sense, show a nonlin-
ear logarithmic dependency on perceived intensity and physical excitation. For the
range of stimuli, for which Weber’s law is valid according to its original formula
(equ. 3.2) a new dependency can be formulated. This dependency, named Fechner’s
law,

Ψ = k log φ (3.4)
provides a linearized measure Ψ of the perception amplitude.

Today Fechner’s law has mainly a historical significance. In 1975 it was replaced
by S.S. S TEVENS, suggesting a law describing the intensity of a stimulus by an
exponential relation:

Ψ = kφ a (3.5)
This relation is called Power-law and allows comparisons of numerous perception-
dependencies by a look at its constants a and k. If a = 1, the equation 3.5 gives a
linear dependency. At values for a > 1 the law gives a dependency increasing with
increased stimulus, at a < 1 a damping of the perception with increased stimulus is
resulting. When logarithmizing equation 3.5, an interdependency easy to display on
diagrams with logarithmic axis (fig. 3.9) can be obtained.

log Ψ = log k + a log φ (3.6)

with y-axis log k and a slope of a.
Table 3.1 gives an extraction of S TEVENS’ published data [242] of the coefficient
a according to equation 3.5.

3.2.1.3 Mean Values and Percentiles

The analysis of psychophysical measures is always laborious due to large variances

in results either between individual subjects or as in certain tests, among specifically
trained test-persons. As a consequence, statistical design and the application of sig-
nal detecting algorithms should be considered for any such experiments. For details
3.2 Haptic Perception 45

Fig. 3.9 Logarithmic display of hypothetic measurements according to equation 3.6 [66].

Table 3.1 Exponents a of the power function 3.5 relating sensory magnitude to stimulus intensity
- extract taken from [242]

Continuum Measured exponent Valid for stimulus

Sound level 0.67 Acoustic pressure of a 3 kHz signal
Vibration 0.95 Oscillation with 60 Hz of unknown ampltiude at the in-
dex finger
Vibration 0.6 Oscillation with 250 Hz of unknown ampltiude at the
index finger
Taste 1.3 Sugar
Taste 0.8 Sweetener
Coldness 1.0 Metal contact at the arm
Warmness 1.6 Metal contact at the arm
Muscular strength 1.7 Static contraction
Pressure on the palm 1.1 Static force on skin
Finger span 1.3 Thickness of blocks
Angular acceleration 1.4 5-sec rotation
Electrical shock 3.5 Current through finger
... ... ...

of these procedures, literature of statistical experiment design and [66] is suggested.

For a more general perspective the following remarks should be considered:
Frequently, in psychophysics experimental results follow a Gaussian normal dis-
tribution. This happens with regard to a single person as well as with respect to a
larger number of people. Of course, a Gaussian distribution can be characterized
by a mean-value μ and a standard deviation σ . The mean value defines the measure
where exactly 50% of a given set (e.g. of experiments) are above and below that
value.
For the usage with sets not following a normal distribution, the usage of per-
centiles is suggested. Typical examples of their application are anthropometric val-
ues in ergonomics. The x-th percentile gives the point on a scale, where x percent
46 3 Biological Basics of Haptic Perception

of tests of a given set are below that value. In the exceptional case of a normal
distribution the fiftieth percentile is identical to the mean-value (fig. 3.10).

Fig. 3.10 Gaussian distribution with medium value μ and standard deviation σ compared with an
arbitrary distribution characterized by percentiles.

3.2.2 Frequency Dependency

As mentioned in section 2.2, every kinaesthetic interaction has a tactile component.

We know from the analysis of “grasping a glass of water” in section 3.1 that tac-
tile components are part of the interaction’s innermost feedback-loop. As a result
the requirements for their dynamic properties are extraordinarily high. This section
discusses the perception thresholds and difference-limens as identified in neurology
and psychophysics from an engineering perspective. It is therefore a preparation
for the identification of requirements for technical systems interfacing the sense of
touch.
The identification of haptic perceptional dynamics can be performed either with
psychophysical methods or with neurological tools. When focussing on the recep-
tors only, the analysis of tactile and kinaesthetic sensors can be done indepen-
dently from each other. Neuronal potential on nerve fibres can be measured via
interventional implanted electrodes, and even positioned during electrode recording
[250, 131].
In [122] several tactile sensory types (compare fig. 3.2 on page 37) have been
analyzed as to their frequency-dependency (dynamics) and their thresholds for the
detection of skin deformation (fig. 3.11). Frequency areas of slow-adapting (SA) and
rapid-adapting (RA) sensors complement and overlap each other. The SA-II sensor
especially affects a range of ≈ 8 Hz. According to this study the mean threshold of
3.2 Haptic Perception 47

the isolated sensors shows a maximum in sensitivity at around ≈ 300 Hz with an

elongation of 10 μ m.

Fig. 3.11 Frequency dynamics and thresholds for the detection of skin deformations according to
[122], summarized by [119] .

W ILKINSON performed a study [287] on thresholds on isolated Golgi-tendons

receptors (fig. 3.12). The results show an almost linear dependency between the
response of receptors in mV and the stimulation in μ m over frequency. The relevant
frequency range of these receptors is lower than the range of the tactile receptors of
figure 3.11, especially as the masses and stiffnesses of limbs show distinct low-pass
characteristics. High frequency components of forces and elongations are damped
anyway and therefore the kinaesthetic sensory system does not have to be able to
measure it.

Fig. 3.12 Frequency dependent thresholds of isolated Golgi-tendon-receptors [287].

48 3 Biological Basics of Haptic Perception

For the design process of haptic systems the focus point does not lie on sin-
gle biological receptors but rather on the human’s combined perception resulting
from the sum of all tactile and kinaesthetic sensors. In this area numerous studies
have been performed, three of which are given here showing the range of results.
In 1935, H UGONY already published a study about the perception of oscillations
dependent on frequency of mechanical stimuli [101]. Additionally he quantified
different stimuli-levels which are defined from the absolute perception threshold
to pain-thresholds (fig. 3.13a). To complement this general study TALBOT added
details about the interdependency of the isolated biological sensor and perception
(fig. 3.13b). Both scientists showed that the sensitivity of perception increases to
a frequency of ≈ 200 Hz (H UGONY) and ≈ 300 Hz (TALBOT). These two studies
along with several others were compiled by H ANDWERKER [219] resulting in a
combined curve of haptic perception thresholds (fig. 3.14).

Fig. 3.13 Thresholds for the perception of oscillations after H UGONY 1935 [101] (a) and study
from TALBOT about the perception of oscillations [249](b).

A source for the analysis of haptic perception worth to be recommended can

be found in the publications by G ESCHEIDER. He followed a stringent analysis-
methodology and discussion of haptic sensory systems. Beginning in 1970 until at
last in 2002 a series of hypotheses and measurements in numerous publications is
documented. Another source worth to be considered is the work by B ÉKÉSY [23]
and by J OHANSSON.
Next to the already known dependency of haptic perception on frequency, an-
other dependency exists connected to the surface area transmitting the mechanical
oscillations: Large areas of force transmission (A > 1 cm2 ) and small areas of force
transmission (A < 1 mm2 ) differ significantly according to their absolute perception
3.2 Haptic Perception 49

Fig. 3.14 Combined curve of thresholds of haptic perception [219].

thresholds (fig. 3.15). When focusing on kinaesthetic devices, usually a large area
of force transmission exists. With tactile devices smaller force transmission areas
have to be considered. The perception curve is a combination of the four tactile sen-
sor types and shows a minimum (point of maximum sensitivity) at ≈ 350 Hz. The
frequency-dependency is obvious and undoubted. Only the precise shape and the
exact position of the minimum vary in the range of ≈ 100 Hz depending on mea-
surement, author and publication. Additionally it can be noted that the perception
of very low frequencies below 0.1 Hz was not subject to many studies. Typically the
perception curves are assumed to stay constant to lower values from a frequency of
approximately 1 Hz.
Besides frequency-dependency there exist two additional dependencies affecting
haptic perception. Ongoing mechanical stimuli result in a reversible desensitization
of receptors. This time dependency is used in [69] to mask single receptor classes in
order to study the part of other receptor classes in overlapping frequency areas. The
time dependency of perception curves Δ K in dB can be approximated according to

Δ K(t) = 12 · (et )12 . (3.7)

As a result desensitization happens in a time frame of a second (spectral com-
ponents below 10 Hz). As a consequence desensitization is not a matter to be nec-
essarily considered for the design of haptic devices, telemanipulation systems or
simulators due to their large ratio between usage vs. desensitization time frame. A
steady state can be considered for almost all relevant applications. In practical ap-
plication this approximation is not necessarily adequate. For example when tactile
devices based on pin- or shear-force systems are used, there is some evidence, ac-
cording to the author’s purely subjective observation, that the mentioned effect still
happens after minutes of usage.
50 3 Biological Basics of Haptic Perception

Fig. 3.15 Frequency-dependency of the perception of oscillations, acquired with two different
sizes A1 and A2 of the contact area [67] (note: PA-, RA-, SA-receptors from fig. 3.2).

The amplitude-resolution (DL) of haptic perception shows a logarithmic depen-

dency analogue to the visual and acoustic perception. The perception of smallest
changes dependent on frequency with varying base excitation was studied in [68].
Measurements were taken at two frequencies (25 Hz, 250 Hz) and with white noise.
The approved dependency of DL of the amplitude of the base excitation is nonlin-
ear with a maximum difference of ≈ +3 dB. It is larger for smaller amplitudes of
the base excitation. This allows the conclusion that the Power-Law (section 3.2.1.2,
equ. 3.6) can be used for the description of perception. However its coefficients have
to be identified for every contact situation independently.

3.2.3 Characteristics of Haptic Interaction

Besides the dynamics’ curves in the prior section there are numerous insular values
from experiments documenting the possibilities of haptic interaction. The results
can be divided into two groups. In the table of haptic perception (tab. 3.2) the pa-
rameters from a receptive perspective are summarized. In the table of active move-
ments (tab. 3.3) border values of the capabilities of the active parts of motor systems
are summarized. The tables are based on a collection by D OERRER [46] and have
been extended by selected additional sources. However, when considering their ap-
plication, a very important statement of B URDEA [34] still has to be remembered:
“... that it is dangerous to bank on recommendations for the design of haptic devices,
especially when they are taken from different experiments, with varying methods,
and when only a small number of participants took part”. The characteristic values
3.2 Haptic Perception 51

given here can only represent a selection of the analyses presented in literature. For
quite an actual and a very compelling summary [118] is recommended.

3.2.3.1 Characteristic Values of Perception

Table 3.2 Characteristic values of the perceptive capabilities of the human hand.

Base item Characteristic value Body part Value Source

Skin-deformation, absolute Fingertip
10 μ m (b) [123]
value (a) (tactile)
Fingertip
Static 2-3 mm (d, e, f) [34][123]
Two-point threshold (c) (tactile)
elongation /
(Spatial resolution) Palm [123]
position 10-11 mm
(tactile) [230]
finger joint
2.5 ◦ [252]
Position-resolution, (kinaesthetic)
Difference-limen (DL)(g) Wrist
2.0 ◦ [252]
(kinaesthetic)
Frequency, upper limit Finger
5-10 kHz [31][34]
(tactile perception) (tactile)
frequency, upper limit whole body
20-30 Hz [31]
(kinaesthetic perc.) (kinaesthetic)
Fingertip, palm [15][25]
Dynamic Maximum sensitivity at 200-300 Hz
(tactile) [123]
elongations
0.1-0.2 μ m at
(vibration) amplitude, absolute Fingertip, palm [15][25]
200-300 Hz
threshold (tactile) (h, i, j) [123]
Amplitude-resolution, fingertip
10-25 % [25]
difference-limen (DL)(g) (tactile)
Frequency-resolution, Fingertip
8-10 % (k) [25]
difference-limen (DL) (tactile)
Fingertip (l)
0.8 mN [34]
Force, (tactile)
absolute threshold Palm
1.5 mN [34]
(tactile)
Force and
Force, Total body 5-10 % (ca.
pressure [192]
Difference-Limen (DL) (kinaesthetic) 7 %) (m, n, o)
Pressure, Finger
0.2 N/cm2 (p) [229]
Absolute threshold (tactile)
Pressure, Wrist
4-19 % (q) [252]
Difference-limen (DL) (kinaesthetic)
Thumb, index
Torque Difference-limen (DL) finger 12.7 % (r) [114]
(kinaesthetic)
Thumb, index
Elasticity Difference-limen (DL) finger 5-15 % (s, t) [251]
(kinaesthetic)
52 3 Biological Basics of Haptic Perception

Annotations and additions to table 3.2

(a) absolute threshold: starting from this threshold a stimulus is perceived.
(b) If movement is permitted, surface structures of 0.85 μ m height can be perceived [141].
(c) Two stimuli with a spatial distance larger than the two-point threshold can be distinguished
from each other. The spatial resolution is the reciprocal value of the two-point threshold. The
table of two-point thresholds at the human body are compelling. In a general the values
increase from the fingertip at 1 mm up to belly and back with 10 mm. The importance of
two-point threshold is discussed contradictory, as the perception of shear forces is getting
more attention today. Nevertheless the two-point threshold gives a good qualitative
impression of the receptor density on the body surface.
(d) With vibrotactile stimulation the two-point threshold is at the lower limit of 2 mm [123].
(e) The two-point threshold decreases, if the two stimuli are presented short after another [123].
(f) A position change of a stimulus can be resolved spatially up to ten times better than the static
two-point threshold [123].
(g) see section 3.2.1.1
(h) The perception threshold is strongly dependant of the vibration frequency, the location of the
stimulus and the size of the contact area [25][123][230].
(i) Amplitudes larger than 0.1 mm are perceived as annoying at the fingertips[25].
(j) A stimulation with constant frequency and amplitude results in a desensitization, increasing
up to a numb feeling which may last several minutes after the end of the stimulation.
[35][122].
(k) The capability to differ stimuli is reduced after 320 Hz [31].
(l) The absolute-threshold of the force perception is perceived by tactile mechanoreceptors
transmitted by deformations of the skin.
(m)The value of 7 % was identified almost independent of the muscle-system and the absolte
value of the force [192].
(n) A special experiment with forces between thumb and index finger gabe a JND of 5-10 % for
reference forces between 2.5 and 10 N at a finger-distance of 45 to 125 mm. For smaller
forces a larger JND is expected [192].
(o) In an experiment to distinguish masses of different objects JNDs of 10 % were identified
[15].
(p) The tactile perception of the human is very sensitive for pressure gradients and especially for
edges [252].
(q) Experiment made at a reference pressure of 1.8 N/cm2 . JND increased strongly with reduced
contact area: 4.4 % at 5.06 cm2 , 18.8 % at 1.27 cm2 [252].
(r) Experiment with a reference torque of 60 mNm.
(s) Experiment with a reference elasticity of 4 mm/N and a spring elongation of 15 to 35 mm.
(t) The mechanical work lost during compression and deformation of object surfaces is relevant
for perceiving elasticity [34].
3.2 Haptic Perception 53

3.2.3.2 Characteristic Values of Movement

Table 3.3 Characteristic values of the output capabilities of the human hand.

Base item Characteristic value Body part Value Source

Finger (joint nearest
17 rad/s (a) [87]
to the palm)
Velocity Maximum velocity
Finger
18 rad/s (a) [87]
(middle joint)
Finger 5-10 Hz (b) [31]
Arm, unexpected
Bandwidth (Movement 1-2 Hz [239]
Movement reaction
and force)
Arm, periodic
2-5 Hz [239]
movement
Learned trajectory bis 5 Hz [239]
Reflex bis 10 Hz [239]
Index-, middle-,
40-50 N (c) [34]
ringfinger
Thumb 85-100 N [87]
Maximum force Wrist 35-65 N [34]
between 2 fingers:
45-65 N (e) [34]
Force fingertip-grasp (d)
between 2 fingers:
76-109 N [34]
key-grasp (f)
Fingertips (with
40 mN (h) [239]
visual feedback)(g)
Movement precision
Fingertips (without
11-15 % (i) [239]
visual feedback)
Finger (palm-nearest 370-500 Ncm
(j) [87]
joint)
Torque Maximum torque 120-289 Ncm
Finger (middle joint) (j) [87]
Finger (most distal
40-85 Ncm (j) [87]
joint)
54 3 Biological Basics of Haptic Perception

Comments and annotations to table 3.3

(a) For slow movements velocities of 3-6 rad/s for palm-nearest and middle finger-joints are
documented [87].
(b) The bandwidth of the ouput capacity is situation dependent: unexpected signas: 1-2 Hz,
periodic signals: 2-5 Hz, known movement patterns: ca. 5 Hz, reflex-reactions: approx. 10 Hz
[31].
(c) With stretched fingers and movements only at the palm-nearest joint. These values are quite
independent from the actual joint position (flection in between 0 and 80 ◦ ) [34].
(d) Several grasp-positions of the finger have to be distinguished. Fingertip-grasp means, that
only the fingertips are touching the object. It is also known as precision grasp. Details can be
found in [34] and chapter 4.2.1.
(e) The maximum force can not be exerted continuously. At a periodic force excitation of 25 %
of the maximum force for 15 s with a pause of additional 15 s results in an impression of
fatigue after 10 min. At 15 % of the maximum force this impression is reached not until
104 min [286].
(f) With the key-grasp the thumb touches the lateral part of the index finger.
(g) The test subjects were asked to exert a constant force. A monitor was used to give a visual
feedback of the actual force [239].
(h) This medium error was identified to be quite independent of the exerted force between 0.25
and 1.5 N. The force had been kept constant for 14 s [239].
(i) Without visual feedback the error increased significantly and was dependant of the exerted
force [239].
(j) The values were calculated of finger-forces and -geometries[87].

3.3 Conclusions from the Biology of Haptics

Next to studying the pure characteristic values of haptic perception we should keep
an eye on the real meaning of μ m-elongations and frequencies of 1 kHz and more
and on its impact on real technical systems which is a small “finger exercise” for
you to be prepared for the challenges of the design of haptic systems; the idea of
this is based on a talk given by N IEMEYER at the Eurohaptics conference in 2006.

3.3.1 Stiffnesses

Already the initial touch of a material gives us information about its haptic proper-
ties. A human is able to immediately discriminate, whether he or she is touching a
wooden table, a piece of rubber or a concrete wall with his or her finger tip. Besides
the acoustic and thermal properties, especially the tactile and kinaesthetic feedback
plays a large role. Based on the simplified assumption of a double-sided fixed plate
its stiffness k can be identified by the usage of the E-modulus according to equation
[158]
3.3 Conclusions from the Biology of Haptics 55

b h3
k=2 ·E (3.8)
l3
Figure 3.16a shows the calculation of stiffnesses for a plate of an edge length of
1 m and a thickness of 40 mm of different materials. In comparison, the stiffnesses
of commercially available haptic systems are given in (fig. 3.16b). It is obvious
that these stiffnesses of haptic devices are factors of ten lower than the stiffnesses
of concrete, every-day objects like tables and walls. However, stiffness is just one
criterion for the design of a good, haptic system and should not be overestimated.
The comparison above shall make us aware of the fact that a pure reproduction of
solid objects can hardly be realized with a single technical system. It rather takes a
combination of stiff and dynamic hardware, for especially the dynamic interaction
in high frequency areas dominates the quality of haptics, which has extensively been
discussed in the last section.

Fig. 3.16 Comparision between stiffnesses of a 1x1x0.04 m3 plate of different materials (a) and
realizable stiffnesses by commercial haptic systems (b).
56 3 Biological Basics of Haptic Perception

3.3.2 One Kilohertz - Significance for the Mechanical Design?

Haptic perception ranges to a frequency of 10 kHz, whereby the area of highest sen-
sitivity lies between 100 Hz and 1 kHz. This wide range of haptic perception enables
us to perceive microstructures on surfaces with the same accuracy as enabling us to
identify the point of impact when drumming with our fingers on a table. For a rough
calculation a model according to figure 3.17 is considered to be a parallel circuit
between a mass m and a spring k. Assuming an identical “virtual” volume V of ma-
terial and taking the individual density ρ for a qualitative comparison, the border
frequency for a step response can be calculated according to
 
1 k 1 k
fb = = . (3.9)
2π m 2π V ρ
Figure 3.17 shows the border frequencies of a selection of materials. Only in case
of rubber and soft plastics border frequencies of below 100 Hz appear. Harder plastic
material (Plexiglas) and all other materials show border frequencies above 700 Hz.
One obvious interpretation would state that any qualitatively good simulation of
such a collision demands at least such bandwidth of dynamics within the signal
conditioning elements and the mechanical system.

1.000

F
3dB - border-frequency [Hz]

m k 100

10

fb
1
er

m
te
d
e

l
s

te

s
y

on

ea
la
oo
en

la

lo
bb

re

nu
ni

lic
G

St
ig

Al
l

ra
W

nc

de
py
Ru

ex

Si
G
Co
ro

yb
Pl
lyp

ol
M
Po

Fig. 3.17 3 dB border frequency fb of an excitation of a simple mechanical model parametrized as

different materials.

As a consequence, a frequent recommendation for the design of haptic systems

is the transmission of a full bandwidth of 1 kHz (and in some sources even up to
10 kHz). This requirement is valid with respect to software and communications-
engineering, as sampling-systems and algorithmic can achieve such frequencies eas-
ily today. Considering the mechanical part of the design, we see that dynamics of
3.3 Conclusions from the Biology of Haptics 57

1 kHz are enormous, maybe even utopian. Figure 3.18 gives another rough calcula-
tion of oscillating force amplitude according to

F0 = |x · (2π f )2 m|. (3.10)

The basis of the analysis is a force source generating an output force F 0 . The
load of this system is a mass (e.g. a knob) of 10 grams (!!). The system does not
have any additional load, i.e.it does not have to generate any haptically active force
to a user. A periodic oscillation of a frequency f and an amplitude x is assumed.
With expected amplitudes for the oscillation of 1 mm at 10 Hz a force of approx-
imately 10 mN is necessary. At a frequency of 100 Hz there is already a force of
2-3 N needed. At a frequency of 700 Hz the force already increases to 100 N - and
this is what happens when moving a mass of 10 grams. Of course in combination
with a user-impedance as load the amplitude of the oscillation will decrease in areas
of below 100μ m, proportionally decreasing the necessary force. But this calculation
should make aware of the simple fact that the energetic design and power manage-
ment of electromechanical systems with application in the area of haptics needs to
be done very carefully.

Fig. 3.18 Equipotential line of necessary forces in dependency of amplitude and frequency of the
acceleration of a mass with 10 g.

The design of a technical haptic system is always a compromise between band-

width, stiffness, dynamics of signal conditioning and maximum force-amplitudes.
Even with simple systems the design process leads the engineer to the borders of
what is physically possible. Therefore it is necessary to have a good model for the
58 3 Biological Basics of Haptic Perception

user according to his being a load to the mechanical system and according to his
or her haptic perception. This model enables the engineer to carry out an optimized
design of the technical system and its generation is the focus point of the following
chapter.
Chapter 4
Modeling the User

T HORSTEN A. K ERN

In the preceding chapter the frequency-dependency of mechanoreceptors rele-

vant for haptic perception. has been discussed, i.e. the amplitude of a mechanical
oscillation is perceived with varying intensity for different frequencies. Addition-
ally, it has been shown that the dynamics of haptic perception results in demanding
requirements for the design of technical systems. Consequently, a good mechanical
design has to consider the user in his or her mechanical properties.
The first part of the fourth chapter deals with the discussion of the user as a
mechanical load on the haptic device. The corresponding model is split into two
independent elements depending on the frequency range of the oscillation. In the
second part of the chapter a model of frequency dependent perception for high-
dynamic oscillations is presented. The chapter ends with examples of the models’
applications as well as a summary.

4.1 Mapping of Frequency Ranges onto the User’s Mechanical

Model

The frequency range of haptic perception (fig. 4.1) ranges from quasi - static (at least
10 s) to until 1 kHz and more. The area of highest sensitivity is located within the
range of 100 to 300 Hz.

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_4, 59

c Springer-Verlag Berlin Heidelberg 2009
60 4 Modeling the User

In fact haptic perception is far from being static. The lower border frequency how-
ever is not exactly known. A simple, personal experiment can easily be done after
waking up, e.g. in bed or on a couch. Before moving any part of the body try to be
aware of the position of your limbs and the structure of the cover and cushions. You
will notice that after some time of rest the haptic sense will hardly transmit any in-
formation (besides thermal transfer) about your surroundings. As soon as you begin
to move, all information will become immediately available. This experience is a
clear hint that haptic perception does not work statically but requires at least some
dynamic movement.

The area of active haptic interaction - movements, made in a conscious and con-
trolled way by the user - is of limited range. Sources concerning the dynamics of
human movements differ (subsection 3.2.3). The fastest conscious movement per-
formed by humans is done with their fingers. Movements for typing of up to 8 Hz
can be observed1 . As these values refer to a ten-finger interaction, they have to be
modified a bit. However, as the border frequency of a movement lies above the pure
number of a repetitive event, an assumption of the upper border frequency of 10 Hz
for active, controlled movement covers most cases. But please note that this border
is not sharp and depends on the task done. The major part of the spectrum of haptic
perception is passive (passive haptic perception and even passive haptic interac-
tion). The user does not have any active influence or feedback within this passive
frequency range. In fact, the user is able to modify his properties as a mechanical
load by e.g. altering the force when holding a knob. But although this change in-
fluences the higher frequency range, the change itself happens with lower dynamics
within the dynamic range of active haptic interaction.

0.1 1 10 100 1k 10k

haptic perception f [Hz]

active haptic interaction passive haptic interaction

Fig. 4.1 Border-frequencies fb resulting from an excitation of a simple mechanical model for
different materials.

A look at both large classes of haptic systems (tactile and kinaesthetic systems)
shows that the above modelling has slightly different impacts:

1 8 Hz corresponds to a typing speed of 480 keystrokes per minute. 400 keystrokes are regarded as

very good for a professional typist, 300-200 keystrokes are good, 100 keystrokes can be achieved
by most laymen
4.1 Mapping of Frequency Ranges onto the User’s Mechanical Model 61

• The output values of kinaesthetic systems F out (fig. 4.2a) result in two reactions
by the user. First, a spontaneous, not directly controllable movement reaction vspo
happens as a result of the mechanical properties of the finger tip (depending on
the type of grasp, this can be also the complete interior hand and its skin elastic-
ity). Second an additional perception of forces takes place. This perception K 2
is weighted according to the actual situation and results in a conscious reaction
of the motor parts of the body. These induced reactions vind summed up with the
spontaneous reactions result in the combined output value vout of the user.
• The movements of tactile devices vout (fig. 4.2b) and the consciously performed
movement of the user vind result in a combined movement and velocity. This
elongation acts on the skin, generating the output value F out as a result of its
mechanical properties. This conscious movement vind sums up to vout in the op-
posite direction of the original movement, as with opposite movement directions
the skin’s elongation increases and results in a larger force between user and tech-
nical system. Analogously it subtracts with movements in the same direction, as
in this case the device (or the user, depending on the point of view) evades the
acting force trying to keep deformation low and to perceive just a small haptic
feedback. According to this model only the output value F out of the combined
movement is perceived and contributes to a willingly induced movement.

Fout K
Fout perception
primarily
kinaesthetic
skin
haptic User =
device vspo
(force source) xout
1 + + vind muscles K'
a) xout s & joints
vout

xout vout - vind muscles

s
xout + & joints
primarily vout- vind K'
tactile
haptic skin
device User =
(elongation Fout
source) K
perception
b) Fout

Fig. 4.2 User-models as a block-structure from kinaesthetic (a+c) and tactile (b+d) systems.

2 K, a variable chosen completely arbitrarily, is a helpful construct for the understanding of block-
diagrams rather than having a real neurological analogy
62 4 Modeling the User

If you transfer the model of figure 4.2 into an abstract notation, all blocks cor-
respond to the transfer-function GHn . Additionally, it has to be considered that the
user’s reaction K  is a combined reaction of complex habits and the perception K;
therefore a necessity to simplify this branch of the model becomes eminent. For the
purpose of device design and requirement specification, the conscious reaction is
modelled by a disturbing variable only limited in bandwidth, resulting in a block-
diagram according to figure 4.3c for kinaesthetic and according to figure 4.3d for
tactile devices. The transfer function GH3 corresponds to the mechanical admittance
of the grasp above the border frequency of user interaction fg . The transfer func-
tion GH1 is a model of the quantification of haptic perception and will be discussed
comprehensively in section 4.3.

Fout K Fout K
GH1 GH1
GH3
=
GH3 ?
vspo vspo
1 + + vind K' 1 + + vind vmax
xout s
GH2 xout s
vout vout fg
a) c)

xout vout + vind xout vout - vind vmax

s
+ GH2 s
+
vout+ vind K' vout- vind fg
1/GH3 = 1/GH3 ?
K K
GH1 GH1
Fout Fout
b) d)

Fig. 4.3 Transformation of the user-models’ block-structures in transfer-functions including sim-

plifications of the model for the area of active haptic interaction for kinaesthetic (a+c) and tactile
(b+d) systems.

With regard to the application of the presented models there are two necessary
remarks to be considered:
• The notation in figures 4.2 and 4.3 for elongations x and forces F being input
and output values of users is just one approach to the description. In fact an
impedance coupling exists between user and haptic system making it impossible
to distinguish between input and output-values. However, the decoupled haptic
device is designed for being a position or force source. This in fact is the major
4.2 A Model of the User as a Mechanical Load 63

motivation to define input- and/or output values of the user. But there are certain
actuators (e.g. ultrasonic devices) which can hardly be defined as being part of
either one of these classes. As a consequence, when describing either system, the
choice of the leading sign and the direction of arrows should carefully be done!
• The major motivation for this model is the description of a mechanical load for
the optimized dimensioning of a haptic system. For guaranteeing the closed-loop
control engineering stability of a simulation or a telemanipulation system, further
care has to be taken of the frequency range of active haptic interaction below
10 Hz. Stability analysis in this area can either be achieved by more detailed
models or by an observation of in- and output values according to their control-
engineering passivity. Further information on this topic can be found in chapter 7.

The following sections on user impedance (section 4.2) and perception (sec-
tion 4.3) give a concrete model for the transfer function GH3 of figure 4.3.

4.2 A Model of the User as a Mechanical Load

The user’s reaction as part of any haptic interaction combines a conscious, bandwidth-
limited portion - the area of active haptic interaction - and a passive portion, mainly
resulting from the mechanical properties of fingers, skin and bones. The influence
of this second part stretches across the whole frequency range, but emphasizes the
upper area for high frequencies. This section describes the passive part of haptic in-
teraction. The transfer function GH3 of figure 4.3 is a component of the impedance
coupling with force-input and velocity-output and is therefore a mechanical admit-
tance of the human Y H (chapter 16) respectively in its reciprocal value the mechan-
ical impedance Z H .
vspo v − vind 1
GH3 = = out = YH = (4.1)
F out F out ZH
In the following this mechanical impedance of the user will be specified. The
discussion based on mechanical impedances instead of admittances allows a better
approach to further understanding. An impedance is nothing else but the stiffness or
hardness of a mechanics or a material. High impedance means that an object has at
least one of three properties:

1. hard and stiff in the meaning of spring-stiffness,

2. inertial in the sense of inertial force,
3. sticky and tight in the sense of high friction.
In any case a small movement (velocity v) results in a high force reaction F with
high impedances. Low impedance means that the object, the mechanics, is accord-
ingly soft and light. Even high velocities result in small counter forces in this case.
64 4 Modeling the User

It is legitimate to replace the term impedance with the term stiffness while read-
ing, but one should be aware that stiffness (in the linear situation spring stiffness
with a spring constant k) is only one of three characteristic properties of impedance
(chapter 16).

The human’s mechanical impedance is dependent on a number of influence pa-

rameters:
• type of grasp being directly influenced by the construction of the handle
• physiological condition
• grasping force being directly influenced by the will of the user
• skin surface changes, e.g. by sweat
The quantification of human’s mechanical impedance requires taking as many
aspects into account as possible. The type of grasp is defined by the mechanical de-
sign of the device. Nevertheless a selection of typical grasping situations will give a
good overview of typical impedances appearing during human-machine interaction.
The user-individual parameters like physiological condition and skin structure can
be covered best by the analysis of a large number of people of different conditions.
By choosing this approach a spanwidth of percentiles can be acquired covering the
mechanical impedances typically appearing with human users. The “free will” itself,
however, is - similar to the area of active haptic interaction - hard if not impossible
to be modelled. The time dependent and unpredictable user impedance dependency
on the will can only be compensated if the system is designed to cover all pos-
sible impedance couplings of actively influenced touch. Another approach would
be to indirectly measure the will to adapt the impedance model of the user within
the control loop. Such an indirect measure is, in many typical grasping situations,
the force applied between two fingers or even the whole hand holding an object or
a handle. In the simplest design the acquisition of such a force can be done by a
so called dead-man-switch, which in 1988 was already proposed by H ANNAFORD
[84] for the usage in haptic systems. A dead-man-switch is pressed as long, as the
user holds the control handle in his or her hand. It detects the release of the handle
resulting in a change in impedance from Z H to 0.

4.2.1 Types of Grasps

There is a nomenclature for different types of grasps (fig. 4.4). The hand is an ex-
tremity with 27 bones and 33 muscles. It combines 13 (fingers) respectively 15
4.2 A Model of the User as a Mechanical Load 65

(incl. the wrist) degrees-of-freedom3 . Accordingly the capabilities of man to grasp

are extremely versatile.

contact-grasps precision-grasps power-grasps

1 Finger 2 Finger 2 Finger

thumb thumb
opposite crosswise
Thumb 3 Finger 3 Finger

equally -
distributed
Hand 5 Finger 4 Finger

thumb
opposite

Hand-comb equally distributed, hands on Hand

Fig. 4.4 Grasps.

There are three classes of grasps to be distinguished:

• The contact grasp describes the touch of an object using the whole hand or
major parts of it. Keys and buttons are typically actuated by contact grasps. Even
the fingers resting on a keyboard or a piano are called contact grasps. A contact
grasp always blocks one direction of movement for an object (which is one half
of a degree of freedom). Contact grasps can be regarded as linear only in case of
a pre-load high enough. With light touches the point of release and the according
lift-off of the object is always nonlinear.
• The precision grasp describes the grasping with several fingers. Typically a pre-
cision grasp locks at least one degree of freedom of the grasped object by form
closure with one finger and a counter bearing - often another finger. Additional
degrees of freedom are hindered frictionally. Precision grasps vary much in stiff-
ness of coupling between man and machine. At the same time they are the most
frequent type of grasping.
• The power grasp describes an object with at least one finger and a counter bear-
ing, which may be another finger, but frequently is the whole hand. The power

3 thumb: 4 DOF, index finger: 3 DOF, middle finger: 2 DOF (sometimes 3 DOF), ring finger: 2
DOF, small finger: 2 DOF, wrist: 2 DOF. The rotation of the whole hand happens in the forearm
and therefore does not count among the degrees of freedom of the hand itself
66 4 Modeling the User

grasp aims at locking the grasped object in all degrees of freedom by a combina-
tion of form and force closures. Power grasps are - as the name already implies -
the stiffest coupling between humans and machines.
For all classes of grasps, measurements of the human’s impedance can be per-
formed. According to the approach presented by K ERN [130], the measurement
method and the models of user impedance are presented including the correspond-
ing model parameters.

4.2.2 Measurement Setup and Equipment

The acquisition of mechanical impedances is a well-known, so-to-speak day-to-day

problem in measurement technology. The principle of measurement is based on an
excitation of the system to be measured by an actuator, simultaneously measuring
force and velocity responses of the system. For this purpose combined force and
acceleration sensors (e.g. the impedance sensor 8001 from Brüel & Kjær) exist,
whereby the charge amplifier of the acceleration sensor includes an integrator. In
general the impedance of organic systems is nonlinear and time-variant. This non-
linearity is a result of a general viscoelastic behaviour of tissue resulting from a
combined response of relaxation, conditioning, stretching and creeping [63]. These
effects can be reproduced by mechanical models with concentrated elements. How-
ever, they are dependent on the time-history of excitation to the measured object. It
can be expected that measurements based on step excitation are different from those
acquired with a sinoid sweep. Additionally, the absolute time for measurement has
some influence on the measures by conditioning. Both effects are systematic mea-
surement errors. Consequently, the models resulting from such measurements are
an indication of the technical design process and should always be interpreted with
awareness of their variance and errors.
All impedance measures presented here are based on a sinoid-sweep from upper
to lower frequencies. The excitation has been made with a defined force of 2 N
amplitude at the sensor. The mechanical impedance of the handle has been measured
by calibration measurements and was subtracted from the measured values. The
impedance-sensors are limited concerning their dynamic and amplitude resolution,
of course. As a consequence, the maximum frequency up to which a model is valid
depends on the type of grasp and its handle used during measure. This limitation is
a direct result of the amplitude resolution of the sensors and the necessity at high
frequencies to have a significant difference between the user’s impedance and the
handle’s impedance for the model to be built on. The presented model-parameters
are limited to the acquired frequency range and cannot be applied to lower or higher
frequencies. The measurement setup is given in figure 4.6.
4.2 A Model of the User as a Mechanical Load 67

Burster F-Sensor Pen-Sensor

Counter-bearing with integrated grip-force
for static pretension
steal-mass measurements
Dynamic shaker FlexiForce A201 L
Brüel&Kjaer 4810

Uout
Impedance-sensor
Brüel&K jaer 8001

Networkanalyzer Sinus-Sweep

Fig. 4.5 Measurement setup for the acquisition of user-impedances according to [130].

Pen-Sensor
with finger-force measure
FlexiForce A201 L

Burster F-Sensor
Counterbearing static pretension
Ball-Knob
steal-mass
Shaker with finger-force measure
Brüel&Kjaer 4810 FlexiForce A201 H

Impedance-sensor
Brüel&K jaer 8001

Fig. 4.6 Impedance measurement settings for different grasps.

68 4 Modeling the User

4.2.3 Models

In order to approximate the human’s impedance a number of different approaches

were taken in the past (fig. 4.7). For its description mechanical models based on
concentrated linear elements were chosen. They range from models including ac-
tive user reactions represented by force sources (fig. 4.7a), to models with just three
elements (fig. 4.7c) and combined models of different design. The advantage of a
mechanical model compared to a defined transfer function with a certain degree
in enumerator and denominator results from the possibility of interpreting the ele-
ments of the model as being a picture of physical reality. Elasticities and dampers
connected in circuit with the exciting force can be interpreted as the coupling to the
skin. Additionally the mechanical model creates very high rankings by its intercon-
nected elements which allow a much better fit to measurements than free transfer
functions.

F out k
s2 F out k2
s
Fout
m1 k1 d1 d2 m1 k 1 d1 d2
v out v out

a) b)
Hannaford 1988 Jungmann 2004

Fout2 k2 d2 F out

d3
k2b k3 F out

m1 k1 d1
v out
ks1b v out
m1
Kunstmann 1999
Fout1 k1 d1
c) d)
Oguztöreli, Stein 1990

Fig. 4.7 Modelling the user with concentrated elements, (a) [84], (b) [119] (c) [188], (d) [147].

K ERN [130] defined an eight-element model based on the models in figure 4.7
for the interpolation of the performed impedance measures. The model can be char-
acterized by three impedance groups typical for many grasping situations:
Z 3 (equ. 4.4) models the elasticity and damping of the skin being in direct con-
tact with the handle. Z 1 (equ. 4.2) is the central element of the model and describes
the mechanical properties of the dominating body parts - frequently fingers. Z 2
(equ. 4.3) gives an insight into the mechanical properties of the limbs, frequently
hands, and allows to make assumptions about the pre-loads in the joints in a certain
grasping situation.
4.2 A Model of the User as a Mechanical Load 69

Z3
d3 Z1
Fout

m2 k1 d1 k2 d2 Z2
vout k3

m1

Fig. 4.8 Eight-element model of the user’s impedance [130], modeling the passive mechanics for
frequencies >20 Hz.

s2 m2 + k1 + d1 s
Z1 = (4.2)
s
 
s 1 −1
Z2 = + (4.3)
d2 s + k2 sm1
d3 s + k3
Z3 = (4.4)
s

Z B = Z1 + Z2 (4.5)
Combined, the model’s transformation is given as

Z H = Z 3 Z B (4.6)

⎛  ⎞−1
 −1 −1
s s2 m2 + k1 + d1 s s 1
ZH = ⎝ + + + ⎠ (4.7)
d3 s + k3 s d2 s + k2 sm1

4.2.4 Modeling Parameters

For above model (equ. 4.7) the mechanical parameters can be identified by measure-
ment and approximations with real values. For the values presented here approx-
imetely 48 to 194 measurements were made. The automated algorithm combines
an evolutionary approximation procedure followed by a curve-fit with optimisation
based on N EWTON [30] curve fitting, to achieve a final adjustment of the evolution-
arily found starting parameters according to the measurement data. The measure-
70 4 Modeling the User

ments vary according to the mechanical pre-load - the grasping force - to hold and
move the control handles. This mechanical pre-load was measured by force sensors
integrated into the handles. For each measurement this pre-load could be regarded
as being static and was kept by the subjects with a 5% range of the nominal value.
As a result the model’s parameters could be quantified not only dependent on the
grasping situation but also dependent on the grasping force. The results are given in
the following section. The display of the mechanical impedance is given in decibel,
whereby 6 dB equals a doubling of impedance. The list of model values for each
grasping situation is given in appendix chapter 18.

4.2.4.1 Power Grasps

Within the class of power grasps three grasping-types were analyzed. Impedance be-
tween 35 and 45 dB is measured for the grasp of a cylinder (fig. 4.9) and of a sphere
of similar dimensions (fig. 4.10) with the whole hand. It shows an anti-resonance in
the area of 80 Hz, which moves for a grasp of the cylinder slightly to higher frequen-
cies with increased grasping force. The percentiles, especially the 5th, reveal that the
model is based on a large variance and uncertainty. It is likely that the influence of
the subjects’ variability in their physical parameters like the size of hands and fin-
gers influence these measurements a great deal. In case of grasping two rings with
thumb and index finger (fig. 4.11) the results are much more accurate. Impedance
ranges in between 15 and 35 dB. The anti-resonance in the frequency range of 70
to 100 Hz shows a clear dependency on grasping forces. If we look at the param-
eters, this change is a result of a variance within the elasticity coefficients k1 and
m2 , building the central parallel resonance of the model. The mechanical system
“hand” becomes stiffer (k1 increases), but the mass m2 part of the anti-resonance
diminishes. An easy interpretation of this effect is not obvious. At a value of 10 kg
the mass m1 builds an almost stiff counter bearing.

20 log|ZH| [dB]

60
95. percentile 5N
50 45 10N
50. percentile 15N
40 20N
30 40
5. percentile
20

10 35

0 1 2 3 1 2 3
10 10 10 10 10 10
f [Hz]

Fig. 4.9 Impedance with percentiles (a) and at different force levels (b) for power grasps of a
cylinder (Ø25 mm, defined for 20 Hz to 400 Hz).
4.2 A Model of the User as a Mechanical Load 71

20 log|ZH| [dB]
60 95. percentile 44
5N
50. percentile 42 10N
40 15N
40 20N
5. percentile
20 38

36
0
34

-20 1 2 3
32 1 2 3
10 10 10 10 10 10
f [Hz]

Fig. 4.10 Impedance with percentiles (a) and at different force levels (b) for a power-grasp of a
sphere (Ø40 mm, defined for 20 Hz to 600 Hz).

Fig. 4.11 Impedance with percentiles (a) and at different force levels (b) for a two-fingered power-
grasp of rings (Ø25 mm of the inner ring, defined for 20 Hz to 2 kHz).

4.2.4.2 Precision Grasps

Within the area of precision grasps three types of grasps were analyzed. Holding a
measurement cylinder similar to a normal pen in an angle of 30◦ (fig. 4.12), we find
a weak anti-resonance in the area of around 150 to 300 Hz. This anti-resonance is
dependent on the grasping force and moves from weak forces and high frequencies
to large forces and lower frequencies. The general dependency makes sense, as the
overall system becomes stiffer (the impedance increases) and the coupling between
skin and cylinder becomes more efficient resulting in more masses being moved at
higher grasping forces. The general impedance does not change significantly. if the
cylinder is held in a position similar to a máobi Chinese pen (fig. 4.13). However
the dependency on the anti-resonance slightly diminishes compared to the above
pen hold posture. This is completely different to the variant of a pen in a horizontal
position held by a three finger grasp (fig. 4.14). A clear antiresonance with frequen-
cies between 80 and 150 Hz appears largely dependent in shape and position on the
grasping force. All observable effects in precision grasps can hardly be traced back
to the change of a single parameter but are always a combination of many parame-
ters’ changes.
72 4 Modeling the User

Fig. 4.12 Impedance with percentiles (a) and at different force levels (b) for a two fingered
precision-grasp of a pen-like object held like a pen (Ø10 mm, defined for 20 Hz to 950 Hz).

20 log|ZH| [dB]
50 30
2N
28 5N
40
95. percentile 7N
26
10N
30 24
50. percentile
20 22

5. percentile 20
10
18
0 1 2 3
16 1 2 3
10 10 10 10 10 10
f [Hz]

Fig. 4.13 Impedance with percentiles (a) and at different force levels (b) for a two fingered
precision-grasp of a pen-like object held like a “máobi” Chinese pen (Ø10 mm, defined for 20 Hz
to 700 Hz).

Fig. 4.14 Impedance with percentiles (a) and at different force levels (b) for a five fingered
precision-grasp of a pen-like object in horizontal position (Ø10 mm, defined for 20 Hz to 2 kHz).
4.2 A Model of the User as a Mechanical Load 73

4.2.4.3 One-Finger Contact Grasp

All measurements were done on the index finger. Parts of these results were pub-
lished in EHC 2008 [132] and are repeated for completeness here. Direction of
touch, size of touched object and touch-force normal to the skin were varied within
this analysis. Figure 4.15a shows the overview of the results for a touch being ana-
lyzed in normal direction. The mean impedance varies between 10 and 20 dB with a
resonance in the range of 100 Hz. Throughout all measured diameters of contactor
size and forces, no significant dependency of the position of the anti-resonance on
touch forces were noted. However, a global increase in impedance is clearly vis-
ible. Observing the impedance dependent on contactor size, we can recognize an
increase of the anti-resonance frequency. Additionally, it is fascinating to see that
the stiffness decreases with an increase of contact area. The increase in resonance is
probably a result of less material and therefore less inertia participating in generating
the impedance. The increase in stiffness may be a result of smaller pins deforming
the skin more deeply and therefore getting nearer to the bone as a stiff mechan-
ical counter bearing. In comparison, with measurements performed with a single
pin of only 2 mm in diameter (fig. 4.15b), the general characteristic of the force de-
pendency can be reproduced. Looking at the largest contact element of 15 mm,in
diameter, we are aware of a movement of the resonance frequency from 150 Hz to
lower values down to 80 Hz for an increase in contact force. In orthogonal direction
the skin results differ slightly. Figure 4.16a shows a lateral excitation of the finger
pad with an obvious increase of impedance at increased force of touch. This rise
is mainly a result of an increase of damping parameters and masses. The position
of the anti-resonance in frequency domain remains constant at around 150 Hz. The
picture changes significantly for the impedance in distal direction (fig. 4.16b). The
impedance still increases, but the resonance moves from high frequencies of around
300 Hz to lower frequencies. Damping increases too, resulting in the anti-resonance
being diminished until non-existence. At 45◦ (fig. 4.16c) a combination of both ef-
fects appears. Anti- resonance moves to a higher frequency and loses its sharpness
compared to the pure lateral excitation. A first trend of change within the position
of the anti-resonance in frequency domain with higher forces can be identified ad-
ditionally.

4.2.4.4 Superordinate Comparison of Grasps

It is interesting to compare the impedances among different types of touch and

grasps with each other:
1. Almost all raw data and the interpolated models show a decrease of impedance
within the lower frequency range of 20 Hz to the maximum of the first anti-
resonance. As to precision grasps (fig. 4.12 to 4.14), normal fingertip excita-
tion (fig. 4.15) and the power grasp of rings (fig. 4.11) the gradient equals
20 dB/decade resembling a dominating pure elongation proportional effect of
74 4 Modeling the User

20 log|ZH| [dB]
40 28
1N 2mm
95. percentile 2N 26 5mm
30 20 4N
24 10mm
6N
50. percentile 15mm
20 22
15
20
10 5. percentile
10
18
a) 0 1 16 1
2 3 4 1 2 3 4 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10
f [Hz]
20 log|ZH| [dB]
40
1N
30 2N
20
95. percentile 4N
50. percentile 6N
20
15
10
5. percentile 10
0

b) -10 1 1 2 3 4
2 3 4
10 10 10 10 10 10 10 10
f [Hz]
20 log|ZH| [dB]
40
1N
95. percentile 2N
30 20 4N
6N
50. percentile
20
15
5. percentile
10
10
c) 0 1 2 3 4 1 2 3 4
10 10 10 10 10 10 10 10
f [Hz]

Fig. 4.15 Impedance of finger touch via a cylindrical plate for different contact forces (1 to 6 N)
and in dependency from diameter (a), for the smallesplate(Ø 2 mm) and the largest plate (Ø 15 mm)
(defined for 20 Hz to 2 kHz).

force response - elasticity - within a low frequency range. Within this low
bandwidth-area nonlinear effects of tissue including damping seem to be not very
relevant. Looking at this type of interactions we can assume that any interaction
including joint rotation of a finger is almost purely elastic in a low frequency
range.
2. Many models show a clear antiresonance. Its position varies between 70 Hz at
power grasps (fig. 4.9 to 4.11) and 200 Hz or even 300 Hz at finger touch ana-
lyzed in orthogonal direction (fig. 4.16). The resonance is a natural effect of any
system including a mass and elasticity. Therefore it is not its existence which
is relevant for interpretation, but its shape and the position within the frequency
range. As to positions, especially the power grasp of two rings (fig. 4.11), the
precision grasps of a cylinder in a pen-like position (fig. 4.12) and in horizontal
position (fig. 4.14) and the touch of an orthogonal moving plate in distal direc-
tion (fig. 4.16c) and a large plate in normal direction (fig. 4.16c) have a clear
dependence on grasping force. The interpretation is not as obvious as in case
4.2 A Model of the User as a Mechanical Load 75

20 log|ZH| [dB]
20 16
95. percentile 1N
14 2N
15 4N
50. percentile 12 6N
5. percentile 10
10
8
5
6

0 1 4 1 2 3 4
2 3 4 10 10 10 10
a) 10 10 10 10
f [Hz]
20 log|ZH| [dB]
40
25 1N
2N
30 4N
20 6N
95. percentile
20
50. percentile 15
10
5. percentile
10
b) 0 1 2 3 4 1 2 3 4
10 10 10 10 10 10 10 10
f [Hz]
20 log|ZH| [dB]
30 26
1N
24 2N
25 95. percentile 4N
22
50. percentile
6N
20 20
5. percentile
18
15
16
c) 10 1 2 3 4 14 1 2 3 4
10 10 10 10 10 10 10 10

Fig. 4.16 Impedance for finger touch of a plate moving in orthogonal direction to the skin at
different force levels (1 to 6 N) (defined for 20 Hz to 2 kHz). Movement in lateral direction (a),
distal direction (b) and at 45◦ (c)..

one. We assume that the normal touch of the plate shows similarities to the con-
tact situation when touching the rings. Additionally the normal touch is part of
the precision grasps mentioned above. In the case of many subjects grasping the
horizontal cylinder, it could be observed that the thumb was positioned less or-
thogonally but more axially to the cylinder, which could excite it primarily in
distal direction, thus also contributing to this effect.
3. The shape of the anti-resonance is another interesting factor. It can be noted
that especially in the analysis of finger grasps and there at orthogonal excita-
tion (fig. 4.16a), the anti-resonance is very narrow. An interpretation is hard to
be formulated. It seems that with grasps and especially touches involving less
material the anti-resonance becomes narrower in shape.
4. For all measurements, at high frequencies above the anti-resonance, the fre-
quency characteristic becomes linear and constant, which resembles a pure
damping effect. This becomes obvious at the pen-hold posture among the preci-
76 4 Modeling the User

sion grasps (fig. 4.12) and with the lateral displacement in orthogonal direction,
(fig. 4.16a), but is part of any curve and model. Alternatively, inertia could be
assumed to dominate the high frequencies, being represented by a linear increase
of mechanical impedance. Mainly power grasps show a tendency to this increase.
This measured effect is especially relevant, as it confirms common assumptions
that for high frequency haptic playback with kinaesthetic devices, the user can
be assumed as a damping load.
5. A last glance should be taken at the absolute height level and the variance of
height of the impedance due to pre-loads. For all grasps it varies in a range (re-
garding the median curves only) of 20 dB as a maximum. Impedance is higher
for power grasps, slightly lower for precision grasps and very much lower for
touches, which is immediately obvious. The change in the pre-load for one grasp
typically displaces the absolute impedance to higher levels. This displacement
varies between 4 to 10 dB.
If speculations should be made on still unknown, not yet analyzed types of
touches according to the given data, it should be reasonable to assume the following:

A. Power grasp:

The median impedance should be around 36 dB. Model the impedance with a domi-
nating elasticity effect until an anti-resonance frequency of 80 Hz, not varying much
neither in height nor in position of the anti-resonance. Afterwards allow inertia to
dominate the model’s behaviour.

B. Precision grasp:

The median impedance should be around 25 dB. Model the impedance with a dom-
inating elasticity effect until an anti-resonance frequency of around 200 Hz. The
position of the anti-resonance diminishes in an area of 100 Hz due to change in pre-
load. Above that anti-resonance let the impedance become dominated by a damping
effect. The height of impedance changes in a range of 5 dB by the force of the grasp.

C. Finger touch:

The median impedance should be around 12 dB. Model the impedance with a well
balanced elasticity and damping effect until an anti-resonance frequency of around
150 Hz. The position of the anti-resonance is quite constant, with the exception
of large contact areas moving in normal and in distal direction. Above that anti-
resonance let the impedance become strongly dominated by a damping effect. The
4.2 A Model of the User as a Mechanical Load 77

absolute height of impedance changes in an area of up to 10 dB depending on the

force during touch.

4.2.4.5 Comparison with Existing Models

For further insight into and qualification of the results, a comparison with published
mechanical properties of grasps and touches is presented in this section. There are
two independent trends of impedance analysis in the scientific focus: the measure-
ment of mechanical impedance as a side product of psychophysical studies at thresh-
old level, and measurements at higher impedance levels for general haptic interac-
tion. The frequency plots of models and measurements are shown in figure 4.17. In
[110] the force detection thresholds for grasping a pen in normal orientation have
been analyzed. Figure 4.17a shows an extract of the results compared to the pen-
like grasp of a cylinder of the model in figure 4.12a. Whereas the general level
of impedance does fit, the dynamic range covered by our model is not as big as de-
scribed in literature. Analyzing the data as published, we can state that the minimum
force measured by I SRAR is ≈ 60 μ N at the point of lowest impedance. A force sen-
sor reliably measuring at this extreme level of sensitivity exceeds the measurement
error of our setup and may be the explanation of the difference in the dynamic range
covered. In another study [111] the force detection threshold of grasping a sphere
with the finger tips was analyzed. The absolute force level of interaction during these
measurements was in the range of millinewton. A comparison (fig. 4.17b) between
our model of touching a sphere and these data show a difference in the range of 10
to 20 dB. However such small contact forces resemble a large extrapolation of our
model data to low forces. The difference can therefore be easily explained by the er-
ror resulting from this extrapolation. F U [61] measured the impedance of the finger
tip at a low force of 0.5 N. He advanced an approach published by H AJIAN [82]. A
comparison between our model and their data concerning the shape is hardly pos-
sible due to the little number of discrete frequencies of this measurement. However
the impedance is again 10 dB lower than of our touch model of a five millimeter
cylinder at normal oscillations similar to figure 8a. Once more the literature data
describe a level of touch force not covered by our measurements and therefore the
diagram of Fig. 10c is an extrapolation of the model of these low forces. As a conclu-
sion of this comparison, the model presented here cannot necessarily be applied to
measurements done at lower force levels. Publications dealing with touch and grasp
at reasonable interaction forces reach nearer to the model parameter estimated by
our research. YOSHIKAWA [293] published a study of a three element mechanical
model regarding the index finger. The study was based on a time-domain analysis
of a mechanical impact generated by a kinaesthetic haptic device. The measured
parameters result in a frequency plot (fig. 4.17d) which is comparable to our model
of low frequencies, but does neither show the complexity nor the variability of our
model in a high frequency range of above 100 Hz. A similar study in time-domain
was performed by H AJIAN [82] with just slightly different results. Measurements
78 4 Modeling the User

available as raw data from J UNGMANN [120] taken in 2002 come quite close to our
results, although obtained with different equipment. Besides these frequency plots,
the model’s parameters allow a comparison with absolute values published in liter-
ature: S ERINA [224] made a study on the hysteresis of the finger tips’ elongation
vs. force curve during tapping experiments. This study identified a k-factor for pulp
stiffness ranging from 2 N/mm at a maximum tapping-force of 1 N to 7 N/mm at
a tapping force of 4 N. This k-factor is in the area of 3 to 8 times larger than the
dominating k-Faktor k2 in our eight-element model. The results of F U [61] make
us assume that there was a systematic error concerning the measurements of S E -
RINA , as the elongation measured at the fingernail does not exclusively correspond
to the deformation of the pulp. Therefore the difference in k-Factor between our
model and their measurements can become reasonable. Last but not least M ILNER
[177] carried out several studies on the mechanical properties of the finger tip in
different loading directions. In the relevant loading situation a k-factor of 200 N/m
to 500 N/m was identified by him. This is almost perfect within the range of our
model’s stiffness.

Israr, Pen, Impedance at threshold Israr, Knob, Impedance at Threshold, Fu, Index-Finger, Threshold
discrete measurement discrete measurement discrete measurement
30 40
15
20 log|ZH| [dB]

20 30
10

10 20
5

a) 0 1 2 3 4
b) 10 1 2 3 4
c) 1 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10

Yoshikawa, Index Finger, m,k,d - model Hajian, Index Finger, m,k,d-model Jungmann, Index Finger,
20 sweep measurement
35 40
20 log|ZH| [dB]

30
15
30
25
10
20
20
5
15

d) 10 1 2 3 4
e)10 f) 0 1
1 2 3 4 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10
f [Hz]

Fig. 4.17 Comparision of the model from figure 4.8 with data from similar touches and grasps as
published by I SRAR [110, 111], F U [61], YOSHIKAWA [293], H AJIAN [82], J UNGMANN [120].

4.2.4.6 Final Remarks on Impedances

The impedance model as presented here will help with the modelling of haptic per-
ception in high frequency ranges of above 20 Hz. However, it completely ignores
4.2 A Model of the User as a Mechanical Load 79

any mechanical properties below that frequency range. This is a direct consequence
of the general approach to human machine interaction presented in section 4.1 and
has to be considered when using this model. Another aspect to consider is that the
above measurements show a large inter-subject variance of impedances. In extreme
cases they span 20 dB meaning nothing else but a factor of 10 between e.g. the
5th and the 95th percentile. Further research on the impedance models will mini-
mize this variance and allow a more precise picture of impedances. But already this
database, although not yet completed, allows to identify helpful trends for human
load and haptic devices.
80 4 Modeling the User

4.3 Modelling Haptic Perception

At the time of this book going to press no standardized and accredited procedure
for the quantification of dynamic haptic perception exists. Nevertheless there is a
need to transform haptic perception into values showing at least some correlation
with each other, allowing at least some comparison of sources with haptically rele-
vant outputs. When designing a technical system it is of obvious advantage to have
a tool enabling the engineer to judge the quality of the technical design. This may
be achieved by time discrete simulation or via analytical identification of charac-
teristic values. Within optics and acoustics such procedures and methods have been
known for decades. By the aid of v(λ ) and v (λ ) the brightness of a light source
or an illuminated object is weighted in a spectral domain according to a normalized
perception curve of wavelengths. In acoustics the spectrum of a dynamic pressure
is calculated according to the perception curve in phone and sone. Surprisingly no
such method has been established for the haptic sense. Most likely this is a direct
consequence of the diversity of haptic interaction. The large number of possible
touch situations and the changing load and perception-situations which are already
considerable, are further increased by the time dependent change of the interaction
itself. Only the haptic sense feeds back on the object generating the sense, something
which also happens to a certain extent to the visual or acoustic sense but mostly can
be ignored4 .
It has to be clearly distinguished between parameters describing the quality of a
technical system, like the transparency (section 2.2.3), and a value quantifying hap-
tic perception. The parameters of the technical system can be part of the optimiza-
tion of a device’s haptic performance, but they are just one part of the technical
transmission.
Freeing oneself from the definitions made in different disciplines and focusing on
the engineering aspect of haptic perception, we could regard the following defini-
tion as a possible approach:

4 Obviously, electromagnetic radiation at visual wavelengths is perceived e.g in a crowd by a

spectator absorbing photons, which will afterwards no longer be available to another spectator
further behind. People cast shadows - yet this is typically trivial and not relevant for the design
of a visual display. Similarly, the acoustics of a room is changed by the presence of a person in
the room. This effect is considered for professional acoustic installations especially in theatres or
operas as being of high importance for their performance. But in every- day-life situations and for
the majority of artificial sound sources the presence of the auditor is not considered at all.
4.3 Modelling Haptic Perception 81

“The sensual impression of an acting force describes the perception of the presence
of a mechanical energy, being not yet perceived as a position change.” (T.A. Kern)

The skin, muscles and joints function like force sensors. Smallest elongations are
a direct result of acting forces. These deformations are measured directly by the
mechanoreceptors known from chapter 3. The skin itself, however, - in contrast to the
design aims of any technical system - is highly nonlinear in its force/displacement
curve and the distribution of the primary sensors.

In analogy to force sensors, the haptic perception of skin, muscles and joints has
to include transfer characteristics describing the frequency dependency of the per-
ception of the input value “force”. Additionally, it should be possible to identify
a mechanical load resulting from the force sensor for the system to be measured.
Finally there has to be a minimum measurement value as being the result of the
noise on the sensors, defining a minimum measurable force and an absolute and rel-
ative error. The existence of all these properties has already been known from haptic
perception, although hidden within different terminologies. The transfer character-
istics are known as a perception curve of mechanical oscillations. The borderline
of minimum measurement error is known as JNDs for different contact situations.
The DL and the power-law provide us with values for the absolute and relative mea-
surement errors. The crux of the quantification of haptic perception lies in the fact
that characteristic values and measurement methods have developed independently
in several scientific disciplines throughout decades. They show gaps in the docu-
mentations of experiments, in the statistic data base or in the number of performed
experiments and in the observation of influence parameters. Additionally, these val-
ues scatter widely, as it is the case with many psychophysical experiments, requiring
a large number of experiments for a large number of touch situations. Consequently
it is necessary to formulate working hypotheses taking from the number of pub-
lications which need an elaborate and long-term verification with very controlled
experiments. Among others, this kind of research is done especially at two insti-
tutes5 . The finalization and verification of the method described in the following
chapters will take several more years for a concluding proof. Nevertheless it has
already shown some very helpful qualities for the technical design process and for
assessing the resulting haptic properties of a product. This is why it is described
here. Another reason is the author’s intention to encourage readers to pursue this
way of thinking and to make use of its results.

As already speculated in section 4.2.4.3, it is most likely that the deformation of

the skin is proportionally linked to the intensity of haptic perception. G ESCHEIDER
and many others have proved by measurements of the perception thresholds of me-
chanical oscillations (section 3.2.2), that there is a functional coherence - similar to
a transfer characteristic - between physically measurable amplitude and a physio-
5 Institute for Electromechanical Design, Technische Universität Darmstadt, Germany; Haptic In-
terface Research Laboratory, Purdue University, Indiana, USA
82 4 Modeling the User

logical and/or psychological perception. Looking at the structure of skins and the
orientation of receptors, we can state that deformations and elongations have to be
responsible for the perception of mechanical stimuli.

As begun in section 4.1, the transfer function GH1 shall put the physical force
F out and the perception of this force K into relation with each other.
K
GH1 = (4.8)
F out
The characteristic curve known best (fig. 3.15 on page 50) merely describes the
perception of mechanical oscillations. There is a transfer function according to
K
GFIP = , (4.9)
x
whereby the index FIP describes the conversion factor between an impression of
an oscillation and the mechanical oscillation itself, independently of whether it is an
elongation x or a force-amplitude F.
Figure 3.15 in section 3.2.2 shows exactly this perception threshold for the haptic
perception of mechanical oscillations at the finger in dependency of the contact area.
The curves can be approximated (fig. 4.18) and be normalized on their particular
maximum. For a contact area of A = 2.9 cm2 the following approximation can be
applied

K (1 + Tn1 s) (1 + Tn2 s)4 (1 + Tn3 s)3

GFIP A2.9 = (4.10)
(1 + Tp1 s) (1 + Tp2 s)9 (1 + Tp3 s)2 (1 + Tp4 s)2
with coefficients taken from table 4.1. For a contact area of A = 0.008 cm2 an ana-
logue approximation can be made

K (1 + Tn1 s)2
GFIP .008 = (4.11)
(1 + Tp1 s)2 (1 + Tp2 s)9 (1 + Tp3 s)8

with coefficients taken from table 4.2.

Within the frequency range of 10 Hz to 10 kHz the resulting curves are appro-
priate to weight oscillations according to haptic perception. As the user impedance
models of section 4.1 are based on the perception of forces, it is necessary to intro-
duce a conversion factor to get from equation 4.9 to 4.8.
Our own experiments and the results obtained by others [111] allow the assump-
tion about the user’s impedance Z H as being appropriate as conversion factor. The
conclusion drawn from the modelling in section 4.2 shows that the transfer function
of equation 4.1 is nothing else but the user’s mechanical admittance, as can be seen
in
1 F
GH3 = mit Z H = out (4.12)
ZH vspo
4.3 Modelling Haptic Perception 83

10

0 A2= 0.008 cm²

-10

x [d B ]
-20

K
-30 A1=2.9 cm²

-40

-50 Measurements data from Gescheider

Fitted, normalized polynom
-60
-1 0 1 2 3 4
10 10 10 10 10 10
f [Hz]

Fig. 4.18 Approximation of the characteristic curves according to G ESCHEIDER from figure 3.15.

With this coherence and the knowledge that x = vs , one result of equation 4.9 is
K Ks K s ZH
GFIP = = = (4.13)
x vspo FH
and consequently for GH1 (with F out = −F H ) the equation is
K G
GH1 = = − FIP (4.14)
F out s ZH
Equation 4.14 now provides the possibility to quantify the perceived quality of
a technological haptic system with an output force based on assumptions on user
impedance Z H and the characteristic curve of haptic perception GFIP .

Table 4.1 Parameters for the approximation in figure 4.18 for a contact area of 2, 9 cm2 according
to equation 4.10.

Parameter Value
K 5 10−3
Tn1 (2 π 2)−1
Tn2 (2 π 80)−1
Tn3 (2 π 320)−1
Tp1 (2 π 15)−1
Tp2 (2 π 200)−1
Tp3 (2 π 420)−1
Tp4 (2 π 1000)−1
84 4 Modeling the User

Table 4.2 Parameters for the approximation in figure 4.18 for a contact area of 0.008 cm2 accord-
ing to equation 4.11.

Parameter Value
K 0.0785
Tn1 (2 π 3)−1
Tp1 (2 π 11)−1
Tp2 (2 π 3000)−1
Tp3 (2 π 4000)−1

4.4 Application Remarks

The model in figure 4.3 can be transferred to a mechanical circuit based on concen-
trated elements (fig. 4.19). This simplifies the formulation of interaction-interdependencies
between the human and the technological system, and allows the analysis of some
simple examples.

4.4.1 Kinaesthetic System

For a kinaesthetic system (fig. 4.19) a simplified model with an ideal force source
(F 0 ) and an unknown mechanical impedance of the device Z D is used.

Fout ZH
F0

FD
vspo
ZD vind
vD

haptic device user

Fig. 4.19 Mechanical representation of the model of haptic interaction (fig. 4.3) with the example
of a device combining an ideal force source and a complex impedance.

The nodal equation for the model (fig. 4.19) results in

F 0 = F D + F out . (4.15)
By the usage of equation 4.14 the perception of an output force generated by a
technical system gives the relation
4.4 Application Remarks 85

F out
K=− G . (4.16)
s Z H FIP
Equation 4.15 resolved according to F out and put into equation 4.16 results in
Z D vD − F 0
K= GFIP (4.17)
s ZH
The movement velocity vD of the device is the sum of vind and vspo , being identi-
cal to the velocity vout in figure 4.3. The solution of the model in figure 4.19 accord-
ing to equation 4.17 is adequate to quantify the haptic quality of any device resem-
bling a force source. If the intention is to take the opposite approach - to calculate
the design of a system based on the modelled interaction - equation 4.17 should
be left dependent on vind . This can be achieved most easily by a mesh equation of
velocities

vD = vspo + vind . (4.18)

Replacing the device’s velocity vD with the impedance and integrating vspo =
xspo s results in
FD
= xspo s + vind . (4.19)
ZD
Using equation 4.15 and with F out = Z H xspo s equation 4.19 can be transferred to

F 0 − vind Z D = xspo s ( Z D + Z H ) (4.20)

This equation - resolved according to xspo and connected with the perception K
by a variation of equation 4.16 - results in a complete description of an arbitrary
haptic system with force source based on user-induced movements:
vind Z D − F 0
K = −xspo GFIP = G (4.21)
s ( Z D + Z H ) FIP

4.4.2 Tactile System

For a tactile system (fig. 4.20) a simplified model with one ideal velocity source v0
and an unknown mechanical device impedance Z D is used.
The solution of this model (equ. 4.25) is applicable to quantify the haptic quality
of any device which can be approximated by a velocity- or elongation source. The
perception itself is dominated by a spontaneous elongation reaction based on the
user’s impedance. The corresponding velocities are

vspo = v0 − vind − vZD . (4.22)

86 4 Modeling the User

ZD Fout ZH

vZD vspo
vind
v0 vD

haptic device user

Fig. 4.20 Mechanical representation of the model of haptic interaction (fig. 4.3) with the example
of a device combining an ideal velocity source and a complex impedance.

After integration and insertion into equation 4.9 the formula

v0 − vind − vZD
K = GFIP x = GFIP . (4.23)
s
becomes available. This one, somewhat more structured, and written in depen-
dency of the output force F 0 results in
v0 vind F
K=( − − 0 ) GFIP , (4.24)
s s ZD s
and in a notation of elongations:
F0
K = (x0 − xind − )G (4.25)
Z D s FIP
Equations 4.24 and 4.25 can frequently be simplified for common tactile systems,
as
• with tactile systems, typically, frequency ranges are of relevance, in which the
human user does not interact with the technical system making vind = 0,
• additionally the stiffness of the technical system is quite high.
The first assumption leads to a simplification of equation 4.24, i.e. to a much
clearer relationship
v0 F
K=( − 0 ) GFIP . (4.26)
s ZD s
And with the second assumption equation 4.26 becomes as simple as
v0
K= G = x0 GFIP , (4.27)
s FIP
which reflects the origin of the transfer function GFIP as being a factor which is
proportional to the perception of an elongation.
4.4 Application Remarks 87

4.4.3 Examples in Time and Frequency Domain

Based on above models, figures 4.19 and 4.20 and the corresponding systematic
descriptions with equations 4.17, 4.21 and 4.25, a number of questions concerning
the design and analysis of haptic systems can be analyzed.

4.4.3.1 The Influence of Discretization of Force Signals on Haptic Perception

In many applications the force F 0 is generated by a time-discrete system. Imme-

diately the question arises whether time discretization is applicable and at which
frequency the lower limits of the technical requirements are. Each discretisation of
arbitrary dynamic signals generates steps in the rhythm of the output frequency.
Such a step shows its maximum amplitude when the transmitted signal has its maxi-
mum gradient. Being of rectangular shape, it contains frequency components which
diminish to the power of two starting at the discretization frequency. At a discretiza-
tion of e.g. 100 Hz single steps are perceived very explicitly6 . However, it is not
immediately obvious at which frequency the ripples in the perceived force are small
enough not to be perceived. Figure 4.21 shows the calculations based on the method
presented before, for a base sinus of 2 Hz with amplitude of 2 N of a simple mechan-
ical system. It is composed of a mass with 20 g and a friction coefficient of 0.1 Ns/m.
The force is discretized with frequencies of 100 Hz and 10 kHz (a), resulting in the
FFT (b) with a dominating base frequency at 2 Hz and peaks of the integer multiple
of the discretization frequency. Within the perception domain K - filtered by the hap-
tic dynamics - the force discretized with 100 Hz shows much smoother force steps
and the discretization at 10 kHz does not show any artifacts at all (c). This is also
reproduced in the FFT (d). This example shows distinctively the damping properties
of the human haptic perception for high dynamic oscillations. By the application of
the method described in the preceding chapters, a visualization and quantification
of steps within time discrete systems is possible.

4.4.3.2 The Influence of Discretization-Width on the Haptic Perception

Another side effect of discretisation is the dead time resulting from the digital to
analogue conversion. Dead times are of course nonlinear and therefore can only be
approximated within the linear analyzing methods used here. Such an approxima-
tion is the description of the dead time Tt by a PT1 element with a time constant Tt
according to
1
GT = . (4.28)
1 + TT s

6 One reason, why such a slow force output is not done at all
88 4 Modeling the User

F0

Force-Domain K-Domain
t [s]
-5
F0 2 K(F0) x 10
2

1
1

0
0

-1
-1
F0 discretized
a) -2 c)
@100Hz
0 0.5 1 -2
0 0.5 1
t [s] F0 discretized t [s]
20 log |F0 | @10kHz
120 20

100 0
20 log |K(F0) |
80 -20

60 -40

40 -60

20 -80

0 -100

b) -20 0 5
d) -120 0 5
10 10 10 10
f [Hz] f [Hz]

Fig. 4.21 Visualization in time- and frequency-domain of the haptic perception of a sinoid force,
discretized with 100 Hz and 10 kHz.

In general, first-order lag elements within a signal conditioning chain are identi-
cal to a reduction of dynamics within the system. This effect can be observed prac-
tically when rendering hard contacts. A collision, e.g. in a simulation of two objects
and the haptic feedback of the acting forces with a device, is perceived stiffer if the
frequency range displayed is wider. System-engineers report that the upper limit of
frequency range is not located at the typically assumed limit of tactile interaction
of around 1 kHz. Instead they say that the rendering of calculated force with up
to 10 kHz results in a definite increase in haptically perceived stiffness. This effect
should possibly be confirmed by an analysis of the method shown here. The basis
for the graphs given in figure 4.22 is the same technical setup as before. It is excited
with a step of 2 N at 0.5 s. The discretization does not affect the amplitude of the
system; the effect analyzed in the preceding example should not affect the quality
of the perception. However, the PT1 element being the approximation of the dead-
time element does not react as dynamically as it could. This is an effect which does
not much affect the FFT of the physical force (b), but makes an obvious difference
in the display of perception K in time domain (c) and frequency domain (d). The
function discretized at 100 Hz has a very smooth but visible inclination. Compared
to the input only slight differences can be observed with the functions discretized
at 10 kHz and 1 kHz. The FFT (d) additionally shows that the frequency range at
10 kHz covers a 20 to 30 Hz more dynamic range. This is an effect which pays off
4.4 Application Remarks 89

in tactile applications, especially as the differences lie exactly within the range of
highest tactile sensitivity.

F0 2.5
2
Force-Domain 1.5
K-Domain
1

0.5

0 -5
F0 0 0.5 1 x 10
t [s]

K(F0)
2
4

1 2

a) c) 0
0 F0 discretized
0.49 0.5 0.51
0.5 0.52 0.54 0.56 0.58
t [s] @100Hz t [s]
20 log |F0 | 20 log |K(F0) |
100
F0 discretized 0

80 @ 1kHz -20

60 F0 discretized -40

40 @10kHz -60

20 -80

0 -100

b) -20 0
10 10
5 d) -120
10
0
10
5

f [Hz] f [Hz]

Fig. 4.22 Visualization in time- and frequency-domain of the haptic perception of a force step at
time interval t = 0.5 s in dependency from the discretization frequencies 100 Hz, 1 kHz, and 10 kHz
.

4.4.3.3 Frequency Domain Analysis in the Design Phase

Another possible application of the above method lies in the frequency analysis dur-
ing the design process of haptic systems. Figure 4.23shows a frequency response of
the model of a kinaesthetic system according to figure 4.19. As an input a PT2 ap-
proximation of a user-induced velocity vind with a border frequency of 10 Hz and an
actuator output force F 0 with a border frequency of 100 Hz was assumed. Addition-
ally two cases were observed, one with the maximum output force F 0 and another
one with just vind and the rest of the system being in an idle situation. The technical
system includes a mass of 20 g and a perception threshold at -138 dB, a value which
can be calculated by the approximation of the perception curve according to equa-
tion 4.10. The observed situation describes the excitation of a finger with a pestle
in a large area (compared to the finger-tips area). Remarkably, even in idle state the
perception curve Kmin of this quite light and extremely dynamic system is below the
perception threshold at frequencies below 2 Hz, only. The area between the curves
Kmin and Kmax is a measure for the haptically available power of the system. It is a
factor of the quality of a haptic device. Comparable visualisations of haptic devices
allow analyzing the validity of a technical design in advance.
90 4 Modeling the User

-50
K max
-100

Magnitude (dB)
-100
-150 -110

-120
-200 -130

threshold K min -140

-250
-150

-160
-300 0.1
10
0.4
10
0.7
10
f [Hz]
-350 0
10
2
10 10
4 2 Hz
f [Hz]

Fig. 4.23 Frequency plot of a kinaesthetic system according to equation 4.21 with maximum force-
ouput and in idle mode.

4.4.3.4 Haptic Quality Estimations of Passive Haptic Devices

The method presented above may also be used to analyze the haptic quality of pas-
sive haptic devices like buttons and knobs; and even textures and surfaces. In a
preliminary study: five visually identical, standard household light switches were
chosen and modified in their haptic switching characteristics. By the exchange of
two springs in the switch, the slope of the force-increase during down movement
and the sharpness of the event of key press was modified (fig. 4.24).
Force [N]

position [mm]

Fig. 4.24 Plot of force vs. position curve of two different switches.

The switches were presented to subjects asked to order them according to their
haptic “explicitness” when pressed. This question was consciously chosen to be
4.4 Application Remarks 91

unsharp in its meaning, as the subjects should only make themselves aware of and
judge the haptic quality when pressing the switches. The subjects should not focus
on anything specific - as this will not happen in e.g. a typical shopping situation,
either. As a result a user rating of the five switches according to figure 4.25 came
up. This user rating can be regarded as a reference to our further analysis.

6
rating (5 being best)

5
4
3
2
1
0
A B C D E
Switch

Fig. 4.25 Mean user rating for the haptic quantification of household light switches with standard
deviation for five switches A-E.

For measurement purposes the switches were pressed at a frequency of 2 Hz with

an electrodynamic actuator. Forces were measured with an ATI nano17 force/torque
sensor. Displacements were measured with a Keyence LK-G32 laser triangulation
system and sound pressure with a Brüel & Kjær condenser microphone. All switches
were mounted on the measurement setup and pressed four times; excited with a
trapezoidal force increase with 3 s period time. Force, displacement and sound pres-
sure were sampled simultaneously with a sampling rate of 44 kHz. Except for a
subsonic filter with a corner-frequency of 22.4 Hz in the microphone amplifier no
filtering was used in recording the signals. Auditory noise was measured and re-
moved from the recorded signal by subtracting the spectra respectively for future
analysis of multimodal effects on perception. The measurements were analyzed in
three different domains according to three characteristic values, each. Within the
physical domain the absolute height of the tactile switching event (switch force),
and the relative height between this point and the lowest point afterwards (switch
difference) were identified (fig. 4.26a). Additionally, the click-ratio as the quotient
of switch difference to switch-force was calculated. The click-ratio is a standardized
way to characterize industrial switches according to DIN 41636. In a bandwidth-
limited domain the original force-measurement data were filtered by a band-pass
filter ranging from 10 Hz to 2 kHz, focusing on the range of tactile perception only.
The maximum and minimum were identified within the resulting data (fig. 4.26b)
giving three characteristic values: filtered maximum, filtered difference and filtered
click-ratio.
92 4 Modeling the User

K(F0) x 10 -4
F [N]
Max
8 1

6 Diff
0
Max
4 Diff
-1
a) 0.8 0.85 0.9 t[s] b) 0.8 0.85 0.9 t[s]

Fig. 4.26 Example for the identification of characteristic values in force-domain (a) and whithin
the K-domain (b).

In the K-domain the bandwidth-limited data were additionally weighted accord-

ing to the described method as stated in equation 4.14. The resulting data were
analyzed identically to the bandwidth-limited domain. The identified values were
normalized according to their maximum and set into relation to the normalized
mean values in user ratings. All values were evaluated within a frequency range
for the measurement data up to 2 kHz. Additionally, measurement data were artifi-
cially limited in bandwidth to 50 Hz, which makes them comparable to setups used
in industry to measure force/displacement curves. Figure 4.27 gives the quadratic
error sum of all five switches against user rating for all nine values.

0,8
0,725
0,694
0,7
0,6
0,5

0,4 0,352

0,3 0,265 0,263

0,2 0,158
0,125
0,1 0,033 0,034
0,000
0,0
g

.
[]

or
.
N]
N]

tio
[N

iff
[N

ax
tin

io

kt
.[
.[

rD
Ra

M
e

e
Ra

at

Fa
ax

iff
rc

nc

o
or
-R

K-
M
Fo

ct
er

re

ic

ct
k

Fa
Cl
lic

e
Us

ffe

Fa

io
rc
h

rc

K-

at
itc

Fo

d
Di

K-
Fo
n

re

R
Sw
ea

k-
lte
d
rc

re
M

lic
re

Fi
Fo

lte

C
lte

Fi
Fi

Fig. 4.27 Quadratic error between characteristic value and mean user rating for bandwidth limited
measurement data with an upper border frequency of 2 kHz.

As a result any analysis including wide bandwidths gives a better rating than
existing reference methods. Additionally, the K-domain analysis has the lowest
4.5 Summarizing Remarks on the Application of the Method 93

quadratic error compared to all other analyses.

This experiment proves the applicability of this method for the analysis of mea-
surement data taken from passive haptic devices like switches.

4.5 Summarizing Remarks on the Application of the Method

According to the above method the frequency dependent parameter K is a measure

for the haptic perception of mechanical oscillations. The quality and meaning of K
depends largely on the quality of GFIP and Z H , i.e. the performance of the models
of the frequency dependency of perception and the mechanical load of the user.
At the time of this book going to press both functions and their dependencies are
subject to ongoing research activities. When using the presented method, we are
recommended to make use of the most actual findings available. Special care should
be taken of selecting the best fits of data of contact and grasp situations. In case such
values are not available, the data presented in sections 4.2 and 4.3 are a good first fit
for an approximation. Nevertheless, they should only be used with sensitivity to the
context and the theoretical basis of this methodology.
“This page left intentionally blank.”
Chapter 5
Internal Structure of Haptic Systems

T HORSTEN A. K ERN ; O LIVER M ECKEL

In the preceding chapters the focus of the discussion was .haptic perception with
regard to the human user. In the following chapter the technical realization of haptic
systems will come to the fore. Consequently, the general view will change from
a user centralized perspective to a device specific viewpoint. The understanding
gained as well as the methods for the quantification of haptic perception will still be
used for the analysis of the quality of technological solutions.

5.1 System Engineering

When starting the design of haptic devices, the engineer has to to deal with the
general structures they can be composed of. Haptic devices of similar functionality
can consist of very different modules. There are four big classes of possible system
designs::
1. “open-loop admittance controlled systems”
2. “closed-loop admittance controlled systems”
3. “open-loop impedance controlled systems”
4. “closed-loop impedance controlled systems”
Impedance controlled systems are based on the transfer characteristics of a me-
chanical impedance Z = Fv and are typical of the structure of many kinaesthetic de-
vices. They generate a force as output and measure a position as input. Admittance
controlled systems instead, are based on the definition of a mechanical impedance
Y = Fv , describing a transfer characteristics with force-input and velocity-output.
These systems generate a position change as haptic feedback and get a force reac-
tion from the user as input source. In the situation of a closed-loop controlled system

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_5, 95

c Springer-Verlag Berlin Heidelberg 2009
96 5 Internal Structure of Haptic Systems

this force is measured and used for the correction of the position. These analyses can
be regarded as analog in the case of torque and angle replacing force and position
for rotary systems. Nevertheless, for readability purposes, the following descrip-
tions concentrate on translational movements and devices, only. In the subsequent
chapters we will take a look at the details of all four variants.

5.1.1 Open-Loop Impedance Controlled

Open-loop impedance controlled systems are based on a quite simple structure

(fig. 5.1). A force signal SF is transferred via a driver GED into a force-proportional
energy form E F . This energy is then altered into the output force F 0 by an actuator
GD1 . This output force interferes with a disturbing force F noise . This noise is a re-
sult of movements generated by the user xout and the mechanical properties of the
kinematic design GD3 . Typically, such disturbing forces are friction and inertia. The
sum of both forces is the actual output force F out of the impedance controller sys-
tem. Usually, there is an optional part of the system, a sensor GD2 , which measures
the movements and the actual position of the haptic system.

SF EF F0 K
Fout
GED GD1 GH1
+
Fnoise
GH3 ?
GD3
vspo
Sx 1 + + vind vmax
GD2
xout s
vout fg

haptic device user

Fig. 5.1 Block-diagram of an open-loop impedance controlled haptic system.

Examples

Open-loop impedance controlled systems are the most frequently available devices
on the market. As a result of their simple internal design, a few standard components
can already be used to build a quite useful design with adequate haptic feedback, if
at least some care is taken of the minimization of friction and masses. Among the
cheapest designs available on the market today, the PHANTOM Omni (fig. 5.2a),
connected via Fire-Wire to the control unit, is among the best known. It is frequently
5.1 System Engineering 97

used in research projects and for the creative manipulation of 3D-data during model-
ing and design. In the higher-price segment there are numerous products, e.g. the de-
vices of the company Quanser. These devices are usually equipped with a real-time
MatLab based control station, adding some flexibility to the modifications of the in-
ternal data processing by the end customer. The doubled Pantograph-kinematics of
the “Haptic Wand” (fig. 5.2b) allows force feedback in up to five degrees of freedom
with three translations and two rotations. Although all these devices may be open-
loop impedance controlled, the software usually includes simple dynamic models
of the mechanical structures. This allows some compensation of inertial and fric-
tional effects of the kinematics based on the continuous measurement of position
and velocities.

a) b)

Fig. 5.2 Example of an open-loop impedance controlled system with a) serial-kinematic (PHAN-
TOM Omni, SensAble) and b) parallel-kinematic (5 DOF Haptic Wand, Quanser) structure.

5.1.2 Closed-Loop Impedance Controlled

Closed-loop impedance controlled systems (fig. 5.3) differ from open-loop impedance
controlled systems in such a manner that the output force F out is measured by a force
sensor GFSense and is used as a control variable to generate a difference value Δ SF
with the nominal value. An additional component typically is a controller GCD in the
control path, optimizing the dynamic properties of the feedback loop. The closed-
loop makes it possible to compensate the force F noise resulting from the mechanics
of the systems. This has two considerable advantages: On the one hand, at idle state
the system behaves in a much less frictional and more dynamic way compared to
similar open-loop controlled systems. Additionally, as the closed-loop design al-
lows some compensation of inertia and friction, the whole mechanical setup can be
98 5 Internal Structure of Haptic Systems

designed stiffer. But it has to be noted that of course part of the maximum output
power of actuators will then be used to compensate the frictional force, which makes
these devices slightly less powerful than an open-loop design.

SS
GFSense

SF - CSF EF F0 Fout K
GCD GED GD1 GH1
+
DSF Fnoise
GH3 ?
GD3
vspo
Sx 1 + + vind vmax
GD2
xout s
vout fg

haptic device user

Fig. 5.3 Block-diagram of a closed-loop impedance controlled system with force-feedback.

Example

Closed-loop impedance controlled systems are usually used in research projects and
as special purpose machines. The delta-series of ForceDimensions (fig. 5.4) is one
example, as it is a commercial system with the option to buy an impedance con-
trolled version. In this variant, force sensors are integrated into the handle, able to
measure interaction forces in the directions of the kinematics’ degrees of freedom.
Closed-loop impedance controlled systems are technologically challenging. One the
one hand they have to comply with a minimum of friction and inertia, on the other
hand, with little friction, the closed loop tends to become instable, as an energy ex-
change between user and device may build up. This is why controllers, typically,
monitor the passive behavior of the device. Additionally, the force sensor is a cost-
intensive element, at present. In case of the delta-device, the challenge to minimize
moving masses has been faced by a parallel-kinematics design.

5.1.3 Open-Loop Admittance Controlled

Open-loop admittance controlled systems (fig. 5.5) provide a positional output. Pro-
portionally to the input value Sx , a control chain with energy converter GED and
kinematics GD1 provides a displacement x0 . This displacement interferes with a
disturbance variable xnoise which is dependent on the mechanical properties of the
kinematics GD3 and a direct reaction to the user’s input F out . In practice an open-
5.1 System Engineering 99

Fig. 5.4 Example of a parallel-kinematic closed-loop impedance controlled system (delta3,

ForceDimension).

loop admittance controlled system typically shows a design which allows to neglect
the influence of the disturbance variable. Another optional element of open-loop
admittance controlled systems is the measurement of the output force with a force
sensor F Sense without closing the control loop.

Sx Ev x0 xout vout + vind vmax

GED GD1 s
+ +
xnoise vout+ vind fg

1/GD3 1/GH3 ?
SS K
GFSense GH1
Fout

haptic device user

Fig. 5.5 Block-diagram of an open-loop admittance controlled haptic system.

Example

Open-loop admittance controlled systems are used especially with haptic devices in
the area of tactile displays. Many tactile displays are based on pin arrays, meaning
that they are generating spatially distributed information by lifting and lowering pins
out of a matrix. These systems’ origins are Braille devices (fig. 5.6) coding letters
100 5 Internal Structure of Haptic Systems

in a tactile, readable, embossed printing. For actuation of tactile pin-based displays

a variety of actuators are used. There are electrodynamic, electromagnetic, thermal,
pneumatic, hydraulic and piezoelectric actuators and even ultrasonic actuators with
transfer media.

Fig. 5.6 Example of an open-loop admittance controlled system being a Braille row (Wikipedia).

5.1.4 Closed-Loop Admittance Controlled Devices

Closed-loop admittance controlled devices (fig. 5.7) provide a positional output and
a force input to the controlling element identical to impedance controlled devices.
The mandatory measurement of the output force F out is used as control variable
SS for calculating the difference Δ SF with the commanding value SF . This differ-
ence is then fed through the controller GCD into the control circuit. As a result, the
displacement xout is adjusted until an aspired force F out is reached.
A variant of a closed-loop admittance controlled device is shown in figure 5.8.
Closed-loop admittance controlled devices show considerable advantages for many
applications requiring large stiffnesses. However, the force sensors GFSense are quite
complex and consequently expensive components, especially when there are nu-
merous degrees of freedom to be controlled. As a variant, the system according to
figure 5.8 does not use a sensor but just a force-proportional measure, e.g. a current,
as control variable. When using e.g. a current with electrodynamic actuators, we
can identify even the reaction of the user generating an induction as an additional
velocity dependent value.

Examples

At present, closed-loop admittance controlled systems are the preferred approach

to provide high stiffnesses with little loss in dynamic properties. The idea to hap-
5.1 System Engineering 101

Sx
GD2

SF DSF Ev x0 xout vout + vind vmax

GCD GED GD1 s
+ +
- CSv vout+ vind fg
xnoise

1/GD3 1/GH3 ?
SS K
GFSense GH1
Fout

haptic device user

Fig. 5.7 Block-diagram of a closed-loop admittance controlled haptic system with force-feedback
loop for control.

Sx
GD2

GFalt-sense
SSalt
SF - x0 xout vout + vind vmax
GCD GED GD1 s
+
DSF CSv + vout+ vind fg
Ev xnoise

1/GD3 1/GH3 ?
SS K
GFSense GH1
Fout

haptic device user

Fig. 5.8 Block-diagram of a closed-loop admittance controlled haptic system with a feedback loop
measuring an internal force-proportional value.

tically hide the actual mechanical impedance from the user by closing the control
loop makes it possible to build serial kinematics with a large workspace. The FCS
HapticMaster (fig. 5.9a) is such a one meter high system with three degrees of free-
dom and a force of up to 100 N. It includes a force sensor at its handle. The axes are
controlled by self-locking actuators. The device’s dynamics is impressive, despite
its size. However, a damping has to be included in the controller for security reasons
resulting in a limitation of bandwidth depending on the actual application. Realisa-
tions of the variant of closed-loop admittance controlled devices are the Virtuose-
systems from Haption (fig. 5.9b). In these devices the current is measured at elec-
102 5 Internal Structure of Haptic Systems

trodynamic electronic commutated actuators and fed back as a control value. The
devices show impressive qualities for the display of hard contacts, but have limited
capabilities in the simulation of soft interactions, e.g. with tissues. Therefore, the
application area of Haption-systems is mainly the area of professional simulation of
assembly procedures for manufacture-preparation.

a) b)

Fig. 5.9 Examples of closed-loop admittance controlled systems in variants with a) direct force
measurement (HapticMaster, Moog) and b) measurement of the actual current (Virtuose 6D35-45,
Haption).

5.1.5 Qualitative Comparison of the Internal Structures of Haptic

Systems

As the haptic human-machine interaction is based on an impedance coupling, it is

always the combination of action and reaction, be it via force or position, which
has to be analyzed. In fact, without any knowledge about the internal structure of
a device, it is impossible to find out whether the system is open-loop impedance
controlled, closed-loop impedance controlled or closed-loop admittance controlled.
With experience of the technological borders of the most important parameters like
dynamics and maximum force, an engineer can make a well-founded assumption
about the internal structure by simply using the device. But concerning the abstract
interface of in- and output values, all the devices of the above three classes are ab-
solutely identical to the user as well as to the controlling instance. Despite this fact
the technical realizations of haptic systems differ widely in their concrete technical
design of course. The parameters influencing this design have to be balanced against
each other. Such parameters are:

• Number of components
• Maximum impedance to be achieved at slow motion
5.1 System Engineering 103

• Minimum impedance to be achieved at fast motion

• Force-resolution
• Impedance of mechanical components (e.g. inertia of kinematics)

These parameters and their mapping onto the technical designs are given quali-
tatively. In figure 5.10 the impedance generated by a device in absolute values and
the impedance range covered may be one criterion for the performance of a device.
Analyzing the systems according to this criterion shows that open-loop admittance
controlled systems may have high impedance, which shows smaller variability in
tighter borders. Closed-loop admittance controlled systems extend these borders by
their ability to modulate the impedance due to the feedback loop. Depending on the
design, closed-loop admittance controlled systems vary in the width of this modula-
tion. In the lower area of possible, realizable impedances the open-loop impedance
controlled systems follow. They stand out more by simplicity in their design than by
large impedance ranges covered. In comparison to the closed-loop admittance con-
trolled systems they gain some impedance width at the lower border of impedances.
In order to be able to equally cover lower as well as higher impedances, the choice
should be made of closed-loop impedance controlled systems.

maximum output impedance

impedance
output

minium output impedance

closed loop admittance controlled closed loop impedance controlled

open loop admittance controlled open loop impedance controlled

tactile
kinaesthetic

device's internal impedance device's internal impedance

dynamic dynamic

pos. resolution force-resolution

Fig. 5.10 Qualitative comparison of the application areas for different device-structures.

Normally, pure open-loop admittance controlled systems are suitable for tactile
devices only, as, with tactile devices, usually there is no active feedback by the user
to be measured. The haptic interaction is limited to tensions being coupled to the
skin of the user’s hands. Such devices show high internal impedance (Z D ). The dy-
namics and the resolution concerning the displacement are very high.
104 5 Internal Structure of Haptic Systems

Kinaesthetic devices can be built with systems allowing a modulation of the dis-
played impedances. The closed-loop admittance controlled systems excel due to
the possibility to use mechanical components with high impedances. The dynam-
ics of these systems are accordingly low (<100 Hz) and the force-resolution is, due
to the typical frictions, not trivial when realized. Open-loop impedance controlled
systems show a wider dynamic range due to the missing feedback loop with, at the
same time, limited dynamic range. Only closed-loop impedance systems allow cov-
ering a wide impedance range from lowest to very high impedances., whereby with
increasing requirements of force resolution the dynamics of the maximum veloc-
ities achieved by the control loop are limited and limitations of the measurement
technology become noticeable.
The decision on the design of a haptic system has significant influence on the
application range and vice versa. On the one hand, it is necessary to identify the
requirements to make such a decision. For this purpose, it is necessary to ask the
right questions and to have an insight into possible technical realizations of single
elements of the above structures. This is the general topic of the second part of this
book. On the other hand, it is necessary to formulate an abstract system descrip-
tion of the device. An introduction of how to achieve this is given in the following
section.

5.2 Abstraction of the System Components

In the preceding section the general structures available for haptic systems have been
introduced. In this section an abstraction of the individual system structure will be
given as a basis for the design and analysis of the system’s transfer characteristics.
The content of this section will be further elaborated and detailed in chapter 7.

5.2.1 Basic System Analysis

All the approaches presented for the design of haptic systems, may they be impedance-
or admittance controlled, can be abstracted and broken down into their major system
components. For the overall system with its borders the complete mechanism and
electronics of a haptic device are regarded. The basic idea behind the haptic system
is to give a human operator, the user, an impression of the load and reactions of an
apparatus or process. Yet in an almost trivial start of analysis, all necessary parts
relevant for the structural analysis should be mentioned:

• a haptic device, - whether impedance or admittance based - providing a haptic

feedback to the user on a value within a process,
• the process the user is going to influence and to change by specific actions,
5.2 Abstraction of the System Components 105

• the user who must be regarded as being part of the overall system and who is
coupled to the process by his reactions to the haptic feedback.
• a telemanipulator which supports the user in his influence on this process. This
element may be totally part of a simulation and in such a case existing only in a
time-discrete simulation engine.
• interface modules with signal conditioning and internal control circuits, inter-
facing the haptic device by simulated or measured process values according to
defined conditions.
• additional control structures on a hidden level which will have an important role
to play in guaranteeing the desired functioning of the system.
Each component itself is coupled to the overall system via interfaces, whereas
these interfaces generate and receive system variables. Figure 5.11 extends the focus
on a complete telemanipulation system.

Control demand:
- Position
Process states Signal transmission Haptic feedback
- Force
- Speed

Actuator system with Interface module with

Schnittstellenmodul mit Actuator system with
Process or - Sensors - Demand
Sollwertgenerierung
generator - Sensors
Objekt - Automatic control - Automatic
Regelung control - Automatic control User
- Actuators - Kommunikation
-Communication - Actuators
- Communication - Kommunikation - Communication

Process- Control demand:

manipulation - Position
- Force Signal transmission User input
- Speed

Fig. 5.11 Abstract representation of a telemanipulation-system.

A deeper analysis of the components in a telemanipulator is done on the basis of

a known transfer characteristics of each system component. In the following section
an approach will be given showing how to understand the different system compo-
nents and describe them systematically.

5.2.2 Analysis of System Components

Within the total telemanipulation system there are numerous coupled system el-
ements, differing significantly in their transfer characteristics. Each of them offers
different opportunities and challenges as a result of the physical phenomena they are
based on. In comparison with all other components, actuators of the haptic device
or the manipulator offer the best chance to be described by mathematical formula-
tions, e.g. differential equations. A general approach is given in the block structure
of figure 5.12.
106 5 Internal Structure of Haptic Systems

z1 z2

w y
GR GS
–

GM

z3

Fig. 5.12 Block-diagram of a closed-loop controlled actuator system.

For this example the actuator is placed in a closed control circuit. The control
transfer function
G R GS
GF = (5.1)
1 + GR GS GM
totally describes the behavior of a controlled actuator. The transfer function of the
controller is given by GR , the transfer function of the actuator by GS and the trans-
fer function of the sensor, e.g. a force sensor within admittance controlled systems,
by GM . This sensor is placed in the feedback loop of the control circuit. Beside
the in- and output values - the desired value or reference variable w and the con-
trolled variable y - there are numerous disturbing variables zi affecting this small
subsystem. Dependent on the subsystem they have different origins and influences.
In this example the disturbing variables are additive. Frequently nonlinear influences
and dependencies on other values not directly part of the control circuit can be de-
scribed by such disturbance variables. When considering many details of disturbing
variables, we can see that a highly complex description evolves from even simple
models, which sometimes hinders further analysis instead of helping to understand
the system elements. It is necessary to choose a good balance between a detailed
and an efficient description of the modeled system.
The descriptions of different principles of the generation of haptic feedback give
an insight into the wideness of the solution space. Nevertheless, taking individual
solutions and reducing them to their abstract description in a control engineering
sense makes them surprisingly similar again:
• The general behavior of control systems, especially of electromechanic actua-
tors, can frequently be reduced to a proportional value. For example an electric
actuator is a transformator between an electric current I as input and a mechan-
ical torque M as output. With a constant current as input, a constant torque as
5.2 Abstraction of the System Components 107

output will be generated. In case of a second closed-loop circuit being placed

around this actuator, e.g. with the task to reach a defined position, the whole
system will show an integral behavior. Such a simple example already includes
many elements of a system description. Beside the transfer characteristics of the
pure actuator, elements of the power train have already to be considered.
• All systems show characteristic dynamic properties. This is a direct result of
the inevitably existing energy storages (in a mechanical sense these are inertia
and elasticities, in an electrical sense these are capacities and inductance). These
energy storages have a direct impact on the system’s response characteristics,
which can be approximated roughly by the usage of delay elements of n’th order.
• Easily distinguishable from the delay elements the system components show a
dead time. The impact of dead time elements on closed circuits is serious and
can frequently be observed with haptic devices as a result of discrete signal pro-
cessing (section 4.4.3). Transmission time of signals is vital to specific tasks of
haptic research, e.g. with medical teleoperation and haptic feedback or telema-
nipulation over large distances in space or combat. More classic and trivial ex-
amples of obvious dead time elements like conveyor belts are very rare in haptic
applications.
The term actuating system already implies the assumption that the actuating unit
is controlled via an electronic circuit and that it is part of a superordinate hierarchy
level. The typical inner disturbance variables of such an electromechanic system
(varying frictional states, temperature dependency on electrical resistances, varia-
tions of source voltages, etc.) are regarded as being compensated within the system
itself. In addition, new disturbance variables appear, resulting from the mechanical
task itself (process loads, counter forces, positional changes, etc.). The disturbance
variables can be modeled differently:

• An additive formulation is a comparably simple way to model the disturbance

variable.
• If the process variable is known to be nonlinear and can be modeled by a non-
linear formulation the whole system is going to become nonlinear. Analysis and
descriptions of nonlinear systems are discussed in chapter 7.

The actuators in a completely mechanical telemanipulator, are, on the one hand,

part of the end-effector, on the other hand part of the haptic system. As a conse-
quence, the interface structures of both applications are different. Figure 5.13 visu-
alizes such an interface for the end-effector part with signal conditioning and control
unit.
As its input, the actuating unit at this end uses a position as commanded value -
typically an open-loop admittance controlled device with force measurement. It may
also be structured as closed-loop impedance or admittance controlled element with
force input. However, such a structure is mainly used with lightweight soft robotic
manipulators and still is quite seldom. The output values are given by measured
actual values and/or control deviations and/or inner system variables. As already
108 5 Internal Structure of Haptic Systems

z1
Nonmeasurable process input
and disturbances

Controlled actuator system

w y
Control demand GR GS Haptic feedback
for the user
– - Position
- Force
- Speed

Feedback to
interface module Measurable process input
and disturbances
z2

Fig. 5.13 Actuator-system in a telemanipulator.

mentioned the actuator system experiences additional disturbance and input values
Z due to being integrated into a complex system.

z
Nonmeasurable process input
and disturbances

Controlled actuator system

w y
Control demand GR GS Haptic feedback
for the user
– - Position
- Force
- Speed

Measurable user input and

interaction
Feedback to
interface module

Fig. 5.14 Actuator-system in a haptic device.

Figure 5.14 shows the interface of an actuator system in a haptic device. The
interface of the central control module for data processing is similar to the one be-
tween telemanipulator and control unit. Once again commanded values are transmit-
ted, in the case of a haptic device they are forces, typically. Additionally, measured
values, e.g. velocities or positions, are fed back to the controlling unit.
The central controlling unit between these actuating units is a node connecting
the haptic device and the end-effector or its time-discrete simulation to a telema-
5.2 Abstraction of the System Components 109

nipulation system. It is the core of the complete setup with a number of important
functionalities:

• Commanding-value generation for the end-effector - an interpretation of the mea-

sured values acquired from the user has to take place. An analysis of the mea-
surement may happen to filter unwanted spectral components which may excite
the system in resonance areas. Other processing of the data may happen with the
aim to keep the control loop stable.
• Commanding-value generation for the haptic device - an interpretation of the
measured (or simulated) values acquired from the end-effector has to take place.
During this value generation mathematical and logical operations may be used to
modify this parameter. Any actual or historic values available within the system
may be helpful to achieve the general aim: minimizing control deviation without
its becoming instable.
• If there is a need to have access to as many actual values as possible, it may be
helpful to integrate parts of the control circuits, from near the actuator, into the
central controlling unit.

Given these functions, this central control and signal-conditioning module is an

important part of the total system. Its transfer characteristics have to be subject to
a design process identical to all other components, from actuators to sensors and
electrical and software interfaces. However, the solution space is almost boundless
making a systematic approach mandatory!
The peripheral surroundings of the telemanipulation system, e.g. the manipulated
object or the subject, is reproduced by the process and the type of interaction. The
preceding sections 4.2 and 4.3 provide a model of the user as a mechanical load
with respect to the grasping situation and the grasping force. However, this model
focuses on the higher frequency range of passive haptic interaction. For stability
issues the lower frequency range in the area of below 10 Hz becomes relevant. The
description of the user in this range of active, haptic interaction can only be done to
a very limited extent. We recommend dealing with the user as a disturbance variable
with defined amplitude and dynamic range, which can be guessed by an analysis of
the task and user’s typical performance. Nevertheless such an approximation has to
be handled with care when being used as the basis for a design process and for the
dimensioning of transfer characteristics of the complete system. Therefore more ad-
vanced systems monitor the device’s behavior and adjust control values depending
on its state (subsection 7.3.3). Describing the process and/or the manipulated object
- in direct comparison - is in many cases much easier to be achieved. There are
of course situations of different complexity depending on the universality intended
to be achieved by the telemanipulator. However, most object interaction processes
will be nonlinear at least, as grasp and collision and release actions include instant
changes of load situations.
Simple examples of well known situations are e.g. the movement of masses
within free space, which can be easily modeled with a differential equation. How-
ever, the formulation of resistive force of the propulsion of bore rods for gas and oil
110 5 Internal Structure of Haptic Systems

exploration is almost unpredictable and subject to sudden changes depending on the

geological layers passed. A help for modeling may be force measurement principles
as discussed in chapter 10, which may be used not only for measurement during tele-
manipulation, but may also be part of measurement setups to optimize mathematical
models of mechanical objects. The design of peripheral influence values is intended
to support the major system components in the telemanipulation system. As a gen-
eral statement, the design of the transfer functions of any component, may it be the
central control and signal conditioning unit as being described in detail in chapter 12
or the description of disturbance variables will not be able to cover reality in its full
complexity. The challenge is to create a model covering major influence parameters
of the selected components. This demands an insight into the requirements of the
task, the interaction being performed, and the limitations and possibilities inherent
in the technical components. How to gain the knowledge necessary to achieve this
will be the subject of the second part of this book.
 Part II
Designing Haptic Systems
“This page left intentionally blank.”
 113

In the first part of the book the basics of a general understanding of the design
of haptic systems are given. There, the emphasis is on the discussion of the biolog-
ical background and on a closer look at resulting technical questions. In the second
part more definite technological subjects are examined and more concrete help for
frequent challenges of the design process is offered. The chapters in this part are
ordered according to the classic task list to be realized during any technical design
process. They start with more general questions concerning the overall system and
proceed to specialized questions concerning specific subcomponents. The chapters
are intentionally ordered in such a way that the ones dealing with questions whose
solution spectrum is limited to a great extent are discussed earlier than those which
provide more flexible solutions applicable to many situations.

• Chapter 6: The acquisition of requirements for the technical design process is

discussed. The design of haptic systems covers a plurality of technological ques-
tions. Especially the challenges concerning the high dynamics to be achieved
make a systematic approach mandatory for identifying the requirements.
• Chapter 7.1: After the basic requirements have been identified, a superordinate
view of the structure of the system to be built is necessary for which a method-
ology is given in this chapter. The resulting analysis does not only aim at the
decision on the control structure of the device, but also defines the technological
subproblems to be addressed during the following design process.
• Chapter 8: Especially kinaesthetic, but also multidimensional tactile systems
have to combine multiple degrees of freedom to fulfill the requirements for cer-
tain tasks. This leads to the systematic design of the kinematics of the device.
This chapter provides the necessary knowledge on kinematic design and covers
specific and sometimes surprising problems of mechanical transfer functions for
haptic devices in parallel kinematics.
• Chapter 9: In this most comprehensive chapter of the second part, typical actu-
ator principles with respect to their application in haptic devices are discussed.
The sections cover all popular actuation principles in an overview, as well as the
details of selected principles for a design from scratch.
• Chapter 10: Especially closed-loop controlled haptic devices need force sensors.
Furthermore, telemanipulation systems - besides simulators - are the second most
relevant group of high fidelity haptic devices. A haptic telemanipulator requires
force sensors at its end-effector, at least. This chapter covers the selection process
and the design of force sensors according to the physical principles lying in their
specifications within a haptically relevant range.
• Chapter 11: For a complete haptic interaction each system requires a position
measurement. Technological solutions for this subordinate technological chal-
lenge are discussed in this chapter, whereby different positioning- and other
movement-sensors are presented.
• Chapter 12: Typically, haptic devices are interfaced with time discrete systems
with digital signal processors, may they be controllers for universal interfaces
or complete simulation systems. This chapter deals with the interfaces between
114

these computing units and gives insight into the performance of different realiza-
tions.
• Chapter 13: The most frequent application of complex haptic systems is the per-
ception of virtual realities in simulations. For this purpose, it is necessary to
calculate the simulated physical properties by a software system running on a
programmable device. This chapter deals with a selection of algorithms used for
such calculations and gives an overview of the technological challenges. It also
provides a terminological basis for mechanics and hardware developers in order
to make them understand software engineers.
• Chapter 14: Final remarks on all previous chapters are made and the most im-
portant recommendations for the design of haptic systems are summarized.
Chapter 6
Identification of Requirements

T HORSTEN A. K ERN

6.1 The Right Questions to Ask

At the beginning of a technical design process the requirements for the product
which, usually, are not clear and unambiguous, have to be identified. Frequently,
customers formulate wishes and demands respectively solutions instead of require-
ments. A typical example is a task of the kind: “to develop a product just like product
P, but better/cheaper/nicer”. If an engineer accepts such a kind of order without get-
ting to the bottom of the original motivation the project will be doomed to failure.
Normally, the original wish of the customer concerning the product has to fulfill two
classes of requirements: The product shall have
• a certain function
• in a distinct technical and market oriented framework
The formulations of the market oriented requirements are manifold yet not in
the focus of the following analysis (for details of a general systematic product de-
velopment see [189]). They may be motivated by an existing product P to compete
with, but usually they are much more comprehensive and cover questions of budget,
timeframe of development, personal resources and qualifications and customers to
address.
With regard to the technical framework, the customer typically gives just unspe-
cific details. A statement like “a device shall provide a force on a glove” is not a
definition of a requirement but already a solution on the basis of existing knowledge
on the part of the customer. The complexity of a real technological solution spans
from a single actuator to provide e.g. a vibration to complex kinematics addressing
single fingers. If you question the customer’s original statement, it may even be that
his intention is, e.g. to simulate the force impression when switching the gears of a

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_6, 115

c Springer-Verlag Berlin Heidelberg 2009
116 6 Identification of Requirements

clutch in a passenger car. The knowledge about the actual application and follow-
ing that knowledge about the interaction itself allows the developer a much broader
approach, leading to a more optimized technical solution. The following sections
identify classes of possible technical realizations and collect the necessary questions
in clusters of possible applications. The list does not make a claim to be complete,
but is the essence of requirement specifications of dozens of developments achieved
during the last few years.

Type of haptic interaction?

Simulation and Telepresence of objects Haptic coding of information

Number of dimensions for the

Usage of a tool Direct haptic
haptically coded informations:
interaction
1
One Dimension
(intensity or time) 3

Direction-vector
in a 2D-space 4 + 3

Altering texture Altering shape Direction-vector

4 + 3
in 3D-space
2 1
Spatial distributed 2
2D Information

x = cluster of technological solution Spatial distributed 5

3D Information

Fig. 6.1 Structure for identifying relevant clusters of questions by analyzing the intended haptic
interaction.

In figure 6.1 a decision tree for the identification of clusters of questions has
been sketched. It is recommended to follow the tree from top to bottom in order to
identify the correct application and the corresponding cluster of questions. Subsec-
tion 6.1.1 describes the usage of the decision tree, whereas in the following sections
the clusters of questions are highlighted in detail.
6.1 The Right Questions to Ask 117

6.1.1 Interaction as a Classification Criterion

The core of the requirements’ identification is the definition of the type of haptic
interaction.

Type of Haptic Interaction

The first question asked should always refer to the type of interaction with the tech-
nical system. Is it a simulation of realistic surroundings, the interaction with physi-
cally available objects in terms of telepresence; or is the focus of the interaction on
the pure communication of abstract information? In the former cases the variants
are less versatile than in the latter.

Simulation and Telepresence of Objects

Does the interaction aim at touching virtual or via telepresence available objects? If
this is the case, does the interaction take place directly via fingers, hands or skin, or
is a mediator, e.g. a tool the interacting object? Does the user hold a specific tool
- a pen, a screw driver, a surgical instrument, a joystick of a plane, in his hands
and control one or more other objects with it, or does the user touch a plurality of
objects during the interaction with his or her hands? In the case of a tool-interaction
the chosen solution can be found in cluster 1 “kinaesthetics”, in the case of a direct
interaction another detail has to be considered

Direct Haptic Interaction

By touching physical objects, the user can notice the differences in all physical at-
tributes of the volume like mass, elasticity, plasticity and inner friction, and their
texture (chapter 2). In the case of interacting with shapes, the questions of cluster
1
“kinaesthetics” remains relevant, in the case of interacting with textures the ques-
tions of cluster 2 “surface-tactile” have to be considered. This is not necessarily an
alternative decision, however, with the same object interaction, both aspects can be
required at the same time or one after the other.

Haptic Coding of Information

In the case of abstract, not physical, object- oriented information communication

via the haptic sense the question of the dimension of information becomes relevant:
118 6 Identification of Requirements

• Does the interaction include a single event which occurs from time to time (e.g.
the call of a mobile phone) or is some permanently active information (e.g. a
distance to a destination) haptically communicated? These questions are one-
dimensional1 and covered by cluster 3 “vibro-tactile”.
• Is the interaction dominated by directional information coding an orientation in a
surface (directional movement) or in a space? In this case the questions covered
by cluster 4 “vibro-directional” are relevant. In such applications frequently
time respectively distal information is included, also making the questions in
cluster 3 become relevant.
• Does the interaction aim at the communication of data distributed within a two-
dimensional information layer, like geological maps, road-maps or texts on a
page? In these cases the questions of cluster 2 “surface-tactile” have to be an-
swered.
• In case there is volumetric information - the electrical field of an atomic bonding
or medical data sets - to be haptically transmitted, the questions of cluster 5
“omnidimensional” are to be be considered.
In the following section the questions in the clusters are further discussed and
some examples are given for the range of possible solutions to the questions aimed
at.

6.1.2 Cluster “Kinaesthetic”

Cluster 1 has to be chosen either when an interaction between fingers and shapes
happens directly or when the interaction takes place between tool and object. Both
cases are technical problems of multidimensional complexity2 . Each possible di-
mension movement corresponds to one degree of freedom of the later technical sys-
tem. Therefore the questions to be asked are quite straightforward and mainly deal
with the requirements for these degrees of freedom of tools and users:
• Which degrees of freedom do the tool /movement show? → rotatory, translatory,
combinations 3
• How large is the volume covered by the movements? → Maximum and minimum
values of angles and translations
• How dynamic is the movement? → Identification of maximum and minimum
velocities and accelerations. Usually, this question cannot be answered imme-
diately. A close observation of the intended interaction will help and - as far
1 as the information includes only one parameter
2 A tool interaction can be a one-dimensional task, but such an assignment concerning the technical
complexity can be regarded as an exception
3 In the case of a finger movement it has to be noted that not necessarily all movement directions

have to be equipped with haptic feedback to provide an adequate interaction capability. Frequently
it is even sufficient to provide the grasp-movement with haptic feedback, solely.
6.1 The Right Questions to Ask 119

as possible - measurements of movement velocities of instruments and fingers

should be made e.g. with the aid of videos,.
• Which forces / torques happen at each degree of freedom?4 ? → Definition of
maximum and minimum forces and torques.
• What is the maximum dynamics of the forces and torques? → Bandwidth of
forces and torques in frequency range, alternatively maximum slope in time-
domain (“from 0 to Fmax in 0.1 s”). Usually, this question cannot be answered
directly. Frequently, measurements are difficult and time-consuming, as an in-
fluence of the measurement on the interaction has to be eliminated. Therefore
it is recommended to do an analysis of the interaction itself and the objects the
interaction happens with. If it is soft - something typical of surgical simulation,
simple viscoelastic models can be used for interpolating the dynamics. The most
critical questions with respect to dynamics often address the initial contact with
an object, the collision. In this case especially the stiffness of the object and the
velocity of the contact describe the necessary dynamics. But is has to be stated
that the resulting high demands are not seldom in conflict with the technical pos-
sibilities (section 3.3).

6.1.3 Cluster “Surface-Tactile”

Haptic texture represents the microstructure of a surface. The lateral movement be-
tween this microstructure and the finger tip results in shear forces generating the
typical haptic impression of different materials. Haptic-bumps on the keyboard-keys
J and F are a special form of texture. Another variant of texture are Braille-letters
carrying additional abstract information. Cluster 2 has to be chosen when there is
a need to present information on any surface via the tactile sense. This can be either
coded information on a geological map on a more or less plane surface, but it can
also be object specific features like the material itself. The resulting questions for
the technical task are:
• Which body parts perform the interaction? → This trivial question has a signif-
icant impact, as the body part selected defines the resolution available on user
side and consequently the requirements for the size of the texture-generating el-
ements.
• Is the form of the texture-carrying shape subject to changes? If so, how much and
in which areas? → If the shape changes a lot, it is likely that the unit providing
the texture information has to be adapted to e.g. each finger (e.g. as a pin-array
or piezoelectric disc), as the fingers will have to be positioned independently of
each other. In this case it may even be necessary to provide a lateral movement

4 Frequently the customer will not be able to specify these values directly. In this case creative
variants of the question should be asked, e.g. by identifying the moving masses, or by taking
measurements with one’s own tools. These aids will be addressed within section 6.2.
120 6 Identification of Requirements

between finger and texture-unit to generate shear forces in the skin. In case of the
shape being fix, e.g. in the case of a map, a relative movement may happen by the
fingers themselves and the texture unit can be designed with less size restrictions.
• How fast does the displayed information change? → Textures change rarely dur-
ing the simulation of objects and display of maps. This is dramatically different
when e.g. texts or the influences of fluids on textures have to be displayed. The
answer to this question has a significant impact on the technical system.
• Which intensity range is covered by the texture? → In the simplest situation the
answer can be given by definite displacements and a resolution in bit. Usually,
only qualitative values of the properties of objects for interaction are available.
These hints have to be complemented by one’s own experiments. With regard
to the definition of these requirements it is very important to make sure that
the planned spatial addressability and maximum intensity change does not ex-
ceed the corresponding resolution of the user. A research on the corresponding
psychophysical experiments is highly recommended, as otherwise it may not be
possible to transmit the intended information density.

6.1.4 Cluster “Vibro-Tactile”

Cluster 3 is a solution space for simple one-dimensional technical problems and

corresponding questions. It covers independent dimensions of information (e.g. cod-
ing an event in a frequency and the importance in the amplitude). In this cluster,
distributions of intensity variations and /or time dependent distributions of single
events are filed. Technological solutions are usually vibrational motors or tactons,
as being used in mobile phones or game-consoles. But even if the technical solution
itself seems quite straightforward, the challenge lies in the coding of information
with respect to intensity and time and an appropriate mechanical coupling of the
device to the user.

• Which mechanical interface for the transmission of haptic information to the user
is planned? → More specifically: Is this interface influenced by mechanical limits
like housings?
• Which design space is available? → Frequently, vibro-tactile solutions are lim-
ited as to the available space at an early stage of the design due to requirements
for mobility.
• Which resolution is expected for the planned intensity variation? → The criteri-
ons are similar to those of the “surface-tactile” cluster. As the “vibro-tactile” clus-
ter frequently deals with oscillating systems, the dependence of the perception of
oscillations on its frequency has to be taken into account. The user’s perception
is the limiting factor for intensity variations, which themselves are dependent on
the mechanical coupling between device and user, too.
6.1 The Right Questions to Ask 121

6.1.5 Cluster “Vibro-Directional”

Vibro-tactile systems code one-dimensional information in the form of intensities.

It is obvious that by the combination of multiples of such information sources di-
rectional information can be transmitted. This may happen two-dimensionally in
a plane surface, but also three-dimensionally. Cluster 4 deals with such systems.
One possible technical solution for directional surface information would be to lo-
cate a multitude of active units in the shape of a ring around a body part, e.g. like a
belt around the belly. The direction is coded in the activity of single elements. This
approach can also be transferred to a volumetric vector, whereby in these cases a
large number of units is located on a closed surface, e.g. the upper part of the body.
The activity of single elements codes the three dimensional direction as an origin of
a normal vector on this surface. In addition to the questions of cluster 3 this cluster
deals with the following questions:
• What is the intended resolution on the surface /in the space? → As well as before
dependent on the body surface used, it is likely that the human perception rep-
resents the limit for the achievable resolution. Corresponding literature [51, 267]
has to be checked carefully before the technical requirements can be met.
• What number of simultaneously displayed vectors is expected? → The fact that
users will be able to identify one direction does not guarantee that with a parallel
display of two points the user will perform equally well. Simultaneous display of
information frequently results in masking-effects hard to be quantified. Experi-
ments and analyses of the intended application are strongly recommended.
• Which frame of reference is used? → The information displayed is usually em-
bedded in a frame of reference, which is not necessarily identical with the user’s
frame of reference and his or her body. The user may change his position for
example in a vehicle, which results in a loss of the position of the elements fixed
to the body and their orientation in the vehicle. It is necessary to be aware of the
active frame of reference (local user-oriented, or vehicle-oriented, or maybe even
world-oriented) and to provide measurement equipment for identifying changes
in user positions and frame of reference. Additionally, it may become necessary
to present a haptic reference signal to the user, which calibrates the user’s per-
ception to the frame of reference, e.g. a “north”-signal.

6.1.6 Cluster “Omni-Dimensional”

Cluster 5 deals with systems coding real volumetric information. Within such
a three-dimensional space each point either includes intensity information (scalar
field) or vector information (vector field). The sources of such data are numerous
and frequent, may it be medical imaging data, or data of fluid mechanics, of atomic
physics, of electrodynamics, or of electromagnetics. Pure systems of haptic inter-
122 6 Identification of Requirements

action with such kinds of data are seldom. Frequently, they are combinations of
the clusters “kinaesthetic” and “vibro-tactile” for scalar fields, respectively “kinaes-
thetic” with six active haptic degrees of freedom for vector fields5 . Consequently,
the specific questions of this cluster add one single aspect to already existing ques-
tions of the other clusters:
• Does the intended haptic interaction take place with scalar fields or with vec-
tor fields?→ For pure vector fields kinaesthetic systems with the correspond-
ing questions for six active degrees of freedom should be considered. In the
case of scalar fields, an analysis of vibro-tactile systems in combination with
three-dimensional kinaesthetic systems and the corresponding questions should
be considered. Then the property of the scalar value corresponds to the dynamics
of the coded information.

6.1.7 Cluster “always”

For any development process there are several questions which always have to be
asked. They often refer to the time-frame as well as to the resources available for
the development. For haptic devices two specific questions have to be focused on,
as they can become quite limiting for the design process due to specific properties
of haptic devices:
• Which energy sources are available? → It is not a necessary prerequisite that
electrical actuators have to be used for haptic devices, especially in the case of
telemanipulation systems. The usage of pneumatic and hydraulic energy sources,
especially for tactile devices is a real alternative and should be considered.
• The design, how expensive may it be? → The prices of current kinaesthetic haptic
systems reach from 200 EUR of mass-products to 1.500 EUR of medium scale
products to devices of 25t EUR for small series and. 100t EUR for individual
solutions. These prices only partly result from commercial acquisitiveness, but
mostly from the technical requirements and the efforts which have to be taken.

6.2 Experiments Together With the Customer

The customer formulates requirements - as mentioned before - typically in an inex-

act instead of a specific way. Additionally, there is the problem of a very unspecific
terminology with regard to the design of haptic systems For the description of haptic
sensual impressions there are numerous adjectives difficult to quantify, like: rough,

5 The haptic interaction with objects in a mathematical abstraction always is an interaction with
vector fields. In the vectors, forces of surfaces are coded, which themselves are time dependent,
e.g. from movements and /or deformations of the objects themselves.
6.2 Experiments Together With the Customer 123

soft, smooth, gentle, mild, hard, viscous, as well as others derived from substantives
such as: furry, silky, hairy, watery, and sticky which can be compared to real objects.
So what could be more obvious than asking the customer to describe his / her haptic
impressions by comparisons?

Ask the customer to describe the intended haptic impression with refer-
ence to objects and items in his/her environment. These items should be
easily at hand, like e.g. vegetables and fruits which offer a wide spec-
trum of textures and consistencies for comparison.

Sometimes the customer first needs to develop a certain understanding of the

haptic properties of objects and items. This can best be achieved by his/her directly
interacting with them. Examples of haptically extreme objects have to be included
in a good sample case, too.

Provide a sample case including weights and springs of different size,

even marbles, fur, leather and silk. Depending on the project, add sand-
paper of different granularity. Use these items to explain haptic param-
eters to the customer and help the customer to optimize the description
of the product performance expected!

From practical experience, we can recommend also to take spring balances and
letter balances or electronic force sensors with you to customer meetings. Fre-
quently, it is possible to attach a handle directly to the items and ask the customer
to pull, until a realistic force is reached. This enables customers of non-technical
disciplines to quickly get an impression of the necessary moments and forces.

Take mechanical measurement instruments with you to the customer

meetings and allow the customer to touch and use them! This gives him
/ her a good first impression of the necessary force amplitudes.

In order to give a better impression of texture, mechanical workshops may pro-

duce patterns of knurls and grooves of different roughness on metals. Alternatively,
sandpaper can be used and, by its defined grade of granularity, can provide a stan-
dardized scale to a certain extent.

Use existing materials with scales to describe roughness and simulate

the impression of texture.
124 6 Identification of Requirements

What does not work:

A normal customer without expertise in the area of haptics will not be able to
give statements concerning resolutions or dynamics of the haptic sense. This kind
of information has to be derived from the type of interaction and the study of the
psychophysical knowledge of comparable applications. Therefore, the experience
of the developing engineer is still indispensable despite all the systematizations in
the technical design process.

Do not confuse the customer by asking questions about the physical

resolution! This is necessarily the knowledge of the haptic engineer.
However, learn about the dynamics of the interaction and try to assess
the application, e.g by asking about the frame rate of a simulation, or
the maximum number of load changes per seconds of a telemanipulator.

6.3 Requirement Specifications of a Haptic System

The customer ’s answers result in a functional requirement list. Depending on this

list, a system specification sheet is formulated. This sheet will differ in its structures
depending on whether tactile or kinaesthetic systems are addressed. In table 6.1 the
most relevant technical parameters of a general haptic system are collected. This
table is meant to be an orientation and has to be adapted to the specific situation by
removing irrelevant points and adding application-specific aspects.
6.3 Requirement Specifications of a Haptic System 125

Table 6.1: Example of a system specification for a haptic device

S YSTEM SPECIFICATIONS OF A HAPTIC DEVICE

Rev.: 01.01.2009
R/W D ESCRIPTION VALUE S OURCE /C OMMENT
R/W6 Especially kinaesthetically motivated parameters
Shall give an idea of DOFs,
R Number of DOFs 2x rot., 1x transl.
name them!
Minimum of workspace to be
R Workspace 100x50x50 mm3
achieved
Maximum workspace neces-
W Maximum Workspace 150x100x100 mm3
sary
Maximum force in DOF
R 5N
“name”
Maximum force in DOF Always define a range of
W 7N
“name” forces!
Minimum force in DOF
R 0.2 N
“name”
Minimum force in DOF Always define a range of
W 0.1 N
“name” forces!
Maximum dynamics (band- Shows (among other things)
R width) for DOF “name” in a 100 Hz e.g. the maximum dynamics of
blocked situation the driver electronics
Maximum dynamics (band-
Shows the bandwidth the cus-
W7 width) for DOF “name” in a 200 Hz
tomer dreams of.
blocked situation
There may be applications
with pure dynamic movements
Smallest border frequency
R static without a static portion. This
when movement is blocked
makes this question interest-
ing.
This is a question regarding se-
Maximum velocity of move- curity too, as it defines the me-
R 10 mm/s
ment in idle mode chanical energy stored in the
system.
The change of velocity, which
Maximum bandwidth of the is the acceleration of the sys-
R 10 Hz
velocity change tem, has a large influence on
the energy the system requires.
This is an alternative represen-
Maximum haptic impedance at tation to the independent defi-
R 10 Ns/m
the output nition of force and velocity for
dynamic (but passive) systems!
Continuation on the next page ...

6 R: requirement, W: wish
7 The combination of requirements and wishes (R and F) may be used for almost any element of
the system specification. It is recommended to make use of this method, but due to clarity in the
context of this book this approach of double-questions is aborted here.
126 6 Identification of Requirements

Continuation:
RT /W D ESCRIPTION VALUE S OURCE /C OMMENT
This is an alternative represen-
Minimum haptic impedance at tation to the independent defi-
R 0.01 Ns/m
the output nition of force and velocity for
dynamic (but passive) systems!
Smallest position resolu- Usually measurement of the
R tion/measurement insecurity 0.1 mm position is self-evident for hap-
for DOF “name” tic interaction.
Smallest position resolu-
W tion/measurement insecurity 0.05 mm
for DOF “name”
Type of the mechanical inter-
R Button/pen/none Is there a handle?
face
Has influence on weight, size
R Mechanical reference point Grounded, worn
and energy
R/W Especially tactile motivated parameters
Direction(s) of the tactile stim- An alternative would be laterial
R Normal to the skin
ulation or a combinationof both
Is especially relevant for
Maximum displacement- pin-displays, but may be also
R amplitude of the tactile 1 mm understood as oscillation-
elements amplitude of vibrational
elements.
Minimum amplitude resolution May include several levels for
R digital (on/off)
of displacements the pin to be moved to
2 mm Distance Varies extremely in depen-
R Highes density of stimulation from midpoint ot dency from the chosen skin
midpoint area in contact
Maximum geometrical size of
R 2 mm Diameter
stimulation
Maximum frequency range of Relevant for tactile actuators
R 100 to 300 Hz
stimulation only, of course.
Minimum frequency-
R 1 Hz For vibrotactile actuators
resolution
Pin-based actuators may not
Maximum force during dis- necessarily be stiff. Systems of
R 20 N
placement/stiffness lower admittance may be used
too.
Necessary to identify, whether
Attached to the envi- there is a relative movement
R Connection to the user
ronment / worn between skin (e.g. finger) and
the display.
May have an immensive im-
Maximum number of fingers
pact on the design when e.b.
R simultaneously in contact with 1-10
full-hand exploration is re-
the device
quired.
R/W Digital interface
Continuation on the next page ...
6.3 Requirement Specifications of a Haptic System 127

Continuation:
RT /W D ESCRIPTION VALUE S OURCE /C OMMENT
Usually slightly lower than the
Minimum resolution of the
R 12 bit measurement error of force-
ouput data
and position-measurement.
Usually slightly larger than the
Minimum resolution of the in-
R 12 bit resolution of force- and posi-
put data
tion input-data.
Should be at least two times,
better would be 10 times,
8 larger than the border fre-
R Frequency of the haptic loop 1000 Hz
quency of the design. Has in-
fluence on the perceived stiff-
ness.
Typically the interface to be
W Other interface-requirements Use USB/FireWire... used is subject to company pol-
itics.
As any other hardware a hap-
R Interface driver API tic interface needs an own soft-
ware driver for abstraction.
R/W General parameters
May become very relevant for
Maximum temperature range ◦ actuator principals with little
R 10 C to 50◦ C
for operation efficiency in extreme environ-
ments (automotive).
R Maximum volume 500 · 500 · 200 mm3 Device-size
Especially relevant if the de-
vice is worn. This limit will
R Weight 1 kg
strongly influence the mechan-
ical energy generated.
Very important, devices were
spotted on fairs, which ceased
R Electrical supply Battery/ 110V/ 230V to function due to errors made
when considering AC voltages
of different countries.
Primary power consumption
R Maximum power 50 W
including all losses.
Additionally a system specification includes references to other standards and spe-
cial requirements relating to the product development process. Among others, these
are the costs for the individual device, the design-process itself and the number
R
of devices to be manufactured in a certain time-frame. Additionally the time of
shipment, visual parameters for the design, and security related issues are usually
adressed.

8 A “haptic loop” is a complete cylcle including the output of the control-variable (in case of simu-

lators this variable was calculated the time-step before) and the read operation on the measurement
value
128 6 Identification of Requirements

6.4 Order of Technical Decisions

During the technical development of haptic systems, the decisions on single com-
ponents influence each other intensively. However, this influence is not identical
between all components. For the engineer it is necessary to proceed in the solution
identification for each component, after having gained the knowledge of the require-
ments for the haptic system. It is obvious that, according to a systematic develop-
ment process, each solution has to be compared to the specifications concerning its
advantages and disadvantages. The recommended procedure of how to deal with the
components is the basis of the chapter structure of this section of the book and is
summarized once again for completeness:
1. Decision about the control engineering structure of the haptic system based on
the evaluation of the application (tactile or kinaesthetic), the impedance in idle
state (masses >20 g and friction acceptable) and the maximum impedance (stiff-
nesses > 300 N/m or smaller) (fig. 6.2, chapter 7)
2. Decision about the kinematics based on calculations of the working volume and
the expected stiffness (chapter 8)
3. Based on the well-known mechanical structure, the actuator design can be made.
For this a rough decision can be taken, based on section 9.1 for orientation, fol-
lowed by the concrete dimensioning and the calculations of the actuators accord-
ing to the other sections in chapter 9.1.
4. Dependent on the chosen control engineering structure, the force-sensor design
can be performed parallel to the actuator design (chapter 10).
5. Relatively uncritical for the design is the choice of the positioning sensors (chap-
ter 11).
6. The electronic interfaces are subordinate to all the decisions made before (chap-
ter 12).
7. The software design of the haptic rendering itself ,in many aspects, is so inde-
pendent of prior technical decisions that it can be decoupled in almost all aspects
(chapter 13) from the rest of the design, - however not of the common set of
specifications, of course.
Nevertheless it is vital to note that e.g. the kinematics design cannot be real-
ized, completely decoupled from the available space for the device and the forces
and torques - respectively the actuator. Additionally, kinematics directly influences
any measurement technology as even displacement sensors have limitations on res-
olution and dynamics. The order suggested above for taking decisions has to be
understood as being a recommendation for processing the tasks; it does not free the
design engineer from the responsibility to keep an overview of the subcomponents
and their reciprocal influence.
6.4 Order of Technical Decisions 129

Decision on control architecture

Tactile Kinaesthetic

Zmin "large"
output
in- and output Zmin "small"
only

Zmax <300 N/m Zmax >300 N/m

Open-loop digital input scalar input
admittance
Closed-loop
controlled
impedance
controlled

Open-loop Closed-loop Open-loop Closed-loop

admittance admittance impedance admittance
controlled controlled controlled controlled
with
switch
functionality

Fig. 6.2 Aid for the decision on the choice of the control structure.
“This page left intentionally blank.”
Chapter 7
Control of Haptic Systems

O LIVER M ECKEL

7.1 Introduction

The analysis and investigation of technical systems and processes is based upon the
requirement to establish safe and reliable system behavior, and controllable system
states. By the depiction as systems the analysis is put on an abstracted level which
allows covering many different technical systems described by their fundamental
physics. On this abstracted level a general analysis of closed loop control issues is
possible using several methods and techniques. The resulting procedures are appli-
cable to a various number of system classes. The main purpose of any depiction
and analysis of control systems is to achieve safe system behavior and reliable pro-
cesses. Power is nothing without control, that might be a credo and the fundamental
motivation for any system investigation to prove system reliability and safe system
behavior.
The abstract description that shall be used for a closed loop control analysis starts
with the mathematical formulation of the physical principles the system follows. As
mentioned above, systems with different physical principles are covered up by sim-
ilar mathematical methods. The depiction by differential equations or systems of
differential equations proves widely usable for the formulation of various system
behaviors. Herein analogies allow transforming this system behavior into the dif-
ferent technical context of a different system, provided that there exists a definite
formulation of the system states that are of interest for the closed loop control anal-
ysis. The mathematical formulation of the physical principles of the system, also
denoted as modeling, is followed by the system analysis including dynamic behav-
ior and its characteristic. With this knowledge, a wide variety of design methods for
control systems in closed loops become applicable. Their main requirements are:

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_7, 131

c Springer-Verlag Berlin Heidelberg 2009
132 7 Control of Haptic Systems

• System stability: The fundamental requirement for stability in any technical sys-
tem is the main purpose for closed loop control design.
• Control quality: Tracking behavior of the system states to demanded values every
system is faced with external influences also denoted as disturbances which inter-
fere with the demanded system inputs and disrupt the optimal system behavior.
To compensate this negative influence a control system is designed.
• Dynamic behavior and performance: In addition to the first two issues, the need
for a certain system dynamic completes this requirements-list. Besides the qual-
ity of the control result tracking the demanded values, the system behavior within
the range of changes from these demanded values is focused. Also the control ef-
fort which needs to achieve a certain control result is to be investigated.
The main challenge for closed loop control design in many other engineering dis-
ciplines is to deal with different goals, that are often in conflict with each other.
These conflicts are a direct result of the above characteristics. Often these issues
are weighted in a different way depending on the used design method. Typically a
gained solution is never an optimal one, rather a trade off between system require-
ments.

7.2 System Description

A variety of description methods can be applied for the mathematical formulation

of systems with different physical principles. One of the main distinctions is drawn
between methods for the description of linear and nonlinear systems, summarized
in the following paragraphs.

7.2.1 Linear SISO-System Description

As shown in chapter 5, a formulation of haptic devices as impedance controlled

respectively admittance controlled system leads to a certain specific definition of
system states. Within both methods of formulation, the physical principles can be
described by the corresponding differential equations. As a simple example a second
order system is analyzed to show the similarity of both approaches. Figure 7.1 shows
the model scheme of a damped mechanical oscillator with the parameters mass m,
proportional damping d and linear stiffness k. The basic mechanical principles form
the following differential equation:

mẍ(t) + d ẋ(t) + kx(t) = f (t) (7.1)

Here x denotes the actual position of the mass m. Using this formulation the speed
follows from the gradient dx
dt = v. The corresponding derivative of the differential
equation leads to
7.2 System Description 133

x, x, x

m f(t)

Fig. 7.1 Mechanical oscillator.


mv̇(t) + dv(t) + k v dt = f (t) (7.2)

Looking on the electric RLC-system shown in figure 7.2, a similar formulation by a

linear differential equation is formulated.

di(t) 1
L + Ri(t) + i dt = u(t) (7.3)
dt C
The correspondence of these two differential equations relates the respective kinetic
and potential energy storing mechanical devices – mass and spring – to the electri-
cal energy storing devices – inductance and capacitance. Furthermore the dissipa-
tive mechanical damping is related to the electrical resistance. Another approach of
system formulation leads to a second way how to relate mechanical and electrical
system devices to each other as shown in the following differential equation:

R L C
i

u(t)

Fig. 7.2 RLC-Network.


du(t) 1 1
C + u(t) + u dt = i(t) (7.4)
dt R L
Deriving from this system formulation the electric impedance Z = ui describes
the system transfer behavior. The reciprocal value denotes the electrical admittance
Y = Z1 . This impedance and admittance respectively based system formulation is
134 7 Control of Haptic Systems

transferable from the electrical RLC-systems to mechanical systems. Appendix 16

shows the general impedance or admittance based formulation of translatory and
rotary mechanical systems. The main distinction is based on the definition of the
input and output variables. For mechanical impedances the system input will be de-
picted by a certain speed or differential speed variable, the system output variable
will denote a resulting force. The formulation of a mechanical admittance switches
input and output variable which leads to a system input describing a force affecting
the energy storing devices and influencing the speed as the system output. In many
approaches the description of mechanical systems is based on the equilibrium of
forces, which often leads to formulation of the system behavior as mechanical ad-
mittance. Nevertheless the formulation as mechanical impedance is not a mistake; it
is merely a change of the point of view on the same system principle. The following
example shows both approaches:

va A
b

d m Fe

Fig. 7.3 Example for mechanical admittance.

Within this system Fe denotes the force affecting the mass m as the system input.
As output variable the system speed in point A va is chosen. The system behavior is
described by the following linear differential equation:

mv̇m + dvm = Fe (7.5)

The kinematic of the links provides vm = ba va which leads to

a a
m v̇a + d va = Fe . (7.6)
b b
Switching the point of view and choosing a measurable force in point A as system
output and the speed of the mass vm = ve as system input as shown in figure 7.4
The resulting differential equation is
a a
Fa = m v̇e + d ve . (7.7)
b b
7.2 System Description 135

Fa A
b

d m
ve

Fig. 7.4 Example for mechanical impedance.

The transformation of impedances or admittances into the L APLACE domain ab-

stracts the system analysis, and a flexible investigation of coupled system compo-
nents becomes possible. In many control design methods the L APLACE domain is
used to gain a consistent formulation of system behavior described as L APLACE
transfer function G(s). Especially while dealing with Single-Input-Single-Output
Systems this method of description has proven its strength and allows the direct
applicability of a various number of techniques for stability analysis, system clas-
sification and control system design procedures. Herein the distinction between
impedance and admittance based system structures is not necessary any more. The
transformation to the L APLACE domain of both differential equations from the ex-
ample above brings up the following transfer functions

Fa (s) a
= · (d + m · s) (7.8)
ve (s) b
va (s) b 1
= · . (7.9)
Fe (s) a d +m·s

It becomes quite obvious that one of these transfer function is the reciprocal
of the other. The corresponding phase plots point out the main difference between
both formulations. Describing the exemplary system as admittance as described in
equation 7.6, va (s) as output signal shows a phase delay to the input signal Fe (s),
whereas the description of the system using impedances leads to exactly the opposite
behavior. For comparison the step responses of both formulations are investigated.
Due to the different choice of input and output variables the system shows a dif-
ferent transfer behavior. This behavior described by equation 7.9 shows as expected
the typical step response of a PT1 -System whereas the formulation of equation 7.8
leads to a DT1 -behavior.
136 7 Control of Haptic Systems

Force Fa
Speed va

Time t Time t

Fig. 7.5 Step response for system analysis.

In figure 7.5 on the left side the PT1 step response in the speed va due to a
force Fe is shown. The right side describes the step response of the force Fa due
to a speed ve . The differences of these step responses become quite obvious, each
corresponding to the based system formulation, and can be justified by the different
particular solution of the differential equation.
The shown methods to establish a mathematical formulation of a system are com-
pletely equivalent. Consequently the use of L APLACE transfer function Gi (s) per-
mits a nearly arbitrary combination of different system components, which can be
independently modelled as impedance or admittance. For further information about
the theoretical background on the L APLACE domain and the corresponding trans-
formation, [264], [60] and [165] are recommended.

7.2.2 Linear State Space Description

Besides the formulation of system characteristics through transfer functions, the de-
scription of systems using the state space representation in the time domain allows
to deal with arbitrary linear systems too. For Single-Input-Single-Output-Systems,
a description using an nth order ordinary differential equation is transformable into
a set with n first order ordinary differential equations. In addition to the simplified
usage of numerical algorithms for solving this set of differential equations, the ma-
jor advantage is the applicability to Multi-Input-Multi-Output-Systems. A correct
and systematic model of their coupled system inputs, system states, and system out-
puts is comparably easy to achieve. On the contrary to the system description in the
L APLACE domain by transfer functions G(s), the state space representation formu-
lates the system behavior in the time domain. Two set of equations are necessary
for a complete state space system representation. These are denoted as the system
equation
7.2 System Description 137

ẋ = Ax + Bu (7.10)
and the output equation
y = Cx + Du. (7.11)
The vectors u and y describe the multidimensional system input respectively system
output. Vector x denotes the inner system states.
As an example for state space representation the 2nd order mechanical oscillating
system as shown in figure 7.6 is examined. Assuming the existence of time invariant
parameters the description by a 2nd order differential equation is:

mÿ + d ẏ + ky = u (7.12)

Figure 7.7 shows the corresponding block diagram of this PT2 -System.

y, y, y

m f(t)

Fig. 7.6 2nd order oscillator.

.. .
1 y y
u
–
—
m ò ò y

–
d

Fig. 7.7 PT2 block diagram.

The transformation of the 2nd order differential equation 7.12 into a set of two
1st order differential equation is done by choosing the integrator outputs as system
states:
138 7 Control of Haptic Systems

x1 = y ⇒ ẋ1 = x2
k d 1
x2 = ẏ ⇒ ẋ2 = − x1 − x2 + u (7.13)
m m m
Thus the system equation for the state space representation is as follows:
   
ẋ1 0 1 x1 0
= + 1 u (7.14)
ẋ2 −m −m
k d x2 m

The general form of the system equation is:

ẋ = A x + B u (7.15)

This set of equations contains the state space vector x. Its components describe all
inner variables of the process that are of interest and that have not been examined
explicitly using a formulation by transfer function. The system output is described
by the output equation. In the given example as shown in figure 7.6 the system
output y is equal to the inner state x1

y = x1 (7.16)

which leads to the vector representation of


  x1
y= 1 0 (7.17)
x2

The general form of the output equation is:

y = Cx + Du (7.18)

which leads to the general state space representation that is applicable for Single or
Multi Input and Output systems. The structure of this representation is depicted in
figure 7.8:

ẋ = A x + B u
y = Cx+Du

Although not mentioned in this example, matrix D denotes a direct feedthrough

which occurs in systems whose output signals y are directly affected by the in-
put signals u without any time delay. Thus these systems show a non delayed step
response. For further explanation on A, B, C und D the reference [164] is recom-
mended.
7.2 System Description 139

x0 (t)

u (t) x× (t) x (t) y (t)

B ò C

Fig. 7.8 State space description.

7.2.3 Nonlinear System Description

A further challenge within the formulation of system behavior is to imply nonlin-

ear effects, especially if a subsequent system analysis and classification is needed.
Although a mathematical description of nonlinear system behavior might be found
fast, the applicability of certain control design methods is an additional problem.
Static nonlinearities can be easily described by a serial coupling of a static nonlin-
earity and linear dynamic device to be used as a summarized element for closed
loop analysis. Herein two different models are differentiated. Figure 7.9 shows the
block diagram consisting of a linear element with arbitrary subsystem dynamics fol-
lowed by a static nonlinearity. This configuration also known as W IENER-Model is

u ~
u y
G(s) f(.)

Fig. 7.9 W IENER-Modell.

described by

ũ(s) = G(s) · u(s)

y(s) = f (ũ(s)).
140 7 Control of Haptic Systems

In comparison, figure 7.10 shows the configuration of the H AMMERSTEIN-Model

changing the order of the underlying static nonlinearity and the linear dynamic sub-
system. The corresponding mathematical formulation of this model is as described

u ~
u y
f(.) G(s)

Fig. 7.10 H AMMERSTEIN-Modell.

by

ũ(s) = f (u(s))
y(s) = G(s) · ũ(s).

More complex structures appear as soon as the dynamic behavior of a system

is affected by nonlinearities. Figure 7.11 shows as an example a system with an
internal saturation of a states derivative. For this configuration both models cannot
be applied as easily as for static nonlinearities furthermore if a system description
is needed usable for certain methods of system analysis and investigation.

u y
K ò
–

Fig. 7.11 System with internal saturation.

Typical examples for systems showing that kind of nonlinear behavior are elec-
trical motors whose torque current characteristic is affected by saturation effects,
and thus whose torque available for acceleration is limited to a maximum value.
This kind of system behavior is one example of how complicated the process of
system modeling may become, as ordinary linear system description methods are
not applicable to such a case. Nevertheless it is necessary to gain a system formu-
lation in which the system behavior and the system stability can be investigated
7.3 System Stability 141

successfully. To achieve a system description taking various system nonlinearities

into account, it is recommended to set up a nonlinear state space descriptions. They
offer a wide set of tools applicable to the following investigations. Deriving from
equation 7.10 and 7.11 the nonlinear system description for single respectively multi
input and output systems is as follows:

ẋ = f(x, u,t)
y = g(x, u,t).

This state space description is most flexible to gain a usable mathematical for-
mulation of a systems behavior consisting of static, dynamic and arbitrarily coupled
nonlinearities.

7.3 System Stability

As mentioned in section 7.1 one of the most important goals of the control design
is the stabilization of systems or processes during their life cycle, while operative or
disabled. Due to the close coupling of haptic systems to a human user via a human
machine interface, safety becomes most relevant. Consequently the focus of this
chapter lies on system stability and its analysis by using certain methods applicable
to many systems. It has to resemble the system’s behavior correctly, and has to
be aligned with applied investigation technique. For the investigation of systems,
subsystems, closed looped systems, and single or multi input output systems, a wide
variety of different methods exists. The most important ones shall be introduced in
this chapter.

7.3.1 Analysis of Linear System Stability

The stability analysis of linear time invariant systems is easily done by the investi-
gation of the system poles or roots derived from the eigenvalue calculation of the
system transfer function G(s). The decisive factor is the sign of the real part of these
system poles. A negative sign in this real part indicates a stable eigenvalue, a posi-
tive sign denotes an unstable eigenvalue. The correspondence to the system stability
becomes obvious while looking at the homogenous part of the solution of the or-
dinary differential equation describing the system behavior. As example a system
shall be described by
T ẏ(t) + y(t) = Ku(t). (7.19)
The homogenous part of the solution y(t) is derived using

1
yh = eλ t with λ =− . (7.20)
T
142 7 Control of Haptic Systems

As it can be seen quite fast the pole λ = − T1 has a negative sign only if the time con-
stant T has a positive sign. In this case the homogenous part of y(t) disappears for
t → ∞, while it rises beyond each limit exponentially if the pole λ = − T1 is unstable.
This section will not deal with the basic theoretical background of linear system sta-
bility, as these are basics of control theory. Focus of this section is the application of
certain stability analysis methods. Herein it will be distinguished between methods
for a direct stability analysis of a system or subsystem and techniques of the closed
looped stability analysis. For direct stability analysis of linear system the investiga-
tion of the poles placement in the complex plane is fundamental. Besides the explicit
calculation of the system poles or eigenvalues the ROUTH H URWITZ criterion of-
fers to determine the system stability and the system pole placement with explicit
calculation. In many cases this simplifies the stability analysis. For the analysis of
the closed loop stability the determination of the closed loop pole placement is also
a possible approach. Additional methods leave room for further design aspects and
extend the basic stability analysis. Well-known examples of such techniques are
• Root locus method
• N YQUIST´s stability criterion.
The applicability of both methods will be discussed in the following without
looking at the exact derivation. The root locus offers the opportunity to investigate
the pole placement in the complex plane depending on certain invariant system pa-
rameters. As example of invariant system parameters changing time constants or
variable system gains might occur. The gain of the open loop is often of interest
within the root locus method for closed loop stability analysis and control design.
In equation 7.21 GR denotes the transfer function of the controller, GS describes the
behavior of the system, plant or process to be controlled.

− Go = GR GS (7.21)

Using the root locus method, it is possible to apply predefined sketching rules when-
ever the dependency of the closed loop pole placement on the open loop gain K is
of interest. The closed loop transfer function is depicted by equation 7.22
GR GS
Gg = (7.22)
1 + GR GS
As an example an integrator system with a second order delay (IT2 ) described by
1 1 1
GS = · · (7.23)
s 1 + s 1 + 4s
is examined. The control transfer function is GR = KR . Thus we find as open loop
transfer function
KR
− Go = GR GS = . (7.24)
s(1 + s)(1 + 4s)
7.3 System Stability 143

Using the sketching rules which can be found in various examples in literature [165],
[264], the root locus graph as shown in figure 7.12 is derived. The graph indicates,
Imaginary axis

Real axis

Fig. 7.12 IT2 root locus.

that small gains KR lead to a stable closed loop system since all roots have a negative
real part. A rising KR leads to two of the roots crossing the imaginary axis and
the closed loop system becomes unstable. This simplified example proves that this
method can easily be integrated in a control design process, as it delivers a stability
analysis of the closed loop system only processing an examination on the open loop
system. This issue is also one of the advantages of the N YQUIST stability criterion.
Additionally the definition of the open loop system is sufficient to derive a stability
analysis of the system in a closed loop arrangement. This section will concentrate
on the simplified N YQUIST stability criterion investigating the open loop frequency
response described by

− Go ( jω ) = GR ( jω )GS ( jω ). (7.25)

The N YQUIST stability criterion is based on the characteristic correspondence of

amplitude and phase of the frequency response. As example we use the already
introduced IT2 -system controlled by a proportional controller GR = KR . The Bode
plot of the frequency response is shown in figure 7.13:
The stability condition which has to be met is given by the phase of the open loop
frequency response, with ϕ (ω ) > −180◦ in case of the frequency response’s ampli-
tude A(ω ) being above 0 dB. As shown in figure 7.13, the choice of the controller
144 7 Control of Haptic Systems

Bode diagramm
40

20

0
Magnitude [dB]

-20

-40

-60

-80

-100

-120

-90

-135
Phase [deg]

jR
-180

-225

-270
-2 -1 0 1 2
10 10 10 10 10

Frequency [rad/sec]

Fig. 7.13 IT2 frequency response.

gain KR transfers the amplitude graph of the open loop frequency response vertically
without affecting the phase of the open loop frequency response. For most applica-
tions the specific requirement of a sufficient phase margin ϕR is compulsory. The
resulting phase margin is also shown in figure 7.13. All such requirements have to
be met in the closed control loop and must be determined to choose the correct con-
trol design method. In this simplified example the examined amplitude and phase of
the open loop frequency response is dependent on the proportional controller gain
KR , which is sufficient to establish system stability including a certain phase mar-
gin. More complex control structures such as PI, PIDTn or Lead Lag extend the
possibilities for control design to meet further requirements.
This section showed the basic principle of the simplified N YQUIST criterion be-
ing applicable to stable open loop systems. For an investigation of unstable open
loop systems the general form of the N YQUIST criterion must be used, which it-
self is not introduced in this book. For this basic knowledge it is recommended to
consult [165] and [264].
7.3 System Stability 145

7.3.2 Analysis of Nonlinear System Stability

The application of all previous approaches for the analysis of system stability is
limited to linear time invariant systems. Nearly all real systems show nonlinear ef-
fects or consist of nonlinear subsystems. One approach to deal with these nonlinear
systems is the linearization in a fixed working point. All further investigations are
focused on this point, and the application of the previously presented methods be-
comes possible. If these methods are not sufficient, extended techniques for stability
analysis of nonlinear systems must be applied. The following are examples of rep-
resenting completely different approaches:
• Principle of the harmonic balance
• Phase plane analysis
• P OPOV criterion and circle criterion
• LYAPUNOV´s direct method
• System passivity analysis
Without dealing with the mathematical background or the exact proof the prin-
ciples and the application of chosen techniques shall be demonstrated. At this point
a complete explanation of this topic is too extensive due to the wide variety of the
underlying methods. For further detailed explanation study [58], [56], [57], [265],
[135] and [234].
As an introductionary example the analysis of closed loop systems can be done
applying the P OPOV criterion respectively the circle criterion. Figure 7.14 shows the
block diagram of the corresponding closed loop structure of the system that is going
to be analyzed:

w ~
u y
G(s)
–

f(.)

Fig. 7.14 Nonlinear closed loop system.

The bock diagram consists of a linear transfer function G(s) with arbitrary dy-
namics and a static nonlinearity f (.). The state space formulation of G(s) is as
follows:
146 7 Control of Haptic Systems

ẋ = Ax + Bũ
y = Cx

Thus we find for the closed loop system description:

ẋ = Ax − B f (y)
y = Cx.

In case that f (y) = k · y this nonlinear system is reduced to a linear system whose
stability can be examined with the evaluation of the system’s eigenvalues. With an
arbitrary nonlinear function f(y) the complexity of the problem is extended. So first
constraint to f(y) is that it exists only in a determined sector that is limited by a
straight line through the origin with a gradient k. Figure 7.15 shows an equivalent
example for the nonlinear function f (y). This constraint is depicted by the following
equation:

0 ≤ f (y) ≤ ky. (7.26)

The P OPOV criterion provides an intuitive handling for the stability analysis of the
presented example. The system is asymptotically idle state (ẋ = x = 0) stable if:
• the linear subsystem G(s) is asymptotically stable and fully controllable,
• the nonlinear function meets the presented sector condition as shown in fig-
ure 7.15,
• for an arbitrarily small number ρ ≥ 0 there exists a positive number α , so that
the following inequality is satisfied:
1
∀ω ≥ 0 Re[(1 + jαω )G( jω )] + ≥ρ (7.27)
k
Equation 7.27 formulates the condition also known as P OPOV inequality. With

G( jω ) = Re(G( jω )) + jIm(G( jω )) (7.28)

equation 7.27 leads to

1
Re(G( jω )) − αω Im(G( jω )) + ≥ρ (7.29)
k
With an additional definition of a related transfer function

G∗ = Re(G( jω )) + jω Im(G( jω )) (7.30)

equation 7.29 states that the plot in the complex plane of G∗ , the so called P OPOV
plot, has to be located in a sector with an upper limit described by y = α1 (x + 1k ).
Figure 7.16 shows an example for the P OPOV plot of a system in the complex plane
constrained by the sector condition.
7.3 System Stability 147

f(y)

1 y

Fig. 7.15 Sector condition.

The close relation to the N YQUIST criterion for the stability analysis of linear
systems becomes quite obvious here. While the N YQUIST criterion examines the

f(y)

1/a

-1/k
1 y

G*(jw)

Fig. 7.16 P OPOV plot.

148 7 Control of Haptic Systems

plot of G( jω ) referred to the critical point (-1|0), the location of the P OPOV plot is
checked for a sector condition defined by a straight line limit.
The application of the P OPOV criterion has the excelling advantage, that it is
possible to gain a result out of the stability analysis without an exact formulation of
the nonlinearity within the system. All constraints for the nonlinear subsystem are
restraint to the sector condition and the condition to have memoryless transfer be-
havior. The most complicated aspect within this kind of analysis is how to formulate
the considered system structure in a way, that the P OPOV criterion can be applied.
For completeness the circle criterion shall be mentioned whose sector condition is
not represented by a straight line, rather

f (y)
k1 ≤ ≤ k2 . (7.31)
y
defines the new sector condition. For additional explanation on these constraints and
the application of the circle criterion it is recommended to consider [265], [135] and
[234].
As second example for stability analysis of nonlinear systems the direct method
by LYAPUNOV is introduced. The basic principle is that if both linear and nonlinear
stable systems tend to a stable steady state, the complete system energy has to be
dissipated continuously. Thus it is possible to gain result from stability analysis
while verifying the characteristics of the function representing the state of energy
in the system. LYAPUNOV´s direct method generalizes this approach to evaluate the
system energy by the generation of an artificial scalar function which can describe
not only the energy stored within the considered dynamic system, further it is used
as an energy like function of a dissipative system. These kinds of functions are called
LYAPUNOV functions V (x).
For the examination of the system stability the already mentioned state space
description of a nonlinear system is used:

ẋ = f(x, u,t)
y = g(x, u,t).

By the definition of LYAPUNOV´s theorem the equilibrium at the phase plane origin
ẋ = x = 0 is globally, asymptotically stable if
1. a positive definite scalar function V (x) with x as the system state vector exists,
meaning that V (0) = 0 and V (x) > 0 ∀ x = 0,
2. V̇ is negative definite, meaning V̇ (x) ≤ 0,
3. V (x) is not limited, meaning V (x) → ∞ as x → ∞.
If these conditions are met in a bounded area at the origin only, the system is locally
asymptotically stable.
As a clarifying example the following nonlinear first order system

ẋ + f (x) = 0 (7.32)
7.3 System Stability 149

is evaluated. Herein f (x) denotes any continuous function of the same sign as its
scalar argument x so that x · f (x) > 0 and f (0) = 0. Applying this constraints a
LYAPUNOV function candidate can be found described by

V = x2 . (7.33)

The time derivative of V (x) provides

V̇ = 2xẋ = −2x f (x). (7.34)

Due to the assumed characteristics of f (x) all conditions of LYAPUNOV´s direct

method are satisfied thus the system has globally asymptotically stable equilibrium
at the origin. Although the exact function f (x) is not known, the fact that it exists
in the first and third quadrant only is sufficient for V̇ (x) to be negative definite. As
second example a multi-input multi-output system is examined depicted by its state
space formulation

ẋ1 = x2 − x1 (x12 + x22 )

ẋ2 = −x1 − x2 (x12 + x22 ).

In this example the system has an equilibrium at the origin too. Consequently the
following LYAPUNOV function candidate can be found

V (x1 , x2 ) = x12 + x22 . (7.35)

Thus the corresponding time derivative is

V̇ (x1 , x2 ) = 2x1 x˙1 + 2x2 x˙2 = −2(x12 + x22 )2 . (7.36)

Hence V (x1 , x2 ) is positive definite and V̇ (x1 , x2 ) is negative definite. Thus the equi-
librium at the origin is globally, asymptotically stable for the system.
A quite difficult aspect when using the LYAPUNOV´s direct method is given
by how to find LYAPUNOV function candidates. No straight algorithm with a de-
termined solution exists, which is a big disadvantage of this method. S LOTINE
[234] proposes several structured approaches to gain LYAPUNOV function candi-
dates namely
• K RASOVSKII´s method,
• the variable gradient method.
Besides these S LOTINE provides additional possibilities to involve the system´s
physical principles in the procedure for the determining of LYAPUNOV function
candidates while analyzing more complex nonlinear dynamic systems.
150 7 Control of Haptic Systems

7.3.3 Passivity in Dynamic Systems

As another method for the stability analysis of dynamic systems the passivity for-
malism is introduced within this subsection. Functions can be extended to system
combinations by using LYAPUNOV´s direct method, and evaluating the dissipation
of energy in dynamic systems. The passivity formalism also is based on nonlinear
positive definite storage functions V (x) with V (0) = 0 representing the overall sys-
tem energy. The time derivative of this energy determines the system´s passivity. As
example the general formulation of a system

ẋ = f(x, u,t)
y = g(x, u,t).

is considered. This system is passive concerning the external supply rate S = yT u if

the inequality condition
V̇ (x) ≤ yT u (7.37)
is satisfied. K HALIL distinguishes several cases of system passivity depending on
certain system characteristics (Lossless, Input Strictly Passive, Output Strictly Pas-
sive, State Strictly Passive, Strictly Passive) [135]. If a system is passive concerning
the external supply rate S, it is stable in the sense of LYAPUNOV.
The combination of passive systems using parallel or feedback structures inherits
the passivity from its passive subsystems. With the close relation of system passivity
to stability in the sense of LYAPUNOV, the examination of the system stability is
possible by verifying the subsystem’s passivity. Based on this evaluation it can be
concluded that the overall system is passive. Always with the assumption that a
correct system structure was built.
As an illustrating example the RLC circuit taken from [135] is anaylzed in the
following. The circuit structure is shown in figure 7.17:

uR2=f(iL)
L

iL

U(t) uC
C

i1=f(uR1) i3=f(uR3)

Fig. 7.17 Passivity analysis of an RLC-Network.

The system´s state vector is defined by

7.3 System Stability 151

iL = x1
uC = x2 .

The input u represents the supply voltage U, as output y the current i is observed.
The resistors are described by the corresponding voltage current characteristics:

i1 = f1 (uR1 )
i3 = f3 (uR3 )

For the resistor which is coupled in series with the inductor the following behavior
is assumed
UR2 = f2 (iL ) = f2 (x1 ). (7.38)
Thus the nonlinear system is described by the differential equation:

Lx˙1 = u − f2 (x1 ) − x2
Cx˙2 = x1 − f3 (x2 )
y = x1 + f1 (u)

The presented RLC circuit is passive as long as the condition

 t
V (x(t)) −V (x(0)) ≤ u(τ )y(τ )d τ (7.39)
0

is satisfied. In this example the energy stored in the system is described by the
storage function
1 1
V (x(t)) = Lx12 + Cx22 . (7.40)
2 2
Equation 7.39 leads to the condition for passivity:

V̇ (x(t), u(t)) ≤ u(t)y(t) (7.41)

which means, that the energy supplied to the system must be equal or higher than
the time derivative of the energy function. Using V (x) in the condition for passivity
provides

V̇ (x, u(t)) = Lx1 x˙1 +Cx2 ẋ2

   
= x1 u − f2 (x1 ) − x2 + x2 x1 − f3 (x2 )
 
= x1 u − f2 (x1 ) + x2 f3 (x2 )
 
= x1 + f1 (u) u − u f1 (u) − x1 f2 (x1 ) − x2 f3 (x2 )
= uy − u f1 (u) − x1 f2 (x1 ) − x2 f3 (x2 )

and finally
152 7 Control of Haptic Systems

u(t)y(t) = V̇ (x, u(t)) + u f1 (u) + x1 f2 (x1 ) + x2 f3 (x2 ). (7.42)

In case that f1 , f2 and f3 are passive subsystems, i.e. all functions describing the cor-
responding characteristics of the resistors exist only in the first and third quadrant,
so V̇ (x, u(t)) ≤ u(t)y(t) is true, hence the RLC circuit is passive. Any coupling of
this passive system to other passive systems in parallel or feedback structures again
results in a passive system. For any passivity analysis and stability evaluation this
method implements a structured procedure and shows a very high flexibility.
In conclusion it is necessary to mention, that all methods for stability analysis
introduced in this section show certain advantages and disadvantages concerning
their applicability, information value and complexity, regardless whether linear or
nonlinear systems are considered. When a stability analysis is expected to be done,
the applicability of a specific method should be checked individually. This section
only can give a short overview on the introduced methods and techniques, and does
explicitly not claim to be a detailed description due to the limited scope of this
section. For any further study the reader is invited to consult the proposed literature.

7.4 Control Design for Haptic Systems

As introduced in the beginning of this chapter, control design is a fundamental and

necessary aspect within the development of haptic systems. Besides the techniques
for system description and stability analysis the need for control design and the ap-
plicable design rules become obvious. Especially for the control design of a haptic
system it is necessary to deal with several aspects and conditions to be satisfied dur-
ing the design process. The control design of the haptic device has the focus on the
analysis of the control loop closed by the user. In a system for telemanipulation this
focus must be extended to the manipulated process or object. In latter case the efforts
to be taken are doubled and the aspects to be considererd become more complex.
Especially the mechanical coupling of human, may it be on the user’s site or on the
process’ site or both, demands the use of control structures to establish the stability
and performance of the overall system continuously and sufficiently. The following
sections present several control structures and design schemes in order to set up a
basic knowledge about the toolbox for analytic control design of haptic systems.
This also involves some of the already introduced methods for system formulation
and stability analysis, as these form the basis for most control design methods.

7.4.1 Structuring of the Control Design

As introduced in chapter 5 various different structures of haptic systems exist, whose

characteristics will be repeated in this subsection:
7.4 Control Design for Haptic Systems 153

• Open-loop impedance controlled: The user experiences an impression of force

which is directly commanded via an open loop based only on a demand value. In
chapter 5 the basic scheme of this structure is shown by figure 5.1.
• Closed-loop impedance controlled: As it can be seen in figure 5.3 , the user also
experiences an impression of force which is fed back to a controller. Here a
specific control design will be needed.
• Open-loop admittance controlled: In this scheme on comparison to the previous
methods the user erxperiences an impression of a defined position. In the open
loop arrangement this position again is directly commanded based only on a de-
mand value. Figure 5.5 shows the corresponding structure of this haptic scheme.
• Closed-loop admittance controlled: This last version as it is depicted in figure 5.7
shows its significant difference in the feedback of the force the user applies to the
interface. This force is fed back to a demand value. This results in a closed loop
arrangement that incorporates the user and his or her transfer characteristics. In
difference to the closed loop impedance controlled scheme this structure uses
a force as demanded value SF compared with the detected SS , but the system
output is still a position xout . This results in the fact that the incorporation of the
user into the closed loop behavior is more complex than it is in a closed loop
impedance controlled scheme.
All of these methods can be basically structured as it is depicted in figure 5.11.
In this structure all necessary control loops of the overall telemanipulation system
become evident.
• On the haptic display and interface site a control loop is closed incorporating
the user which is valid as long as the user´s reaction is fed back to the central
interface module for any further data processing or control.
• On the process site also, a closed loop exists if measurable process signals (reac-
tions, disturbances) are fed back to the central interface module for data process-
ing or control.
• Underneath these top-level control loops various subsystem control loop exists
which have a major impact on the overall system too. As an example, each elec-
trical actuator will most likely be embedded in a cascaded control structure with
current, speed and position control.
It becomes obvious that the design of a control system for a telemanipulation
system with a haptic interface is complex and versatile. Consequently a generally
valid procedure for control design cannot be given. The control structures must be
designed step by step involving the following controller:
1. design of all controllers for the subsystem actuators,
2. design of top level controller for the haptic display and interface,
3. design of top level controller for the process or object manipulator.
This strict separation proposed above might not be the only way of structuring
the overall system. Depending on the application and functionality, the purposes of
the different controller and control levels might be in conflict to each other or simply
154 7 Control of Haptic Systems

overlap. Therefore it is recommended to set up the underlying system structure and

define all applied control schemes corresponding to their required functionality. As
examining example the following control structure shall be looked at: The altitude
control of a passenger airplane takes into account several process variables e.g. the
actual altitude, actual speed, fuel consumption, amount of freight, weather, air traf-
fic control clearances. On this level the controller generates demand values for the
control loops underneath that are responsible for all stability issues and maneuvers
of the airplane. These control loops affect the actuation systems that will move the
necessary elevator, ailerons, rudders, thrust control just to name few.
While looking at the control of haptic systems, a similar structure can be es-
tablished. For both the control of the process manipulation and the haptic display
or interface the central interface module will have to generate demand values for
force or position, that are going to be followed by the controllers underneath. These
demand values derive from a calculation predefined by designed control laws. To
gain such control laws a variety of methods and techniques for structural design and
optimization can be applied depending on certain requirements. The following sub-
sections give an overview of typical requirements to closed control loop behavior
followed by examples for control design.

7.4.2 Requirement Definition

Besides the fundamental need for system stability with sufficient stability margins
additional requirements can be set up to achieve a certain system behavior in a
closed loop scheme such as dynamic or precision. A quantitative representation of
these requirements can be made by the achievement of certain characteristics of the
closed loop step response.
Figure 7.18 shows the general form of a typical closed loop step response and its
main characteristics: As it can be seen the demanded value is reached and the basic
control requirement is satisfied. Additional characteristics are discussed and listed
as follows:
xd, max - maximum overshoot
Tmax - point of time for xdmax
Tε - time frame in which the residuum to the demanded value
remains within a predefined scope ε
Tan - point of time when the demanded value is reached for the first time
For all mentioned characteristics a quantitative definition of certain requirements
is possible. For example the number and amplitude of overshoots shall not extend a
defined limit or have a certain frequency spectrum that is of special interest for the
control design in haptic systems. As it is analyzed in sections 4.2 and 4.3, the user´s
impedance shows a significant frequency range which must not be excited within the
control loop of the haptic device. Nevertheless a certain cut-off frequency has to be
reached to establish a good performance of the dynamic behavior. All these issues
7.4 Control Design for Haptic Systems 155

maximum overshoot xd, max

e
reference w(t)
output y(t)

Tres Tmax Te e.g. 95% time t

Fig. 7.18 Closed loop step response requirements .

are valid for the requirements to the control design of the process manipulation. In
addition to the requirements from the step response due to changes of the setpoint, it
is necessary to formulate requirements concerning the closed loop system behavior
considering disturbances originating from the process. Especially when interpreting
the user´s reaction as disturbance within the overall system description a require-
ment set up for the disturbance reaction of the control loop has to be established.
As it can be seen in figure 7.19 similar characteristics exist to determine the distur-
bance reaction quantitatively and qualitatively. In most cases both the step response
behavior and the disturbance reaction cannot satisfy all requirements, as they often
come into conflict with each other which is caused by the limited flexibility of the
applied optimization method. Thus it is recommended to estimate the relevance of
step response and disturbance reaction in order to choose an optimization approach
that is most beneficial. Although determined quantitatively, it is not possible to use
all requirements in a predefined optimization method. In most cases an adjustment
of requirements is necessary to be made, to apply specific control design and op-
timization methods. As an example the time TAn as depicted above cannot be used
directly, and must be transferred into a requirement for the closed loop dynamic
characterized by a definite pole placement.
Furthermore simulation techniques and tests offer iteration within the design pro-
cedure to gain an optimal control law. However, this very sufficient way of analyzing
system behavior and test designed control laws suggests to forget about the analytic
system and control design strategy and switch to a trial an error algorithm.
156 7 Control of Haptic Systems

maximum overshoot xd, max

deviation
xd(t) e
0

disturbance z(t)

TcompTmax Te e.g. 5% time t

Fig. 7.19 Closed loop disturbance response requirements.

7.4.3 Control Law Design

This section shall present some possible types of controllers and control structures
that might be used in the already discussed control schemes. For optimization of the
control parameters several methods exists. They are introduced here. Depending on
the underlying system description several approaches to set up controllers and con-
trol structures are possible. This section will present the classic PID-control, addi-
tional control structures e.g. compensation, state feedback controllers, and observer
based state space control.

7.4.3.1 Classic PID-Control

Maybe one of the most frequently used controllers is the parallel combination of
a proportional (P), an integrating (I) and a derivative (D) controller. This combi-
nation is used in several variants including a P-controller, a PI combination, a PD
combination or the complete PID structure. Using the PID structure all advantages
of the individual components are combined. The corresponding controller transfer
function is described by
 1 
GR = KR 1 + + TV s . (7.43)
TN s
Figure 7.20 shows the equivalent block diagram of a PID controller structure. Ad-
justable parameters in this controller are the proportional gain KR , the integrator
time constant TN and the derivative time TD .
7.4 Control Design for Haptic Systems 157

xd 1 u
KR ¾¾
Tns

Tvs

Fig. 7.20 PID block diagram.

With optimized parameter adjustment a wide variety of control tasks can be han-
dled. This configuration offers on the one hand the high dynamic of the proportional
controller and on the other, the integrating component guarantees a high precision
step response with a residuum xd = 0 for t → ∞. The derivative finally provides an
additional degree of freedom that can be used for a certain pole placement of the
closed loop system.
As major design techniques the following examples shall be introduced:
• Root locus method: This method has its strength by the determined pole place-
ment for the closed loop system, directly taken into account the dependence on
the proportional gain KR . By a reasonable choice of TN and TD the additional
system zeros are influenced which affects directly the resulting shape of the root
locus and thus the stability behavior. Besides this the overall system dynamic can
be designed.
• The second method for the optimization of the closed loop system step response
or disturbance reaction by the minimization of an integral criterion is possible.
Minimizing the underlying integral criterion the controller parameters are ad-
justed to optimized values. The basic procedure for this method is as the follow-
ing: The tracking error xd due to changes of the demanded set point or a process
disturbance is integrated (eventually weighted over time). This time integral will
have to be minimized in dependence on all adjustable parameters. In case of
convergence of this minimization, the result is the set of optimized controller
parameters.
For any additional theoretical background concerning controller optimization the
reader is invited to consult the literature on control theory and control design [165],
[164].

7.4.3.2 Additional Control Structures

In addition to the described PID controller additional control structures extend the
influence on the control result without having an impact on the system stability.
158 7 Control of Haptic Systems

The following paragraphs therefore shall present a disturbance compensation and a

direct feedforward of auxiliary process variables.

Disturbance Compensation

The basic principle of disturbance compensation assumes that if a disturbance on

the process is measurable and its influence is known, this knowledge can be used
to establish compensation by corresponding evaluation and processing. Figure 7.21
shows a simplified scheme of this additional control structure.

measrurable
disturbance z

Disturbance transfer
GD function

compensator GC

GR GS

Fig. 7.21 Simplified disturbance compensation.

In this scheme a disturbance signal is assumed to affect the closed loop via a dis-
turbance z transfer function GSZ . Measuring the disturbance signal and processing
the compensator transfer function GC results in a compensation of the disturbance
interference. Assuming an optimal design of the compensator transfer function this
interference caused by the disturbance is completely erased. The optimal design of
a corresponding compensator transfer function is depicted by
−GSZ
GC = . (7.44)
GS

This method assumes that a mathematical and practicable inversion of GSZ exists.
For those cases where this assumption is not valid, the optimal compensator GC
must be approximated. Furthermore Figure 7.21 states clearly, that this additional
control structure does not have any influence on the closed loop system stability
and can be designed independently. Besides the practicability the additional effort
should be taken into account. This effort will definitively increase just by the sensors
7.4 Control Design for Haptic Systems 159

to measure the disturbance signals and by the additional costs for realization of the
compensator.

Auxiliary Input Feedforward

A similar structure compared to the disturbance compensation is the feedforward

of auxiliary input variables. This principle is based on the knowledge of additional
process variables that are used to influence the closed loop system behavior without
affecting the system stability. Figure 7.22 shows an example of the feedforward of
the demanded setpoint w to the controller signal u using a feedforward filter function
GFF .

Feedforward filter
GFF

w u y
GR GS

Fig. 7.22 Feedforward of auxiliary input variables.

7.4.3.3 State Space Control

Corresponding to the techniques for the description of multi-input multi-output sys-

tems discussed prior in this chapter, the state space control provides additional fea-
tures to cover the special characteristics within those systems. As described in sec-
tion 7.2.2 multi-input multi-output systems are preferably depicted as state space
models. Using this mathematical formulation enables the developer to implement
a control structure that controls the internal system states to demanded values. A
big advantage is that the design methods for state space control use an overall ap-
proach for control design and optimization instead of a control design step by step
for each system state. With this approach it becomes possible to deal with pro-
foundly coupled multi-input multi-output systems with high complexity, and design
a state space controller simultaneously. This section will present the fundamental
state space control structures. This will cover the state feedback control as well as
160 7 Control of Haptic Systems

the observer based state space control. For further detailed procedures as well as
design and optimization methods the reader is referred to [164] and [265].

State Feedback Control

As it is shown in figure 7.23 this basic structure for state space control uses a feed-
back of the system states x.

x0 (t)

w (t) u (t) x× (t) x (t) y (t)

V B ò C

Fig. 7.23 State feedback control.

Similar to the depiction in figure 7.8 the considered system is presented in state
space description using the matrices A, B, C and D. The system states x are fed back
gained by the matrix K to the vector of the demanded values that were filtered by
matrix V. The results represent the system input vector u. Both matrices V and K do
not have to be square matrices for a state space description is allowed to implement
various dimensions for the state vector, the vector for the demanded values and the
system input vector.

Observer Based State Space Control

The state space control structure discussed above requires a complete knowledge
of all system states, which is nothing else but that they have to be measured and
processed to be used in the control algorithm. From a practical point of view this
is not possible all the time due to technical limits as well as costs and effort. As
a result the developer is faced with the challenge to establish a state space control
without the complete knowledge of the system states. As a solution those system
7.4 Control Design for Haptic Systems 161

x0 (t)

w (t) u (t) x× (t) x (t) y (t)

V B ò C

L
-

D*

x× (t)* x (t)*
B* ò C*
y (t)*

A*

Fig. 7.24 Observer based state space control.

states that cannot be measured due to technical difficulties or significant cost factors
are estimated using a state space observer structure that is shown in figure 7.24.
In this structure a system model is calculated in parallel to the real system. As
exact as possible this system model is described by the corresponding parameter
matrices A∗ , B∗ , C∗ and D∗ . The model input also is represented by the input vec-
tor u. Thus the model provides an estimation of the real system states x∗ and an
estimated system output vector y∗ . By comparison of this estimated output vector
y∗ with the real output y, which is assumed to be measurable, the estimation error
is fed back gained by the matrix L. This results in a correction of the system state
estimation x∗ . Any estimation error in the system states or the output vector due to
varying initial states is corrected and the estimated states x∗ are used to be gained
by the equivalent matrix K and fed back for control.
This structure of an observer based state space control uses the L UENBERGER
observer. In this configuration all real systems states are assumed not to be measur-
162 7 Control of Haptic Systems

able thus the state space control refers to estimated values completely. Practically,
the feedback of measurable system states is combined with the observer based esti-
mation of additional system states. In [164] und [265] examples for observer based
state space control structures as well as methods for observer design are discussed
in more detail.
7.5 Conclusion 163

7.5 Conclusion

The control design for haptic devices faces the developing engineer with a complex
manifold challenge. According to the fundamental requirement, to establish a safe
reliable and determined influence on all structures, subsystems, or processes the
haptic system is composed of, an analytical approach for control system design is
not negligible anymore. It provides a wide variety of methods and techniques to
be able to cover many issues that arise during this design process. This chapter
intends to introduce the fundamental theoretical background. It shows several tasks,
functions and aspects the developer will have to focus on, as well as certain methods
and techniques that are going to be useful tools for the system’s analysis and the
process of control design.
Starting with an abstracted view on the overall system, the control design process
is based on an investigation and mathematical formulation of the system’s behavior.
To achieve this a wide variety of methods exists, that can be used for system descrip-
tion depending on the degree of complexity. Besides methods for the description of
linear or linearized systems, this chapter introduced techniques for system descrip-
tion to represent nonlinear system behavior. Furthermore the analysis of multi-input
multi-output systems is based on the state space description, which is presented here
too. All of these techniques on the one hand are aimed at the mathematical repre-
sentation of the analyzed systems as exact as possible, on the other hand they need
to satisfy the requirement for a system description that further control design pro-
cedures are applicable to. These two requirements will lead to a trade off between
establishing an exact system formulation that can be used in analysis and control
design procedures without extending the necessary effort unreasonably.
Within system analysis the overall system stability is the most important aspect
that has to be guaranteed and proven to be robust against model uncertainties. The
compendium of methods for stability analysis contains techniques that are applica-
ble to linear or nonlinear system behavior, corresponding to their underlying princi-
ples that of course limit the usability. The more complex the mathematical formu-
lation of the system becomes, the higher the effort gets for system analysis. This
comes in direct conflict to the fact that a stability analysis of a system with a sim-
plified system description can only provide a proof of stability for this simplified
model of the real system. Therefore the impact of all simplifying assumptions must
be evaluated to guarantee the robustness of the system stability.
The actual objective within establishing a control scheme for haptic systems is
the final design of controller and control structures that have to be implemented
in the system in various levels to perform various functions. Besides the design of
applicable controllers or control structures the optimization of adjustable parame-
ters is also part of this design process As shown in many examples in the literature
on control design a comprehensive collection of control design techniques and opti-
mization methods exists, that enable the developer to cover the emerging challenges,
and satisfy various requirements within the development of haptic systems as far as
automatic control is concerned.
“This page left intentionally blank.”
Chapter 8
Kinematic Design

A NDREAS RÖSE

This chapter addresses the basics of the kinematic design of haptic devices. Only
few haptic devices generate forces in just one dimension. In such a case the kine-
matic structure of the device is simple. It is much more challenging and interesting
to measure displacements and angles and generate forces and torques in two, three
or more degrees of freedom (DOF). For the transformation of forces in numerous
spatial directions mechanisms are necessary, which are able to transmit forces from
the actuators to the point of action, e.g. the hand. In theory and with some exam-
ples two basically different kinematic structures are discussed in this chapter: serial
kinematic and parallel kinematic ones.
The chapter focusses on the kinematic design process necessary for haptic applica-
tions. It does not cover a universal approach for any form of mechanism. Especially
advanced considerations of static and dynamic analysis should be taken from litera-
ture recommended within this chapter. Although it excludes advanced optimisation
techniques it is still an introduction to the choice and design of kinematic structures
for haptic systems.

8.1 Basics

Before taking a look at specific versions of kinematic designs some general termi-
nology and basic mathematical tools have to be understood. This section introduces
this knowledge to the reader, starting with simple examples and extending them to
a formal analysis of kinematic structures.

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_8, 165

c Springer-Verlag Berlin Heidelberg 2009
166 8 Kinematic Design

8.1.1 Mechanisms and their Classification

Kinematics - the tenet of movement - deals with the mathematical calculation of

movement patterns in mechanical systems. In contrast to kinetics, inner and outer
forces of the system are neglected. It is a pure concentration on the movement it-
self. The results of a kinematic observation are ideal transfer characteristics of the
observed, massless system.
In general a kinematic structure capable to generate or transfer spatial movements
is required. Thus this section deals with the design and kinematic description of -
nonlinear transferring - mechanisms with multiple degrees of freedom.
Usually the design process starts with the definition of a topology for the mecha-
nism. Afterwards its transfer characteristics are calculated and - if required - ad-
ditionally analyzed and optimized. The term “topology” refers to the placement
of joints and rigid bodies within the mechanism. Afterwards the geometrical di-
mensioning follows with respect to specific requirements such as reaching a certain
workspace.
A driven mechanism - a robot or robot arm - is made of driven joints - motors,
actuators, drives - and un-driven joints - passive joints -. A driven mechanism con-
tains as many actuators, as it has degrees of freedom to move. Special realisations
include more actuators than DOFs, what increases the structure’s stiffness. Such
designs are not typical for haptic devices and will not be part of this chapter’s fo-
cus. Two basic mechanism configurations are distinguished: parallel kinematic and
serial-kinematic. A combination of both configurations is called hybrid-kinematic.
Figure 8.1 gives examples for all three mechanisms. Purely serial mechanisms in-
clude no passive joints; all actuators are in serial order within one single kinematic
chain. Parallel mechanisms give the possibility to place all actuators at the frame,
minimizing the moving masses. The small inertia makes them highly relevant for
haptic applications. The “frame” in the context of kinematic design is a rigid, non-
moving body.

Examples

There is a large number of different robots, positioning systems and mechanisms.

They all have in common that they are composed of drives and rigid bodies. In some
robots the driving power is transmitted via tendons, such as in some haptic devices
(fig. 8.2). As the process of dimensioning is identical to all mechanisms two exam-
ples are given here, although they do not have an explicit haptic background.
Figure 8.3 shows a fully parallel mechanism with six legs called hexapod by the
company Physik Instrumente (PI) GmbH & Co. KG. Within literature this structure
is often referred to as S TEWARD-G OUGH-platform, as it has been invented by D.
S TEWARD and V.E. G OUGH and also by K.L. C APPEL independently from each
other at almost the same time. The tool-center-point (TCP) can be moved in six de-
8.1 Basics 167

Tool Center Point (TCP)

Actuated joint (actuator)
Passive joint

a) b) c)
Fig. 8.1 Serial a), parallel b) und hybrid c) mechanism. In the case of a parallel kinematic mecha-
nism the actuators can be placed fixed to the frame.

Parallel kinematic
TCP mechanism

Base

Tackles to drive the

mechanism

Fig. 8.2 Haptic control element Falcon by the company Novint with parallel kinematic structure.
The drives are coupled to the mechanism via tendons. source: Novint Technologies, Inc.

grees of freedom. The movement is generated by six linear actuators connecting the
frame to the TCP. Such a structure is easily applicable to haptic devices, as its kine-
matic relations have been solved analytically [103]. The presented hexapod by the
company PI however is a positioning system without the dynamics being necessary
for haptic applications.
Figure 8.4 shows a SCARA-robot (Selective Compliance Assembly Robot Arm) by
the company Mitsubishi Electric. This kind of robot is used frequently within as-
sembly automation for pick and place tasks. As it is purely serial each actuator has
to carry and move all following actuator’s weight.
168 8 Kinematic Design

Tool Center Point (TCP)

Linear actuators

Base

Fig. 8.3 Parallel kinematic hexapod by Physik Instrumente (PI) GmbH & Co. KG. The mechanism
contains six DOF, which are driven by liear actuators. source: company PI (www.pi.ws)

Rotatory drives

Base

Tool Center Point (TCP)

Fig. 8.4 SCARA robot by the company Mitsubishi Electric for use as a pick and place robot.
source: Mitsubishi Electric

8.1.2 Calculating Kinematics

The kinematic description of a mechanism for haptic applications can be reduced to

the description of in- and output values like force, torques, translations and angles
at the handle (tool-center-point, TCP) and the actuators.
8.1 Basics 169

May a = (a1 , a2 , ..., an ) be the vector combining the actuator displacements and
angles ai , and x = (x1 , x2 , ..., xm ) the vector combining the displacements and angles
of the handle x j . The direct kinematic problem (also: direct kinematics or forward
kinematics) can be formulated as

x = f (a) (8.1)

and its inverse - the inverse kinematic problem (also: reverse kinematics or inverse
kinematics) - to

a = f −1 (x) (8.2)

The forward kinematics according to equation 8.1 is necessary for most haptic ap-
plications. It contains the relation for the identification of movements. The hand
moves the TCP (x-vector) directly. The actuators are moved at the same time via
the transmission of the mechanism. For input purposes the position of the TCP shall
be identified. If the position of the actuators is measured via sensors (e.g. rotary
encoders) the forward kinematics can be used to calculate the position of the hand.
Besides position measurement the force output is also based on the forward kine-
matics too. The calculation of forces can be done by the aid of the Jacobian-matrix
(chapter 8.1.3).
For serial mechanisms the derivation of the forward kinematics according to equa-
tion 8.1 can quite easily be achieved with the method of D ENAVIT-H ARTENBERG-
parameters (section 8.2.2). For parallel mechanisms the inverse kinematics can
frequently be solved by vector analysis, the forward kinematics however is usu-
ally challenging to be solved analytically. Luckily for a number of parallel kine-
matic mehanisms the forward kinematics is already solved within common litera-
ture [257].
Annotation:
The before mentioned is relevant for the largest class of haptic devices, which are
open- or closed-loop impedance controlled. For open- or closed loop admittance
controlled devices the inverse kinematic problem plays an important role. In these
cases a change in force is generated by a position change of the TCP relative to the
hand. The calculation of in actor positions to achieve a desired change in the posi-
tion is given by the inverse kinematics.

Example

The Example in figure 8.5 is a Delta-robot. It has a parallel kinematic structure (sec-
tion 8.3). The Delta Robot has a total number of three DOFs, being able to move
in three independent Cartesian coordinates (movements along orthogonal axis). The
kinematic problem of Delta-robots is solved and available within literature [257]. It
is actuated by three rotatory actuators. When using a Delta-robot for haptic applica-
170 8 Kinematic Design

tions the TCP can be used similar to a joystick for the control of a telemanipulator
or simulation. Being used as an impedance controlled parallel kinematic device the
forward kinematic solution (equ. 8.1) is necessary to do both, the calculation of the
position of the Hand based on angular measurement at the actuators, and actuating
the drives to achieve the desired force at the TCP.

Rotatory drives

Bars and passive joints

Tool Center Point (TCP)

Fig. 8.5 The Delta-Robot has 3 rotational actuators and 3 DOF at its TCP which are best described
in catesian coordinates. For a haptic device the TCP would serve as handle. (source: company
ABB)

8.1.3 Transfer Characteristics and Jacobian Matrix

To understand this section it is necessary to be aware of the fact that we are dealing
with nonlinear transmitting gears with multiple degrees of freedom. Their trans-
fer characteristics - the kinematic dependencies - are impossible to be solved by a
system of linear equations. The transmission of displacements and forces changes
depending on the position in the workspace. Of course for a haptic device this non-
linear dependency has to be considered and implemented within the control.
When moving a mechanism just a little around an operating point the transfer char-
acteristics can be linearized. The gear in figure 8.6a has a nonlinear transmission
function. The input is a rotational drive with angle α and the the output is the dis-
placement x of the driving rod is the output. The forward kinematics is calculated by

x = a · sin(α ) (8.3)

with its functional plot is given in figure 8.6b.

8.1 Basics 171

Ma
Fx

a) b)
Fig. 8.6 Nonlinear transmitting gear with one degree of freedom. a) functional sceme, b) forward
kinematics

Linearizing the function at an arbitrary point α0

dx
dx|α0 = |α · d α = cos(α0 ) · d α = b0 · d α (8.4)
dα 0
gives the transmission ratio b0 for differential movements.
The derivation b0 at α0 contains the displacement transmission ratio at the input
value α0 as well as the transmission ratio of the forces. Using the principle of virtual
work δ W x at the drive and δ Wα at the transmission bar is calculated:

δ Wx = δ Wα
Fα · δ α = Fx · δ x
Fα δx
=
Fx δα
Mα · α δx
= (8.5)
Fx δα
Thus the force-torque transmission ratio at the position α0 is given by
Mα · α 1
= (8.6)
Fx b0
which is the reciprocal value of the displacement transfer characteristics. Mα equals
the input torque in axial direction of the actuator and Fx the output force in direction
x.
Remark:
Virtual values are labelled with a δ , to distinguish them from differential values
(d). The principle of virtual work consideres the mechanism in one single position.
An infinitesimal small displacement (just some virtual movement) is assumed. By
this approach the work is independent from the actual displacement and can be ex-
pressed as a force transmission. But even in a differential analysis an infinitesimal
172 8 Kinematic Design

small displacement happens. Consequently the derivation of the displacement is cal-

culated. A singularity is reached when b0 = 0 (e.g. when α = 90◦ ). In this position
the gear cannot generate defined forces in x-direction anymore which makes it use-
less for a haptic applications in this point.

Dealing with Multidimensional Problems

In general problems with multiple degrees of freedom have to be analyzed (Rn →

Rm ). Identical to the one-dimensional situation in these applications a linearization
is performed too. The linearization is expressed by the Jacobian matrix J. Con-
sidering n degrees of freedom at the actuators a = (a1 , a2 , ..., an ) and m degrees of
freedom at the tool center points x = (x1 , x2 , ..., xm ) the Jacobian matrix is calculated
by
⎛ ⎞
∂ x1 /∂ a1 · · · ∂ x1 /∂ an
∂x ⎜ .. .. .. ⎟
J= =⎝ . . . ⎠ (8.7)
∂a
∂ xm /∂ a1 · · · ∂ xm /∂ an

The matrix contains all transfer information of the forward kinematics. The inverse
Jacobian matrix J−1 includes the same information for the inverse kinematics. The
differential displacement behaviour can be expressed by the aid of the Jacobian
matrixes as

dx = J · da
da = J−1 · dx (8.8)

The result is a system of linear equations with coefficient matrix J. When the number
of actuators is equal to the number of DOF of the TCP, the system of equations has a
square coefficient matrix. From linear algebra it is known that a system of equations
with a square coefficient matrix contains an unlimited number of solutions if the
determinant of the matrix equals 0. Consequently one actuator DOF respectively
one DOF of the TCP can be chosen freely if either one of the conditions

det(J) = 0

respectively

det(J−1 ) = 0 (8.9)

is fulfilled. In these cases a defined force transmission is no longer possible. In case

of a haptic device this is identical to a malfunction of the unit. The positions in
which the determinants of J and J−1 equals zero are called singular positions. The
transfer functions have a singularity in this position (in contrast to all other positions
8.1 Basics 173

where the functions are regular, i.e. clearly defined). For the dimensioning of a hap-
tic device the workspace has to be chosen such that all singularities are excluded.
This requirement is important for the design of haptic devices, especially for parallel
mechanism (section 8.3) what makes their design even more challenging compared
to serial mechanisms.

Remark on singularities: The specific situation with a non-square Jacobian matrix

is always the case when the TCP is characterized in more coordinates than number
of drives exist. If there is a mechanism with four actuators controlling four DOFs
of the TCP, it is usually helpful to describe the TCPs movement in six coordinates.
These are three Cartesian directions of movement and three Euler-angles. The TCP
will then move simultaneously in more than one of these coordinates. However the
coordinates will not become independently controllable. In this case we achieve a
6 × 4 Jacobian matrix. The preconditions for a singularity in these situations are
given by the rank of the Jacobian Matrix [30]. Is the rank of either J or J−1 less than
4, the position is singular.

For the derivation of force transfer characteristics a similar approach as in equa-

tion 8.5 can be used for multidimensional positioning systems [257]. The vector
fa = ( fa1 , fa2 , . . . , fan )T is the vector of all forces and/or torques fai on the driven
joints. The vector fx = ( fx1 , fx2 , . . . , fxm )T is the vector of all forces and torques fx j
acting on the tool-center-point. The virtual displacement of the actuators is labelled
δ a = (δ a1 , δ a2 , . . . , δ an ), those of the TCP are named δ x = (δ x1 , δ x2 , . . . , δ xm ).
The indices are identical to those in equation 8.7. Calculating the virtual work

δ Wx = δ Wa
fTx · δ x = fTa · δ a

with the relation dx = J · da from equation 8.8 yields

fTx · J · δ a = fTa · δ a

The concept of virtual work is applicable to any non-singular configuration for any
virtual displacement δ a. Consequently

fTx · J = fTa

is valid. By transposing the equation the final solution is

fa = JT · fx (8.10)

giving the relation between forces and torques on the actuators and the tool-center-
point. The same approach generates the inverted equation

fx = J−T · fa (8.11)
174 8 Kinematic Design

Remarks:

• Equation 8.10 is of highest interest for the control of haptic devices. It explicitly
contains the input-output-relation of forces. When a specific force/torque con-
dition shall be applied at the TCP this equation calculates the necessary control
values for the actuators.
• Equation 8.10 and 8.11 are calculated from a purely kinematic analysis and ne-
glect dynamic forces of inertia or friction and static forces of e.g. gravity. For
haptic devices with comparably heavy mechanics these forces have to be consid-
ered [257]. This is always the case when the forces from accelerated masses or
gravity get into ranges which cannot be neglected anymore. Example: If a mass
of m = 5 g is accelerated linear in t = 0.1 s over a distance of s = 1 mm a force of
F = m · s/t 2 = 0.5 N can be reached easily.

8.1.4 Optimizing the Transfer Characteristics

A specific topology consisting of actuators, rods and passive joints within a mul-
tidimensional transmitting gear shows a nonlinear transfer characteristic, which is
influenced mainly by the aspect ratios of its components. According to M ERLET
parallel mechanisms with well designed dimensions can perform better than one
with better suited topology but worse dimensions [174]. Optimization in this con-
text is defined as the change of specific geometrical dimensions within the same
mechanism to achieve a specific performance regarded as optimum. This may be
a maximized workspace with homogeneous transfer characteristics of forces from
the TCP to the actuators. A systematic and frequently numerical optimization is
applied mainly on parallel kinematic mechanisms. In direct comparison serial kine-
matic displays can be designed according to geometrical requirements quite easily.
Additionally singularities of serial kinematic chains are obvious. They appear when
the kinematic chains are stretched. Nevertheless the transfer characteristics of serial
kinematics can still be altered and optimized by changing the length of rods.
To perform an optimization the following steps should be taken:
1. Definition of parameters and their span of values (e.g. rod lengths)
2. Analytical or numerical description of the optimization problem
3. Mathematical optimization, e.g. via a gradient approach or evolutionary algo-
rithms
These steps are discussed in [185] and covered in detail within [138]. The key chal-
lenge is the formal description of the optimum. In [24] several optimizations are
given using the singular values of the Jacobian matrix as a measure. Within theses
approaches the measurement value for an optimum has to be determined by scan-
ning the complete workspace and optimizing relevant parameters of the mechanism
between each scanning process. Usually this approach needs a significant amount
8.2 Serial Mechanisms 175

of time. In practical work it is highly recommend to intensively study literature for

possible analysis and optimization of the chosen mechanism, as an own optimiza-
tion will mostly need unreasonable efforts and time. The optimization of complex
mechanisms with multidimensional optimization and evolutionary algorithm is still
- although often discussed - subject of research work and not necessarily successful
in every case.

8.2 Serial Mechanisms

This section describes the approach for a design of a serial kinematic structure for
haptic devices. The design is structured in three steps:
• Topological synthesis - the design of an assembly of rods and actuators
• Calculation of the kinematics - Mathematical description of displacements pro-
vided by the user and the output of forces
• Dimensioning - Defining the geometrical dimensions (section 8.1.4)

8.2.1 Topological Synthesis

A serial mechanism is not less or more than a sequence of rods and actuators,
whereas the actuators can be regarded as driven joints. Whether the actuators are
linear or rotary is of no importance for the complexity of the kinematic problem. For
the workspace and the orientation of the tool-center-point however this aspect is of
highest importance. A spacial serial mechanism with three rotatory drives changes
the orientation of its TCP all over its workspace. If it is not intended to generate
a torque as output to the user, the handle attached to this serial mechanism has to
be equipped with a passive universal joint. Such a realisation as haptic device can
be found in figure 8.7. Torques are decoupled from the hand. The handle does not
have to be placed exactly in the TCP, as the moments are eliminated by the passive
joints. Force vectors can be displaced arbitrarily within space. As a result the Hand
experiences the same forces like the TCP.

When dimensioning a serial haptic device several aspects should be considered dur-
ing the selection of joints:

• Linear joints tend to cant with loads orthogonal to the direction of movement.
This requires either a design minimizing orthogonal forces on linear joints, or
these joints must be designed adequately to work under orthogonal load.
In an idle situation and in the case of impedance-controlled devices actuated
joints are moved passively by the user. This subordinates them to the same re-
quirements as passive linear joints.
176 8 Kinematic Design

Handle

Passive joints to
decouple torques
from the hand
TCP

Actuated joints

Base

Fig. 8.7 The PHANTOM Omni haptic device by SensAble Technologies is an example for a spa-
cial working serial kinematic haptic device. The hand is decoupled from rotational movements by
passive joints. Thus no torques are induced to the hand.

• Rotatory joints are not able to transmit forces, when their kinematic chain is in a
stretched position. Within the example in figure 8.6 this situation happens for α =
π /2. The passive joint at the outer area of the disk reaches a stretched position.
As obvious from equation 8.5 no defined haptic feedback can be generated in this
position. The same applies to the forces in radial direction to the prismatic joint.
• A general recommendation for impedance-controlled devices is the minimization
of friction within the joints. One option to achieve this is the use of ball bearings
for rotational joints. When used for linear joints the roll motion is unsteady and
therefore noticeable.
As human beings are equipped with many serial kinematic chains (e.g. arms, legs)
the working area of a serial kinematic chain can be understood intuitively. This
makes it simple to design a corresponding haptic control unit. This is however not
the only criterion and will be further addressed in section 8.4.
The design can be done geometrically “with circle and ruler”, however the following
should be considered:
• Actuators add inertia and moving masses to the mechanism. In serial mechanisms
all actuators are placed in series. This has negative influence on the dynamics of
force transmission. Approaches for a dynamic analysis can be found in [257] and
[102].
8.2 Serial Mechanisms 177

• A simple design criterion could be, to place actuators as near as possible to the
base platform of the system and use transmission elements to the points/joints of
actuation.

8.2.2 Calculation of the Kinematic Problems

Section 8.1.3 introduced the importance of kinematic transfer functions for the de-
sign of haptic devices. Calculating the force transfer characteristics the forward
kinematics is necessary for impedance controlled devices, and the inverse kinemat-
ics for admittance controlled devices. The following section describes a method to
calculate the forward kinematics for any serial mechanism called the “D ENAVIT-
H ARTENBERG-parameters” method. Literarily this method is based on climbing
along a serial kinematic chain step by step. This is done by multiplying a homoge-
neous coordinate transformation matrix containing the coordinates of each element
with the combined product of its predecessors. This transformation matrix is dom-
inated by four parameters, called the D ENAVIT-H ARTENBERG-parameters. Math-
ematically moving through the whole kinematic chain and using this approach the
kinematic correlation between the TCP and the frame - the forward kinematics - is
obtained.
An adequate homogeneous coordinate transform matrix is given by the matrix T
[102]:
⎛ ⎞
  R1,1 R1,2 R1,3 t1
Rt ⎜ R2,1 R2,2 R2,3 t2 ⎟
T= =⎜ ⎟ (8.12)
0 1 ⎝ R3,1 R3,2 R3,3 t3 ⎠
0 0 0 1

with homogeneous coordinate transformation matrices rotations, translations, scal-

ing and perspective transformations can easily be calculated [257]. For robotic ap-
plications only rotations and translations are relevant, which are included in the
specific matrix given in equation 8.12.
Remarks:
• Matrix multiplications are associative, but they are not commutative. Thus the or-
der for multiplication is highly important for calculating the forward kinematics
(see below).
• The numbers (0, 1) in the last row of the matrix guarantee that rotations and
translations do not influence each other. With this feature a simple algorithm
can perform rotations and translations with a single matrix multiplication. This
increases the clarity of an implementation and may be one reason why homoge-
nous coordinate transformations are widespread within robotics and even virtual
reality programming.
178 8 Kinematic Design

Using the homogenous transformation matrix T a vector p - itself being homoge-

nous too - can be transformed via scalar multiplication from one coordinate system
(0) in another coordinate system (1):

p1 = T · p0 (8.13)

In order to perform multiplications with a 4 × 4 matrix the vectors have to be of

type 4 × 1, which is mandatory for homogenous vectors. For robotic applications a
vector p is formed as

p = (px , py , pz , 1)T (8.14)

where px , py and pz are Cartesian coordinates.

The calculation of the forward kinematics happens in the following steps:
1. Definition of a coordinate system for each body attached to its lower end begin-
ning at the coordinate system attached to the frame and ending in a coordinate
system attached to the TCP. The coordinate systems have to be oriented in such
way that the coordinate system (i+1) can be reached by its predecessor coordinate
system (i) with the combination of the following changes:
• Rotation θ around the zi -axes (rotation according to the right-hand rules,
clockwise looking in the direction of the rotational axes)
• Displacement d along the zi -axes
• Displacement a along the xi+1 -axes
• Rotation α along the xi+1 -axes (right-hand rule)
2. Composition of a table with all D ENAVIT-H ARTENBERG-parameters (θ , d, a, α )
of the whole chain. The chain includes n coordinate systems.
3. Formulation of a transformation matrix (Ti ) for each group of D ENAVIT-
H ARTENBERG-parameters (see below).
4. Multiplication of the transformation matrixes to calculate the total transformation
Ttot = Tn · . . . · Ti · . . . · T0
The resulting total transformation matrix Ttot transforms the base coordinate system
p0 to the TCP coordinate system pTCP and represents the forward kinematics.

pTCP = Ttot · p0 (8.15)

Each matrix Ti is composed of the D ENAVIT-H ARTENBERG-parameters by the fol-

lowing scheme:
⎛ ⎞
cos(θi ) −sin(θi ) · cos(αi ) sin(θi ) · sin(αi ) a · cos(θi )
⎜ sin(θi ) cos(θi ) · cos(αi ) −cos(θi ) · sin(αi ) a · sin(θi ) ⎟
Ti = ⎜⎝ 0
⎟
⎠
sin(αi ) cos(αi ) d
0 0 0 1
(8.16)
8.2 Serial Mechanisms 179

The matrix Ttot = ∏0i=n Ti depends on every Denavit-Hartenberg-parameter. It is the

serial execution of n + 1 transformations in three dimensional space with six degrees
of freedom. Therefore Ttot is given by

T =
⎛tot ⎞
cosα · cosβ cosα · sinβ · sinγ − sinα · cosγ cosα · sinβ · cosγ + sinα · sinγ x
⎜ sinα · cosβ sinα · sinβ · sinγ + cosα · cosγ sinα · sinβ · cosγ − cosα · sinγ y ⎟
⎜ ⎟
⎝ −sinβ cosβ · sinγ cosβ · cosγ z⎠
0 0 0 1
(8.17)

Where x,y and z are the Cartesian coordinates of the TCP relative to the origin. α , β
und γ are roll-, pitch- and yaw angle, equivalent to a rotation around the x-axis (roll,
α ) a rotation around the new y-axis (pitch, β ) and a rotation around the resulting
z-axis (yaw, γ ). They can be extracted out of the matrix by some simple calcula-
tions. Instead of roll-, pitch- and yaw angle other conventions like Euler-angles can
be used, of course. Accordingly the interpretation of the matrix Ttot will have to
change.
In contrast to the direct kinematics the inverse kinematics is hard to derive for serial
kinematic problems. Some methods can be obtained from [257]. The inverse kine-
matics is only relevant for admittance based devices and of minor importance for
impedance-based devices.

8.2.3 Example of a Serial Mechanism

Figure 8.8 gives an example for a serial mechanism with four driven joints q1 to q4 ,
three of them rotary and one linear. Coordinate systems are located in the centres
180 8 Kinematic Design

x
TCP
z
q4 z
x 5

4
q3 z l4
3

x l3

l2

x 2
x q2
z
0 z
l1
1
x

q1

Fig. 8.8 Example of a serial mechanism

of each actuator such as in the TCP (1 to 5). As a reference a world coordinate

system (0) marks a fixed point on the frame. The coordinate systems are placed
such that each of them can be reached by a combined movement according to the
four D ENAVIT-H ARTENBERG-parameters from the preceding one (section 8.2.2).
In this example the forward kinematics of the mechanism shall be calculated. First
the necessary parameters are collected within table 8.1.
then the transformation matrixes T0 to T4 can be generated.
8.2 Serial Mechanisms 181

Table 8.1 D ENAVIT-H ARTENBERG-parameters of the mechanism in fig. 8.8

n θ d a α
(rotation (displacement in (displacement in (rotation
around zn ) zn -direction) xn+1 -direction) around xn+1 )
0 90◦ q1 0 90◦
1 0 l1 0 −q2
2 90◦ l2 0 −q3
3 −q4 l3 0 0
4 0 l4 0 0

⎛ ⎞
001 0
⎜1 0 0 0 ⎟
T0 = ⎜ ⎟
⎝ 0 1 0 q1 ⎠
000 1
⎛ ⎞
1 0 0 0
⎜ 0 cos(−q2 ) −sin(−q2 ) 0 ⎟
T1 = ⎜ ⎟
⎝ 0 sin(−q2 ) cos(−q2 ) l1 ⎠
0 0 0 1
⎛ ⎞
0 −cos(−q3 ) sin(−q3 ) 0
⎜1 0 0 0⎟
T2 = ⎜ ⎟
⎝ 0 sin(−q3 ) cos(−q3 ) l2 ⎠
0 0 0 1
⎛ ⎞
cos(−q4 ) −sin(−q4 ) 0 0
⎜ sin(−q4 ) cos(−q4 ) 0 0 ⎟
T3 = ⎜ ⎟
⎝ 0 0 1 l3 ⎠
0 0 01
⎛ ⎞
100 0
⎜0 1 0 0 ⎟
T4 = ⎜ ⎟
⎝ 0 1 0 l4 ⎠
000 1

The total forward-kinematics Ttot is a result of the product

Ttot = T4 · T3 · T2 · T1 · T0

and is dependent only on the geometrical dimensions (l1 , l2 , l3 , l4 ) and the positions
of the driven joints (q1 , q2 , q3 , q4 ).
Remarks:
• The resulting analytical solution will become quite bulky due to the large num-
bers of multiplications of sine and cosine terms, but will be fast to computate
with given values for q1 to q4 .
182 8 Kinematic Design

• The Jacobian-matrix of the forward kinematics will be even more complex, but
will have to be calculated for the force transfer characteristics of impedance
based devices. A recommendable help for the calculation are programs for the
solution of symbol based mathematical problems like the symbolic math toolbox
of MatLab R
.

8.3 Parallel Mechanisms

In contrast to serial assemblies parallel kinematic mechanisms consist of coupled

passive and active joints. In general each active joint (each actuator) contributes to
one degree of freedom of the TCP’s movement. Hence all previous considerations
from section 8.1.3 can be applied to parallel mechanisms. This section coveres the
design of parallel kinematic mechanisms which is devided in the following steps:
• Topological synthesis - the design of an assembly of odys and joints
• Calculation of the kinematics - Mathematical description of displacements pro-
vided by and the output of forces to the user
• Dimensioning - Defining the geometrical dimensions (section 8.1.4)

8.3.1 Topological Synthesis

The G RÜBLER Formula

If active and passive joints are coupled by bodies with the aim, that the resulting
mechanism offers as many degrees of freedom as it contains actuators, then the
number of degrees of freedoms from the joints will have to be exactly defined and
distributed according to specific rules. The absolute number of passive joint DOFs
to be distributed can be calculated by the formula of G RÜBLER which can easily be
described [174][102][24]:
A mechanism with n rigid bodies (including the frame) without a coupling of the
elements in a three-dimensional space has a total number of

F = 6 · (n − 1)

degrees of freedom. Each body has six DOFs in space. The frame itself is fixed to
space, it has no DOF. When coupling any free body with g joints of DOF number fi
(i = 1 . . . g) F is reduced by (6 − fi ). Thus F can be calculated as
8.3 Parallel Mechanisms 183

g
F = 6 · (n − 1) − ∑ (6 − fi )
i=1
g
= 6 · (n − g − 1) + ∑ fi (8.18)
i=1

and is called gear-DOF. F is the number of relative movements which have to be

constrained, to constrain the whole mechanism. Calculating the number of joint-
DOFs according to equation 8.18 a mechanism results, which can be actuated by F
actuators in F DOFs. Within the calculation the so called identical conditions fid
and constraints s have to be corrected to exactly define the number of joints.
g
F = 6 · (n − g − 1) + ∑ fi − fid + s (8.19)
i=1

F Gear-DOF
n Number of bodys
g Number of joints
fi Degree of freedom of the ith joint
fid Sum of identical condition
s Sum of constraints

Remark:
An identical condition is given for example when a rod has universal joints at both
of its ends. The rod will be able to rotate around its axis, without violationg any
constrains. Another example are two coaxial oriented linear joints. Constraints ap-
pear whenever conditions have to be fulfilled to enable the movement [121]. If 5
joint axes have to be parallel to a 6th axis to enable a movement, then z = 5. Another
example for a passive condition are two driving rods which have to be placed in
parallel to enable a motion.

Calculation of the Joint DOFs to be distributed

Usually in the state of the design of the mechanism topology the number of gear-
elements and joints is not yet defined. The origin for the design process is the ques-
tion about the total degrees-of-freedom F the mechanism has to achieve. This value
for F is identical to the total number of actuated joints. A parallel kinematic mech-
anism can be classified according to its level of parallelism [102]. If the mechanism
contains as many legs as it contains DOFs - an example is the hexapod in figure 8.3
- then it is called fully parallel. Other designs are called partly parallel or high-grade
parallel. In this context the focus lies on fully parallel designs. The correlation be-
tween the number of chains k, joints g and elements n is defined according to

n = g−k+2 (8.20)
184 8 Kinematic Design

Frame and TCP are taken into account by the right hand summand “2”. By insertion
into equation 8.19 a relation between chain number k and gear-DOF F and the sum
of all joint-DOFs ftot = ∑gi=1 fi is achieved.
g
ftot = ∑ fi = F + 6 · (k − 1) + fid − s (8.21)
i=1

Remarks:
• For fully parallel mechanisms every chain - a connection between the frame to
the TCP - contains exactly one actuator with one DOF.
• The number of chains k is identical to the number of DOFs for fully parallel
mechanisms. Hence equation 8.21 simplifies to ftot = 7 · F − 6 + fid − s in this
case

Kinematic Scheme and Joint Distribution

A “kinematic scheme” is used to clarify the distribution of joint-DOFs (including

the actuators) within the design of a mechanism. The kinematic scheme of a Delta-
robot (fig. 8.5) is shown in figure 8.9. This robot consists of k = 3 chains, has a total
number of F = 3 DOFs and does not include identical DOFs nor any constraints.
According to equation 8.21

ftot = F + 6 · (k − 1) = 15

joint DOFs have to be distributed among three chains. This is often done symmet-
rically - such as in this case too. Each chain consists of 5 DOFs. Each leg includes
one rotary actuator representing one DOF. The remaining DOFs are realized as uni-
versal joints resulting in a joint-distribution as shown in figure 8.9b.

Tool Center Point

n Passive joint with n degrees
2 2 2 of freedom

5 5 5 2 2 2
m Actuator with m degrees of
1 1 1 freedom

Base Base
a) b)
Fig. 8.9 Kinematic scheme of the delta robot from fig. 8.5. The mechanism contains 15 joint-DOF
distributed to 3 legs (a). Every leg contains a rotary actuator with 1 DOF (b). The joint DOFs are
denoted in the circles.

Many other distributions of joint-DOFs are possible and will result in a different
8.3 Parallel Mechanisms 185

topology of the mechanism, of course. Usually the design process itself is quite
intuitive. As a simple rule of thumb the following advices can be used:
• Concentrate joint-DOFs: Universal joints and ball joints simplify the design and
the transfer characteristics compared to serially placed joints with just one DOF
each. Especially the kinematic transfer characteristics are much easier to formu-
late with concentrated joints. Additionally the play within joints will be concen-
trated which will reduce the play of the total mechanism.
• Use actuators which themselves are fixed to the frame: The drives of the hexapod
are placed in the rods and move during positioning of the platform. Due to the
little workspace covered this does not influence the dynamics of the hexapod sig-
nificantly. However still with the actuators being fixed to the frame the dynamic
properties of the design would be increased, as their mass does not count to the
inertia of the device anymore.
• DOFs should be distributed symmetrically: A symmetrical design should be pre-
ferred compared to an asymmetrical one. The kinematic transfer functions get
significantly simpler. For an asymmetrical design it is very likely that no closed
analytical solution will be found.
For the first draft of a parallel kinematic haptic system it is recommend basing on
known parallel kinematic mechanisms. Generally there is a wide range of known
systems in literature. Examples of typical structures can be found in [185] [24] and
[174].

8.3.2 Calculation of the Kinematic Transfer Functions

In contrast to serial kinematics the inverse kinematic problem of a parallel mecha-

nism is easier to calculate than the forward kinematics. Even with simple designs it
is challenging or even impossible to analytically solve the direct kinematics, which
is the base for the force transfer function of impedance based haptic devices. Gener-
ally speaking systems with symmetries and concentrated degrees of freedom within
passive joints are easier to solve than other designs. In [259] and [258] analytical so-
lutions of the forward kinematic problem of simple manipulators with three degrees
of freedom are shown. Other designs can be found in [257] and [103]. The proce-
dure of calculating the inverse kinematic problem can be split up into the following
three steps:
1. Formulation of closed vector chains for each leg, starting at the reference coor-
dinate system (0) enclosing the TCP coordinate system (0) and going back to the
reference coordinate system..
2. Splitting the vector chains in all - Cartesian - movement directions of the indi-
vidual leg.
3. Solving the resulting system of equations according to the TCP’s coordinates.
186 8 Kinematic Design

An example for this calculation is given in section 8.3.3. The transfer function of
parallel kinematic mechanisms and especially their forward kinematic calculation
include singularities (section 8.1.3). Usually they are a result of stretched orienta-
tions of passive joints. A problem that is obvious within serial mechanisms. It is
mandatory to consider this problem during device design.

8.3.3 Examples of a Parallel Mechanism

Figure 8.10 shows one example of a planar parallel mechanism. The actuated joints
are located at q1 and q2 and summarized to the vector q. The lengths of both rods
on the right side are identical (a) and are placed at the same distance on both sides
(b), resulting in a parallel linkage. This prevents the TCP from rotating and makes it
move within planar Cartesian coordinates x = (x, y). The corresponding coordinate
system is labelled with T.

TCP d
a

a b
T
b
x=(x,y) b
a
q1 y q2
x a
0
Fig. 8.10 Example of a parallel mechanism

Following the G RÜBLER method for the total number of mechanism degrees of
freedom (equ. 8.21) this parallel linkage generates a constraint. Additional five con-
straints are given by the requirement, that all five axis of the rotary joints are placed
in parallel to the axis of the 6th rotatory joint. Resulting in s = 6 constraints. The
mechanism is composed of k = 3 chains, is expected to have F = 2 degrees of free-
dom (movement in the x-y-plane without rotating). Identical DOFs (section 8.3.1)
are not existent ( fid = 0). Equation 8.21 calculates
8.3 Parallel Mechanisms 187

ftot = F + 6 · (k − 1) + fid − s = 8

DOFs, which have to be distributed among the joints to achieve a mechanically

defined mechanism. Two of these DOFs are given by the linear drives q1 and q2 .
The remaining 6 DOFs have to be realized as passive joints. Here, this is achieved
by 6 passive rotational joints. The following paragraph deals with the calculation of
the inverse kinematic problem q = f −1 (x), which is not too complicated in this case.
The inverse kinematics would be suitable for an admittance controlled haptic device.
It can be calculated with the aid of two closed vector chains in world-coordinates
via the tool-center-point. These vector chains include both legs one the right and left
side and are split up according to their x- and y-coordinate.
  2

 c − d2 − x
d
x = c − a · cos(α ) − ⇒ sin(α ) = 1 − (8.22)
2 a
b b
y = q2 + a · sinα + ⇒ q2 = y − a · sin(α ) − (8.23)
2  2
  2
 x − d2
d 
x = a · cos(β ) − ⇒ sin(β ) = 1 − (8.24)
2 a
y = q1 + a · sinβ ⇒ q1 = y − a · sin(β ) (8.25)

By combining 8.22 with 8.23 and 8.24 with 8.25 the solutions for q1 and q2 can be
found:
 

 x − d 2
q1 = y − a · 1 − 2
a
  2

 c − d
− x
q2 = y − a · 1 −
b
2
−
a 2

These are the inverse kinematic transfer functions.

Remarks:
• Additionally the solution of the forward kinematic problem x = f (q) is necessary
for calculating the position of the TCP based on position measurements taken at
the actuators. If the forward kinematic problem is not known, one option could
be to include additional sensors within the haptic device to gain additional infor-
mation about position and orientation. 
• The solution of the equations include sin(φ ) = 1 − cos2 (φ ). The radicand of
the square root has to stay positive for real solutions. This is a plausible indicator
that the mechanism contains singularities.
188 8 Kinematic Design

8.4 The Complete Process of Kinematic Design

Within this chapter the basics for the kinematic design process were discussed. This
final section will summarize the whole process and repeat it in context. Figure 8.11
shows the rough structure.

Selection of a kinematic structure (serial or parallel)

Design of an adapted topology

Solution of the kinematic problems

Optimisation of the dimensions

Fig. 8.11 The complete process of kinematic design.

Selection of a Structure

Basically two different structures exist: “serial kinematic” or “parallel kinematic”.

A mix of both is called “hybrid kinematic”, combining some degrees of freedom
with parallel elements, while others are purely serial. This variant combines the ad-
vantages and disadvantages of both concepts. The haptically relevant properties of
each variant are given in table 8.2 and compared qualitatively. Serial and parallel
kinematics were discussed in section 8.2 and 8.3.

Table 8.2 Interesting properties of serial kinematic and parallel kinematic mechanisms for haptic
applications

Property Serial Mechanisms Parallel Mechanisms

Mechanical impedance high low
Ratio workspace/size high low
Calculation of the forward kinematic problem easy difficult
8.4 The Complete Process of Kinematic Design 189

The most important criterion of table 8.2 is given with the mechanical impedance.
For haptic devices - especially open- or closed-loop impedance controlled devices
- the mechanism’s impedance should be as low as possible. This guarantees a
good transformation of driving forces to the TCP. The easiest way to achieve low
impedance is given with parallel kinematics with light rods and joints with low fric-
tion. The actuators should be fixed to the frame in a parallel concept and located near
the frame in a serial mechanism. A transmission of movement energy via tendons
should be considered. The haptic device “Delta.3” from the company ForceDimen-
sion (fig. 8.12) with the kinematic structure of a delta-robot (fig. 8.5) fulfils these
recommendations and is therefore an example of a well chosen kinematic structure.
In the haptic device “Falcon” of the company Novint (fig. 8.2) the same kinematic
configuration is chosen with variants in the design of joints and the quality of com-
ponents. The intensive usage of the delta structure shows its suitability for haptic
applications. Concerning the concepts of open- or closed-loop control (section 7)
all kinematic principles are suitable to be used.

Fig. 8.12 The haptic device “Delta.3” by the company ForceDimension uses a parallel kinematic
structure with rotatory actuators fixed to the frame. The bodys are constructed lightweight to
achieve a low mechanical impedance.

Design of a Topology

The design process of an adequate topology has been discussed within the sec-
tion 8.2.1 and 8.3.1. Frequently the topology can be defined intuitively when dealing
with serial structures.
The topology-design for parallel structures is more challenging. First the number of
necessary DOFs have to be identified and a kinematic scheme has to be designed
or selected (section 8.3.1). Afterwards a single leg stretching from the frame to the
tool center point should be dimensioned. Adding additional legs to the first one con-
190 8 Kinematic Design

strains movement capabilities of the mechanism, for example by constrained parallel

movements. For fast progress especially with parallel mechanism a detailed review
of existing literature of industrial robots is recommended [174] [185] [257].

Calculation of the kinematic problems

Depending on the design of the topology the direct kinematics may be analytically
unsolvable. Even the inverse kinematics, the kinematic description for admittance
based devices, may not be analytically solvable in some cases. Section 8.2.2 gives
a general method for the calculation of forward kinematics of serial mechanisms.
When an analytical solution can not be found, numerical real-time solutions are an
interesting option. A simple Netwon-Raphson-approximation method is able to find
roots for closed vector chains. But even on a PC with several GHz of computing
power these approximation algorithms are not fast enough for haptic applications
yet. Adjusted hardware, e.g. field programmable gate arrays (FPGA) are suitable
for these types of calculations and are able to solve such kinematic problems at an
update rate considerably below 1 ms.

Optimization of Structural Dimensions

Section 8.1.4 gave already a short insight into the challenges for automated op-
timization of mechanisms. Even if no complete optimization can be done, a fun-
damental thing to do is an analysis for singularities within the workspace, and an
intuitive and iterative optimization of the structure to guarantee a force transmission
in any point of the workspace.
Chapter 9
Actuator Design

T HORSTEN A. K ERN , M ARC M ATYSEK , S TEPHANIE S INDLINGER

The preceding chapters were focused on the basics of control engineering and
kinematic design. They covered topics of structuring and fundamental character.
This and the following chapters deal with the design of technical components as
parts of haptic devices. Actuators are the most important elements of every haptic
device, as their selection respectively their design influences the quality of the hap-
tic impression significantly. This chapter deals with frequently used actuators struc-
tured according to their physical working principle. Each actuator type is discussed
as to its most important physical basics, with examples of their dimensioning, and
one or more applications given. Other, rarely used actuator principles for haptic sys-
tems are mentioned in section 9.6 “special forms”, or are classified in the overview
of physical working principles in section 9.1.

Experience teaches us that actuators for haptic applications can rarely be found
“off-the-shelf”. Their requirements always include some outstanding features in ro-
tational frequency, power-density or geometry. These specialties make it necessary
and helpful for their applicants to be aware of the capabilities and possibilities of
modifications of existing actuators. Hence this chapter addresses both groups of
readers: the users who want to choose a certain actuator and the mechanical engi-
neer who intends to design a specific actuator for a certain device from scratch.

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_9, 191

c Springer-Verlag Berlin Heidelberg 2009
192 9 Actuator Design

9.1 General Facts about Actuator Design

T HORSTEN A. K ERN

Before a final selection of actuators is made, the appropriate kinematics and the
control-engineering structure, according to the previous chapters, should have been
fixed. However, in order to handle these questions in a reasonable way, some basic
understanding about actuators is mandatory. Especially the available energy densi-
ties, forces and displacements should be estimated correctly for the intended haptic
application. This section provides some suggestions and guidelines to help and pre-
select appropriate actuators according to typical requirements.

9.1.1 Overview of Actuator Principles

There are a certain number of approaches to transform an arbitrary energy source

into mechanical energy. Each of these ways is one actuation principle. The best
known and most frequently used principles are:

• Electrodynamic principle - A force, so called L ORENTZ-force, acting upon a

conductor conducting a current
• Electromagnetic principle - A force, acting upon a magnetic circuit to minimize
the enclosed energy
• Piezoelectric principle - A force, acting upon the atomic structure of a crystal
and deforming it.
• Capacitive principle - A force, resulting from charges trying to minimize the
energy in a capacitor
• Magnetorheological principle - Viscosity change within a fluid resulting from
particles trying to minimize the energy contained within a magnetic circuit.
• Electrochemical principle - displacement of or pressure within a closed system,
whereby a substance emits or bounds a gas and consequently changes its volume
due to the application of electrical energy.
• Thermal principle - Change of length of a material due to controlled tempera-
ture changes, making use of the material’s coefficient of thermal expansion.
• Shape-memory alloy - Sudden change of an object’s shape made of special ma-
terial due to relatively small temperature changes (≈ 500◦ C). The object trans-
forms into a root shape embossed during manufacture by the application of high
temperatures (≈ 1000◦ C).
9.1 General Facts about Actuator Design 193

Each of these principles is used in different embodiments. They mainly differ

in the exact effective direction of e.g. a force vector1 or a building principle2 . As
a consequence a wide-spread terminology exists for naming actuators. The major
terms are given here:
• Electric motor - the most general term at all. It may describe any electrome-
chanic transformer. However, in most cases it refers to an actuator rotating con-
tinuously whose currents are commutated (mechanically or electronically), or
which is equipped with a multiphase alternating current unit. Typically, it is a
synchronous motor, a drive with a rotor moving synchronously to the rotating
electromagnetic field. In a more general understanding the term includes hys-
teresis motors and squirrel-cage rotors, too. But the latter has not yet reached any
relevance for haptic systems, even not in very exceptional cases.
• EC-motor - Specific embodiment of the synchronous motor and very common
to haptic applications. Motor according to the electromagnetic or electrody-
namic principle with an electronic control unit for the rotating field (electronic-
commutated, electronically commutated).
• DC-motor - Another specific form of a synchronous motor and used among
haptic applications because of its cheapness and simplicity. This is an actuator
according to the electromagnetic or electrodynamic principle with a mechanical
control unit for the rotating filed using switching contacts (mechanically com-
mutated).
• Resonance-actuator - Generic term for a whole class of actuators with very dif-
ferent actuation principles. The term describes an actuator containing one com-
ponent which is driven in its mechanical resonance mode (or nearby its resonance
mode). Typically, parts of this component make an elliptic oscillation driving a
second component in small steps via a frictional coupling. As a result of the
high frequency, the movement of the second component seems uninterrupted.
The term is most frequently applied to piezoelectric actuators.
• Ultrasonic-actuator - Resonance-actuator performing steps at a frequency within
ultrasonic ranges (>15kHz). These actuators are built almost always according to
the piezoelectric principle.
• Voice-coil-actuator - Drive according to the electrodynamic principle. Mainly
consisting of a conductor wrapped around a cylinder. The cylinder itself is placed
in a magnetic circuit, resulting in a Lorentz-force when a current is applied to the
conductor. There are two major embodiments, one with a “moving coil”, another
variant with a “moving magnet”.
• Shaker - Another form of a voice-coil-actuator with an elastic suspension of
the coil. When current is applied to the coil an equilibrium condition between
the suspension’s spring and the Lorentz-force is achieved at a specific displace-
ment. Actuators according to this structure are frequently used for fast and dy-
1 The electromagnet principle for instance is divided into magnetic actuators and actuators ac-
cording to the reluctance principle; the piezoelectric principle is subdivided into three versions
depending on the relative orientation of electrical field and movement direction.
2 e.g. resonance drives vs. direct drives
194 9 Actuator Design

namic movements of masses for vibration testing (this is where its name comes
from). It is nothing else but a loudspeaker principle without a spatially expanded
membrane for sound generation. Another embodiment of this principle primarily
aiming at measurement applications is called ampere balance. In this case a given
load is applied against the elasticity resulting in a displacement. The current is
increased until the displacement is reset to zero. The value of the current is a di-
rect measure for the load applied. This measurement principle of forces/torques
is also known as “T HOMSON balance” or “electrodynamic balance”.
• Plunger-type magnet - Actuator according to the electromagnetic principle. A
rod made of ferromagnetic material is pulled into a magnetic circuit equipped
with a coil. These actuators have very nonlinear force-displacement characteris-
tics.
• Stepper-motor - Generic term for all actuation principles moving forward step
by step. In contrast to the resonance-drives no component of a stepper-motor is
driven in resonance mode. Their step-frequency is below any resonance of the
system and may vary. These actuators may even be used in a “microstep mode”,
interpolating movement between so called “full-steps”, which are original to their
mechanical design. The term is most frequently used for rotatory drives. Espe-
cially for those working according to the reluctance-principle or another electro-
magnetic actuation principle.
• Pneumatic and Hydraulic - These actuation principles do not have a direct elec-
tric input value. They transform pressure and volume-flow into displacement and
force. The media for pressure transmission is air in case of pneumatics and a
fluid, typically oil, in case of hydraulics. Usually, the pressure itself is generated
via actuators, e.g. electrical actuators attached to a compressor, and controlled
via electrical valves.
• Bending-actuator - Actuator with an active layer, frequently made of a piezo-
electric material - attached to a passive mechanical substrate. By actuating the
active layer, mechanical tensions between this layer and the substrate build up,
resulting in a bending movement of the whole actuator.
• Piezoelectric stack - A larger number of piezoelectric layers mechanically con-
nected in series. Small displacements of each layer sum up to a large usable
displacement of the whole actuator.
• Piezoelectric motor/drive - Generic term for all drives according to the piezo-
electric principle. It frequently refers to drives moving a rotor or translator with
a frictional coupling. However, this movement does not need to happen in reso-
nance mode.
• Capacitive actuator - Actuator according to the capacitive principle and fre-
quently used in microtechnology/micro engineering. Usually equipped with a
comb-like structure of electrode-pairs, generating forces in milli-newton with
micrometers of displacement.
• Shape-memory wire - Wire on the basis of shape-memory alloys capable to
shorten in the range of percents (≈ 8% of its total length) when changing its
temperature (e.g. by controlling a current flowing through the wire. The current
heats up the wire according to its thermal loss at the wire’s electrical resistance).
9.1 General Facts about Actuator Design 195

• Surface-wave actuators - Generic term for a group of actuators generating high-

frequency waves in mechanical structures or exciting the resonance-modes of
structures. This actuator is frequently based on piezoelectric principles and has
been used for the generation of haptic textures for some years.

Each of the above actuation principles can be found in tactile and/or kinaesthetic
systems. To simplify the decision process for a new design, all actuators can be
grouped into classes. Most of the physical working principles can be grouped either
into “self-locking” or “free-wheeling” systems. These groups are identical to:

• Positional-sources (x) respectively angular-sources (α )

• Force-sources (F) respectively torque-sources (M)
According to the basic structures of haptic systems (chapter 5) it is likely that
both classes are used within different haptic systems. The correlation between basic
structures of haptic systems and actuators is depicted in table 9.1. This table shows
a tendency towards typical applications; however by adding mechanical elements
(springs, dampers) it is possible to use any actuator for any basic structure of haptic
systems.

Table 9.1: “Classic” Application areas for actuator principles in haptic

systems
X: is frequently used by many groups and even commercialized;
(X): some designs, especially in research;
-: very rare to almost none, an if it is used, it is only in the context of research);
T YPE: Gives an idea about which actuator design (translatory or rotatory) is used
more often. If the actuator is unusual but does exist, the marker is set into brackets.

Control-type: Admittance Impedance

Type Actuator closed-l. open-l. open-l. closed- l.
Rot. electric motor3 X X (X)4 -
Rot. & transl. EC-motor - - X X
Rot & transl. DC-motor - - X X
Rot & transl. Resonance-actuator X X (X) -
Rot & transl. Ultrasonic-actuator X X (X) -
Transl. Voice-Coil - - X X
Transl. Shaker X X - -
Transl. Plunger-type magnet X - - -
Rot. (& transl.) Stepper-motor X X - -
Transl. (& rot.) Pneumatic (X) X - -
Transl. (& rot.) Hydraulic - X - -
Transl. Bending-actuator - X - -
3 in the meaning of a mechanically commutated drive with a power between 10 to 100 W
4 by high frequency vibrations of the commutation
196 9 Actuator Design

Control-type: Admittance Impedance

Type Actuator closed-l. open-l. open-l. closed- l.
Transl. Piezo-stack (X) X - -
Transl. & rot. Piezo-actuator X X X -
Transl. Capacitive - (X) - -
Transl. Shape-memory - (X) - -
Transl. Surface wave - (X) - -

9.1.2 Actuator Selection Aid Based on its Dynamics

Different actuator designs according to the same physical principle still cover wide
performance ranges regarding their achievable displacements or forces. Based on
the author’s experience, these properties are put into relation to the dynamical ranges
relevant for haptic applications. In figure 9.1 the most important actuation principles
are visualized in squares scaled according to their achievable displacements (a)5 and
typical output forces and torques (b). The area covered by a specific version of an
actuator is typically smaller than the area shown here. The diagram should be read in
such a way that e.g. for haptic applications, . electromagnetic linear actuators exist,
providing a displacement up to 5 mm at ≈ 50 Hz. These designs are not necessarily
the same actuators which are able to provide ≈ 200 N, as with electromagnetic sys-
tems the effectively available force increases with smaller maximum displacement
(section 9.3). The diagrams in figure 9.1 visualize the bandwidths of realization-
possibilities according to a certain actuation principle and document the preferred
dynamic range of their application. Using the diagrams, we have to keep in mind
that the borders are fluent and have to be regarded in the context of the application
and the actuator’s individual design.
Torque [Nm]

hydraulic
stroke [mm]

Force [N]

Hydraulic
El.-dyn. rotatory ultrasonic Piezo-staple
linear E.-magnetisch
100 100 linear reluctance
E.-magnetic
(step-motor)
rotatory
E.-magnetic rotatory
reluctance rotatory
rotatory (step-motor)
10 10 El.-dyn. rotatory

reluctance
linear El.-dyn. linear
(step-motor) E.-magnetic El.-dyn. linear
linear piezo-bending
Piezo-Bending
1 1 ultrasonic
reluctance linear
capacitive (step-motor)
linear
Piezo-staple
a) b) capacitive

1 10 100 1000 f [Hz]

f [Hz] 1 10 100 1000

Fig. 9.1 Order of the actuator principles according to the achievable displacement (a) and forces
resp. torques (b) in depedency from their dynamics.
5 for continuously rotating principles all displacements are regarded as unlimited
9.1 General Facts about Actuator Design 197

9.1.3 Gears

In general machine engineering the usage of gears is a matter of choice for adapting
actuators to their load and vice versa. Gears are available in many variants. A sim-
ple lever can be a gear; complex kinematics according to chapter 8 are a strongly
nonlinear gear. For haptic applications specialized gear designs are discussed for
specific actuation principles in the corresponding chapters. However, there is one
general aspect of the application of gears with relevance to the design of haptic
systems which has to be discussed independently: the scaling of impedances.

ain
aout xout
Min
Mout ain Fout
Min xout
Fout r1
r2 r1 xin
a) b) l2 Fin c)
l1

Fig. 9.2 Simple gear design with wheels (a), a level (b) and a Bowden-cable (c).

There is no principal objection to the use of gears for the design of haptic
systems. Each gear (fig. 9.2) may it be rotatory/rotatory (gearwheel or frictional
wheel), translatory/translatory (lever with small displacements), rotatory/translatory
(Bowden-cable) has a transmission “tr”. This transmission ratio scales forces and
torques neglecting loss due to friction according to
Fout l2
= tr = , (9.1)
Fin l1
Mout r2
= tr = , (9.2)
Min r1
Fout 1
= tr = , (9.3)
Min 2π r1
and displacements resp. angles according to
xin l2
= tr = , (9.4)
xout l1
αin r2
= tr = , (9.5)
αout r1
αin 1
= tr = . (9.6)
xout 2π r1
198 9 Actuator Design

The velocities and angular velocities scale analogously to the differential of

above equations. Assuming the impedance of the actuator Z transl = Fv resp. Z rot = αM ,
one consequence of the load-condition of a driven impedance Z out from the perspec-
tive of the motor is:
F in F 1 1
Z transl = = out 2 = Z transl out 2 (9.7)
vin vout tr tr
M in M 1 1
Z rot = = out = Z rot out 2 (9.8)
α α  tr2 tr
The transmission-ratio tr is quadratic for the calculation of impedances. From
the perspective of an actuator, the driven impedance of a system is getting small
with a gear with a transmission-ratio larger than one. This is favorable for the ac-
tuating system (and the original reason for the application of gears). For haptic ap-
plications, especially for impedance controlled ones, the opposite case has to be
considered. In an idle situation and with a transmission ratio larger than one6 the
perceived mechanical impedance of a system Z out increases to the power of two
with the transmission-ratio. Another aspect makes this fact even more critical, as
the increase of output force changes only in a linear way with the transmission ratio,
whereas e.g. a motor’s unwanted moment of inertia is felt to increase quadratically.
This effect is obvious to anyone who has ever tried to rotate a gear-motor with a high
transmission ratio (e.g. tr > 100) at its output. The inertia and the internal frictions
upscaled by the gear is identical to a self-locking of the actuator.

As a consequence, the usage of gears with force-controlled haptic systems makes

sense only for transmission ratios of 1 to 20 (with emphasis on the lower transmis-
sion ratios between 3 to 6). For higher transmission ratios gear designs according
to figure 9.2c and equation 9.6 have proved valid. They are used in many com-
mercial systems, as with the aid of the definition tr = 2π1r1 and the included factor
2π a comparably high transmission ratio can be achieved easily. In combination
with rotatory actuators (typically EC-drives) with low internal impedances this de-
sign shows very impressive dynamic properties. Some advice may be given here
out of practical experience: Gearwheels are applicable for haptic systems but tend
to generate unsteady and waving output torques due to their toothing. By a care-
ful mechanical design, this unsteadiness can be reduced. The mechanical backlash
should be minimized (which is typically accompanied by an increase in friction) for
example by material combinations with at least one soft material. At least one gear
should have spur/straight gearing, whereas the other one can keep involute gearing.

6 which is the normal case, as typically the fast movement of an actuator is transmitted into a
slower movement
9.2 Electrodynamic Actuators 199

9.2 Electrodynamic Actuators

T HORSTEN A. K ERN

Electrodynamic actuators are the most frequently used type of drives for haptic
applications. This popularity is a result of the direct proportion between their output
value (force or torque) from their input values (the electrical current). In case of ki-
naesthetic applications they are typically used as open-loop controlled force sources.
But even with tactile applications these very dynamic actuators are frequently ap-
plied for oscillating excitements of skin areas. They are equally often used rotatory
and translatory. Depending on the design either the electrical coil or the magnet
is the moving component of the actuator. This section gives a short introduction to
the mathematical basics of electrodynamic systems. Afterward some design variants
are discussed in more details. The final subsection deals with the drive electronics
necessary to control electrodynamic systems.

9.2.1 The Electrodynamic Effect and its Influencing Variables

Electrodynamic actuators are based on the L ORENTZ-force

FLorentz = i · l × B, (9.9)
acting upon moving charges in magnetic fields. The L ORENTZ-Force is depen-
dent on the current i, the magnetic induction B such as the length of the conductor l,
which is typically formed as a coil. This subsection deals with optimization of each
parameter for the maximisation of the generated output force FLorentz . Any electro-
dynamic actuator is made of three components:

• generator of the magnetic filed (coil, or most frequently a permanent magnet)

• magnetic flux conductor (iron circuit, magnetic core)
• electrical conductor (frequently formed as coil or a more complex winding)

After a shallow look a recommendation for the maximization of the output force
could be to simply increase the current i in the conductor. However with a given and
limited available space for the conductor’s length l (coil’s cross section), and a flux
density B with an upper border (0.8 to 1.4 T), the effectiveness of this change has to
be put into question. This can be shown with a simple calculation example:
200 9 Actuator Design

coil FLorent z
pole shoe i

S
N

magnet
magnetic yoke

Fig. 9.3 Moving-coil actuator AVM by adrive-technology building the basis for the following
example.

9.2.1.1 Efficiency Factor of Electrodynamic Actuators

For example a straight-forward design of an electrodynamic actuator similar to the

AVN 20-10 (fig. 9.3) is analyzed. It contains a wound coil with a permanent-magnet
in a ferromagnetic core. The electrical power loss Pel of this electrodynamic system
is generated mainly in a small moving coil with a pure ohmic resistance Rcoil =
3.5 Ω and a nominal current i = 0.78 A:

Pel = Rcoil i2 = 3.5 Ω · 0.78 A2 = 2.13 W. (9.10)

With this electrical power loss, and at a flux density B=1.2 T, and with an orthog-
onal conductor orientation, and a conductor length within the air gap l=1.58 m, the
actuator itself generates the force

FLorentz = i l B = 0.78 A · 1.58 m · 1.2 T = 1.48 N . (9.11)

Assuming the system being driven in idle mode - working against the coil’s own
mass of m=8.8 g only - being accelerated from idleness, and performing a displace-
ment of x=10 mm, above electrical power Pel is needed for a period of
 
x xm
t= 2 = 2 = 0.011 s (9.12)
a F
seconds. The electrical energy loss sums up to

Wel = Pel · t = 23, 4 mJ. (9.13)

This gives an efficiency factor of WelW+W

mech
mech
= 38% for idle mode and continu-
ous acceleration. Assuming now that the same actuator shall generate a force of
1 N against a finger tip for a period of e.g. two seconds, an electrical power of
9.2 Electrodynamic Actuators 201

Wel = 2.13 W · 2s = 4.26 J is needed. This would be identical to an efficiency factor

well below 1%. And indeed the efficiency factor of electrodynamic actuators in hap-
tic applications lies in the area of low percentages due to the common requirement to
generate quite static forces without much movement. This simple calculation points
to one major challenge with electrodynamic actuators: The electrical power lost due
to heat transmission extends the mechanically generated power by far. Consequently
during the design of electrodynamic actuators attention has to be paid to the opti-
mization of power within the given actuator volume and to the thermal management
of energy losses.

9.2.1.2 Minimization of the Power Loss

Typical designs of electrodynamic actuators either have a winded conductor which

in itself is self-supportive, or which is wound on a coil-carrier (fig. 9.4). The avail-
able space for the electrical coil within the homogenous magnetic flux is limited
(ACoil ). The number of coil-turns NConductor is limited too within this area due to the
cross-sectional surface a single turn needs AConductor . This cross-sectional surface
is typically more than the actual cross-section of the conductor, as the winding will
have gaps in between single turns (equ. 9.15). Additionally the actual conducting
core with the cross sectional surface ACore will be smaller than the cross-section of
the conductor itself due to its isolation. Both parameters describing the geometrical
losses in cross sections which are available within tables of technical handbooks
[187] and are assumed as factors k ≥ 1 according to equation 9.14. The length l of
the conductor can be easily calculated by multiplication with the numbers of turns
and the mean circumference Circ of the coil (equ. 9.16).

i ASpule

B pole shoe of the

magnetic circuit

Acore
Aconductor
FLorent z
cylindrical coil in cross
section

Fig. 9.4 Cross-section through a cylindric electrodynamic actuator according to the moving coil
principle.

The choice of the conductor’s diameter influences the resistance of the coil via
the conducting area ACore . The specific length-based resistance Rspezf of a conductor
is given according to equation 9.17. Big conducting diameters with large cross-
202 9 Actuator Design

sections ACore allow coils able to conduct high currents at low voltages but - due
to the limited volume available - few windings. Small diameters limit the necessary
currents at high voltages and carry more windings. By a careful choice of wire-
diameter the winding can be adjusted as a load to the corresponding source to drain
the maximum available power.
The power loss PLoss (equ. 9.18) acceptable within a given winding is limited.
This limit is defined by the generated heat being able to dissipate. The technical
solutions are dependent on the time of continuous operation, the thermal capacity
resulting from the volume of the actuator and the materials it consists of, and a po-
tential cooling system. A calculation of heat transmission is specific to the technical
solution and can not be solved in general within this book. But the dependency of
L ORENTZ-force on power loss can be formulated:

AConductor = k · ACore (9.14)

ACoil
NConductor = (9.15)
AConductor
lConductor = NConductor ·Circ (9.16)
lConductor ρ
Rspezf. = (9.17)
AConductor
PLoss = i2 · RCoil (9.18)
From equation 9.18 follows 
PLoss
i= (9.19)
RCoil
With equation 9.17 there is 
PLoss ACore
i= (9.20)
ρ lConductor
put into equation 9.9 (keeping the direction of current flow ei ) there is

PLoss ACore lConductor
FLorenz = ei × B (9.21)
ρ

by considering equation 9.15 to 9.16 the result is


PLoss ACoil N Circ
FLorenz = ei × B (9.22)
ρk

The equations 9.15 to 9.18 put into equation 9.9 gives a precise picture of the
influence values on the L ORENTZ-force (equ. 9.22). The level of Lorentz-force is
given by the power loss PLoss acceptable within the coil. If there is room for modifi-
cations to the geometrical design of the actuator, the cross-sectional area of the coil
and the circumference of the winding should be maximized. Additional a choice of
9.2 Electrodynamic Actuators 203

alternative materials (e.g. alloy instead of copper) may minimize the electrical re-
sistance. Furthermore the filling-factor k should be reduced. One approach could be
the use of wires with a rectangular cross section to avoid empty spaces between the
single turns.

The question for the maximum current itself is only relevant in combination with
the voltage available and in the context of adjusting the electrical load to a specific
electrical source. In this case for iSource and uSource the corresponding coil-resistance
has to be chosen according to equation 9.24.

PSource = uSource · iSource = i2Source · RCoil (9.23)

PSource
RCoil = (9.24)
i2Source
Surprisingly from the perspective of a realistic design an increase in current is
not necessarily the preferred option to increase the L ORENTZ-force according to
equation 9.22. The possibility to optimize PLoss by adding cooling, or to analyze
the temporal pattern of on- and off-times is much more relevant. Additionally the
flux-density B has a - compared to all other influence factors - quadratically higher
influence on the maximum force.

9.2.1.3 Maximization of the Magnetic Flux-Density B

For the optimization of electrodynamic actuators a maximization of the flux density

B is necessary within the area where the conducting coils are located. This place
is usually called air-gap and resembles an interruption of the otherwise closed fer-
romagnetic core conducting the magnetic flux. The heights of the magnetic flux
density is influenced by
• the choice of ferromagnetic material for the magnetic core,
• the field winding/exciter winding of the static magnetic field, and
• the geometrical design of the magnetic core.
In the context of this book some basic design criteria for magnetic circuits are
given. For an advanced discussion and optimization process source [125] is recom-
mended.

Basics for the Calculation of Magnetic Circuits

Calculating magnetic circuits show several parallels to the calculation of electrical

networks. As shown in table 9.2 several analogies between electrical and magnetic
variables can be defined.
The direct analogy to the magnetic flux φ is the electrical charge Q. For the appli-
cation of equations however it is helpful to regard electrical currents I as counterpart
204 9 Actuator Design

Table 9.2 Analogies between electric and magnetic values.

Description Electric Magnetic
flux charge Q [C=As] magnetic flux φ [Vs]
differential flux I = dQ
dt [A]
flux-value dielectric
 flux density D [C/m2 ] flux density
 B [T=Vs/m2 ]
Q = DdA φ = BdA
A A
current 2
 density J [A/m ]
I = JdA
A

El.-mag. coupling voltage U [V] flux/ampere turns Θ [A]

formerly: Electromotive force Magnetomotive force
(e.m.f.) (m.m.f.)
induction-laws U = −N ddtφ Θ = N dQ
dt
Θ =NI
(N= turns)
field-values el. field strength E [V/m] magn. field strength H [A/m]
differential-values voltage U [V] magnetic voltage V [A]
 
U = ab Eds V = ab Hdl
mesh-equations Uges = ∑ Ui Θ = ∑ Vi
i i

resistances el. resistance R [Ω ] magn. resistance Rm [A/Vs]

reluctance
U
R= I Rm = Vφ
coupling factors permittivity ε = ε0 εr permeability μ = μ0 μr
(ε0 = 8, 854 · 10−12 C/Vm) (μ0 = 1, 256 · 10−6 Vs/Am)
coupling between D=εE B=μH
field- and flux-values
power [W] Pel = U · I
energie [J] Wel = Wel t Wmag = φ V
Wmag = ∑ Hn ln · Bn An
n

to the magnetic flux. Please note that this is an aid for thinking and not a mathemat-
ical reality, although it is very common. The actual direct analogy for the current
I would be a time dependent magnetic flux ddtφ , which is usually not defined with
an own variable name. The great exception with this model is the magnetomotive
force Θ , which resembles the sum of all magnetic voltages V identical to a rotation
within an electrical network. Nevertheless it is treated differently, as many appli-
cations require generating a magnetomotive force Θ to be generated by a certain
number of winding-turns N and a current I, often referred to as ampere turns. The
coupling between field- and flux-variables is given by the permittivity ε in case of
electrical values and by the permeability μ in case of magnetic values. It is obvious
9.2 Electrodynamic Actuators 205

that the field-constants ε0 differs from μ0 by the factor 106 . This is the main reason
for the electromagnetic effect being the preferred physical realization of actuators
in macroscopic systems7 .
However there is another speciality with the field-constants. The electrical per-
mittivity can be regarded as quite constant (section 9.5) even for complex actuator
designs, and can be approximated as linear around an operating point. The perme-
ability μr of typical flux-conducting materials however shows a strong nonlinear
relationship, the materials are reaching saturation. The level of magnetic flux has to
be limited to prevent saturation-effects in the design of magnetic core.

B [T]

1 gray cast iron

2 Sintered iron
3 X12CrMosS17 1.4104
4 X4CrMoS annealed 1.4105
5 ST37 1.0254
6 CoFe-alloy

H [A/cm]

Fig. 9.5 Saturation curve of typical magnetic materials [125].

Magnetic Circuits

For the maximization of the magnetic flux density it is necessary to either analyze
the magnetic circuit mathematical-analytically and/or do a numerical simulation of
it. For the simulation of magnetic fields common CAD and FEM products are avail-
able8 . For classification of the mathematical problem three solution levels exist:

7 In micro-mechanical systems the energy-density relative to the volume becomes more important.

The manufacture of miniaturized plates for capacitive actuators is much easier to realize with batch
processes than the manufacture of miniaturized magnetic circuits.
8 For the very beginning there are several free or open software-projects available for electrical

and magnetic field simulation, e.g. for rotatory or planar systems a program from DAVID M EEKER
named “FEMM”.
206 9 Actuator Design

stationary, quasi-stationary, and dynamic magnetic fields. With stationary magnetic

fields there is no time dependent change of the magnetic circuit. A steady state of
flux density is assumed. With quasi-stationary field the induction is being consid-
ered resulting from changes within the current generating the magnetic field or a
linearized change within the geometry of the magnetic circuit (e.g. a movement of
an anchor). Dynamic magnetic fields consider additional effects covering the dy-
namic properties of moving mechanical components up to the change of the ge-
ometry of the magnetic circuit and the air gaps during operation. By dealing with
electrodynamic actuators the analysis of static magnetic circuits is sufficient for a
first dimensioning. The relevant dynamic drawbacks for electrodynamic actuators
are presented in section 9.2.1.4.
There are two principle possibilities to generate the magnetic flux densities
within the volume of the conducting coil:

1. Generation via winded conductors with another coil (exciter winding).

2. Generation via a permanent magnet.

Both approaches show specific pros and cons: With a winded conductor the flux
density B = μ (N I − HFe lFe ) can be raised without any theoretical limit. In practi-
cal application the flux-conducting material will reach saturation (fig. 9.5) actually
limiting the achievable maximum flux density. Additionally the ohmic resistance
of the winding will generate electrical power losses, which will have to be dissi-
pated in addition to the losses resulting from the electrodynamic principle itself (sec-
tion 9.2.1.1). Abandoning any flux-conducting material and using exciter-windings
with extremely low electrical resistance extraordinary high field-densities can be
reached9 . Till now such a technological effort for haptic devices is not made yet.

Building a magnetic circuit with a permanent magnet, the practical border for
the flux density is given by the remanence flux density Br of the magnetic material.
Such a magnet can be compared to a source providing a certain magnetic power.
The flux density - being the relevant quality for electrodynamic actuators - is not
independent from the magnetic load attached to the permanent magnet. Addition-
ally the relevant properties of the magnetic material are temperature-dependent, and
wrong use of specific magnet-materials may harm its magnetic properties10 .

Nevertheless modern permanent-magnetic materials made of “rare earths” are the

preferred source to generate static magnetic fields for electrodynamic actuators. The
following section gives some basics on the calculation for simple magnetic circuits.
In extension to what is shown here a more precise analytical calculation is possible
[125]. However it is recommended to use simulation tools early within the design

9 Magnet-resonance-tomographs for medical imaging generate field densities of 2 Tesla and more
within air gaps of up to 1 m diameter by the use of supra-conducting coils and almost no magnetic
circuit at all.
10 e.g. when removing AlNiCo magnets out of their magnetic circuit after magnetization, they may

drop below their coercive field strength actually losing performance.

9.2 Electrodynamic Actuators 207

process. Especially leakage fields are a great challenge for the design of magnetic
circuits. And especially beginners should develop a feeling for the look of these
fields with the aid of simulation tools.

Direct Current Magnetic Field

Figure 9.6a shows a magnetic circuit out of iron with a cross section A and an air-
gap with the length ξG (G=Gap). The magnetic circuit has a winding with N turns
conducting a current I. The medium length of the magnetic circuit is lFe . For calcu-
lation the circuit can be transformed into a magnetic equivalent network (fig. 9.6b).
According to the analogies defined in table 9.2 the magnetic induction generates a
magnetomotive force Θ as a differential value. In combination with two magnetic
resistances of the iron circuit RmFe and the air gap RmG a magnetic flux φ can be
identified.

lFe f RmFe
I

N B xG Q RmG

a) b)

Fig. 9.6 Magnetic field generation B via a current-conducting coil with N turns (a), and derived
equivalent circuit representation (b).

For the calculation of the flux density B in the air gap, it is assumed that this
magnetic flux φ is identical to the flux within the iron part of the circuit. Leakage
fields are disregarded in this example11 .
φ
B= (9.25)
A
The magnetic resistance of materials and surfaces are dependent on the geometry
and can be found in special tables [125]. For the magnetic resistance of a cylinder
of the length l and the diameter d a resistance according to equation 9.26 is given.

11 Considering leakage fields would be identical to a parallel connection of additional magnetic

resistors to the resistance of the air gap.

208 9 Actuator Design

4l
Rm = (9.26)
μ π d2
For the magnetic circuit the magnetic resistances RmFe and RmG can be regarded
as known or at least calculable. The magnetic flux is given by
Θ
φ= , (9.27)
RmFe + RmG
and the flux density by
Θ
B= . (9.28)
(RmFe + RmG ) A
Using this procedure a clever approximation of the magnetic resistances of any
complex network of magnetic circuits can be made. In this specific case of a simple
horseshoe-formed magnet an alternative approach can be chosen. Assuming that the
magnetic flux density in the air-gap is identical to the flux density in the iron (no
leakage fields, see above) the flux-density B is given by:

B = μ0 μr H (9.29)
Assuming that μr is given either as a factor or with a characteristic curve (like in
fig. 9.5) only the magnetomotive force Θ within the iron has to be calculated. With
B B
Θ = HFe lFe + HG ξG = lFe + ξG (9.30)
μ0 μr μ0
the flux density
1
B=Θ , (9.31)
lFe
μ0 μr + ξμG0
results and can be written down immediately.

Permanent Magnets Generating the Magnetic Field

As stated earlier the typical approach to generate the magnetic field within an elec-
trodynamic actuator is the choice of a permanent magnet. Permanent magnets can
not simply be regarded as flux- or field-sources. Therefore some basic understand-
ing of magnet technology will be necessary.

As a simple approach a magnet is a source of energy which is proportional to the

volume of the magnet. Magnets are being made out of different magnetic materials
(tab. 9.3) differing in the maximum achievable flux density (remanence flux den-
sity Br ), the maximum field-strengths (coercive field strength Hc B and Hc J )), and
their energy density BHmax , such as the temperature coefficient. Additionally iden-
tical materials are differentiated according to being isotropic or anisotropic. With
isotropic magnets its substance is made of homogeneous material which can be
magnetized in one preferred direction. With anisotropic material a magnetic pow-
9.2 Electrodynamic Actuators 209

der was mixed with a binding material (e.g. epoxy) and formed via a casting or
injection-molding process. Latter approach enables almost unlimited freedom for
the magnet’s geometry and a very large influence concerning the pole-distribution
on the magnet. However anisotropic magnets are characterized by slightly worse
characteristic values in energy density such as maximum field-strengths and flux
densities.

Table 9.3 Magnetic properties of permanent-magnet materials [125]

Matrerial Br Hc B (BH)max
[T] [kA/m] [kJ/m3 ]
AlNiCo (isotrop) 0,5 ... 0.,9 10 ... 100 3 ... 20
AlNiCo (anisotrop) 0,8 ... 1,3 50 ... 150 30 ... 70
Hard ferrite (isotrop) 0,2 ... 0,25 120 ... 140 7 ... 9
Hard ferrite (anisotrop) 0,36 ... 0,41 170 ... 270 25 ... 32
SmCo (anisotrop) 0,8 ... 1,12 650 ... 820 160 ... 260
NdFeB (anisotrop) 1,0 ... 1,47 790 ... 1100 200 ... 415

Figure 9.7 shows the second quadrant of the B-H-characteristic curve (only this
quadrant is relevant for an application of a magnet within an actuator) of differ-
ent magnetic materials. The remanence flux density Br equals the flux density with
short-circuit pole shoes (a magnet being surrounded by ideal iron as magnetic cir-
cuit). When there is an air gap within the magnetic circuit (or even by the magnetic
resistance of the real magnetic circuit material itself), a magnetic field strength H
appears as a load. As a reaction an operation point is reached, which is shown here
as an example on a curve of NdFeB for a flux-density of ≈200 kA/m. The actually
available flux density at the poles is decreased accordingly. As electrodynamic ac-
tuators for haptic applications face high requirements according to their energy den-
sity, there are almost no alternatives to the usage of magnet materials based on rare
earths (NdFeB, SmCo). This is very accommodating for the design of the magnetic
circuit, as nonlinear effects near the coercive field strength such as with AlNiCo or
Barium-ferrite are of no relevance12 . Rare earth magnets allow an approximation of
their B/H-curve with a linear equation, providing a very nice relationship for their
magnetic resistance (fig. 9.8c):

V Hc lMag
RMag = = (9.32)
φ Br A

12 The small coercive field strength of these materials e.g. result in the effect, that a magnet magne-
tized within a magnetic circuit does not reach its flux density anymore once removed and even after
re-assembly into the circuit again. This happens due to the temporary increase of the air gap, which
is identical to an increase of the magnetic load to the magnet beyond the coercive field strength.
Additionally the temperature-dependency of the coercive field strength and of the remanence flux
density is critical. Temperatures just below the freezing point may result in a demagnetization of
the magnet.
210 9 Actuator Design

B [T]
1 Barium-ferrite
2 AlNiCo 500 Br
3 NdFeB resin bonded
4 SmCo5
5 Sm2Co17
6 NdFeB (N35)
7 NdFeB (N50)

H [kA/m] Hc

Fig. 9.7 Demagnetization curves of different permanent-magnet materials [125].

With this knowledge the magnetic circuit of figure 9.8a and the corresponding
equivalent circuit (fig. 9.8b) can be calculated identical to an electrically excited
magnetic circuit.

N lMag
S
lFe c)

N f RmFe
lMag B xG
S
HClMag RmG

RMag
a) b)

Fig. 9.8 Magnetic field generation B via permanent magnets (a), derived equivalent circuit (b), and
dimensions of the magnet (c).

The flux density within the iron is once again given by

φ
B= (9.33)
A
For the given magnetic circuit the resistances RmFe and RmG are assumed as
known or calculable. From equation 9.32 the magnetic resistance of the permanent
9.2 Electrodynamic Actuators 211

magnet is known. The source within the equivalent circuit is defined by the coercive
field strength and the length of the magnets Hc lMag . These considerations result in
Hc lMag
φ= , (9.34)
RmFe + RmG + RMag
and the flux density
Hc lMag
B= . (9.35)
(RmFe + RmG + RMag ) A
Slightly rearranged and RMag included gives
lMag
Br Hc A
B= lMag
. (9.36)
(RmFe + RmG ) BR + Hc A

l
Equation 9.36 states by the factor Br Hc Mag
A that it is frequently very helpful for
achieving a maximum flux density B in the air gap to increase the length of a magnet
with at the same time minimized cross-sectional area of the magnetic circuit - which
is of course limited by the working distance within the air gap and the saturation
field strengths of the magnetic circuit.

9.2.1.4 Additional Effects in Electrodynamic Actuator

To do a complete characterization of an electrodynamic actuator there are at least

three more effects, whose influences will be sketched within the following para-
graphs.

Induction

For a complete description of an electrodynamic actuator beside the geometrical

design of its magnetic circuit and the mechanical design of its winding and the
considerations concerning the electrical power losses, its other dynamic electrical
properties have to be considered. For this analysis the electrodynamic actuator is
regarded as a bipolar transformator (fig. 9.9).
A current i0 generates via the proportional constant B l a force F 0 , which moves
the mechanical loads attached to the actuator. The movement itself results in a ve-
locity v0 which is transformed via the induction law and the proportional constant
to an induced voltage u1 . By measurement of u1 and a current source the rotational
velocity or the movement velocity v can be measured, with a voltage source the
measurement of i0 provides a force- or torque-proportional signal. This is the ap-
proach taken by the variant of admittance controlled devices as a control value (see
section 5.1.4).
212 9 Actuator Design

i0 R L F0
F0=B l i0 k
m d
u0 u1 v0
u1=B l v0

Fig. 9.9 Electrical and mechanical equivalent circuit of an electrodynamic actuator as being a
transformer.

The induction itself is a measurable effect, but should not be overestimated. Typ-
ically electrodynamic actuators are used within haptic systems as direct drives at
small rotational or translational velocities. Typical coupling factors with rotatory
drives are - depending on the size of the actuators - in an area between 100 to 10
revolutions
sV . At a rotational speed which is already fast for direct drives of 10 Hz, in-
duced voltage amplitude |u1 | of 0.1 to 1 V can be achieved. This is around 1% to
5% of the control voltage’s amplitude.

Electrical Time Constant

Another aspect resulting from the model according to figure 9.9 is the electrical
transfer characteristics. Typical inductances L of electrodynamic actuators lie in the
area of 0.1 mH to 2 mH. The ohmic resistance of the windings is largely depending
on the actual design, but as a rule of thumb values between 10 Ω to 100 Ω can
i
be assumed. The step-response of the electrical transfer system u0 shows a time-
0
constant τ = RL = 10 μ s to 30 μ s and lies within a frequency range »10 kHz, which
is clearly above the relevant dynamic area of haptics.

Field Response

A factor which can not so easily be neglected when using electrodynamic actuators
for high forces is the feedback of the magnetic field generated by the electromag-
netic winding on the static magnetic field. Taking the actuator from the example at
the beginning (fig. 9.3 on page 200) positive currents generate a field of opposite
direction to the field generated by the magnet. This influence can be considered by
substitution of both field sources. Depending on the direction of current this field
either enforces or weakens the static field. With awkward dimensioning this can
result in a directional variance of the actuator properties. The problem is not the
potential damage to the magnet, modern magnetic materials are sufficiently stable,
but a variation of the magnetic flux density available within the air gap. An intended
application of this effect within an actuator can be found in an example according
9.2 Electrodynamic Actuators 213

to figure 9.29 on page 235.

A deeper discussion about electrodynamic actuators based on concentrated ele-

ments can be found in [158].

9.2.2 Actual Actuator Design

As stated earlier electrodynamic actuators are composed of three basic compo-

nents: coil/winding, magnetic circuit, and magnetic exciter. The following section
describes a procedure for the design of electrodynamic actuators based on these
basic components. As the common principle for excitation a permanent magnet is
assumed.

9.2.2.1 Actuator Topology

The most fundamental question for the design of an electrodynamic actuator is its
topology. Usually it is known whether the system shall perform rotary or translatory
movements. Afterward the components magnetic circuit, the location of magnets,
pole-shoes and the coil itself can be varied systematically. A few quite common
structures are shown in figure 9.11 for translational actuators, and in figure 9.10
for rotatory actuators. For the design of electrodynamic actuators in any case the
question should be asked, whether the coil or the magnetic circuit shall move. By
this variation apparently complex geometrical arrangements can be simplified dras-
tically. Anyway it has to be considered that a moving magnet has more mass and
can typically be moved less dynamically than a coil. On the other hand there is no
contact- or commutating problem to be solved with non-moving windings.

Moving Coils

Electrodynamic actuators according to the principle of moving coils with a fixed

magnetic circuit are named “moving coil” in the case of a linear movement and
“iron-less rotor” in the case of a rotatory actuator. They always combine few moving
masses and as a result high dynamics. The translatory version shows displacements
of a few millimeters, and is used especially within audio applications as loudspeaker.
Actuators according to the principle of “moving coils” have two disadvantages:
• As the coil is moving, the electrical contact is subject to mechanical stresses.
Especially with high displacements the contact has to be mechanically robust.
• If there is the idea to design moving coils as pure force sources with large dis-
placements, always only a small area of the conducting coil is within the air-
gap and therefore contributes to the force generation. With large displacements
214 9 Actuator Design

FLorent z FLorent z FLorent z

FLorent z
pole shoe

coil

S
magnetic
yoke S S S N S N
N

N N

a) b) c) d)

Fig. 9.10 Variants of electrodynamic actuators for translational movement with moving magnets
(a), moving coils (b), as plunger-type (c), and as flat-coil (d).

rotating, self supported

rotating, self supported coil
coil
I

S
N N

magnetic
yoke fixed magnet S
a) b) N

Fig. 9.11 Variants of electrodynamic actuators for rotatory movements with self-supportive wind-
ing (a), and with disc-winding.

moving coils show an even lower efficiency factor. This can be compensated by
switching the active coil areas, which again results in the necessity to have more
contacts.
A similar situation happens with rotatory systems. Based on the electrodynamic
principle there are two types of windings applicable to rotatory servo-systems: the
FAULHABER and the M AXON-winding of the manufacturers with identical names.
These actuators are also known as “iron-less” motors. Both winding principles al-
low the manufacture of self-supportive coils. A diagonal placement of conductors
and a baking process after winding generates a structure sufficiently stable for the
9.2 Electrodynamic Actuators 215

centrifugal forces during operation. The baked coils are connected with the rotating
axis via a disk. The complete rotor (fig. 9.12) is build of these three components.
By the very small inertia of the rotor such actuators show impressive dynamic prop-
erties. The geometrical design allows placing the tubular winding around a fixed
diametral-magnetized magnet. This enables another volume reduction compared to
conventional actuators as its housing has to close the magnetic circuit only instead
of providing additional space for magnets.

Connector
cables

Cap of brushes
Gear 2

Housing
Commutation-
carrier
self supporting
Korbwicklung
winding

spindle

Permanent-
magnet

Gear 1

Fig. 9.12 Design of an electrodynamic actuator with self-supportive winding according to the
FAULHABER-principle [97].

Within the self-supportive winding there are areas of parallel lying conductors
combined to poles13 . With moving coils there is always the need for a special-
ized contactor, either via contact rings, or electronic commutation or via mechani-
cal switching. Depending on the number of poles all coils are contacted at several
points. In case of mechanical switching these contacts are placed on the axis of the
rotor and connected via brushes with the fixed part of the actuator named “stator”.
This design enables a continuous movement of the rotor, whereas a change of the
current flow is made purely mechanically by the sliding of the brushes on the contact
areas of the poles on the axis. This mechanical commutation is a switching proce-
dure with an inductance placed in parallel.
As such an actuator can be connected directly to a direct current source, they are
known as “DC-drives”. As stated within section 9.1 the term “DC-drive” is not only
limited to actuators according to the electrodynamic principle but is also frequently
applied to actuators following the electromagnetic principle (section 9.3).

13 The FAULHABER and the M AXON excel by a very clever winding technique. On a rotating
cylinder respectively a flatly pressed rectangular winding poles can be combined by contacting
closely located areas of an otherwise continuous wire.
216 9 Actuator Design

Moving Magnet

In case of translatory (fig. 9.10a) systems actuators according to the principle of a

moving magnet are designed to provide large displacements with compact wind-
ings. The moving part of the actuator is composed almost completely of magnetic
material. The polarity direction of this material may vary in its exact orientation.
Actuators according to this principle are able to provide large power, but are ex-
pensive due to the quantity of magnet material necessary. Additionally the moving
magnet is heavy; the dynamics of the actuator is therefore smaller than in the case
of a moving coil.
In case of a rotatory system a design with moving magnet is comparable to a design
with moving coil. Figure 9.13 shows such a drive. The windings fixed to the stator
are placed around a diametral magnetized magnet. It rotates on an axis, which fre-
quently additionally moves the magnetic circuit too. Providing the right current feed
to the coil the orientation of the rotor has to be measured. For this purpose sensors
based on the Hall-effect or optical code wheels are used.

digital
stationary rotating
hall-sensors
self-supported winding iron core
i1

windings- Permanent magnet i3

connectors glued on shaft
i2

Fig. 9.13 Components of a EC-drive and equivalent circuit.

Electrodynamic actuators with moving magnet are known as EC-drives (electronic-

commutated). This term is not exclusive to electrodynamic actuators, as there are
electronic-commutated electromagnetic drives too. EC-drives - whether they are
electrodynamic or electromagnetic - combined with the corresponding driver elec-
tronics are frequently known as servo-drives. Typically a servo-drive is an actuator
able to follow a predefined movement path. Servo-drives are rarely used for haptic
devices. However the usage of EC-drives for haptic application is very frequent, but
then they are equipped with specialized driver electronics.

9.2.2.2 Commutation in the Context of Haptic Systems

The necessary commutation of the current for rotating actuators has a big influence
upon the quality of force- respective torque-output.
9.2 Electrodynamic Actuators 217

Mechanically Commutating Actuators

With mechanically commutating actuators the current flow is interrupted suddenly.

Two effects of switching contacts appear: The voltage at the contact point increases,
sparks may become visible - an effect which is called electrical brush sparking. Ad-
ditionally the remaining current flow induces a current within the switched-off part
of the winding which itself results in a measurable torque. Figure 9.14b shows a
measurement of a force acting upon a rod driven by a mechanically commutated
actuator according to figure 9.14a. The force ripples excite the base force of 1.3 N
by≈ 100 mN, which is around 8%. A transformation of the measurement into the
perceptional space K according to the concept of quantification haptic perception
(section 4.3) amplifies the difference even further. The perception-peaks resulting
from commutation exceed the continuous perception by 25%. This amplification is
a result from the lower frequencies being damped whereas higher frequency com-
ponents are enhanced. The high frequencies of the commutating flanks dominate
disproportionately high haptic perception.

2 2x10
-6

1.8
Fab [fOsc]
Fab [N]

1.6
Fab
1.4

Travel 1.2
MMotor
1 1x10-6
10 20 30 Travel [mm]
a) b) c) 10 20 30 Travel [mm]

Fig. 9.14 Measured and perceived force of a mechanically commutated actuator.

a) Measurement setup, b) measured force plot, c) perceived force according to the FIP-method.

The current- and torque changes can be reduced by the inclusion of resistors
and capacitors into the coil. However this results into high masses of the rotor and
worse dynamic properties. Beside that a full compensation is impossible. Neverthe-
less mechanically commutating actuators are in use for inexpensive haptic systems.
The Phantom Omni from SensAble and the Falcon from Novint use such actuators.

Electronic Commutated Electrodynamic Actuators

Electronic commutated electrodynamic actuators differ from mechanically commu-

tated actuators by the measurement technology used as a basis for switching cur-
rents. There are four typical designs for this technology:
218 9 Actuator Design

• In sensor-less designs (fig. 9.15a) an induced voltage is measured within a coil.

At zero-crossing point one pole is excited with a voltage after an interpolated 30◦
phase delay dependent on the actual revolution speed of the rotor. In combination
of measurement of the inductance followed by a switched voltage, a continuous
rotation with batch-wise excitation is realized. This procedure can not be applied
to low rotation speeds, as the induced voltage becomes too low and accordingly
the switching point can hardly be interpolated. Additionally the concept of using
one to two coils for torque generation results in a high torque variations at the
output of up to 20%, making this approach not useful for haptic systems.
• Block-commutating procedures (fig. 9.15b) are based on the usage of simple hall-
switches or field-plates for position detection of the rotor. Three sensors located
at 120◦ angular phase shift allow the detection of six different rotor positions.
Reducing positioning information to six orientations per revolution makes this
approach equally inappropriate for haptic applications, as the torque varies in a
range of >15% for one revolution.
• Sinus-commutating procedures with analogue hall-sensors are based on the mea-
surement of the rotor position by at least two sensors. They are placed with an
angle of either 120◦ or 90◦ at the front of the rotor. They provide voltages in
an angular phase shift according to their geometrical position. By analyzing the
polarity and the absolute height of the voltages absolute positioning information
can be obtained and used for commutating the windings. If the phase lag between
both sensor signals is identical to the phase lag between the poles of the wind-
ing a direct control of current-drivers can be performed without the need for a
digitalisation or a specific calculation step.
• Sinus-commutating with digital code-wheels (fig. 9.15c) are based on the mea-
surement of rotor position by the use of - usually optical - code discs. By reflec-
tive or transmittive measurement the rotor position is sampled with high resolu-
tion (e.g. 128 positions / turn). This relative positioning information can be used
for position measurement after an initial calibration. Depending on the code-
wheels resolution a very smooth sinoid commutation can be achieved with this
method.

The sinus-commutating methods are the preferred solutions used for haptic appli-
cations due to the little torque variations and their applicability for slow revolution
speeds typical to direct drives.

9.2.3 Actuator Electronics

Electrodynamic actuators require some specific electrical circuits. In the following

section the general requirements on these electronics are formulated.
9.2 Electrodynamic Actuators 219

Hall-Sensor 1
Uind Hall-Sensor 2

Hall-Sensor 3

2 U 1-2
a) 3
U 2-3

U 3-1

Hall-Sensors

c) b)
N

Fig. 9.15 Technologies for different commutation methods: sensor-less (a), block-commutation
(b) and optical code-wheel (c)

9.2.3.1 Driver Electronics

Driver electronics are electrical circuits transforming a signal of low power (several
volts, some milli-ampere) into a voltage- or current level appropriate to drive an
actuator. For electrodynamic actuators in haptic applications driver electronics have
to provide a current in a dynamic range from static up to several kilohertz. This
paragraph describes general concepts and approaches for such circuits.

Topology of Electric Sources

Driver electronics for actuators - independently from the actuation principle they are
used for - are classified according to the flow of electrical energy (fig. 9.16). There
are four classes of driver electronics:
220 9 Actuator Design

• 1-quadrant controllers are capable of generating positive output currents and volt-
ages. An actuator driven by them is able to move in one direction. These con-
troller use only the first quadrant according to figure 9.16a.
• Switched 1-quadrant controllers are capable of a direction change by the input of
a logical signal. They are working within the 1st and 3rd quadrant according to
figure 9.16a. The switching point is a nonlinear step in their characteristic curve.
• Real 2-quadrant controllers are capable of providing a characteristic curve which
is steady around the zero point. They function in the 1st and 3rd quadrant ac-
cording to figure 9.16a, but are not capable to conduct currents and voltages with
opposite directions.
• 4-quadrant controllers function within all four quadrants of figure 9.16a. They
are able to control currents and voltages in any combination of directions. Four-
quadrant controllers allow energy recovery by induced currents to an energy stor-
age, which is especially relevant for mobile applications.

For haptic application the switched 1-quadrant controller is frequently met, as

many haptic systems do not have the necessity to control the device near the voltage-
or current-zero point. However for systems with high dynamics and low impedance
the 2-quadrant and the 4-quadrant controller are relevant, as the unsteadiness near
the zero-point is perceivable with high quality applications.

2-quadrant controller
U
2. quadrant 1. quadrant switched U
generator mode motor mode 1-quadrant controller
direction 1 direction 1 I
U
I M
I
3. quadrant 4. quadrant
motor mode generator mode
direction 2 direction 2

a) b)

Fig. 9.16 Visualization of the four quadrants of an electric driver, formed by the directions of
current and voltage

Pulse-Width-Modulation and H-Bridges

With the exception of some telemanipulators, the sources controlling the actuators
are always digital processors. As actuators need an analogue voltage or current to
generate forces and torques some transformer between digital signals and analogue
control value is necessary. There are two typical realizations of these transformers:
9.2 Electrodynamic Actuators 221

1. Usage of a digital-analog converter (D/A-converter)

2. Usage of a pulse-width-modulation (PWM)
The use of D/A-converters as external components or integrated within a micro-
controller is not covered further in this book, as it is, if necessary to use, extremely
simple. It just requires some additional efforts in circuit layout. Latter results in it
being not used much for the control of actuators.

With electrodynamic actuators the method of choice are driver electronics based
on PWM (fig. 9.17a). With the PWM a digital output of a controller is switched with
a high frequency (>10 kHz14 ). The period of the PWM is given by the frequency.
The program controls the duty cycle between on- and off-times. Typically one byte
is available to provide a resolution of 256 steps within this period. After filtering the
PWM, either via an electrical low-pass or via the mechanical transfer-characteristics
of an actuator, a smoothed output signal becomes available.
Pulse-width-modulation is frequently used in the combination with H-bridges
(fig. 9.17b). The term H-bridge results from the H-like shape of the motor sur-
rounded by four switches. The H-bridge provides two operation modes for two
directions of movement and two operation modes for braking. If according to fig-
ure 9.17b the two switches S2 and S5 are on, the current I will flow through the
motor in positive direction. If instead switches S3 and S4 are switched on, the cur-
rent I will flow through the motor in negative direction. One additional digital signal
acting upon the H-bridge will change the direction of movement of the motor. This
is the typical procedure with switched 1-quadrant controllers. Additional switching-
states are given by switching the groups S2 and S3 respectively S4 and S5. Both
states results in short-circuit of the actuator and stops its movement. Other states
like simultaneously switching S2 and S4 respectively S3 and S5 results in short-
circuit of the supply voltage, typically destroying the integrated circuit of the driver.
To combine the H-bridge with a PWM either switch-groups S2 and S5 can be
switched according to the timing of the PWM, or additional switches S1 (fig. 9.17c)
can be placed in series to the H-bridge modulating the supply voltage U. In a prac-
tical realization latter is the preferred design, as the timing of the switches S2 to S5
is very critical to prevent likely short circuits of the supply voltage. The effort to
perform this timing between the switching is usually higher than the costs of an-
other switch in series. The practical realization of H-bridges is done via field-effect
transistors. The discrete design of H-bridges is possible, but not easy. Especially the
timing between switching events, the prevention of short-circuits, and the protec-
tion of the electronics against induced currents is not trivial. There are numerous
14 Typical frequencies lie in between 20 kHz to 50 kHz. However especially within automotive
technology for driving LEDs, PWMs for current drivers with frequencies below 1 kHz are in ap-
plication. Frequencies within this range are not applicable to haptic devices, as the switching in
the control value may be transmitted by the actuator and will therefore be perceivable especially in
static conditions. Typical device designs show mechanical low-pass characteristics even at frequen-
cies in the area of 200 Hz already. However due to the sensitivity of tactile perception in an area of
100 to 200 Hz, increased attention has to be paid on any switched signal within the transmission
chain.
222 9 Actuator Design

Umax

Uout/Uin
µC U
f

t
a) Iout
S1

S2 S3 S2 S3
I I
U M U M

S4 S5 RSense S4 S5

b) c)

USense ~ I

Fig. 9.17 Principle of puls-width-modulation (PWM) at a digital μ C-output (a), h-bride circuit
principle (b), and extended h-bridge with PWM (S1) and current measurement at (RSense )(c)

integrated circuits available at the market which already include appropriate pro-
tective circuitry and provide only a minimum of necessary control lines. The ICs
L6205 (2A), L293 (2.8A) and VNH 35P30 (30A) are some examples common with
test-bed developments. For EC drives there are specific ICs performing the timing
for the field-effect transistors and reducing the number of necessary PWMs from
the microcontroller. The IR213xx series switches three channels with one external
half-bridge per channel built up from N-MOS transistors with a secure timing for
the switching events.

The PWM described above with an H-bridge equals a controlled voltage source.
For electrodynamic systems such a control is frequently sufficient to generate an
acceptable haptic perception. Nevertheless for highly dynamic haptic systems a
counter induction (section 9.2.1.4) due to movement has to be expected, resulting in
a variation of the current within the coils generating an uncontrolled change of the
L ORENTZ-force. Additionally the power-loss within the coils (paragraph 9.2.1.1)
may increase the actuator’s internal temperature resulting in a change of conductiv-
ity of the conductor’s material. The increasing resistance with increasing tempera-
tures of the conductor results in less current flow at a constant voltage source. An
electrodynamic actuator made of copper as conductive material would generate a re-
duced force when operated. With higher requirements on the quality of haptic output
a controlled current should be considered. In case of a PWM a resistor with low re-
9.2 Electrodynamic Actuators 223

sistance (RSense in fig. 9.17c) has to be integrated, generating a current-proportional

voltage USense , which itself can be measured with an A/D input of the controller.
The control circuit is closed within the microcontroller. However the A/D transfor-
mation and the closing of the control circuit can be challenging for state of the art
electronics with highly-dynamic systems with border frequencies of some kilohertz.
Therefore analog circuits should be considered for closed-loop current controls too.

Analogue Current Sources

Analogue current sources are - to make it simple - controlled resistors within the
current path of the actuator. Their resistance is dynamically adjusted to provide the
wished current flow. Identical to classical resistors analogue current sources trans-
form the energy which is not used within the actuator into heat. Consequently in
comparison to the switched H-bridges they are generating a lot of power loss. By the
use of a discrete current control (fig. 9.18a) analogue current sources for almost any
output currents can be built by the choice of one to two field-effect-transistor (FET).
For heat dissipation they are required to be attached to adequate cooling elements.
There are only little requirements on the operational amplifiers themself. They con-
trol the FET within its linear range proportional to the current-proportional-voltage
generated at RSense . Depending on the quadrant used within operational mode (1 or
3) either the N-MOS transistors or the P-MOS transistor is conductive. An alterna-
tive to such discrete designs is the use of power-amplifiers (e.g. LM675, fig. 9.18b).
It contains fewer components and is therefore less dangerous to make errors. Real-
ized as non-inverting or inverting operational amplifier with a resistor for measure-
ment RSense , they can be regarded as a voltage controlled current source.

U+
RSense +
10k

- U-
PMOS
+ -
1k
+
1k U+ RSense = 4 W
+ Iout Uin 10k
Uin
Iout
- NMOS M
1k M

a) RSense - b)
U-

Fig. 9.18 Discrete closed-loop current control [255] (a), and closed-loop current control with a
power-operational-amplifier (b).
224 9 Actuator Design

9.2.3.2 Monitoring Temperature

Resulting from the low efficiency factor and the high dissipative energy from elec-
trodynamic actuators it is useful to monitor the temperature nearby the coils. Instead
of including a measuring resistor PT100 nearby the coil, another approach monitors
the electrical resistance of the windings themselves. Depending on the material of
the windings (e.g. cooper, Cu) the conductivity changes proportional to the coil’s
temperature. With copper this factor is 0.39% per Kelvin temperature change. As
any driver electronics either works with a known and controlled voltage or current,
measurement of the other component immediately provides all information to cal-
culate resistance and consequently the actual coil temperature.

9.2.4 Examples for Electrodynamic Actuators in Haptic Devices

Electrodynamic actuators are most frequently used as force and torque sources
within kinaesthetic systems. Especially EC-drives can be found in the products of
Quanser, ForceDimensions, Immersion, and SensAble. Mechanically commutated
electrodynamic actuators are used within less expensive devices, like the Phantom
Omni or the Novint Falcon. For tactile applications electrodynamic actuators ap-
pear only with linear moving coils or magnets as oscillation sources. The possibility
to control frequency and amplitude independently from each other makes them in-
teresting for tactors. Tactors (fig. 9.19) are small, disk-like actuators which can be
integrated e.g. in clothes or mobile devices to transmit information via tactile stim-
ulation in small areas of the skin.

Fig. 9.19 Electrodynamic tactor by Audiological Engineering Inc., with a frequency range from
100 Hz to 800 Hz.
9.2 Electrodynamic Actuators 225

9.2.4.1 Cross-Coil System

Beside self supportive coils electrodynamic actuators according the design of cross
coils are one possibility to generate defined torques. Continental VDO developed a
haptic rotary actuator device being a central control element for automotive appli-
cations (fig. 9.20). It contains a diametral magnetized NdFeB-magnet. The magnet
is surrounded by a magnetic circuit. The field-lines reach from the magnet to the
magnetic circuit. The coils surround the magnet in an angular phase of 90◦ , and the
electrodynamic active winding section lies in the air-gap between magnetic circuit
and magnet. The rotary position control is made via two hall-sensors placed in a
90◦ position. The actuator is able to generate a ripple-free torque of ≈25 mNm at a
geometrical diameter of 50 mm, which is additionally increased by an attached gear
to ≈100 mNm torque output.

cross-coil winding
magnetic yoke

magnet
N S

conductor
board
hall-sensors

Fig. 9.20 Electrodynamic cross-coil system with moving magnet as haptic rotary actuator.

9.2.4.2 H ANNAFORD- Flat-Coil System

The flat coil system (fig. 9.21) designed from the workgroup of H ANNAFORD is part
of the fingertip-haptic-display. The alloy wire (reduced specific resistance compared
to copper) is wound in a coil in one plane and fixated by a backing process. For
better conductance of the thermal loss it is integrated in a carrier made of alloy,
being pivot in one axis. The magnetic circuit is closed above two blades of the coil
with a diametral magnetized thin magnet. The system is able to generate torques of
160 mNm and more. The power management to dissipate the thermal loss and the
analysis for convection is detailed in [160].
226 9 Actuator Design

magnetic poles

winding N axis
carrier for
windings I

Fig. 9.21 Electrodynamic flat-coil system for the generation of forces a the finger-tips[160].

9.2.4.3 K UNSTMANN-Telemanipulator

The electrodynamic haptic device of the K UNSTMANN telemanipulator is part of

a micro-assembly workplace. It aims at generating forces to two finger tips, and
provides one additional force for gravity generation. A solution (fig. 9.22) of two
rectangular wounded flat coils moving in an iron brace was chosen. Via diametral
magnetized thin magnets a field normal to the brace is generated. Therefore the long
sides of the coils are active for force generation, as they lie within the air gap. Forces
up to 3 N can be generated by this device in continuous operation.

Magnetpole
N
S

N
S
Spule Drehachse
Eisenrückschluss

Fig. 9.22 Electrodynamic rectangular flat-coil system similar to a moving coil [147].

9.2.4.4 D OERRER-HapKeys

The HapKeys from D OERRER have been mentioned earlier in section 2.2.5.2 . The
electrodynamic linear actuators building the basis of this device are equipped with
9.2 Electrodynamic Actuators 227

friction type bearings, and moving magnets with pole-shoes within cylindrically
wound fixed coils. The coils have an inner diameter of 5.5 mm and an outer diameter
of 8 mm. The magnetic circuit is decoupled from other nearby elements within the
actuator-array. It is made of a tube with a wall thickness of 0.7 mm of a cobalt-iron
alloy with very high saturation flux density. Each actuator is able to generate 1 N in
continuous operation mode.

brass
N
S linear guides

flux conductors
windings
(4x)
N
S
magnets (3x)
NeFeBr

Fig. 9.23 Electrodynamic linear actuator with moving magnet [46].

9.2.5 Conclusion about the Design of Electrodynamic Actuators

Electrodynamic actuators are the preferred actuators used for kinaesthetic impedance-
controlled haptic devices due to their proportional correlation between the control
value “current” and the output-values “force” or “torque”. The market of DC -and
EC-drives offers a wide variety of solutions, making it possible to find a good com-
promise between haptic quality and price for many applications. If there are special
requirements to be fulfilled, the design, development, and start of operation of spe-
cial electrodynamic actuator variants are quite easy. The challenges by thermal and
magnetic design are manageable, as long as some basic considerations are not for-
gotten. The examples of special haptic systems seen in the preceding section prove
this impressively. Just driver electronics applicable to haptic systems and its re-
quirements are still an exceptional component within the catalogs of manufactures
from automation-technology. They must either be paid expensively or be built by
oneself. Therefore commercial manufacturers of haptic devices, e.g. Quanser, offer
228 9 Actuator Design

their haptic-applicable driver electronics independent from the own systems for sale.

For the design of low-impedance haptic systems currently no alternative to elec-

trodynamic systems is at hand. Other actuation principles which are discussed
within this book need a closed-loop control to overcome their inner friction and
nonlinear force/torque-transmission. This always requires some kind of measure-
ment technology such as additional sensors or the measurement of inner actuator
states. The efforts connected with this are still a big advantage for electrodynamic
actuators, which is gained by a low efficiency factor and as a consequence the rela-
tively low energy density per actuator-volume.
9.3 Electromagnetic Actuators 229

9.3 Electromagnetic Actuators

T HORSTEN A. K ERN

Electromagnetic actuators are the most frequently used actuator type within gen-
eral automation-industry. Due to their simple manufacture and assembly they are
the matter of choice. Additionally they do not necessarily need a permanent magnet,
and their robustness against exterior influences is very high. They are used within
coffee-machines, water pumps and for redirecting paper flow within office printers.
But nevertheless their applicability for haptic devices, especially kinaesthetic de-
vices, is limited. Their main fields of application are tactile systems. This can be
reasoned by several special characteristics of the electromagnetic principle. Within
this chapter the theoretical basics for electromagnetic actuators are given. Techni-
cal realizations are explained with examples, whereas first the general topology and
later the specific designs are shown. The chapter closes with some examples of hap-
tic applications of electromagnetic actuators.

9.3.1 Magnetic Energy

The source responsible for the movement of a magnetic drive is the magnetic energy.
It is stored within the flux-conducting components of the drive. These components
are given by the magnetic core (compare paragraphs 9.2.1.3 on page 203) and the
air gap, such as all leakage fields - which are neglected for the following analysis.
It is known from table 9.2 on page 204, that stored magnetic energy is given by the
products of fluxes and magnetic voltages in each element of the magnetic circuit:

Wmag = ∑ Hn ln · Bn An (9.37)
n

As every other system does, the magnetic circuit tries to minimize its inner en-
ergy15 . Concentrating on electromagnetic actuators, the minimization of energy al-
most always refers to the reduction of the air gap’s magnetic resistance RmG . For
this purpose two effects may be used, which can be found within electrostatic for
electrical fields too (section 9.5):
• Electromagnetic longitudinal effect (fig. 9.24a) (also: reluctance effect)
• Electromagnetic transversal effect (fig. 9.24b)

The forces respectively torques generated with the individual effects are the
derivations of the energy according to the corresponding direction,

15 Minimizing potential energy is the basis for movements in all actuator principles. Actuators may

therefore be characterized as “assemblies aiming at the minimization of their inner energy”.

230 9 Actuator Design

xL
I I
xQ
N N B
B

a) b)

Fig. 9.24 Electromagnetic transversal- (a) and longitudinal-effect (b).

d Wmag
Fξ = , (9.38)
dξ
being equal to a force in the direction of the change of the air gap
1 d RmG
Fξ = − φ 2 . (9.39)
2 dξ

Example: Transversal effect

The magnetic resistance of an arbitrary homogenous element of length l between

two boundary surfaces (fig. 9.25a) with the surface A is calculated as
l
Rm = . (9.40)
μA
This gives the stored energy Wmag within the magnetic resistance:

l
Wmag = (B A)2 . (9.41)
μA
The flux density B is dependent on the length of the material. Assuming that the
magnetic core contains one material only the magnetomotive force Θ is calculated
as
B
Θ= = N I, (9.42)
μl
which gives
μ
B =NI. (9.43)
l
This equation used to replace the flux density in equation 9.41, several variables
canceled, finally results in the magnetic energy
9.3 Electromagnetic Actuators 231

1
Wmag = (N I)2 Aμ . (9.44)
l
With the assumption about the magnetic energy concentrating within the air gap
- which is identical to the assumption that the magnetic core does not have any
relevant magnetic resistance - the approximation of the force for the transversal
effect in the direction of l can be formulated
1 1
Fl = − (N I)2 Aμ 2 . (9.45)
2 l
The force shows an anti-proportional quadratic coherence (fig. 9.25b) to the dis-
tance l. The closer the poles become, the higher the force attracting the poles in-
creases.

Fl

B l

a) b)
l

Fig. 9.25 Electromagnetic transversal effect in the air-gap (a) and with a qualitative force plot (b).

Example: Longitudinal effect, reluctance effect

The same calculation can be repeated for the longitudinal effect. Assuming that the
surfaces A from equation 9.44 is rectangular and its edges’ lengths are given by a
and b, further assuming that a flux-conducting material is inserted into direction a,
the forces in longitudinal direction can be calculated as
1
Fa = (N I)2 bμ , (9.46)
l
and in direction b as
1
Fb = (N I)2 aμ . (9.47)
l
The reluctance effect is - in contrast to the transversal effect - linear (fig. 9.26b).
The force is dependent on the length of the mobbing material’s edge only. Conse-
quently the stored energy within the magnetic circuit is necessary for the design
of an electromagnetic actuator. Above examples have the quality of rough estima-
232 9 Actuator Design

tions. They are sufficient to evaluate the applicability of an actuation principle -

no more, no less. The magnetic networks sufficient for a complete dimensioning
should contain effects with magnetic leakage fields and the core’s magnetic resis-
tance. Therefore it is necessary to further deal with the design of magnetic circuits
and their calculation.

b
a

B Fa, b

l x

a) b)
x

Fig. 9.26 Electromagnetic longitudinal effect in the air gap (a) and as qualitative force plot (b).

9.3.2 Design of Magnetic Circuits

The basic interdependencies for the design of magnetic circuits have already been
discussed within paragraph 9.2.1.3 in the context of electrodynamic actuators. Taken
from the approach of longitudinal- and transversal effect, several basic shapes
(fig. 9.27) can be derived applicable to electromagnetic actuators. In contrast to
electrodynamic actuators the geometrical design of the air gap within electromag-
netic actuators is freer. There is no need to guide an electrical conductor within the
air gap anymore. Beside the designs shown in figure 9.27 there are numerous other
geometrical variants too. For example all shapes can be transferred into a rotational-
symmetrical design around one axis. Additional windings and even permanent mag-
nets can be added. There are just two limits to their design:
• A sufficient cross section of the flux conducting elements has to be guaranteed,
to prevent the material from getting into saturation.
• A sufficient distance between flux-conducting elements has to be kept to prevent
a magnetic leakage-field via the air.

9.3.2.1 Cross Section Surface Area - Rough Estimation

The calculation of the cross section surface area for dimensioning the magnetic
core is simple. A common, easily available and within precision engineering and
9.3 Electromagnetic Actuators 233

Fig. 9.27 Basic shapes of electromagnetic actuators.

prototype-design gladly used material is steal ST37. The B/H-characteristic curve

with its saturation is given in figure 9.7 on page 210. For this example we choose
a reasonable flux density of 1.2 T. This equals a field intensity of H ≈1000 A/m.
Within the air gap a flux density of 1 T should be achieved. The magnetic flux within
the air gap is given as

φ = AG BG . (9.48)
As the magnetic flux is conducted completely via the magnetic core - neglecting
leakage fields and other side bypasses - the relation

AIron BIron = AG BG , (9.49)

is given, and consequently with the concrete values from above:

AIron BG
= = 0.833 . (9.50)
AG BIron
At its tightest point the magnetic core may have 83% of the cross section of the
air gap. Whereas more surface of the cross-section results in lower field intensities,
which should be aimed at if is geometrically possible. Please note that AG ≤ AIron is
with almost all technical realization, as the boundary surface of the magnetic core
is always one pole of the air gap.
234 9 Actuator Design

9.3.2.2 Magnetic Energy in the Magnetic Core and Air Gap

Within the preceding examples the assumption was made that the energy stored
within the magnetic core is clearly less than the energy within the air gap. This
assumption should be checked now for validity. Calculating the magnetic resistance
of an arbitrary element
l
Rm = . (9.51)
μA
the relation of two elements with identical length and cross-section scales via the
magnetic resistance with the permeability μ :
Rm1 μ2
= . (9.52)
Rm2 μ1
The permeability μr = HBμ0 is given by the relation between flux density vs. field
strength relatively to the magnetic constant. It is nonlinear (fig. 9.28) for all typical
flux-conducting materials within the flux-density areas relevant for actuator design
in between 0.5 and 2 T. It is identical to the inverse gradient of the curves given in
diagram 9.7. The maximum permeability values are given in tables frequently, but
refer to field strengths within the material only. They range from 6,000 for pure iron
over 10,000 for nickel alloys up to 150,000 for special soft-magnetic materials.

Mechanical processing of flux-conducting materials and the resulting thermal

changes within its microstructure will result in a considerable degradation of its
magnetic properties. This change can be restored by an annealing process.

Generally speaking however even outside an optimum value for flux density the
stored energy within typical materials is always several orders of magnitudes below
the energy within the air gap. This legitimates to neglect this energy portion for
the rough estimations in actuator design, but does show too that there is potential
within the optimization of electromagnetic actuators. This potential can be used by
the application of FEM-software, which is typically available as module for CAD
software 16 .

9.3.2.3 Permanent Magnets in Electromagnetic Actuators

Permanent magnets do not differ significantly in their properties from coils con-
ducting a DC current. They generate a polarized field, which - in combination with
another field - provide attraction or repulsion. For calculating combined magnetic
circuits a first approach can be taken by substituting the sources within the mag-
netic equivalent circuit (neglecting saturation effects). The calculation is analog to

16 or as free software, e.g. the tool “FEMM” from DAVID M EEKER

9.3 Electromagnetic Actuators 235

mr

mr max

Fig. 9.28 Qualitative change of the permeability for common flux-conducting materials.

the methods read in the chapter about electrodynamic actuators (paragraph 9.2.1.3).
A permanent magnet within a circuit either allows
• the realisation of a state held without any current applied
• or switching between two states with just one winding powered
A good example for a currentless held state [125] shows the calculation of a
polarized magnetic clamp (fig. 9.29). With non-active winding the flux is guided
through the upper anchor and held securely. With active coil the magnetic flux via
the upper anchor is compensated. The magnetic bypass above the coil prevents the
permanent magnet from being depolarized by a counter field beyond the kink in the
B/H-curve.

N N N
S S S

a) b) c)

Fig. 9.29 Permanent magnet in the magnetic circuit in shape (a), field-lines with inactive coil b)
and field-lines with active coil (c), releasing the anchor.
236 9 Actuator Design

9.3.3 Examples for Electromagnetic Actuators

Electromagnetic actuators are available in many variants. The following section

gives typical designs for each principle and shows the corresponding commercial
products. The knowledge about the designs will help to understand the freedom in
the design of electromagnetic circuits more broadly.

9.3.3.1 Claw-pole Stepper Motor

The electromagnetic claw-pole stepper motor (fig. 9.30) is one of the most fre-
quently used rotatory actuation principle. These actuators are made of two stamped
metal sheets (1,2) with the poles - the so called claws - bended by 90◦ to the interior
of the motor. The metal sheets are the magnetic core for conducting the flux of one
coil each (3). The permanent-magnet rotor (4) with a pole subdivision equalizing the
claw pattern orientates to the claws in currentless state. In stepper mode the coils are
powered subsequently, resulting in a combined attraction and repulsion of the rotor.
The control of the coils’ currents may happen either by simply switching them or
by a microstep mode with different current levels being interpolated between dis-
crete steps. Latter generates stable states for the rotor not only in the positions of the
claws but also in between.
Claw-pole stepper motors are available with varying number of poles, different num-
bers of phases and varying loads to be driven. As a result of the permanent-magnet
they show a large holding torque with respect to their size. The frequency of single
steps may reach up to 1 kHz for fast movements. By counting the steps the po-
sition of the rotor can be detected. Step-losses - the fact that no mechanical step
happens after a control signal - are not very likely with a carefully designed power
chain. Claw-pole stepper motors are the working horses of the electrical automation
technology.

1
3

a) b) c)
2
4

Fig. 9.30 Two-phase stepper motor made of stamped metal sheets and with a permanent magnet
rotor in a 3D-sketch (a), cross-section (b), and with details of the claws-poles (c) [218].
9.3 Electromagnetic Actuators 237

9.3.3.2 Reluctance Drives

The rotatory reluctance drives (fig. 9.31) are based on the electromagnetic longitudi-
nal effect. By clever switching of the windings (2) it is possible to keep the rotor (3)
in a continuous movement with just minimal torque ripples. To make this possible
the rotor has to have fewer poles than the stator. The rotor’s pole-angle βr is larger
than the pole-angle βs . Reluctance drives can be used as stepper motors by the inte-
gration of permanent magnets too. Generally speaking it excels by high robustness
of the components and a high efficiency factor with - for electromagnetic drives -
comparably little torque ripples.

a) b) c)

Fig. 9.31 Switched reluctance drive with pole- and coil-layout (a), in cross section (b), and with
flux-lines of the magnetic excitation (c) [218].

9.3.3.3 Electromagnetic Brakes

Electromagnetic brakes (fig. 9.32) are based on the transversal effect. They make
use of the high force increase at electromagnetic attraction to generate friction on a
spinning disk (1). For this purpose usually rotational-symmetrically flux conducting
magnetic cores (2) are combined with embedded coils (3). The frontal area of the
magnetic core and braking disk (1) itself is coated with a special layer to reduce
abrasion and influence positively the reproducibility of the generated torque. The
current/torque characteristic of electromagnetic brakes is strongly nonlinear. At the
one hand this is the result of the quadratic proportion between force and current of
the electromagnetic transversal effect, but on the other hand this is also a result of
its friction pairing. Nevertheless they are used in haptic devices for the simulation
of “hard contacts” and stop positions. A broad application for haptic devices is nev-
ertheless not visible. This is likely a result of their limits in reproducibility of the
generated torque, the resulting complex control of the current and the fact, that they
can only be used as a break (passive) and not for active actuation.
238 9 Actuator Design

1 3
2 Flight-Depot.com

a) b)
1

Fig. 9.32 Electromagnetic brake in cross section (a) and as technical realization for an airplane
model (b).

9.3.3.4 Plunger-type Magnet

Electromagnetic plunger-type magnets (fig. 9.33) are frequently based on the elec-
tromagnetic transversal effect. Their main uses are switching and control applica-
tions, requiring the actuation in specific states. With a magnetic core frequently
made of bended iron steel sheets (2) a coil-induced (3) flux is guided to a central
anchor, which itself is attracted by a yoke (4). The geometry of the yoke influences
significantly the characteristic curve of the plunger-type magnet. By varying its ge-
ometry a linearization of the force-position curve is possible within certain lim-
its. Even strongly nonlinear pulling-force characteristics can be achieved by such
a modification. Plunger-type magnets are available with additionally magnets and
with more coils. In these more complex designs they provide mono- and bi-stable
switching properties. By variation of the wires diameter and the number of turns
they can be adapted easily to any power level.

2 3 3

2
1

a) b) 1

Fig. 9.33 Plunger type magnet (a) with altered force-position curve (4), and realization as pulling-
anchor (b) with metal-sheet-made magnetic circuit (2).
9.3 Electromagnetic Actuators 239

9.3.4 Magnetic Actuators in Haptic Devices

For haptic applications electromagnetic actuators are mainly used within tactile de-
vices. Nevertheless admittance controlled devices can be found providing impres-
sive output forces of high quality even by the use of stepper-motors. Beside the
commercial systems such as “HapticMaster” of Moog (section 5.1.4) especially an
idea of L AWRENCE attracted attention within the past few years.

9.3.4.1 Spring-Tendon Actuator

In [157] L AWRENCE describes an inexpensive actuator for kinaesthetic haptic sys-

tems (fig. 9.34) based on an electromagnetic stepper motor coupled via a tendon to
a pen and with a spring mechanically connected in parallel. Analog to other haptic
devices the pen is the interface to the user. Between pen and tendon and spring there
is a bending body with DMS as a force sensor. To additionally minimize the torque
ripples of the stepper drive resulting from the latching of the poles, a high resolu-
tion external encoder has been attached to the motor, and a torque/angle curve was
measured. A mathematical spline fit of this curve was used for the actuator’s con-
trol to compensate the torque oscillations. Beside this compensation the closed-loop
control of the actuator via the force sensor near the pen includes a compensation
of frictional effects too. The result of all these efforts is a force source, providing a
force transmission with little noise and high stiffness up to 75 kN/m with movements
of limited dynamics.

Fig. 9.34 Electromagnetic stepper motor with spring, actuated in closed-loop admittance control
mode [157].
240 9 Actuator Design

9.3.4.2 Electromagnetic Pin Array

The usage of electromagnetic actuators for the control of single pins in array design
is very frequent. The earliest usages for haptic applications go back to printer-heads
of dot matrix printers being used in the 80th an earl 90th of the last century. Modern
designs are a lot more specific to haptics and make use of manufacturing technolo-
gies available from microtechnology. In [45] an actuator array (fig. 9.35) is shown
made of coils with 430 windings each and a 0.4 mm wide iron core. Above of it
a magnet is embedded in a polymer-layer, being attracted by the flux induced into
the core. With such an actuator and a diameter of 2 mm a maximum force of up to
100 mN is possible. A position measure is realized via inductive feedback.

a) b)

Fig. 9.35 Electromagnetic monostable actuator with permanent magnet: principle sketch (a) and
actuator design (b) [45].

9.3.4.3 Electromagnetic Plunger-type Magnet for the Tactile Transmission of

Speech Information

One fundamental motivation for the design of haptic devices is the partially sub-
stitution of lost senses. Especially methods to communicate information from the
sense of sight or hearing by the aid of tactile devices have some tradition. B LUME
designed and tested an electromagnetic plunger-type magnet according to the reluc-
tance effect 1986 at the University of Technology, Darmstadt. Such actuators were
attached to the forearm and stimulated up to eight points by mechanical oscillations
encoded from speech signals. The actuator (fig. 9.36) was made of two symmetrical
plunger-type magnets (on the horizontal axis) acting upon a flux conducting ele-
ment integrated into the plunger. The whole anchor takes a position symmetrically
within the actuator due to the integrated permanent magnet. In this symmetrical
position both magnetic circuits conduct a magnetic flux resulting in identical reluc-
tance forces. In active mode either the flux in the upper or the lower magnetic circuit
9.3 Electromagnetic Actuators 241

is amplified depending on the direction of current flow. The reluctance forces on the
amplified side pull the anchor in a current proportional position working against the
magnetic pretension from the permanent magnets, the mechanic pretension from the
springs and the load of the actuator. The plunger is displaced in the direction of the
weakened magnetic field. At a diameter of 20 mm this actuator covers a dynamic
range of 500 Hz at a efficiency factor of 50%. The forces lie in the range of ≈4 N
per ampere.

Cover Moving Pin

Coils

Magnetic
yoke
Spring- Anchor
bearing

a) b)

Fig. 9.36 Electromagnetic actuator according to the reluctance principle in a “counteractive

plunger type” design with permanent magnet: cross-section (a) and design incl. driver-electronics
(b) [25].

9.3.5 Conclusion about the Design of Magnetic Actuators

Electromagnetic actuators are - identical to electrodynamic systems - mainly force

sources. In rotary drives especially the reluctance effect is used to generate a con-
tinuous movement. With linear drives mainly plunger-type magnets are used based
on the nonlinear transversal effect. Whereas there are exception to both showing
some surprising properties (paragraph 9.3.4.3). The translational systems are usu-
ally used to actuate as either bistable switches between two discrete states or monos-
table against a spring (plunger-type, break, and valve). There are applications within
haptics based on either of both effects. Whereas reluctance based actuators can be
found equally often within kinaesthetic applications as drives and in an admittance
controlled application in tactile systems as vibrating motor, switching actuators are
almost exclusively found in tactile devices with single pins or pin arrays. In con-
trast to the highly dynamic electrodynamic drives, electromagnetic actuators excel
in less dynamic applications with higher requirements on torque and self-holding.
During switching in between two states however the acceleration and deceleration
at the mechanical stop are a highly dynamic but almost uncontrollable action. The
dynamic design of switching actions were not subject to this chapter, but are usually
242 9 Actuator Design

based on modeling a nonlinear force source of the electromagnet and assuming the
moving parts as concentrated elements of masses, springs and dampers. Due to their
relatively high masses within the moving parts, the hard to control nonlinearities
of fluxes and forces, and the low efficiency factor of the transversal effect in many
designs, electromagnetic actuators occupy niches within haptic applications only.
However in those niches there is no way around their usage. If an appropriate area
has been found they excel by an extremely high efficiency factor for the specific
design and a big robustness against exterior influences.
9.4 Piezoelectric Actuators 243

9.4 Piezoelectric Actuators

S TEPHANIE S INDLINGER , M ARC M ATYSEK

Next to the very frequently found electrodynamic actuators, the past few years
piezoelectric actuators were used for a number of device designs. Especially their
dynamic properties in resonance mode allow an application for haptics, which is
very different from the common positioning application they are used for. As vari-
able impedance a wide spectrum of stiffnesses can be realized. The following chap-
ter gives the calculation basics for the design of piezoelectric actuators. It describes
the design variants and their application in haptic systems. Beside specific designs
for tactile and kinesthetic devices approaches for the control of the actuators and
tools for their dimensioning are presented.

9.4.1 The Piezoelectric Effect

The piezoelectric effect was discovered by JACQUES and P IERRE C URIE first. The
term is derived from the Greek word “piedein - piezo” = “to press” [116].
Figure 9.37 shows a scheme of a quartz crystal (chemical: SiO2). With force
acting upon the crystal mechanical displacements of the charge-centers can be ob-
served within the structure, resulting in microscopic dipoles within its elementary
cells. All microscopic dipoles sum up to a macroscopic measurable voltage. This
effect is called “reciprocal piezoelectric effect”. It can be reversed to the “direct
piezoelectric effect”. If a voltage is applied on a piezoelectric material a mechani-
cal deformation happens along the crystal’s orientation, which is proportional to the
field strength in the material [7, 8, 225].

F
+ + + + + +
+ +
- Si - -
electrodes O O- d0 O Si
O d d0
Si Si Si Si
+ O + + O +
- -
- - - - - -
F

Fig. 9.37 Crystal structure of quarz in initial state and under pressure [225]
244 9 Actuator Design

Piezoelectric materials are anisotropic - direction dependent - in their proper-

ties. Consequently the effect depends on the direction of the electrical field applied,
and on the angle between the direction of the intended movement and the plane of
polarization. For the description of these anisotropic properties the directions are
labeled with indices. The index is defined by a Cartesian space with the axes being
numbered with 1, 2 and 3. The plane of polarization of the piezoelectric material is
typically orientated on direction 3. The shear at the axes is labeled with indices 4, 5
and 6.
Among all possible combinations, there are three major effects (fig. 9.38), com-
monly used for piezoelectric applications: longitudinal- , transversal- and shear-
effect.

metallic electrodes
T3,S3 U
x0
l
b
x0
T1,S1

U l U h h
x0
b b
h l 3
2
T5,S5
1

Fig. 9.38 Effects during applied voltage: longitudinal effect (left), transversal effect (center), shear
effect (right)[225]

The longitudinal effect acts in the same direction as the applied field and the
corresponding field strength E3 . As a consequence the resulting mechanical tensions
T3 and strains S3 lie within plane 3 too. With the transversal effect mechanical
actions show normal to the electrical field. As a result from a voltage U3 with the
electrical field strength E3 the mechanical tensions T1 and strains S1 appear. The
shear-effect happens with the electrical voltage U applied along plane 1 orthogonal
to the polarization plane. The resulting mechanical tensions appears tangential to the
polarization - in the direction of shear - and are labeled with the directional index 5.

9.4.1.1 Basic Piezoelectric Equations

The piezoelectric effect can be described most easily by state equations:

P = e·T (9.53)
9.4 Piezoelectric Actuators 245

and
S = d ·E (9.54)
with

P = direction of polarization (in C/m2 )

S = deformation (non-dimensional)
E = electrical field strength (in V/m)
T = mechanical tension (in N/m2 )

The piezoelectric coefficients are

• the piezoelectric coefficient of tension (also: coefficient of force) e (reaction of
the mechanical tension on the electrical field)

∂ Ti j
ei j,k = ∂ (9.55)
∂ Ek
• and the piezoelectric coefficient of strain (also: coefficient of charge) d (reaction
of the strain on the electrical field)

∂ εi j
di j,k = ∂ (9.56)
∂ Ek
The correlation of both piezoelectric coefficients is defined by the elastic con-
stants Ci jlm  
ei j,k = ∑ Ci jlm · dlm,k (9.57)
lm

Usually the tensors shown in the equation above are noted as matrix (VOIGT’sche
notation) . In this format, matrices result of six components identical to the defined
axes. The matrix shown below describes the concatenation of the dielectrical dis-
placement D, the mechanical strain S, the mechanical tension T , and the electrical
field strength E.

T1 T2 T3 T4 T5 T6 E1 E1 E3
D1 0 0 0 0 d15 0 ε11 0 0
D2 0 0 0 d15 0 0 0 ε11 0
D3 d31 d31 d33 0 0 0 0 0 ε11
S1 s11 s12 s13 0 0 0 0 0 d31
S2 s12 s11 s13 0 0 0 0 0 d31
S3 s13 s13 s33 0 0 0 0 0 d33
S4 0 0 0 s44 0 0 0 d15 0
S5 0 0 0 0 s44 0 d15 0 0
S6 0 0 0 0 0 2(s11 − s12 ) 0 0 0
246 9 Actuator Design

This matrix can be simplified for the specific cases of a longitudinal and a
transversal actuator. For a longitudinal actuator with electrical contact in direction 3
the following equations are the result .

D3 = ε33
T
E3 + d31 T1 (9.58)

S3 = d31 E3 + sE11 T1 . (9.59)

Accordingly for a transversal actuator the correlation

D3 = ε33
T
E3 + d33 T3 (9.60)

S3 = d33 E3 + sE33 T3 (9.61)

with

D3 = dielectric displacement in C/m2 D=0: open-ended

E3 = field-strength in V/m E=0: short-cut
S1 , S3 = Δ L/L = strains, dimensionless S=0: mech. short-cut
T1 , T3 = mechanical tensions N/m2 T=0: idle mode
ε33
T = relative dielectricity constant at mechanical tension = 0
d31 , d33 = piezoelectric charge constant in C/N
sE11 , sE33 = elasticity constant at field strength = 0

becomes valid.
Therefore the calculation of piezoelectric coefficients simplifies into some handy
equations:
The charge constant d can be calculated for the electrical short-circuit- which is
E = 0- to
D
dE=0 = (9.62)
T
and for the mechanical idle situation - which is T = 0 - to
S
dT =0 = . (9.63)
E
The piezoelectric tension constant is defined as

d
g= . (9.64)
εT
The coupling factor k is given by equation 9.65. It is a quantity for the energy
transformation and consequently for the strength of the piezoelectric effect. It is
used for comparison among different piezoelectric materials. However note that it
is not identical to the efficiency factor, as it does not include any energy losses.
9.4 Piezoelectric Actuators 247

converted energy
k= . (9.65)
absorbed energy
A complete description of the piezoelectric effect, a continuative mathematical
discussion, and values for piezoelectric constants can be found in literature, such as
[29, 104, 158].

9.4.1.2 Piezoelectric Materials

Till 1944 the piezoelectric effect was observed with monocrystals only. These were
quartz, turmalin, lithiumniobat, potassium- and ammonium-hydrogen-phosphat (KDP,
ADP), and potassium sodium tartrate [7]. With all these materials the direction of
the spontaneous polarisation is given by the direction of the crystal lattice [116].
DThe most frequently used material was quartz.
The development of polarization methods made it possible to retrospectively po-
larize ceramics by the application of a constant exterior electrical field in 1946.
By this approach “piezoelectric ceramics” (also “piezoceramics”) were invented.
By this development of polycrystalline materials with piezoelectric properties the
whole group of piezoelectric materials got an increased attention and technical sig-
nificance. Today the most frequently used materials are e.g. barium titanate (Ma-
TiO3) or lead zirconate titanate (PZT) [7]. C 82 is a piezoelectric ceramic suitable
for actuator design due to its high k-factor. However as all piezoelectric ceramic ma-
terials it shows reduced long term stability compared to quartz. Additionally it has
a pyroelectric effect which is a charge increase due to temperature changes of the
material [158]. Since the 60th the semi-crystalline synthetic material polyvinylidene
fluoride (PVDF) is known. Compared to the materials mentioned before, PVDF ex-
cels by its high elasticity and reduced thickness (6 μ m to 9 μ m).
Table 9.4 shows different piezoelectric materials with their specific values.
Looking at these values PZT is most suitable for actuator design due to its high
coupling factor with large piezoelectric charge modulus and still a high Curie tem-
perature . The Curie temperature represents the temperature at which the piezoelec-
tric properties from the corresponding material are lost permanently. The value of
the curie temperature depends on the material (table 9.4).

9.4.2 Designs and Properties of Piezoelectric Actuators

Actuators using the piezoelectric effect are members of the group of solid actuators
(also: solid-state actuators). The transformation from electrical into mechanical en-
ergy happens without any moving parts, resulting in a very fast reaction time and
high dynamics compared to other actuation principles. Additionally piezoelectric
actuators have a high durability. The thickness changes are smaller - compared to
other actuation principles. Although the generated forces are much higher.
248 9 Actuator Design

Table 9.4 Selection of piezoelectric materials with characteristic values[158].

Constant Unit Quartz PZT-4 PZT-5a C 82 PVDF
d33 10−12 2,3 289 374 540 -27
d31 m/V -2,3 -123 -171 -260 20
e33 A·s 0,181 15,1 15,8 28,1 108
e31 m2 -0,181 -5,2 -5,4 -15,4 -
sE33 10−12 12,78 15,4 18,8 19,2 -
sE11 m2 /N 12,78 12,3 16,4 16,9 -
cE33 1010 7,83 6,5 5,3 5,2 -
cE11 N/m2 7,83 8,1 6,1 5,9 -
ε33
T ε33
S
ε0 ; ε0 - 4,68; 4,68 1300; 635 1730; 960 3400; - 12 ;12
ε11
T ε11
S
ε0 ; ε0 - 4,52; 4,41 1475; 730 1700; 830 3100; - -
k33 - 0,1 0,7 0,71 0,72 0,20
k31 - - 0,33 0,34 0,36 0,15
ϑCurie 1
◦C 575 328 365 190 80
ρ kg
m−3
2660 7500 7500 7400 1790

9.4.2.1 Basic Piezoelectric Actuator Designs

Depending on the application different designs may be used. One may require a
large displacement; another one may require self-locking or high stiffnesses. The
most frequently used actuator types are bending actuators and actuator staples. A
schematic sketch of each design is given in figure 9.39 (a) and (c).
Stacked actuator are based on the longitudinal piezoelectric effect. For this pur-
pose several ceramic layers of opposite polarity are stapled above each other. In
between each layer contact electrodes are located for electrical control. A staple is
able to generate high static forces up to several 10 kN. The achievable displacement
of 200 μ m is low compared to other piezoelectric designs. By the use of levers 9.39
(b) the displacement can be significantly increased (s. figure 9.39 (b)). Voltages of
several 100V are necessary to drive a piezoelectric actuation staple.
Bending actuators are based on the transversal piezoelectric effect. Designed ac-
cording to the so called bimorph principle - with two active layers - they are used
in applications requiring large displacements. The transversal effect is characterized
by comparably low controlling voltages [7, 116]. These electrical properties and the
large displacements can be achieved by very thin ceramic layers in the direction of
the electrical fields, and an appropriate geometrical design. Other geometrical de-
signs using the transversal effect are tubular actuators, film actuators, or bending
discs 9.39 (d) to (f). Due to their geometry they equal staple actuators in their me-
chanical and electrical characteristics. The achievable displacements of 50 μ m are
9.4 Piezoelectric Actuators 249

Longitudinal effect::
x
x

3
2
UB UB 1

(a) (b)
Transversal effect::
x x

x x

UB UB
UB UB

(c) (d) (e) (f)

Fig. 9.39 Piezoelectric transducers separated by longitudinal and transversal effect:

longitudinal effect: (a) stack, (b) stack with lever transformation, change of length: x = d33 ·UB
transversal effect: (c) bending actuator, (d) cone, (e) band (f) bending disk, change of length: x =
−d31 ·UB [116]
.

comparably low, whereas the achievable forces excel bending actuators at several
orders of magnitude..
The use of the shear effect is uncommon in actuator design. This is somewhat
surprising as it shows charge modulus and coupling factor which is twice as much
as the transversal effect. Additionally it is possible to increase the elongation x0
in idle mode (displacement without any load) by the optimization of the length to
thickness (l/h) ratio. However the clamping force Fk of the actuator is not influenced
by these parameters.
Table 9.5 summarizes the properties of different geometrical designs. Typical
displacements, actuator forces and control voltages are shown.
250 9 Actuator Design

Table 9.5 Properties of typical piezoelectric actuator designs [116].

Standard de- stack stack with bending ac- tape actuator tubus bending
signs lever tuator discs
actuator dis- 20...200 μ m ≤ 1.000 μ m ≤ 1.000 μ m ≤ 50 μ m ≤ 50 μ m ≤ 500 μ m
placements
actuating ≤ 30.000 N ≤ 3.500 N ≤ 5N ≤ 1000 N ≤ 1000 N ≤ 40 N
forces
control volt- 60...200 V 60...200 V 10...400 V 60...500 V 120...1000 V 10...500 V
ages 200...500 V 200...500 V
500...1000 V 500...1000 V

9.4.2.2 Selection of Special Designs for Piezoelectric Actuators

Beside the standard designs shown above several variations of other geometrical
designs exist. In this section, examples of an ultrasonic drive with resonator, os-
cillatory/traveling waves actuators and piezoelectric stepper motors are discussed.
Ultrasonic actuators are differentiated according to resonators with bar like geome-
try and rotatory ring geometry.

Ultrasonic Actuators with Bar-shaped Resonator

One special variant of piezoelectric ultrasonic actuators is made of a piezoelectric

actuator staple and a mechanical resonator [116, 260, 261, 263, 280]. The resonator
is part of the stator of the drive, and its oscillating end actuates the rotor. As the
movement of the actuator is not continuous and resembles represents a standing
wave, the design is also known as “standing-wave actuator”. ,
The oscillating movement of the stator is transferred on the rotor and makes it
move. Although the amplitudes are in the area of several μ m, rotational speeds of
100 turns / minute are easily possible:
The amplitude at the end of the resonator is 4 μ m. With an oscillation of 25 kHz
the tip moves

v = 4 μ m · 25 kHz = 100 m/sec. (9.66)

Basically two different designs of bar like resonators are available: mono-modal
and bi-modal actuators. Both designs are schematically sketched in figure 9.40.
With mono-modal actuators only one oscillatory resonance mode is builds up. It
is actuated in one frequency, resulting in a longitudal movement of the bar’s tip in
axial direction. The tip itself is beveled. The contact point between bar and rotor is
placed asymmetrically to the rotary axis. The one-dimensional oscillation of the axis
results in a push-movement against the rotor. To change the direction of movement
the position of the bar has to change (or a second bar has to be available) pushing the
9.4 Piezoelectric Actuators 251

clamping clamping rotor

rotor

monomodal resonator bimodal resonator

Fig. 9.40 Design of the monomodal (left) and bimodal (right) piezoelecric actuator with a bar-
shaped resonator [116].
.

rotor below the rotatory axis. With bi-modal actuators the resonator is used in two of
its resonance modes for oscillation: one longitudinal mode and one bending mode.
The control is done with two frequencies being summed up in one signal. With the
frequencies being chosen according to one frequency resembling an integer multiple
of the other one, a closed rotary motion results (L ISSAJOUS-figure). By changing
the phase lag between both frequencies a rotation in both direction is possible to
achieve. Additionally the phase lag is able to alter the rotational speed of the motor.
Figure 9.41 visualizes this correlation.
The geometrical shape of the tip’s movement is defined by the phase Θ . With
a phase angle Θ = 0a optimum contact between tip and rotor is achieved for anti-
clockwise movement. The rotational speed is at its maximum. By increasing Θ , the
movement of the tip and accordingly the contact time and shape between resonator
and rotor change. An example is given for Θ = π4 . With a phase angle of Θ = π2
no driving torque is left. This phase can be used to prevent a self-locking of the
motor in an idle mode. It is somewhat like an active unblocking state, wherein the
friction between rotor and resonator is canceled by the oscillating movement of the
resonator’s tip. Another increase of the phase lag to Θ = 34 π results in a change of
direction for the motor. At Θ = π the clockwise movement direction is done with an
optimum in contact states and a maximum torque being transmitted.
The movement of the rotor tip according to figure 9.41 can be described by the
following equation:

x = x0 sin(ω1t), y = y0 sin(ω2t + Θ ). (9.67)

Whereas f1 dis the frequency of the longitudinal movement and f2 = 2 · f1 the

frequency of the bending mode.
If the bending mode is not designed to be an integer multiple of the transversal
mode, the frequencies will drift during operation. In such cases a defined phase has
to be identified between both control frequencies to keep a defined contact between
tip and rotor. Considerations dealing with the optimization of the contact properties
are given e.g. in [62, 80, 180, 248, 297].
252 9 Actuator Design

Y0 Y0
counterclockwise

-X0 X0 -X0 X0

-Y0 -Y0
1
Q=0
Y0 Q= 4p
standstill

-X0 X0

-Y0

Y0 Q= 1p Y0
2
clockwise

-X0 X0 -X0 X0

-Y0 -Y0
3
Q= 4p Q=p

Fig. 9.41 Motion of the tip of the resonator of the bimodal resonator with varying phase shift. The
contact area is marked. [116].
.

The movements’ amplitudes of piezoelectric actuators with bar-like resonator lie

in the area of several μ m. To increase it the geometrical shape of the resonator has to
be changed. As the resonator is also called “horn”, as it fulfills the same functionality
as the horn of any wind instrument - it changes the amplitude of the input oscillation.
There are three basic shapes common for horns: the stepwise narrowing horn, the
exponential narrowing horn and the linear narrowing horn. The geometrical design
of horns is challenging and goes widely beyond the scope of this book. Additional
information on the shapes of horns can be taken from [16, 260, 261, 263]. Calcu-
lation hints and optimizing criterions for the design of resonators for piezoelectric
actuators are discussed in [48, 228, 262, 280].
A commercially available piezoelectric actuator with mono-modal resonator and
still able to move in two directions is sold by the company Elliptec [168, 268].
9.4 Piezoelectric Actuators 253

The shape of this resonator and the placement of the piezoelectric elements are -
compared to the basic design - varied 9.42.

clamping piezoelectric stack

resonating horn

spring contact area

Fig. 9.42 Drawing of the piezoelectric ultrasonic actuator by Elliptec [168].

.

The motor combines a piezoelectric staple with resonator, whose tip is oscillating
elliptically. There is a groove included in the tip providing a contact surface and
guide between driven element and the resonator. The actuator is pressed by an elastic
element, e.g. a torsion- or flat-spring, at a defined angle against the driven element.
This driven element may either be a disk for rotary movement or a cylinder for
translatory movements.

Table 9.6 Technical properties of the Elliptec-motor X15G [4].

Symbol Parameter Values Unit
min. typ. max.
v0 idle mode 300 350 550 mm/s
a acceleration ftom 0 up to max. velocity at m = 1, 5g 5 ms
FH holding force in zero-current case 0,5 0,8 1,2 N
FB blocking force 200 300 500 mN
F100 actuating force at 100 mm/s 100 200 350 mN
F200 actuating force at 200 mm/s 25 100 200 mN
MB blocking-toraue at 20mm  of teh driven rotor 2 3 5 mNm
M100 torque at 20mm  and 95,5 rpm 1 2 3,5 mNm
M200 torque at 20mm  and 191 rpm 0,25 1 2 mNm
ffw frequency forward mode at vmax 77 81 84 kHz
fbw frequency backward mode at vmax 93 98 108 kHz
fr frequency resolution 0,2 0,6 kHz
fD temperature-depending frequency drift 35 50 70 Hz/◦ C
rs positioning accuracytre var. 5 bis 100 μm
Pp0 energy of driver electronics 1,8 W
Idriver max. current of driver electronics at 5 V 300 450 600 mA
Umotor motor voltage (peak-to-peak) 5 7 10 V

For control of the motor just a single frequency f1 has to be provided. Due to the
actuators geometrical design the elliptic movement changes its direction at a second
frequency f2 resulting in a movement change on the drive side.
254 9 Actuator Design

A typical length of the actuator is 20 mm. The resonator shows a height and width
of 3 mm. The weight is 1,2 g. The major technical specifications of the Elliptec-drive
are given in table 9.6. Due to its components and the very compact driver electronics
it can be obtained for little costs at low quantities.

Ultrasonic Actuators with Circular Resonators

As mentioned before beside actuators providing standing waves another group of

actuators based on traveling waves exists. The traveling wave actuators known best
are circular in their design. The first actuator according to this principle has been
built in 1973 by S ASHIDA [261]. Traveling waves actuators count to the group of
ultrasonic actuators as their control frequencies typically lie in between 20 kHz up
to 100 kHz liegen.
This section is reduced to the presentation of ring-shaped traveling wave actua-
tors with a bending wave. Other design variants for linear traveling wave actuators
can be found in the corresponding literature [54, 92, 93].
Figure 9.43 shows an actuator’s stator made of piezoelectric elements. They have
an alternating polarization all around the ring. The stator itself carries notches actu-
ally enabling the formation of the rotating traveling wave.

ring-shaped rotor rotational direction of the motor

piezoelectric
metal part
elements
friction lining
- + - +
+ -
- +
+ -
- + stator surface propagation direction
+ - of the travelling wave
- + stator
+ -
- + piezo ceramics
+ - + -

Fig. 9.43 Piezoelectric traveling wave motor: left: stator disk with piezoelectric elements. right:
schematic view of the functionality of a ring-shaped piezoelectric traveling wave motor [116]
.

Each point on the surface of the stator performs a local elliptic movement (tra-
jectory). This movement is sketched schematically in figure 9.43 These individual
elliptic movements overlay to a continuous wave on the stator. With a frictional
coupling this movement is transferred on the rotor, resulting in a rotation. The con-
tact between stator exists with the same number of contact points anytime during
operation.
The movement equation of the traveling wave actuator is given by
9.4 Piezoelectric Actuators 255

u(x,t) = A cos(kx − ω t) (9.68)

By reshaping it the following form results in

u(x,t) = A(cos(kx))(cos(wt)) + A(cos(kx − π /2))(cos(kx + π /2)) (9.69)

The second term of equation 9.69 includes important information for the control
of traveling wave actuators. A traveling wave can be generated by two standing
waves being spatially and timely different. Within typical realization the spatially
difference of x0 = λ /4 unis chosen, and a time phase lag of Φ0 = π /2. The usage of
two standing waves is the only practical possibility for generating a traveling wave.
The direction of rotor-movement can be switched by changing the phase lag from
+π /2 to −π /2 [81, 95, 99, 261].
Figure 9.44 shows the practical realization of a traveling wave motor.

Fig. 9.44 Realization of a traveling wave motor by Daimler-Benz AG [279].

.

Big advantages of a traveling wave motor are the high torques possible to achieve
at low rotational speeds. It has a low overall size and is of little weight. This enables
a very thin design as shown in figure 9.44 . In passive mode the traveling mode
motor has a high locking torque of several Nm.
Other advantages are given by the good control capabilities, the high dynamics,
and robustness against electromagnetic noise such as the silent movement [282].
Typical applications of traveling wave actuators are autofocus-functions in cameras.

Uchinomotor

The piezoelectric actuator of K. U CHINO [260, 261] has a geometrically very com-
pact design (fig. 9.45). It is made of a metallic tube which was flattened on two
sides to provide two orthogonal surfaces. Two rectangular piezoelectric elements
are attached to these surfaces.
By actuating plate X the metallic tube experiences a bending torque in the axial
direction of x. Due to the asymmetrical mass distribution resulting from the plate Y
on the cylinder, a second bending movement in y direction is induced. This move-
256 9 Actuator Design

y
metal cylinder
PZT-plate X

elastischer
Zylinder
PZT-plate Y

Fig. 9.45 Schematic view of the piezoelectric ultrasonic actuator by U CHINO [260, 261].

ment is of identical phase with the movement in x-direction. The resulting combined
movement is elliptic. If another rod is placed in the metallic tube, it is driven by the
combined movement in axial translational and rotational direction. To change the
direction of movement the plate Y is actuated.
The actuator is extremely small. Its typical diameter is given with 2.4 mm at a
length of 12 mm. It excels by its simple design and easy and cheap manufacture. The
controlling frequency lies at 61 kHz. In idle mode the achieved rotational speed is
1800 rpm and a torque of 1.8 mNm. The translational velocity can be varied between
1 μ m/s to 10 mm/s at a resolution of 0,5 μ m. The controlling voltages are 80 V.

driven rod ultrasonic motor

finger tip

Fig. 9.46 Photo of the implementation of the piezoelectric U CHINO-motor by Squiggle Motors
[238].

A very small variant of this actuator has the geometrical dimensions of 1,5 x
1,5 mm2 at a length of 6 mm and is manufactured by Squiggle Motors (fig. 9.46). It
is the smallest linear actuator available commercially at the time of print.
9.4 Piezoelectric Actuators 257

Piezoelectric Stepper Motors

Another interesting design can be found with the actuator PI Nexline. It combines
the longitudinal effect with the piezoelectric shear effect, resulting in a piezoelectric
stepper motor.
The principle design is sketched in figure 9.47.

z-longitudinal
elements
P E
driven rod P E x-shear
P E
P E elemts
P E
P E
piezo- P E z
electric P E y
elements x

Fig. 9.47 Piezoelectric stepper motor using the shear effect and the longitudinal effect [134].

The movement of the motor is similar to the inchworm-principle. Drive- and

release-phases of the piezoelectric elements produce a linear movement of the
driven rod. The piezoelectric longitudinal elements generate the clamping force in z-
direction, the shear elements rotated by 90◦ a translational movement in y-direction
is possible too.
The advantage of this design is given by the high positioning resolution. Over the
whole displacement of 20 mm a resolution of 0.5 nm can be achieved. The stepping
frequency is given by - dependent on the control - up to 100 Hz and enables, de-
pending on its maximum step-width, velocities of up to 1 mm/s. The step-width can
be chosen stepless between 5 nm and 8 μ m. The intended position can be achieved
either closed-loop or open-loop controlled. For the closed-loop control a linear en-
coder has to be added to the motor. In open-loop control the resolution can be in-
creased to 0.03 nm in a high resolution dithering mode.
The actuator can generate push- and pull-forces of 400 N maximum. The self-
locking reaches up to 600 N. The typical driving voltage is 250 V. The specifications
given above are based on the actuator N-215.00 Nexline of the company Physik
Instrumente (PI) GmbH & Co. KG (www.pi.ws) [134]. Beside the impressive forces
and positioning resolutions which can be achieved, these actuators have a high dura-
bility compared to other designs of piezoelectric actuators, as no friction happens
between moving parts and stator.
An application of the stepper motor is given in figure 9.48. Six of above drives are
used within in a hexapod parallel kinematic design, which can be used to position
high loads in the presence of strong magnetic fields.
258 9 Actuator Design

Fig. 9.48 Application of the piezoelectric stepper motor in a hexapod [134].

9.4.3 Design of Piezoelectric Actuators for Haptic Systems

Within the preceding section the basic designs of piezoelectric actuators have been
discussed and special variants were shown. This section transfers this knowledge
about the design of piezoelectric actuators focussing on haptic applications now.
First of all the principle approach for designing is sketched. Hints are given about
which designs are appropriate for which applications. Afterward three tools for prac-
tical engineering are shown: description via electromechanic networks, analytical
formulations and finite-element simulations

9.4.4 Procedure for the Design of Piezoelectric Actuators

Figure 9.49 gives the general procedure for the design of piezoelectric actuators.
The choice of a general design based on those shown in the prior section is largely
dependent on the intended application. For further orientation figure 9.50 shows a
decision tree for classifying the own application.
The following paragraph describes the appropriate designs for specific applica-
tion classes according to this scheme. The list has to be regarded as a point for
orientation, but it does not claim to be complete. The creativity of an engineer will
be able to find and realize other and innovative solutions beside those mentioned
here. Nevertheless especially for the design of tactile devices some basic advice can
be given:
1 Braille displays have to act against the finger’s force. At typical resolutions this
requires forces in the area of mN at displacement of around 100 μ m. The re-
quirements on the dynamics are in the lower range of several Hertz. The smallest
resolution of a pixel has to be in the area of 1 x 1 mm2 and is defined by the
9.4 Piezoelectric Actuators 259

clarifying the
requirements

selection of a
basic
transducer

analytical electromechanical FEM-

calculation schematic diagram simulation
Design of the actuator

practical verification

Fig. 9.49 Procedure of designing piezoelectric actuators.

device

tactile kinaesthetic

kind of penetration kind of stimulation

vertical lateral passive

active
to skin to skin

resolution of the display kind of stimulation direction of impression for the user

Braille vibrotactil discrete continuous rotatory linear rotatory linear

1 2 3 4 5 6 7 8

Fig. 9.50 Decision tree for the selection of a type of piezoelectric actuator.
260 9 Actuator Design

lowest resolution at the finger’s tip. Looking at the force-amplitude diagram of

figure 9.51 bending-actuators fit well to these requirements.
2 In comparison to Braille-displays there are vibrotactile displays, which need
higher frequencies and smaller displacements and forces to present a static shape
to the user. With the diagram of figure 9.51 especially bending disc or staple
actuators would be appropriate to fulfil these requirements, although there are
overpowered concerning the achievable forces.
3,4 Such displays are subject to current research and are not yet applied in broad
range. Their design is typically based on bending actuators, as shear the skin re-
quires less force than forces acting normal to the skin by generating a comparable
perception.
5-8 In contrast to tactile displays the actuator selection for kinaesthetic systems is
more dependent on the actual application. Forces, displacement and degrees of
freedom influenced by the kinematics alter in wide ranges. Additionally the ac-
tuator’s volume may be a criterion for selection. Figure 9.51 gives an overview
about piezoelectric actuation principles and has to be interpreted according to the
specific kinaesthetic problem at hand. Generally speaking ultrasonic piezoelec-
tric actuators are usually the matter of choice for kinaesthetic devices, although
they have to be combined with a closed-loop admittance control.
Figure 9.51 gives an overview about piezoelectric actuation principles and has
to be interpreted according to the specific kinaesthetic problem at hand. Generally
speaking ultrasonic piezoelectric actuators are usually the matter of choice for ki-
naesthetic devices, although they have to be combined with a closed-loop admit-
tance control.
Additional reference for actuator selection can be found in chapter 9.4.2 are suit-
able for haptic applications, but still need some care in their usage due to high volt-
ages applied and their sensitivity on mechanical damage. This effort is often re-
warded by piezoelectric actuation principles, which can be combined to completely
new actuators. And the only thing required is some creativity of the engineer.
After choosing the general actuator the design process follows next. For this
purpose three different methods are available, which are presented in the following
and discussed in their pros and cons. In addition some hints on further references
are given.

9.4.4.1 Methods and Tools for the Design Process

There are three different engineering tools for the design of piezoelectric actuators:
• Description via the aid of electromechanical concentrated networks
• Analytical descriptions
• Finite elements simulations
9.4 Piezoelectric Actuators 261

amplitude [µm]

1k
bending
Stapel
actuator
100 bending
Stapel
disk
stack
Stapel
cone with lever
10

1 Stapel
stack

force [N]
1 10 100 1k 10k

Fig. 9.51 Force-amplitude-diagram for the classification of the piezoelectric actuating types.

Description via the Aid of Electromechanical Concentrated Networks

The piezoelectric basic equations from paragraph 9.4.1.1 on page 244 are the ba-
sis for the formulation of the electromechanic equivalent circuit of a piezoelectric
converter.
The piezoelectric actuator can be visualized as an electromechanical circuit. Fig-
ure 9.52 shows the converter with a gyratory coupling (a), alternatively a transfor-
matory coupling (b) is possible too (see appendix 17).

i iw Fw F i Fw F

u Cb
( ) ( )( )
u 0 Y v
i = 1/Y 0 Fw
nK v u Cb
( ) ( )( )
u j/X 0 v
i = 0 jX Fw
nK nC v

(a) (b)

Fig. 9.52 Piezoelectric actuator as a electromechanical schematic diagram (a) gyratory and (b)
transformatory combination [158].

For the gyratory coupling the equations 9.70 to 9.73 summarize the correlations
for the calculation of the values for the concentrated elements. They are derived
from the constants e, c, ε ssuch as the actuator’s dimensions l and A [158].
262 9 Actuator Design

A A
Cb = ε · = (ε − d 2 · c) with v = 0 (9.70)
l l
1 l A
nK = · = s · with U = 0 (9.71)
C A l
1 l s l
Y = · = · (9.72)
e A d A
e2 d2
k2 = = (9.73)
ε ·c ε ·c
With the piezoelectric force constants

d
e = d ·c = (9.74)
s
Which makes for the transformatory coupling
1
X= und nC = Y 2 ·Cb (9.75)
ω Cb ·Y
Figure 9.53 shows the sketch of a element Δ x taken out of a piezoelectric bimorph
bending actuator (dimensions Δ l x Δ h x Δ b) as a electromechanical equivalent cir-
cuit.
9.4 Piezoelectric Actuators 263

0 x1 x1 +Dx
x
l
3 piezoelectric bimorph
2 u P E3,D3
h b in initial state
1 P E3,D3
Dx

DA

M(x1)
M(x1+Dx) separating a bending
W(x1) element in the area Dx
W(x1+Dx)
Dx

u
DCb bending element Dx in
an equivalent network
with 10 pols
(Y)

M(x1) DnRK M(x1+Dx)

W(x1) W(x1+Dx)

F(x1) F(x1+Dx)

v(x1) 2/Dx Dm 2/Dx v(x1+Dx)

Fig. 9.53 Piezoelectric bimorph bending element in a electromechanical schematic view in qua-
sistatic state [158].

It is:
b·Δx
Cb = 4ε33
T
(1 − kL2 )
h
(Δ x)3
Δ nRK ≈ 12 sE11
b · h3
1 1 d31 b·h
=
Y 2 sE11 Δx

The piezoelectric lossless converter couples the electrical with the mechanical
rotatory coordinates first, which are torque M and angular velocity Ω . To calcu-
late the Force F and the velocity v an additional transformatory coupling between
rotatory and translatory mechanical network has to be introduced. As a result the
complete description of a sub-element Δ x of a bimorph is given in a ten-pole equiv-
alent circuit.
264 9 Actuator Design

Analytical Calculations

A first approach for the design of piezoelectric actuators is given by the application
of analytical equations. The advantage of analytical equations lies in the descriptive
visualization of physical interdependencies. The influence of different parameters
on a target value can be derived directly from the equations. This enables a high
flexibility in the variation of e.g. dimensions and material properties. Additionally
the processing power for the solution of equations can - compared to simulations -
be neglected.
A disadvantage of analytical solution results from the fact, that they can only be
applied to simple and frequently only symmetrical geometrical designs. Although
already limited to such designs even simple geometries may result in very complex
mathematical descriptions requiring a large theoretical background for their solu-
tion.
The following collection gives relevant literature for the familiarization with spe-
cific analytical challenges being faced with during the design of piezoelectric actu-
ators:
• Very compelling and complete works on the design of piezoelectric actuators are
[260, 261, 263, 262].
• The theory of the piezoelectric effect and piezoelectric elements are discussed in
[29, 104, 105].
• The mathematical description of traveling wave actuators can be found in [214,
294].
• The contact behavior between stator and rotor with piezoelectric multilayer
bending-actuators is analyzed in [80, 248, 278, 297]
• In [7, 8] the static and dynamic behaviour of multilayer beam bending actuators
is described.
• The description of the mechanical oscillations for resonance shapes is discussed
in [16, 48, 55, 140, 228, 280] elaborately.

Finite Element Simulation

The application of both approaches given before is limited to some limited geometri-
cal designs. In reality complex geometrical structures are much more frequent which
can not be solved with analytical solutions or mechanical networks. Such structures
can be analyzed according to the method of finite element simulation (FEM).
For the design of piezoelectric actuators the use of coupled systems is rele-
vant. One example of a FEM-simulation for a piezoelectric travelling wave motor is
shown in figure 9.54
For the solution of the differential equations the market offers a reasonable num-
ber of simulation tools. At this position only some shall be named explicitly
• ANSYS (www.ansys.com)
• ATILA (www.cedrat.com)
9.4 Piezoelectric Actuators 265

Fig. 9.54 FEM-simulation of the oscillation shape of the stator of a piezoelectric travelling wave
motor (view highly exaggerated) [282].
.

• Comsol Multiphysics (www.femlab.de)

• ProMechanica (www.ptc.com) oder
• GiD (http://gid.cimne.upc.es).
The simulation tools are subject to costs, but free test-versions limited in the
number of nodes or the availability certain functions shapes can also be found. Lit-
erature for the start is offered in form of books like “FEM für Praktiker” [179, 241]
such as manifold documentation and handbooks of the individual software.
266 9 Actuator Design

9.4.5 Piezoelectric Actuators in Haptic Systems

Piezoelectric actuators are among the most frequently used actuation principles in
haptic systems. The designs shown before can be optimized and adapted for a rea-
sonable number of applications. One of the most important reasons for their usage
is their effectiveness at a very small required space, which is identical to a high
power density. To classifiy the realized haptic systems a division into tactile and
kinaesthetic is done for the following paragraphs.

9.4.5.1 Piezoelectric Actuators for Tactile Systems

For the design of any tactile systems the application area is of major importance.
The bandwidth ranges from macroscopic table top devices, which may be used for
embossed printings in Braille being placed below a PC-keyboard, up to highly inte-
grated systems, which may be used for mobile applications. Especially for latter use,
the requirements on volume, reliable and silent operation, but also on low weight
and energy consumption are enormous. The following examples are structured into
two subgroups. Each of them addresses one of two directions of the penetration of
the skin: lateral and normal.

Tactile Displays with Normal Penetration

Braille Devices
A Braille character is encoded by a dot pattern formed by embossed points on a
flat surface. By touching this pattern made of eight dots (two columns with four
rows of dots each) combinations of up to 256 characters can be addressed. Since the
seventies reading tablets for visually handicapped people are developed which are
capable to present these characters with a 2x4 matrix of pins. The most important
technical requirements are a maximum stroke of 0.1 − 1 mm and a counter force
of 200 mN. Already early in this development electromagnetic drives have been re-
placed by piezoelectric bimorph bending actuators. These actuators enable a thinner
design, are more silent during operation and faster. At typical operating voltages
of ±100 − 200 V and a nominal current of 300 mA they additionally need less en-
ergy than the electromagnetic actuators used before. Figure 9.55 shows the typical
design of a Braille character driven by a piezoelectric bimorph actuator. A disad-
vantage of this system is the high price, as for 40 characters with eight elements
each all together 320 bending actuators are needed. Additionally they require still a
large volume as the bending elements have to show a length of several centimeters
to provide the required displacements. This group of tactile devices is part of the
shape-building devices. The statically deflected pins enable the user to detect the
displayed symbol.
9.4 Piezoelectric Actuators 267

Fig. 9.55 Schematic setup of a Braille row with piezoelectric bending actuators

Vibrotactile Devices
With vibrotactile devices the user does not detect the displacement of the skin’s sur-
face but the skin itself is put into oscillations. At smaller amplitude the sensation is
similar to the static elongation. The general design of vibrotactile displays equals an
extension of the Braille-character to an N x N matrix which is actuated dynamically.
The tactile image generated is not perceived by the penetration depth but by the am-
plitude of the oscillation [107]. Another impact factor is the oscillation frequency,
as the tactile perception depends extremely on the frequency. With the knowledge
of these interdependencies optimized tactile displays can be built generating a very
well perceivable stimulation of the receptors. Important for displays according to
this approach is a large surface, as movements performed by the own finger disturb
the perception of the patterns.
The “Texture Explorer” is designed as a vibrating 2x5 pin array [106]. It is used
for the research on the perception of tactile stimulation, such as the overlay of tactile
stimulations with force-feedback. The surfaces displayed change according to their
geometry and their roughness within the technical limits of the device. Figure 9.56
shows a schematic sketch of the actuator array. The contact-pins have a size of 0.5 ×
0.5 mm2 with a point-to-point distance of 3 mm. Each pin is actuated separately by
a bimorph bending actuator at a driving voltage of 100 V and a frequency of 250 Hz.
The maximum displacement of these pins with respect to surface level is 22 μ m and
can be resolved down to 1 μ m resolution.
An even more elaborate system is based on 100 individually controlled pins
[244]. It can be actuated dynamically in a frequency range of 20 − 400 Hz. Fig-
ure 9.57 shows a schematic sketch. 20 piezoelectric bimorph bending actuators
(PZT-5H, Morgan Matrinic, Inc.) in five different layers one above the other are
located in a circuit around the stimulation area. Each bending actuator carries one
stimulation pin, which is placed 1 mm above the surface in idle state. The pins have
a diameter of 0.6 mm and are located equally spaced at 1 mm distance. At a maxi-
mum voltage of ± 85 V a displacement of ± 50 μ m is achieved. A circuit of equally
high passive pins is located around the touch area to mark the borders of the active
display.
268 9 Actuator Design

Fig. 9.56 Schematic setup of the “Texture Explorer” [106]

Fig. 9.57 Schematic setup of the 100-pin-array [244]

Another even more compelling system can be found in [149] (fig. 9.58). The
very compact 5 x 6 array is able to provide static and dynamic frequencies of up
to ≈ 500 Hz. Once again, piezoelectric bending actuators are used to achieve a dis-
placement of 700 μ m. However the locking-force is quite low with a maximum of
60 mN.

Ubi-Pen
The “Ubi-Pen” is one of the highest integrated tactile systems. Inside of a pen both
components, a spinning disc motor and a tactile display, are assembled. The design
of the tactile display is based on the “TULA35” ultrasonic linear drive (Piezoelec-
tric Technology Co). The schematic sketch of the design is given in figure 9.59. The
actuator is made of a driving component, a rod and the moving part. The two piezo-
electric ceramic discs are set to oscillation resulting into the rod oscillating up- and
downwards. The resulting movement is elliptical. To move upwards, the following
procedure is applied: in the faster downward movement the inertial force is excelled
by the speed of the operation and the element remains in the upper position. Whereas
9.4 Piezoelectric Actuators 269

Fig. 9.58 Schematic setup of the 5x6-pin-array [149]

in the upwards movement the speed is controlled slow enough to carry the moving
part up by the frictional coupling between moving element and central rod. The ac-
tuator discs have a diameter of 4 mm and a height of 0.5 mm. The rod has a length of
15 mm and a diameter of 1 mm. It can be used as contact-pin directly. The actuator’s
blocking force is larger than 200 mN and at a control frequency of 45 kHz velocities
of 20 mm/s can be reached.

Fig. 9.59 Schematic setup of the ultrasonic motor “TULA35” [152]

Figure 9.60 shows the design of a 3 x 3 pin array. Especially the very small size
of the design is remarkable: all outer dimensions have a length of 12 mm. The pins
are distributed in a matrix of 3 mm. On the area of 1.44 cm2 nine separate drives are
located. To achieve such a high actuator density the lengths of the rods have to be
different, allowing the moving parts to be placed directly next to each other. If this
unit is placed at the upper border of the display all pins move in- respectively out-of
270 9 Actuator Design

the plane. The weight of the whole unit is 2.5 g. When the maximum displacement
of 1 mm is used, a bandwidth of up to 20 Hz can be achieved.

Fig. 9.60 Tactile 3x3-pin-array [152]

The integration in a pen including another vibratory motor at its tip is shown in
figure 9.61. This additional drive is used to simulate the contact of the pen with a
surface. The whole pen weights 15 g.

Fig. 9.61 Prototype of the “Ubi-Pen” [151]

The Ubi-Pen provides surface structures such as roughnesses and barriers or

other extreme bumpy surfaces. To realize this, vibrations of the pins are superim-
posed with the vibratory motor. If the pen is in contact with a touch-sensitive surface
(touch-panel), the shown graphical image may be displayed in its gray scale values
by the pins of the tactile display. The system has been used for a number of tests
for recognition of information displayed in different tactile modalities [150]. The
results are very good with a mean recognition rate of 80 % with untrained users.
9.4 Piezoelectric Actuators 271

Tactile Displays with Lateral Penetration

Discrete Stimulation
The concept of discrete stimulation is based on an excitation lateral to the skin’s
surface (“laterotactile display”) [91]. On the left hand in figure 9.62 the schematic
sketch of a one-dimensional array of actuators is given. An activation of a piezoelec-
tric element results in its elongation and a deformation of the passive contact comb
(crown). If the skin of a touching finger is stretched by this movement a contact
point is perceived. Figure 9.62 shows a 2-dimensional display on the right. With the
extension from 1D to 2D array it is necessary to consider the more complex move-
ment patterns resulting from it. A more deep analysis of the capabilities of such
a system is found in [161], proving the applicability of a laterotactile display as a
virtual 6-point Braille device.

Fig. 9.62 1D array of the “laterotactile display” [91] and 2D STReSS2 display [161]

Continuous Stimulation
The transfer from the discrete points to a piezoelectric traveling wave drive is shown
in [22]. The touching finger faces a closed and continuous surface. Due to this de-
sign the tactile display itself becomes less sensitive in its performance to movements
of the finger. With the contact surface beyond the skin being excited as a standing
wave, the user perceives a surface texture. With relative movement between wave
and finger even a roughness can be perceived. By modifying the shape of the trav-
elling wave a simulation of a touch-force perceivable by the moving finger can be
achieved. Figure 9.63 shows the schematic sketch of the contact between finger and
traveling wave, such as the corresponding movement direction. In a test bed appli-
cation [22] the stator of a travelling actuator USR60 from Shinsei has been used.
This actuator provides a typical exploration speed by its tangential wave speed of
15 cm/s and forces up to ≈ 2 N. This system enables to generate continuous and
braking impressions by the change of wave shapes. An additional modulation of the
ultrasonic signals with a low frequency periodical signal generates the sensation of
272 9 Actuator Design

Fig. 9.63 Contact area of finger and travelling wave [22]

surface-roughnesses. Actual research is performed on the design of linear ultrasonic

travelling wave displays.

9.4.5.2 Piezoelectric Actuators for Kinaesthetic Systems

Piezoelectric actuators used in kinaesthetic systems are usually part of active sys-
tems (in a control-engineering sense). The user interacts with forces and torques
generated by the actuator. A classic example is a rotational knob which is actuated
by a travelling wave motor. With passive systems the actuator is used as a switching
element, which is able to consume power from an actuator or user in either time-
discrete or continuous operation. Examples are breaks and clutches.

Active Kinaesthetic Systems

Piezoelectric traveling wave actuators show a high torque to mass ratio if they are
compared to other electrical actuation principles. They are predestined for the use
in applications with a high torque at small rotational velocity, as they do not need an
additional bearing or other transmission ratios. Kinaesthetic systems require exactly
these properties. A very simple design can be used to build a haptic knob: a rotation-
ally mounted plate with a handle attached for the user is pressed on the stator of a
traveling wave piezoelectric motor. A schematic sketch of the critical part is shown
in figure 9.64. Due to the properties specific to the traveling wave motor the rota-
tional speed of the rotor can be adjusted easily by increasing the wave’s amplitude
w at the stator. As this actuation principle is based on a mechanical resonance mode,
it is actuated and controlled with frequencies nearby its resonance. Coincidentally,
this is the most challenging part of the design, as the piezoelectric components show
a very nonlinear behavior at the mechanical stator-rotor interface. Hence, the pro-
9.4 Piezoelectric Actuators 273

cedures for its control and its electronics have a large influence on its performance.

Fig. 9.64 Characteristics of the stator’s traveling wave beneath the rotor [72]

Figure 9.65 shows the speed vs. torque characteristics for different wave ampli-
tudes w of the actuator [43]. The torque is highly dependent on the actual rotational
speed and the amplitude. By monitoring the speed and controlling the phase and
the wave amplitude, the system can be closed-loop controlled to a linear torque-
displacement characteristic. In this mode a maximum torque of ≈ 120 mNm can be
achieved with this example. A deeper discussion of the phase control for a piezo-
electric travelling wave motor according to this example is given in [72].

Fig. 9.65 Speed / load torque-characteristics for different wave amplitudes [43]

A specialized design of such a device is used in neurological research for the

application with magneto-resonance tomography [59]. To close-loop control the ad-
mittance of the device the torque has to be measured. In the specific application near
a large magnetic field glass fibers are used, measuring the reflection at a bending
274 9 Actuator Design

polymer body. Preventing disturbance from the device on the switching gradients of
MRI and vice versa the device has been designed from specific non-conductive ma-
terials. It was based on a traveling wave motor in a special MR-compatible version
of the “URS60” from Shinsei.

Hybrid Systems
Another class of kinaesthetic systems are so called hybrid systems. If there is the
need of generating a wide bandwidth of forces and elongations with a single de-
vice, there is no actuator fulfilling all requirements alone. Due to this reason several
hybrid systems are designed with two (or more) components complementing each
other. A typical example is the combination of a dynamic drive with a brake, latter is
used to provide large blocking torques. As seen in the above paragraph the closed-
loop control of a traveling wave motor’s impedance is a challenging task. A simple
approach to avoid this problem is the combination of a traveling wave actuator with
a clutch. The main difference between a traveling wave actuator and other types of
actuators is given by its property to rotate with a load-independent velocity. Provid-
ing a certain torque this system is accompanied with a clutch. Other designs add a
differential gear or a break. Such a system is presented in [37]. If the system experi-
ences a mechanical load, the operation of the break is sufficient to increase friction
or even block the system completely. Consequently, the system provides the whole
dynamic range of the traveling motor in active operation, whereas passive operation
is improved significantly by the break. Due to the simple mechanical design and the
reduction of energy consumption such systems are suitable for mobile applications,
as well.

Passive Kinaesthetic Systems

Objects can be levitated by standing acoustic waves. To achieve this, an ultrasonic

source has to provide a standing wave first. Within the pressure nodes of the wave a
potential builds up, attracting a body placed nearby the node. The size of the object
is important for the effect to take place, as with a large size the influence on the
next node may be too high. A system based on this principle is described in [145].
It shows a design of an exoskeleton in form of a glove with external mechanical
guides and joints. The joints are made of piezoelectric clutches with a schematic
design shown in figure 9.66. In their original state both discs are pressed together by
an external spring generating a holding torque. With the vibrator being actuated the
levitation mode is achieved between rotor and stator creating a gap h. This reduces
the friction drastically allowing to turn both disks almost free from each other.
9.4 Piezoelectric Actuators 275

Fig. 9.66 Schematic setup of the “levitation-clutch” [145].

9.4.5.3 Summary

Tactile systems are distinguished according to their direction of movement. With

a normal movement into the direction of the skin additional differences are made
between passive systems simulating a more or less static surface by their pins, and
active systems - so called vibrotactile systems - providing information by a dynamic
excitement of the skin’s surface. The user integrates this information into a static
perception. The advantages of this approach are given by the reduced requirements
on the necessary force and displacements, as the dynamical perception of oscilla-
tions is higher than the perception of static slow-changing movements. When the
display is not fixed to the finger however its fast movements will be a problem. With
a static display in a fixed frame the user is able to repeatedly touch the display, in-
creasing the dynamics of the haptic impression by own movements. With a dynamic
display this interaction does not work as well anymore as periods of oscillations
from the vibrating elements are lost.
Another alternative are tactile systems with a lateral movement of the skin. With
an appropriate control the human can be “fooled” to feel punctual deformations
analog to an impression of a normal penetration. Systems with a closed surface are
very comfortable variants of such displays, but their dynamic control is demanding
for finger positions moving across larger surfaces. Typically today’s solutions show
smaller contact areas than with other variants, as the actuator elements can not be
placed as close together as necessary.
Kinaesthetic (force-feedback) systems can be distinguished in active and passive
systems according to a control-engineering sense. Active systems are able to gener-
ate counter forces and supportive forces. The spectrum of movements is only limited
by the degrees-of-freedom achieved by the machanical design. A stable control for
active system tends to become very elaborate due to required measurement technol-
ogy and complex control algorithm of sufficient speed. As with all active systems
a danger remains: an error of its functionality of the system may harm the user.
This danger increases for piezoelectric actuators, as the available forces and torques
276 9 Actuator Design

are high. Passive systems with breaks and clutches enable the user to feel the resis-
tance against their own movement as reactive forces. These designs are simpler to
build and less dangerous by definition of passivity. General disadvantages of passive
systems can be found in their high reaction times, the change of their mechanical
properties in long-time applications and their comparably large volume. Hybrid sys-
tems combining both variants - usually including another actuation principle - may
enlarge the application area of piezoelectric actuators. Although the mechanical de-
sign increases in volume and size, the requirements on control may become less
and large holding forces and torques can be achieved with low power consumption.
From a standpoint of haptic quality they are one of the best actuator solutions for
rotating knobs with variable torque/angle characteristics available today.
9.5 Electrostatic Actuators 277

9.5 Electrostatic Actuators

M ARC M ATYSEK

Electrostatic transformers are part of the group of electric transformers, such as

the piezoelectric actuators, too. Electric transformers show a direct coupling be-
tween the electrical value and the mechanical value. This is contrary to electrody-
namic and electromagnetic actuators, which show an intermediate transformation
into magnetic values as part of the actuation process. In principle the transformation
may be used in both directions. Hence, all these actuators can be used as sensors,
as well. Electrostatic field actuators are utilized due to their simple design and low
power consumption. As a result from the technical progress of micro-engineering,
the advantages of integrated designs are fully utilized. Especially for miniaturized
systems the electrostatic field actuators gain increased importance compared to all
other actuator principles. This is even more surprising as their energy density is
significantly lower in macroscopic designs. But during miniaturization the low effi-
ciency factor and the resulting power loss and heat produced become limiting factors
for magnetic actuators [159]. An important subgroup of electrostatic field actuators
is given by a solid state actuators, with an elastomeric dielectric. It has a high break-
down field strength compared to air, builds the substrate of the electrodes and can
simultaneously provide an isolating housing, too. Beside the classic field actuators
mentioned above, electro rheological fluids are part of the electrostatic actuators, as
well. With these actuators an electric field of an arbitrary external source results in
a change of the physical (rheological) properties of the fluid.

9.5.1 Definition of the Electric Field

The following paragraphs define the electric field and relevant variables for the de-
sign of electrostatic actuators.

9.5.1.1 Force on Charge

The magnitude of a force F acting on two charges Q1 and Q2 in a distance r is given

by C OULOMB ’ S-law (equation (9.76)).
1 Q1 Q2
F= (9.76)
4πε0 r2
278 9 Actuator Design

9.5.1.2 Electric Field

The electric field E describes the space where these forces are present. The field
strength is defined as the relation of the force F acting on the charge in the field and
the charge’s magnitude Q.
F
E= (9.77)
Q
The charges cause the electric field; the forces on the charges within an electric
field are the effect. Cause and effect are proportional. With the electric constant
ε0 = 8, 854 · 10−12 C/Vm within vacuum and air equation (9.78) results:

D = ε0 E (9.78)
The electric displacement field D describes the ratio of the bound charges and
the area of the charges. The direction is given by the electric field pointing from
positive to negative charges. If the electric field is filled with an insulating material
(dielectric), the electric displacement field is bound partly due to the polarizing of
the dielectric. Accordingly, the field strength drops from E0 to E (with still the same
electric displacement field). Consequently, the ratio of the weakened field depends
on the polarizability of the dielectric and is called “permittivity” εr = E0 /E.

9.5.1.3 Capacity

The electrical capacity is defined as the ratio of charge Q on each conductor to the
voltage U between them. A capacitor with two parallel-plates charged contrary with
a surface of the plates A and a fixed distance d shows a capacity C depending on the
dielectric:
Q A
C= = ε0 ε r (9.79)
U d

9.5.1.4 Energy Storage

Work must be done by an external influence to move charges between the conduc-
tors in a capacitor. When the external influence is removed, the charge separation
persists and energy is stored in the electric field. If charge is later allowed to return
to its equilibrium position, the energy is released. The work done in establishing the
electric field, and hence the amount of energy stored, is given by equation (9.80) and
for the parallel-plate capacitor by the use of equation (9.79) according to equation
(9.81).

1 1 Q2
Wel = CU 2 = (9.80)
2 2 C
9.5 Electrostatic Actuators 279

1 A
Wel = ε0 εr U 2 (9.81)
2 d
This stored electric energy can be used to perform mechanical work according to
equation (9.82).

Wmech = Fx (9.82)

9.5.2 Designs of Capacitive Actuators with Air-Gap

A preferred setup of electrostatic actuators is given by parallel-plate capacitors with

air-gap. In these designs one electrode is fixed to the frame, while the other one is
attached to an elastic structure, enabling the almost free movement in the intended
direction (DOF). All other directions are designed stiff enough to prevent a signifi-
cant displacement of this electrode. To perform physical work (displacement of the
plate) the energy of the electric field according to equation (9.81) is used. Consider-
ing the principle design of these actuators two basic variants can be distinguished:
the displacement may result in a change of the distance d, or the overlapping area A.
Both variants are subject of discussion in the following paragraphs.

9.5.2.1 Movement Along Electric Field

Fig. 9.67 Parallel-plate capacitor with air-gap

Looking at the parallel-plate capacitor from figure 9.67, the capacity CL can be
calculated with
A
CL = ε0 · (9.83)
d
As shown before the stored energy Wel can be calculated for an applied voltage U:
280 9 Actuator Design

1 1 A
Wel = CU 2 = ε0 U 2 (9.84)
2 2 d
The force between both plates in z-direction can be derived by the principle of
virtual displacement:

∂W 1 ∂C
Fz,el = = U2 (9.85)
∂z 2 ∂z
A 2
Fz,el = −ε0 U ez (9.86)
2d 2
The inhomogeneities of the electric field at the borders of the plates are neglected
for this calculation, which is an acceptable approximation for the given geometri-
cal relations of a large plate surface A and a comparably small plate distance d. A
spring pulls the moving electrode into its idle position. Consequently, the actuator
has to work against this spring. The schematic sketch of this actuator is shown in
figure 9.68. The plate distance d is limited by the thickness of the insulation layer dI .
Analyzing the balance of forces according to equation (9.87) the interdependency
of displacement z and electrical voltage U can be calculated:

Fz (z) = Fspring (z) + Fz,el (U, z) = 0 (9.87)

1 U2
− k · z − ε0 A =0 (9.88)
2 (d + z)2
k
U 2 = −2 (d + z)2 · z (9.89)
ε0 A

Fig. 9.68 Schematic setup of an actuator with variable air-gap

Analyzing the electrical voltage U in dependancy of the displacement z, a maxi-

mum can be identified:
9.5 Electrostatic Actuators 281

dU 2 k
= −2 (d 2 + 4dz + 3z2 ) = 0 (9.90)
dz ε0 A
4 1
z2 + dz + d 2 = 0
3 3
1
z1 = − d; z2 = −d (9.91)
3
To use the actuator in a stable state, the force of the retaining spring has to be
larger than the attracting forces of the charged plates. This condition is fulfilled for
distances z
1
0>z>− d
3
Smaller distances cause attracting forces larger than the retaining force and the
moving plate is strongly pulled onto the fixed plate (“pull-in” effect). As this would
immediately result in an electric short cut, typical designs include an insulating layer
on at least one plate. Equations 9.89 and 9.91 are used to calculate the operating
voltage for the pull-in:

8 k 3
Upull−in = d (9.92)
27 ε0 A
The retention force to keep this state is much less than the actual force at the
point of pull-in. It should be noted that force increases quadratically with decreasing
distance. A boundary value analysis for d → 0 provides the force F → ∞. Conse-
quently, the insulation layer fulfills the purpose of a force limitation, too.

9.5.2.2 Moving Wedge Actuator

A special design of air-gap actuators with varying plate distance is given by the mov-
ing wedge actuator. To increase displacement, a bended flexible counter-electrode
is placed on a base electrode with a non-conductive layer. The distance between
the electrodes increases wedge-like from the fixation to its free end. The resulting
electrical field is higher inside the area where the flexible electrode is closest to the
counter-electrode and it decreases with increasing air-gap. When designing the stiff-
ness of the flexible electrode it has to be guaranteed, that it is able to roll along with
the tightest wedge on the isolation. Figure 9.69 shows the underlying principle in
idle-state and during operation [216].
282 9 Actuator Design

Fig. 9.69 Schematic view of a moving wedge actuator

9.5.2.3 Movement Perpendicular to Electric Field

The major difference compared to the prior design is given by the fact, that the
plates are moving in parallel to each other. The plate distance d is kept constant,
whereas the overlapping area varies. Analog to equation (9.84) the forces for the
displacement can be calculated in both direction of the plane:

∂W 1 ∂C
Fx = = U2 (9.93)
∂x 2 ∂x
1 b
Fx = ε0 U 2 ex (9.94)
2 d
∂W 1 ∂C
Fy = = U2 (9.95)
∂y 2 ∂y
1 a
Fy = ε0 U 2 ey (9.96)
2 d
The forces are independent on the overlapping length only. As a consequence
they are constant for each actuator position. Figure 9.70 shows the moving electrode
being attached to a retaining spring.
If an electrical voltage is applied on the capacitor, the surface A increases along
the border a. Hence, the spring is deflected and generates a counter force FF accord-
ing to

FF = −kxEx (9.97)
The equilibrium of forces acting upon the electrode is given by

Fx (x) = FF (x) + Fx,el (U) (9.98)

From idle position (Fx (x) = 0) the displacement of the electrode in x-direction is
given by:
9.5 Electrostatic Actuators 283

Fig. 9.70 Electrostatic actuator with variable overlapping area

1 b1
x = ε0 U 2 (9.99)
2 dk
Typically, this design is realized in a comb-like structure, with one comb of elec-
trodes being engaged in a counter electrode comb. This equals an electrical parallel
circuit of n capacitors, which is identical to a parallel circuit of force sources com-
plementing each other. Figure 9.71 shows such a design. The area of the overlapping
electrodes is given by a in x-direction and b in y-direction. With the plate distance
d the capacity according to equation (9.100) can be calculated.

Fig. 9.71 Actuator with comb-electrodes and variable overlapping area

ab
CQ = ε0 ·
·n (9.100)
d
By differentiating the energy according to the movement direction the electro-
motive force can be calculated:
284 9 Actuator Design

∂W 1 ∂C 1 2 b
Fx = = U2 = U ε0 · n (9.101)
∂x 2 ∂x 2 d

9.5.2.4 Summary and Examples

For all actuators shown, the electrostatic force acts indirect against the user and
is transmitted by a moveable counter-electrode. A much simpler design of tactile
displays makes use of the user’s skin as a counter electrode, which experiences the
whole electrostatic field force. Accordingly, tactile electrostatic applications can be
distinguished in direct and indirect principles.

Direct Field Force

The simplest design combines one electrode, respectively a structured electrode ar-
ray, and an isolating layer. A schematic sketch is given in figure 9.72. The user
and his finger resemble the counter electrode. With the attractive force between the
conductive skin and the electrodes, a locally distributed increase of friction can be
achieved. It hinders a relative movement and can be perceived by the user. Such sys-
tems can be easily realized and excellently miniaturized. Their biggest disadvantage
is their sensitivity on humidity on the surface, which is brought onto the electrodes
during any use in form of sweat. This leads to a blocking of the electrical by the
conductive sweat layer above the isolation, preventing the user to feel any relevant
force.

Fig. 9.72 Electrostatic stimulator with human finger as coutner electrode [253].

Indirect Field Force

With these systems the field force is used to move an interacting surface. The finger
of the user interacts with these surfaces (sliders) and experiences their movements
9.5 Electrostatic Actuators 285

as a perceivable stimulation. A realization with a comb of actuators moving orthog-

onal to the field direction is given in figure 9.73. The structural height is 300 μ m
providing 1 mN at operating voltages of up to 100 V. The same design with an ac-
tuator made of parallel electrodes can achieve displacements of 60 μ m. The comb-
electrodes shown here displace 100 μ m.

Fig. 9.73 Electrostatic comb-actuator for tangential stimulation [71].

Summary

Electrostatic drives with air gap achieve force in the range of mN to N. As the ac-
tuators are driven by fields, the compromise between plate distance and electrical
operation voltage has to be validated for each individual application. The break-
down field strength of air (approx. 3V/μ m) is the upper, limiting factor. The actua-
tors’ displacement is limited to several μ m. At the same time the operating voltages
reach several hundred volts. Due to the very low achievable displacement the ap-
plication of electrostatic actuators is limited to tactile stimulations only. For the
concrete actuator design it is recommended to deal with the modeling of such actu-
ators, e.g. based on the theory of concentrated networks according to LENK [158].
This allows the analysis of the complete electromechanical system starting from the
applicable mechanical load situation to the electrical control with a single method-
ological approach.

9.5.3 Dielectric Elastomer Actuators

As with many other areas, new synthetic materials replace classic materials such as
metals in actuator design. Thanks to the enormous progress in material development
their mechanical properties can be adjusted to a large spectrum of possible applica-
tions. Other big advantages are given by the very cheap material costs. Additionally,
almost any geometrical shape can be manufactured with relatively small efforts.
286 9 Actuator Design

Polymers are called “active polymers” if they are able to change their shape and
form under the influence of external parameters. The causes for theses changes may
be manifold: electric and magnetic fields, light and even ph-value. Being used within
actuators, their resulting mechanical properties like elasticity, applyable force and
deformation at simultaneously high toughness and robustness are quite comparable
to biological muscles [10].
To classify the large variety of “active polymers” they are usually distinguished
according to their physical working principle. A classification into “non-conductive
polymers”, activated e.g. by light, pH-value or temperature, and “electrical poly-
mers”, activated by an arbitrary electrical source. Latter are called “electroactive
polymers” (EAP) and are further distinguished into “ionic” and “electronic” EAPs.
Generally speaking electronic EAP are operated at preferably high field strengths
near the breakdown field strength. Depending on the layer thickness of the di-
electrics 1 − 20 kV are typical operation voltages. Consequently, very high energy
densities at low reaction times (in the area of milliseconds) can be achieved. In con-
trast, ionic EAP are operated at obviously lower voltages of 1 − 5 V. However, an
electrolyte is necessary for transportation of the ions. It is frequently provided by
a liquid solution. Such actuators are typically realized as bending bars, achieving
large deformations at their tip with long reaction times (several seconds).
All EAP technologies are subject of actual research and fundamental develop-
ment. However, two actuator types are already used in robotics: “Ionic polymer
metal composite” (IPMC) and “dielectric elastomer actuators” (DEA). A summary
and description of all EAP-types is offered by K IM [137]. Their functionality will be
further discussed in the following paragraphs as they affiliate to the group of electro-
static actuators. A comparison between characteristic values of dielectric elastomer
actuators and the muscles of the human is shown in table 9.7. By the use of an
elastomer actuator with large expansion additional mechanical components such as
gears or bearings are needless. Additionally, the use of these materials may be com-
bined to complex designs similar to and inspired by nature. One application is e.g.
the locomotion of insects and fishes within bionic research [11].

Table 9.7 Comparison of human muscle and DEA according to Pei [196]
parameter human muscle DEA
strain (%) 20 − 40 10 bis > 100
stress (MPa) 0, 1 − 0, 35 0, 1 − 2
energy density (kJ/m3 ) 8 − 40 10 − 150
density (kg/m3 ) 1037 ≈ 1000
velocity of deformation (%/s) > 50 450 (acrylic)
34000 (silicone)
9.5 Electrostatic Actuators 287

9.5.3.1 Dielectric Elastomer Actuators - Electrostatic Solid State Actuators

Fig. 9.74 DEA in initial state (left) and charged state (right).

The design of dielectric elastomer actuators is identical to the design of a

parallel-plate capacitor, but with an elastic dielectric (a polymer respectively elas-
tomer) sandwiched by compliant electrodes. Hence, it is a solid state actuator. The
schematic design of a dielectric elastomer actuator is visualized in figure 9.74, left.
In an uncharged condition the capacity and the energy stored is identical to an air-
gap actuator (equations 9.79 and 9.80). A change of this condition happens by the
application of a voltage U and is visualized in figure 9.74, on the right: The charged
capacitor contains more charges (Q + Δ Q), the electrode area increases (A + Δ A),
whereas the distance (z − Δ z) simultaneously decreases. The change of energy after
an infinitesimal change dQ, dA and dz is calculated in equation 9.102:
     
Q 1 Q2 1 1 Q2 1
dW = dQ + dz − dA (9.102)
C 2 C z 2 C A
    
1 1
dW = UdQ +W dz − dA (9.103)
z A
The internal energy change equals the change of the electrical energy by the volt-
age source and the mechanical energy used. Latter depends on the geometry (parallel
(dz) and normal (dA) to the field’s direction). In comparison to the air-gap actuator
in section 9.5.2 an overlay of decreasing distance and increasing electrodes’ area
occurs. This is causes by a material property which is common to all elastomers and
to almost all polymers: the aspect of volume constancy. A body compressed in one
direction will extend in the remaining two other dimensions if it is incompressible.
This gives a direct relation between distance change and the change of electrodes’
area. As a consequence equation (9.104) results

A dz = −zdA (9.104)
simplifying equation (9.103) to
 
1
dW = UdQ + 2W dz (9.105)
z
288 9 Actuator Design

The resulting attractive force of the electrodes can be derived from this electrical
energy. With respect to the electrode surface A the electrostatic pressure pel at dQ =
0 is given according to equation (9.106)
1 dW 1
pel = = 2W (9.106)
A dz Az
and by the application of equation (9.80)
 
1 U2 1
pel = 2 ε0 εr Az 2 = ε0 ε r E 2 (9.107)
2 z Az
Comparing this result with equation (9.85) as a reference for a pressure of an air-
gap actuator with variable plate distance, dielectric elastomer actuators are capable
of generating a pressure twice as high with otherwise identical parameters [198].
Additional reasons for the obviously increased performance of the dielectric elas-
tomer actuators are based on their material, too. The relative permittivity is given
by εr > 1, depending on the material εr  3 − 10. By chemical processing and im-
plementation of fillers the relative permittivity may be increased. However, it has
to be noticed that other parameters (such as the breakdown field strength and the
e-modulus) may become worse, possibly the positive effect of the increased εr gets
lost. Especially the breakdown field strength is one of the most limiting factors.
With many materials an increase in breakdown field strength could be observed af-
ter planar prestrain. In these cases breakdown field strengths of 100 − 400 V/μ m are
typical.
The pull-in effect does not happen at z = 1/3 · z0 (air-gap actuators), but at much
higher deflections. With some materials mechanical prestrain of the actuator allows
to displace the pull-in further, reaching the breakdown field strength before. The
reason for this surprising property is the volume constant dielectric layer show-
ing viscoelastic properties. It complies with a return spring with strong nonlinear
force-displacement characteristics for large extensions. Its working point is dis-
placed along the stress-strain-curve of the material as an effect of the mechanical
prestrain.
For the application in dielectric elastomer actuators many materials may be used.
The material properties cover an extreme wide spectrum ranging from gel-like poly-
mers up to relatively rigid thermoplastics. Generally speaking, every dielectric mate-
rial has to have a high breakdown field strength and elasticity beside a high relative
permittivity. Silicone provides highest deformation-velocities and a high temper-
ature resistance. Acrylics have high breakdown field strength and achieve higher
energy densities. The following list is a selection of the dielectric materials most
frequently used today:

• silicone
HS 3 (Dow Corning)
CF 19-2186 (Nusil)
Elastosil P7670 (Wacker)
• acrylics
9.5 Electrostatic Actuators 289

VHB 4910 (3M)

The most frequently used materials for the elastic electrodes are graphite powder,
conductive carbon, and carbon grease.

9.5.4 Designs of Dielectric Elastomer Actuators

As mentioned before, dielectric elastomer actuators achieve high deformations

(compression in field direction) of 10 − 30%. To keep voltages within reasonable
ranges, layer thicknesses of 10 − 100 μ m are used depending on the breakdown
field strength. The resulting absolute displacement in field direction is too low to be
useful. Consequently, there are several concepts to increase it. Two principle move-
ment directions are distinguished for this purpose: the longitudinal effect in parallel
to the field (thickness change), and the transversal effect orthogonal to field (surface
area change). The importance of this discrimination lies in the volume constancy of
the material: uni-axial pressure load equals a two-axial tension load in the remain-
ing spatial directions. Hence, two transversal tensions within the surface result in a
surface change. For materials fulfilling the concept of “volume constancy”, equation
(9.108) is valid, providing the following properties for the longitudinal compression
Sz and the transversal elongation Sx :
• with a longitudinal compression of 62 % both extensions are identical;
• for smaller values of the longitudinal compression the resulting transversal ex-
tension is smaller;
• for a longitudinal compression > 62 % of the transversal extensions increases
much faster than the longitudinal compression.

1
Sx = √ −1 (9.108)
1 − Sz
The extension of the surface area SA depends on the longitudinal compression Sz
according to equation (9.109):
dA Sz
SA = = (9.109)
A 1 − Sz
The increase of the area with uni-axial compression is always larger than the
change of thickness. Actuators built according to this principle are the most effec-
tive ones. Figure 9.75 shows three typical designs. A roll-actuator (left) being built
as full or tubular cylinder can achieve length changes of more than 10 %. KORN -
BLUH [143] describes an acrylic roll-actuator achieving a maximum force of 29 N
with an own weight of no more than 2, 6 g at a extension of 35 mm. The manufacture
of electrodes with a large area is very simple. On the other hand the rolling of the
actuators with simultaneous prestrain (up to 500 %) can be very challenging. With a
290 9 Actuator Design

stack-actuator (middle) very thin dielectric layers with minimized operational volt-
ages can be achieved, depending mainly on the manufacturing technique. As the
longitudinal effect is used extension are limited to approximately 10 %. However,
due to their design and fabrication process actuator arrays at high density can be
built [119]. The most simple and most effective designs are based on a restrained
foil, whose complete surface change is transformed into an uparching (diaphragm-
actuator, right) [144]. If this actuator experiences a higher external load, such as
from a finger, an additional force source, e.g. a pressure, has to be provided to sup-
port the actuators own tension.

Fig. 9.75 Typical designs of Dielectric Elastomer Actuators: roll-actuator (left), stack-actuator
(center) and diaphrag-actuator (right).

9.5.4.1 Summary and Examples

As with air-gap actuators, a dielectric solid state actuator’s major limit is given by
the breakdown field strength of the dielectric. However, in contrast to the air-gap ac-
tuators, a carefully chosen design can easily avoid any pull-in effect. Consequently,
these actuators show a larger workspace; and with the high number of different de-
sign variants a wide variety of applications can be realized, just depending on the
requirements on displacement, maximum force and actuator density.

Tactile Displays

The simplest application of a tactile display is a Braille-device. Such devices are

meant to display Braille-letters in patterns of small, embossed dots. In standard
Braille six dots are being used, in computer compatible Euro-Braille eight dots.
These points are arranged in a 2x3 or 2x4 matrix, respectively (9.76). In a display-
device 40 to 80 characters are displayed simultaneously. In state-of-the-art designs
each dot is actuated with one piezoelectric-bending actuator (section 9.4.5). This
technical effort is reason for the high price of these devices. As a consequence there
9.5 Electrostatic Actuators 291

are several functional samples existing, which prove the applicability of less ex-
pensive drives with simplified mechanical designs, but still sufficient performance.
Each of the three variants for dielectric elastomer actuators has already been used
for this application.
Figure 9.76 shows the schematic sketch of roll-actuators formed to an actuator
column [206]. Each roll-actuator moves one pin, which itself is pushed up above
the base plate after applying a voltage. The elastomer film is coiled around a spring
of 60 mm length with a diameter of 1.37 mm. With an electric field of 100 V/μ m
applied, the pretensioned spring can achieve a displacement of 1 mm at a force of
750 mN. The underlying force source is the spring with spring-constant 225 N/m
pretensioned by a passive film. The maximum necessary displacement of 500 μ m is
achieved at field strengths of 60 V/μ m.

Fig. 9.76 Presenting a Braille sign with roll-actuators. left: geometry ; right: schematic setup of a
Braille row [206].

The application of stack actuators according to J UNGMANN [119] is schemat-

ically sketched in figure 9.77, left. The biggest advantage of this variant is given
by the extremely high actuator density combined with a simple manufacturing pro-
cess. Additionally, the closed silicone elements are flexible enough to be mounted
on an almost arbitrary formed surface. The surface - in itself being made of sili-
cone - shows an adequate roughness and temperature conductivity. It is perceived as
“convenient” by many users. With a field strength of 30 V/μ m a stack made of 100
dielectric layers displacements of 500 μ m can be achieved. The load of a reading
finger on the soft substrate generates a typical contact pressure of 4 kPa resulting
in a displacement of 25 μ m. This extension is considerably less than the perception
threshold of 10 % of the maximum displacement. For control of the array it has to
be noted, that the actuators are displaced in a negative logic. With applied voltage
the individual pin is pulled downwards.
The design of a Braille display with diaphragm-actuators according to H EYDT
[94] demonstrates the distinct properties of this variant. The increase of elastomer
surface results in a notable displacement of a pin from the device’s contact area.
However, a mechanical prestrain is necessary to provide a force. This can be either
generated by a spring or air pressure below the actuator. Figure 9.77 on the right
292 9 Actuator Design

gives a schematic sketch for a single point being pretensioned by a spring with a
diameter of 1.6 mm. At an operating voltage of 5.68 kV the actuator displaces in
idle mode 450 μ m.

Fig. 9.77 left: Actuator row with stack-actuators [119]; right: Use of diaphrag-Actuators [94].

Artificial Muscles

To advance research and development and to market the future potential of adaptive
material-systems media-friendly, YOSEPH BAR -C OHEN (Jet Propulsion Laborato-
rys of the NASA at “Institute of Technology”) initiated a competition in California
in 2001. In March 2005 during the SPIE’s Smart Structures and Materials Sym-
posium (Conference on Electriactive Polymer Actuators and Devices EAPAD) the
first contest between a 17 years old girl and several machines actuated by EAP
actuators took place. With two actuator variants from DEA and IPMC a synthetic
arm from Environmental Robots Incorporated (ERI) was able to generate sufficient
counter forces for 26 seconds, and achieved a stalemate situation. The second rank
was achieved by the arm of the Suisse “Eidgenössische Material- und Prüfanstalt
(EMPA)”. The used “muscles” have been made of seven roll-actuators and have
lifted a mass of approximately 38 kg.

9.5.5 Electro-Rheological Fluids

Fluids being influenced in their rheological properties (especially the viscosity) by

electrical field varying in direction and strength are called electro-rheological flu-
ids (ERF). Consequently ERF are classified as non-Newton fluids, as they have a
variable viscosity at constant temperature. The electro-rheological effect has been
observed for the first time in 1947 at a suspension of cornstarch and oil by W ILLIS
W INSLOW. Electro-rheological fluids include dipoles made of polarized particles,
9.5 Electrostatic Actuators 293

which are dispersed in a conductive suspension. These particles are aligned in an

applied electrical field. An interaction between particles and free charge carriers
happens. Chain-like microstructures are built between the electrodes [41, 100, 227]
in this process. However, it seems as if this is not the only effect responsible for
the viscosity change, as even when the microstructures [194] were destroyed a sig-
nificant viscosity increase remained. The exact analysis of the mechanism respon-
sible for this effect is subject of actual research. The viscosity of the fluid changes
depending on the strength of the applied electrical field. With an electric field of
1 − 10 kV/mm the viscosity may change up to a factor of 1000 compared to the
field-free state. This enormous change equals a viscosity difference between wa-
ter and honey. A big advantage of this method can be found in the dynamics of
the viscosity change. It is reversible and can be switched within one millisecond.
Therefore, electro-rheological fluids are suitable for dynamic applications, too. If
large field strengths are assumed, the ERF can be modelled as B INGHAM-fluid. It
has a threshold for linear flow characteristics: starting at a minimum tension τF,d
(flow threshold) the fluid actually starts to flow. The fluid starts flowing right below
this threshold. The shear forces τ are calculated according to equation (9.110):

τ = μ γ̇ + τF,d (9.110)
With μ being the dynamic viscosity, γ̇ the shear rate, and τF,d is the dynamic
flow limit. Latter changes quadratically with the electrical field strength (equation
(9.111)). The proportional factor Cd is a constant provided with the material’s spec-
ifications.

τF,d = Cd E 2 (9.111)
For complex calculations modelling the fluid’s transition to and from the state of
flow, the model is extended to a nonlinear system according to equation (9.112) (for
n = 1 this equals equation (9.110))

τ = τF,d + kγ˙n (9.112)

This general form describes the shear force for visco-plastic fluids with flow-
limit according to V ITRANI [273]. For an analysis of idle state with shear rate γ̇ = 0
the static flow-limit τF,s with τF,s > τF,d is introduced. When exceeding the static
flow limit, the idle fluid is deformed. With the specific material constants Cs and
Ere f equation (9.113) can be formulated:

τF,s = Cs (E − Ere f ) (9.113)

The materials used for the particles are frequently metal oxides, silicon anhy-
dride, poly urethane and polymers with metallic ions. The diameter of particles is
1 − 100 μ m, their proportion on the fluid’s volume is 30 to 50 %. As carrier medium
typically oils (such as silicone oil) or specially treated hydrocarbon are used. To ad-
ditionally improve the viscosity change, nanoscale particles are added in the electro-
rheological fluids, too (“giant electro-rheological effect” [75], [283]). In [49] and
294 9 Actuator Design

[202] further mathematical modelling is presented for the dynamic flow behaviour
of ER-fluids. The central property of ERF - to reversibly change the viscosity - is
used for force-feedback devices, haptic displays, and artificial muscles and joints.
As the change in viscosity is mainly a change in counter-forces but not in shape or
direct-forces, ERF actuators are counted to the group of “passive actuators”. For the
characterization of their performance, the ratio between stimulated and idle state is
used. They are built in three principle design variants [33]:

9.5.5.1 Shear Mode

Fig. 9.78 Using ERF to vary the shear force.

The ER-fluid is located in between two parallel plates, one fixed and one moving
relatively to the fixed one. The only constrain is given by a fixed inter-plate distance
d. If a force F is applied on the upper plate, it is displaced by a value x at a certain
velocity v. For the configuration shown in figure 9.78 the mechanical control ratio λ
can be calculated according to equation (9.116) from the ratio of dissipative forces
(field-dependent flow-stresses, equ. 9.115) and the field independent viscosity-term
(equation (9.114)) [199]. η gives the basis viscosity of the ER-Fluid (in idle state)
and τy the low-stress depending on the electrostatic field.
η vab
Fη = (9.114)
d
Fτ = τy ab (9.115)

Fτ τy d
λ= = (9.116)
Fη ηv
9.5 Electrostatic Actuators 295

Fig. 9.79 Varying the flow channel’s resistivity with ERF-actuators

9.5.5.2 Flow Mode

The schematic sketch of this configuration is shown in figure 9.79. Both fixed plates
form a channel, with the fluid flowing through it due to an external pressure differ-
ence p and a volume-flow V̇ . With an electric field E applied between the plates the
pressure loss increases along the channel and the volume flow is reduced. Analog
to the prior design a field independent viscosity based pressure loss pη and a field
dependent pressure loss pτ can be calculated [199]:

12η V̇ a
pη = (9.117)
d3b
cτy a
pτ = (9.118)
d
The mechanical control ratio equals

pτ cτy d 2 b
λ= = (9.119)
pη 12η V̇

At an adequate dimensioning of the fluid, the flow-resistance can be increased by

the electrical field to such a degree, that the fluid stops completely when exceeding
a specific voltage. This makes the channel a valve without any moving mechanical
components.

9.5.5.3 Squeeze Mode

A design to generate pressure is schematically sketched in figure 9.80. In contrast

to the variants shown before, the distance between both plates is subject to change
now. If a force acts on the upper plate, it moves downwards. This results in the fluid
being pressed outside. A plate distance d0 is assumed at the beginning, and a relative
movement of v of the plate moving downwards. The velocity dependent viscosity
force Fη and the field dependent tension term Fτ [117] are calculated according to:
296 9 Actuator Design

Fig. 9.80 Varying the acoustic impedance with ERF-actuators under external forces

3πη vr4
Fη = (9.120)
2(d0 − z)

4πτy r3
Fτ = (9.121)
3(d0 − z)
which gives the mechanical control ratio:

8τy 3(d0 − z)2

λ= (9.122)
9η vr
With pressure (force on the upper plate) the fluid is pressed out of the gap. In this
configuration the force-displacement-characteristics is strongly influenced by the
electrical field strength. An analysis of the dynamic behaviour of such an actuator
is described in [277].

9.5.5.4 Designing ERF-Actuators

The maximum force Fτ and the necessary mechanical power Pmech are the input val-
ues for the design of ERF-actuators from the perspective of an application engineer.
Equations 9.114 and 9.122 can be combined to calculate the necessary volume for
providing a certain power with all three actuator configurations.
η
V =k λ Pmech (9.123)
τy2
Consequently, the volume is defined by the mechanical control ratio, the fluid-
specific values η and τy , such as a constant k dependent on the actual configuration.
The electrical energy Wel necessary to generate the electrostatic field of the actuator
(volume-dependent) is calculated according to equation (9.124).
1
Wel = V ( ε0 εr E 2 ) (9.124)
2
9.5 Electrostatic Actuators 297

9.5.5.5 Comparison to Magneto-Rheological Fluids

Magneto-rheological fluids (MRF) are very similar to electro-rheological fluids.

However, the physical properties of the fluids are influenced by magnetic fields.
All calculations which are shown before are applicable to MRF, too. Looking at
the volume necessary for an actuator according to equation (9.123), considering the
viscosities of electro-rheological and magneto-rheological fluids being comparable,
a volume ratio proportional to the reciprocal ratio of the fluid-tensions’ square ac-
cording to equation (9.125) results:

VERF τ2
= MRF (9.125)
VMRF τERF
2

In a rough but good approximation the flow-stress of a magneto-rheological fluid

is one magnitude larger than of an ERF, resulting in a smaller (approximately factor
100) volume of a MRF actuator compared to the ERF. However, a comparison be-
tween both fluids going beyond the pure volume analysis for similar output power
is hard: for an ERF high voltages at relatively small currents are required. The main
power leakage is lost by leakage-currents through the medium (ERF) itself. With
MRF-actuators smaller electrical voltages at very high currents become necessary
to generate an adequate magnetic field. The energy for a MRF-actuator is calculated
according to equation (9.126) with the magnetic flux density B and the magnetic
field strength H.
1
Wel,MRF = VMRF ( BH) (9.126)
2
The ratio between the energies for both fluids is calculated according to equation
(9.127)

Wel,ERF VERF ε0 εr E 2
= (9.127)
Wel,MRF VMRF BH
With typical values for all parameters the necessary electrical energy for actuator
control is comparable for both fluids. A good overview on the design of actuators
for both types of fluids is given in [36].

9.5.5.6 Summary and Examples

Electro-rheological fluids are also called partly-active actuators, as they are not
transforming the electrical values into a direct movement, but change their prop-
erties due to the electrical energy provided. The change of their properties covers a
wide range. Naturally, their application in haptics ranges from small tactile displays
to larger haptic systems.
298 9 Actuator Design

Tactile Systems

The first application of ERF as tactile sensor in an artificial robot hand has been
made in 1989 by K ENALEY [127]. Starting from this work several ideas developed
to use ERF in tactile arrays for improving systems for virtual reality applications.
Several tactile displays, among them a 5x5 Matrix from TAYLOR [254] and another
one from B ÖSE [32], were built. Figure 9.81 shows the schematic design of such a
tactile element. A piston is pressed in an ERF filled chamber by the user. Varying
counter forces are generated depending on the actuation state of the ERF. Elastic
foam is connected to the piston as a spring to move it back to its resting position.
With an electric field of 3 V/μ m a force of 3.3 N can be achieved at a displacement
of 30 mm. Switching the electrical voltages is realized by light emitting diodes and
corresponding receivers (GaAs-elements) on the backplane.

Fig. 9.81 Schematic setup of a tactile actuator [32]

Haptic Rotary Knob

Another obvious application for ERF in haptic systems is their use as a “variable
brake”. This is supported by the fact that typical applications beside haptic systems
are variable brakes and bearings (e.g. adaptive dampers). There are several designs
with a rotary knob mobbing a spinning disk within an ERF or MRF [162] generating
varying counter torques. Figure 9.82 shows the schematics of a MRF-rotary knob.
In this case the measurement of the rotary angle is solved by a potentiometer. In
dependency on the rotary angle the intended counter force respectively torque is
generated. The user can perceive a “latching” of the rotary knob with a mature
system. The latching depth itself can be varied in a wide range. By the varying
friction hard stops can be simulated, too, such as sticking and of course free rotation.
An extension of the one-dimensional system is presented in [281]. Two systems
based on ERF are coupled to a joystick with two DOF. A counter force can be
generated in each movement direction of the joystick. As ERF are able to generate
9.5 Electrostatic Actuators 299

Fig. 9.82 Schematic setup of a haptic rotary knob [162]

higher torques with less energy required compared to a normal electrical drive, they
are especially suitable to mobile applications like in cars.

Force-Feedback Glove

A force-feedback glove was designed as a component for a simulator of surgeries

[12]. Surgical interventions shall be trained by the aid of haptic feedback. The sys-
tem MEMICO (“Remote Mechanical Mirroring using Controlled stiffness and Ac-
tuators”) shall enable a surgeon to perform the treatment with a robot in telemanip-
ulation, whereas the haptic perception is retained. ERF actuators are used for both
ends: on the side of the end-effector, and for the haptic feedback to the user. The
adjustable elasticity is based on the same principle as with tactile systems. For gen-
erating forces a force source is necessary. A new ECFS actuator (“Electronic Con-
trolled Force and Stiffness”) is used for this application. The schematic design is
shown in figure 9.83. It is an actuator according to the inchworm-principle, wherein
both brakes are realized by the ER-fluid surrounding it. The driving component for
the forward- and backward movement is realized by two electromagnets.
Both actuators are assembled within a haptic exoskeleton. They are mounted
on the rim of a glove to conserve the mobility of the hand. With the actuators in
between all finger-joints arbitrary forces and varying elasticities can be simulated
independently. The ECFS actuators are operated at voltages of 2 kV and generate a
force of up to 50 N.
300 9 Actuator Design

Fig. 9.83 Schematic setup of an ERF-Inchworm motor [12]

9.6 Special Designs of Haptic Actuators 301

9.6 Special Designs of Haptic Actuators

T HORSTEN A. K ERN

The actuation principles discussed so far are the most common approaches to
the actuation of haptic devices. Besides these principles, there are numerous re-
search projects, singular assemblies, and special classes of devices. The knowledge
of these designs is an enrichment for any engineer, yet it is impossible to completely
cover the variety of all haptic designs in a single book. This section, nevertheless,
intends to give a cross section of alternative, quaint and unconventional systems for
generating kinaesthetic and tactile impressions. This cross-section is based on the
authors’ subjective observations and knowledge and does not claim to be exhaustive.
The discussed systems have been selected, as examples suited best to cover one spe-
cial class of systems and actuators, each. They are neither the first systems of their
kind, nor necessarily the best ones. They are thought to be crystallization points of
further research, if specific requirements have to be chosen for special solutions.
The systems shown here are meant to be an inspiration and an encouragement not
to discard creative engineering approaches to the generation of haptic impressions
too early during the design process.

9.6.1 Haptic-Kinaesthetic Devices

Haptic-kinaesthetic devices of this category excel primarily due to their extraordi-

nary kinematics and not to very special actuation principles. Nevertheless, every
engineer may be encouraged to be aware of the examples of this device class and let
this knowledge influence his / her own work.

9.6.1.1 Spidar-System

The Spidar (fig. 9.84) is based on the work of Prof. S ATO and has frequently been
used in research projects [247, 181] as well as in commercial systems. It is com-
posed of an interaction handle - usually a ball - held by eight strings. Each string
is operated by an actuator, which is frequently (but not obligatorily) mounted in the
corners of a rectangular volume. The drives are able to generate pulling forces on
the strings, enabling the generation of forces and torques in six DOFs on the handle.
Typically the actuators used are based on electrodynamic electronic - commutated
units. The Spidar-system can be scaled to almost any size, ranging from table-top
devices to room-wide installations. It convinces by the small number of mechanical
components and the very small friction. As strings are able to provide pull forces
302 9 Actuator Design

only, it is worth noting that just two additional actuators are sufficient to compensate
this disadvantage.

a) http://ascii.jp/elem/000/000/023/23108/
b)

Fig. 9.84 a) Desktop-version of the Spidar with ball-like interaction handle, b) room-size version
“INCA 6D” with 3D visualization environment by Haption.

9.6.1.2 MRI-Compatible Haptic Device

The design of electrodynamic and electromagnetic actuators always includes a mag-

netic field and the corresponding flux guiding materials. Yet, when designing actua-
tors for medical applications to be used near or in a magnet-resonance tomography
(MRT, MRI) the developer is limited in his/her choice of materials. In the presence
of the strong static magnetic fields ferromagnetic materials are not applicable. At
the same time conductive materials cannot be used due to the high frequencies of
the switched gradient-fields inducing currents. For the application in haptic systems
inside or near a MRT, in [208] an electromagnetic actuator (fig. 9.85) was developed
and tested. It uses the existing static magnetic field of MRI as a counterpole to the
variable magnetic field generated by two coils. The torques generated reach impres-
sive 4 Nm at a distance of 1 m to the MRI, without influencing the imaging process
significantly.

9.6.1.3 Magnetorheological Fluids as Three-Dimensional Haptic Display

The wish to generate an artificial haptic impression in a volume for free interaction
is one of the major motivations for many developments. The rheological systems
shown in section 9.5 provide one option to generate such an effect. For several
years the team of B ICCHI has been working on the generation of spatially resolved
areas of differing viscosity in a volume (fig. 9.86) to generate force-feedback on an
exploring hand. Lately, the results were summarized in [20]. The optimization of
such actuators is largely dependent on the control of the rheological fluid [21]. The
psycho-physical experiments performed until today show that the identification of
9.6 Special Designs of Haptic Actuators 303

a) b)

Fig. 9.85 Working-principle of the MRI actuator by [208] (a) and actual design with coils visible
on its left side (b).

simple geometrical structures can be achieved on the basis of a 4x4 pattern inside
the rheological volume.

Fig. 9.86 Magnetorheological actuation principle for full-hand interaction based on a 4x4 pattern
[20].

9.6.1.4 Self-Induction and Eddy Currents as Damping

An active haptic device is designed to generate forces resp. torques in any direction.
By the concept of “active” actuation the whole spectrum of mechanical interaction
objects (e.g. masses, springs, dampers, other force sources like muscles, and moving
objects) is covered. Nevertheless, only a slight portion of haptic interaction actually
304 9 Actuator Design

is “active”. This has the side effect (of control engineering approaches) that active
systems have continuously to be monitored for passivity. An alternative approach to
the design of haptic actuators is given by choosing technical solutions able to dissi-
pate mechanical energy. A frictional brake would be such a device, but its properties
are strongly nonlinear and hard to control. Alternatives are therefore highly interest-
ing. The team of C OLGATE showed in [170] how to increase the impedance of an
electronic - commutated electrodynamic actuator, whereby two windings were by-
passed by a variable resistor. The mutual induction possible by this bypass damped
the motor significantly. In [76] the team of H AYWARD went even further by imple-
menting an eddy current break into a pantograph-kinematics (fig. 9.87). This break
is a pure damping element with almost linear properties. By this method a controlled
dynamic damping up to 250 Hz was achieved.

a) b)

Fig. 9.87 Principle of eddy currents damping a rotating disc (a) and realization as a haptic device
(b) by [76].

9.6.1.5 MagLev - Butterfly Haptics

In the 1990s the team of H OLLIS developed a haptic device [17] based on the elec-
trodynamic actuation principle (fig. 9.88). Since recently, the device has been sold
commercially by “Butterfly Haptics”. It is applied, e.g., to ongoing research projects
on psychophysical analysis of texture perception. Six flat coils are mounted in a
hemisphere with a magnetic circuit each. The combination of L ORENTZ-forces of
all coils allows an actuation of the hemisphere in three translational and three rota-
tional directions. Via three optical sensors - each of them measuring one translation
and one rotation - the total movement of the sphere is acquired. Besides the ac-
tuation within its space, the control additionally includes compensation of gravity
with the aid of all six actuators. This function realizes a bearing of the hemisphere
with Lorentz-forces, only. The air-gap of the coils allows a translation of 25 mm
9.6 Special Designs of Haptic Actuators 305

and a rotation of ±8◦ in each direction. As a consequence of the small mass of the
hemisphere, the electrodynamic actuator principle as a drive and the abandonment
of mechanical bearings, forces of a bandwidth of 1 kHz can be generated.

Interaction-handle

Flotor
Inner Stator
magnet assemblies (6)
Outer stator
a) b) magnet assemblies (6) Optical sensors (3)

Fig. 9.88 The MagLev device with a hand resting on its handle (a) and the sketch of the internal
structure (b) [17].

9.6.2 Haptic-Tactile Devices

Haptic-tactile devices of this category are intelligent combinations of well-known

actuator principles of haptic systems with either high position resolutions or extraor-
dinary, dynamic properties.

9.6.2.1 Thermo-Pneumatic

A classic problem of tactile pin-arrays is given by the high density of stimulator

points to be achieved. The space below each pin for control and reconfiguration of
the pin’s position is notoriously finite. Consequently, a large number of different de-
signs has been tested until today. In [272] a thermo-pneumatic system is introduced
(fig. 9.89) based on tubes filled with a fluid (methyl-chloride) with a low boiling
point. The system allows a reconfiguration of the pins within 2 seconds. however, it
has high power requirements, although the individual elements are very cheap.

9.6.2.2 Aggregate State for Fixation

An alternative approach is pursued in [182] by changing the state of matter from

fluid to fixed (fig. 9.90). The movement of the pins is brought about by air pressure.
306 9 Actuator Design

a) b)

Fig. 9.89 Thermo-pneumatic actuation principle in a schematic sketch (a), and as actual realization
(b) [272].

One part of each pin is sticking in its own pot filled with a metal (U-alloy) of a low
melting point. Each pot can be heated separately. In the fluid state of the metal the
pin is moved by air pressure, in the fixed state the pin is immovable. Based on this
principle, array resolutions of 2 mm are quite easily achievable.

a) b)

Fig. 9.90 Tactile display with a fixation of the individual pins by changing the aggregate state of a
metal: cross-section (a), and realization (b) [182].
9.6 Special Designs of Haptic Actuators 307

9.6.2.3 Fishbone-Illusion

Similar to the visual and acoustic perceptions, the haptic sense can be subject to
impressive illusions. Among the best known illusions is the one consequently re-
searched by NAKATANI [183]: generating the perception of concave or convex
lines when touching completely different geometrical structures. This “Fishbone”-
illusion (fig. 9.91) - according to today’s knowledge - is based on the effect that
shear forces are coupled to neighboring skin areas when touching these structured
surfaces, which show a similarity to tensions and strains generated by real shapes.
H AYWARD, an enthusiastic lateral thinker of haptic science, only recently published
an overview of the best-known tactile illusions [89], suggesting a structure of the
individual effects and their dependence of each other.

a) b)

Fig. 9.91 Fishbone-structure generating the illusion of a line with concave bending: in real pro-
portions (a), and as scheme (b) [183].

9.6.2.4 Piezoelectric Texture Actuator

Besides an application in Braille-related tasks, the design of tactile displays is rel-

evant for texture perception, too. At the World Haptics Conference in 2007, W IN -
FIELD impressively demonstrated a simple tactile texture display on the basis of
a piezoelectric disk in resonance mode [288]. By variation of frequency and am-
plitude it was possible to put the finger touching the disk into different degrees of
levitation, thus actually affecting the frictional coefficient between both. By the aid
of an optical tracking right above the disc, and with a corresponding modulation
of the control signal, perceivable textures with touchable, spatial resolutions were
generated (fig. 9.92).
308 9 Actuator Design

Glas
Piezoelectric disc

a) Mounting ring b)

Fig. 9.92 Piezoelectric disc on a glass substrate (a), and variants of displayable textures (b) [288].

9.6.2.5 Volume-Ultrasonic Actuator

I WAMOTO built tactile displays (fig. 9.93) which are made of piezoelectric actuators
and are actuated in the ultrasonic frequency range. They use sound pressure as a
force transmitter. The underlying principle is given by generating a displacement of
the skin and a corresponding haptic perception by focused sound pressure. Whereas
in the first realization an ultrasonic array had been used to generate tactile dots in a
fluid [112], later developments used the air for energy transmission [113]. The pres-
sures generated by the designs (fig. 9.93) provide a weak tactile impression, only.
But especially the air-based principle works without any mechanical contact and
could therefore become relevant for completely new operation concepts combined
with gesture recognition.

9.6.2.6 Perception of Surface Adhesion - Stickiness Perception

A seldom addressed effect of every day interaction is based on the perception of

adhesion between skin and object. Adhesion may result from pollution of surfaces,
but does frequently happen with soft materials like silicone, rubber, or, of course,
glues. A device generating the effect of adhesion is presented in [291], and used
with characteristic adhesion curves measured before. The device (fig. 9.94) is based
on three vacuum pumps, able to generate negative pressure below the finger in three
independent elliptic rings. By the variation of the vacuum within the concentrically
located cylinders, it is possible to retain the finger in three discrete steps. The com-
bined perception is very similar to the perception of sticking to surfaces with varying
adhesiveness.
9.6 Special Designs of Haptic Actuators 309

a) b)

Fig. 9.93 Tactile display based on ultrasonic sound pressure transmitted by a fluid [112](a) or as
array of senders for a transmission in the air [113] (b).

Fig. 9.94 Sketch of the device for generating a sensation of surface adhesion [291].

9.6.2.7 Tactile Contact Lens

As tactile perception mainly bases on the perception of tensions and strains coupled
to the skin, it is possible to use levers between object and finger to magnify fine
surface structures on the skin. This principle had been used in [136] to manufacture
a tactile contact lens as shown in figure 9.95. The tactile contact lens is a purely
mechanical device, using levers attached to a membrane in order to to extend the
strain in the skin resulting from the touched structures.
310 9 Actuator Design

a) b)

Fig. 9.95 Tactile contact lens for magnification by shear-tensions applied to the skin’s surface:
sketch (a), and realization (b) [136].

9.6.2.8 Tactile Markers by Tangential Forces

The findings that especially tangential forces and surface tensions are responsible
for tactile perception, suggest further research whether controlled shear forces are
able to generate perceptions comparable to a levitated pin on an otherwise flat sur-
face. The team of H AWYARDhas been developing shear-force displays for Braille-
related applications for several years now. The system shown in [161] (fig. 9.96)
uses piezoelectric bending actuators for the generation of tangential forces. The
piezoelectric bending actuators are additionally placed in a half-circuit to improve
the coupling to the skin. The generated tactile impression is very realistic and can
hardly be distinguished from a moving pin below the finger surface. A related tactile
illusion is known as “hair-comb-effect”.

a) b)

Fig. 9.96 Array made of piezoelectric bending actuators for shear-force generation in a planar (a),
and in a curved “STReSS2 ” design[161].
9.6 Special Designs of Haptic Actuators 311

9.6.2.9 Haptic Display for Free-Space Motions

In case of the interaction with large virtual worlds it is frequently necessary to design
devices which are worn on the body. An interesting solution has been shown in
[256], generating a tactile sensation with belts (fig. 9.97) at the palm and at each
finger. The underlying principle is based on two actuators for each belt, generating
a shear force to the skin when being operated in the same direction, and a normal
force when being operated in the opposite direction. This enables us to provide
tactile effects when grasping or touching objects in a virtual world, but without the
corresponding kinaesthetic effects. Nevertheless there are positive user responses to
such devices.

a) b)

Fig. 9.97 Worn tactile device for the generation of normal and shear-forces: sketch for a palm-
version (a), and the actual device (b)[256].

9.6.2.10 Electro-Tactile

As haptic receptors can be stimulated electrically, it is not far-fetched to design hap-

tic devices able to provide low currents to the tactile sense organs. The design of
such devices can be traced back to the 1970s. One realization is presented in [124]
(fig. 9.98). Electro-tactile displays do work - no doubt - however they have the dis-
advantage also to stimulate noci-receptors for pain sensation beside the mechano-
receptors. Additionally, the electrical conductivity between display and skin is sub-
ject to major variations. These variations are inter-person differences due to varia-
tions in skin-thickness, but they are also a time-dependent result of electro-chemical
processes between sweat and electrodes. The achievable tactile patterns and the abil-
ities to distinguish tactile patterns are subject to current research.
312 9 Actuator Design

a) b)

Fig. 9.98 Electro-tactile display worn on the forehead: Electrodes (a), and edge recognition and
signal conditioning principle (b) [124].
Chapter 10
Force Sensor Design

JACQUELINE R AUSCH

The following chapter deals with selection and design of force sensors, which are
implemented in haptic systems. In section 10.1 fundamental problems are discussed,
which are the basis of every sensor design process. A selection of factors, which
have to be taken in account, is made in section 10.1.5. After a short introduction in
basic transfer properties, the sensor characteristics are analyzed according to haptic
aspects and complemented by application examples.

10.1 Constraints

The topology of haptic systems significantly influences the design of force sensors.
The application of the haptic device itself has an extraordinary relevance. All sys-
tems have in common that an user mechanically contacts objects. On this has to be
clarified, which use of the device is intended, e.g. if it is going to be a telemanip-
ulator for medical purposes, or a CAD tool with force feedback. The mechanical
properties of the user itself and in case of telemanipulation systems the mechanical
properties of manipulating objects, have to be analyzed for the sensor development.
All these factors are going to be discussed within this section.

10.1.1 Topology of the Device

The application itself appoints the topology of the haptic device. Taking control
engineering aspects into account haptic systems can be basically classified in four

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_10, 313

c Springer-Verlag Berlin Heidelberg 2009
314 10 Force Sensor Design

types, which are discussed in chapter 5. In the following these topologies are anal-
ysed referring to the measured values:

• Open-loop control of impedance: Measurement of user movements (velocity or

displacement), feedback of a force
• Closed-loop control of impedance: Measurement of both user movements and
interaction, feedback of a force
• Open-loop control of admittance: Optional measurement of user force, feedback
of a position
• Closed-loop control of admittance: Measurement of both user force and move-
ments, feedback of a position
In case of an open-loop control only the mechanical properties of objects has
to be taken into account for force sensor design, independent if objects are physi-
cal or virtual ones. In case of haptic simulators like flight simulators virtual objects
are acting. The mechanical properties are often stored in look-up tables and force
sensors are dispensable. In case of telemanipulation systems like e.g. da Vinci, the
endeffector of the haptic system interacts with physical objects. Their mechanical
properties have to be measured with capable force sensors. Most of the telemanip-
ulation systems are impedance controlled. In case of closed-loop control the me-
chanical impedance of both user and manipulating object are considered. Designing
closed-loop impedance controlled systems force sensors have to be integrated into
the device detecting the user force. Designing closed-loop admittance controlled
systems the user movements have to be measured using e.g. a velocity sensor (chap-
ter 7).
Consequently, the measuring object can be both the user it self and a real, physi-
cal object. Beside its mechanical properties the modality of the interaction with the
haptic systems has to be analyzed to identify fundamental sensor requirements like
dynamic bandwidth, nominal load and resolution. The main factors influencing the
sensor design are both the contact situation and the objects’ mechanical properties.
In the following, they are analyzed by examining separately mechanical properties
and texture of the objects’ surface.

10.1.2 Contact Situation

It is necessary to distinguish between the user of the haptic system and the physi-
cal object due to the different interaction modalities identifying mechanical proper-
ties. If the user is the “measuring object”, interaction forces have to be measured.
Universally valid conclusions relating to amplitude, direction and frequency of the
acting force cannot be done. The mechanical impedance depends on the manner of
grasping the device, age and gender of the user itself (section 4.2). In section 4.2.1
the manners of grasping are classified: power-grasps, precision-grasps and touch-
grasps. In case of power- and precision-grasps finger or palm are used as counter
10.1 Constraints 315

bearing, which results in high absolute value of force up to 100 N [34, 88] and a
stiffer contact.
Additionally, the direction of the force vector has to be taken into account. De-
pending on application of the haptic device and manner of grasping up to six degrees
of freedom result - three force components and sometimes three torques. Neglecting
the torques between user and device three components of force have to be measured.
If the user is in static contact with the device’s handheld, measuring normal force
components with respect to the orientation of the contact plane is sufficient. If the
user is exerting relative movements to the device, also shear forces occur and three
components have to be measured.
Considering the frequency dependency of humans’ haptic perception, both static
and dynamic signal components have to be considered equally (chapter 3). The
lower cut-off frequency of haptic devices tends to quasi-static action at almost zero
Hertz, which may happen when a device is held without movement in free space.
If the force signal is subject to noise or even the slightest drift, the haptic impres-
sion will be disturbed soon (compare perception thresholds in section 3.1). Manner
and pre-load of grasping affect the upper cut-off frequency of the sensor. In case
of power- and precision-grasps, the absolute value of force achieves higher values
which results in an upper cut-off frequency being << 10,000 Hz 1 . Values of about
300 Hz are sufficient (section 3.1). Within the contact grasps the pre-load is much
lower than before enabling high frequency components to be transmitted directly to
the skin up to a range of approximately 1,000 Hz. In case of telemanipulation sys-
tems the endeffector interacts with a real, physical object. The assumptions made for
the measuring object “user” can partially be transferred to this situation. Following
Newton’s law actio et reactio, the absolute value of force depends on intensity and
way of interaction. Possible examples are the compression and lift of objects with a
gripper, or exploration with a stick. For telemanipulation systems in minimally inva-
sive surgery the absolute values of force range from 1 N to 60 N (comp. e.g. [210]).
The most promising approach is given by analyzing the intended application within
preliminary tests and derivation of a model. The dynamics of the interaction, espe-
cially of the upper cut-off frequency, is dominated by the mechanical impedance of
the object itself, which will be described within the following section.

10.1.3 Mechanical Properties of Measuring Objects

As stated for the user in chapter 5, the mechanical impedance of an object can
be subdivided into three physical actions: the elastic compliance n, the damping
d and the mass m. In case of rigid objects made of e.g. metal or ceramics, the prop-
erty of elasticity is dominant. The interaction between haptic systems and objects
can be considered as a rigid contact. Consequently, the force signal includes high-

1Please note: as stated in section 3.1, vibrotactile perception enables to feel frequencies of up to
10 kHz.
316 10 Force Sensor Design

frequency components. The upper cut-off frequency should take a value of at mini-
mum 1,000 Hz, to make sure to cover all dynamics responsible for haptic perception.
Soft objects, such as silicone or viscera have a viscoelastic material performance.
Following K ELVIN viscoelastic behavior can be simulated by a network made of
elastic compliances ni and damping elements di , such as masses mi . Using such an
equivalent network dynamic effects like relaxation and creeping can be modeled
(fig. 10.1 and fig. 10.2).

vin
vin
n2 d2
n2 d2
Fin Fin n1
n1
m

(a) Kelvin model modeling (b) Kelvin model extented by objects weight
dynamically effects for calculating the resonance frequency

Fig. 10.1 Kelvin model (standard linear solid) modeling viscoelastic behavior of objects. For cal-
culating the resonance frequency a mass element has to be added. Adding further damping and
spring elements the dynamic behavior of every object material can be modeled.

First of all the elasticity of mesuring objects has to be analysed during the sen-
sor design. An arithmetic example in section 3.3.1 compares the different cut-off
frequencies of materials. For soft materials such as rubber the upper cut-off fre-
quency takes values below 10 Hz. During interaction with soft materials mainly low
frequency components appear. The upper cut-off frequency is defined by the inter-
action frequency of 10 Hz at maximum [128, 204, 233]. If the measuring object is a
soft one with embedded rigid objects, like e.g. tumors in soft body tissue, an upper
cut-off frequency of about 1,000 Hz should be realized. To get more precise infor-
mation about the frequency requirements it can hardly be done without an analysis
of the interaction object. For a first rule of thumb calculated cut-off frequency as
derived in section 3.3.1 are sufficient. In case of doubt, the frequency range of the
sensor should always be oversized, not to already loose relevant haptic information
already at this very first point in the processing chain.
As the dynamics does, the required force resolution is also depending on a phys-
iological value. The Just Notable Difference (JND) lies in the range between 5 to
10 % of the absolute force value (chapter 3.1). From the JND the sensor character-
istics of measurement uncertainty can be derived. If realized as a constant value -
which is common to many technical sensor solutions - 5 % of the lowest apearing
value of force should be chosen to prevent distortion of the haptic impression of
the object. Nevertheless, there is no actual requirement for haptic applications to
10.1 Constraints 317

stress T

stress T
strain S
time t time t strain S
(a) relaxation (b) creep (c) hysteris

Fig. 10.2 Visualization of viscoelastic phenomena relaxation, creep and hysteresis.

have a constant or even linear sensor resolution. With telemanipulation systems the
interaction of the haptic system and real, physical objects is the main application.
Depending on the type of interaction, frequently the surface structure of objects, the
so called texture become equally or even more important than the object’s elastic
compliance. The resulting challenges for sensor development are discussed within
the following subsection.

10.1.4 Texture of Measuring Objects

The properties, which are relevant for the human perception of texture, are the
geometrical surface structure on the one hand (e.g. the wood grain), on the other
hand some kind of “frequency image” generated by the geometrical structure in the
(vibro-)tactile receptors when being touched by skin. To detect the surface structure
of an object, the variation of force against the contact area can be derived. For static
measurement sensor arrays of single-component force or pressure sensors are a
common technical solution. These arrays are put on the object. The structure of the
object generates different values of contact forces, providing a force distribution on
the sensor surface. The size of array and individual array elements can not be defined
in general, but is depending on the smallest detectable structure on the measurement
object itself. In case of the static measurement sketched above, number and size of
the sensor array elements should be dimensioned slightly smaller than the minimum
structure of the measuring object. The size of each element may be not larger than
half of the size of the smallest structure to be measured. However, even fulfilling
this requirement aberration will appear. Figure 10.3 shows that in case of the width
of the sensor element being larger or identical to the smallest structure the distance
between the elements is detected smaller than in reality. With n sensor elements the
width of the structure element is replayed to n+1 n−1
n and the distance to n . If the num-
ber of sensor elements per surface area increase, the aberration is diminishing and
318 10 Force Sensor Design

the structure is approximated more realistic (fig 10.4). However, with the number of
elements the effort of signal conditioning and analysis is increasing.

l
Dx
position 1 sensor 1 sensor 2 sensor 3

position 2 sensor 1 sensor 2 sensor 3 translational displacement of

the array (Dx)
l l
texture

l l
signal 1

signal 2

Fig. 10.3 Illustration of static and spatially resolved force measurement using as 3xn array. One
sensing element has the same dimension like a texture element. At position 1 the array is optimally
placed. If the array is shifted about Δ x to position 2 the texture is incorrectly detected.

l
Dx
position 1 S1 S2 S3 S4 translational displacement of
the array (Dx)
position 2 S1 S2 S3 S4

l l

texture

l l
signal 1

signal 2 1/2l 3/2l

Fig. 10.4 Illustration of static and spatially resolved force measurement using as 6xn array. The
size of one sensing element is half of a texture element. At position 1 the array is optimally placed.
In case of any other position an aberration occurs. The aberration decreases with increasing number
of sensing elements in an appropriate array.
10.1 Constraints 319

Beside the described aberration an additional disadvantage of static measure-

ments is given by the fact, that the knowledge of the texture is not sufficient to get
information about the object’s material. The complete haptic impression needs fre-
quency information depending on the elastic properties of texture and surface fric-
tion too. To gain these data, a relative movement between object and haptic system
should be performed, to measure the texture dynamically and spatially. Depending
on velocity of the relative movement and speed of the signal detection algorithms
the spatial resolution can be multiplied using the same number of sensor elements
as in the example shown before. Even the use of sensor array with a simultaneous
detection of multiple points becomes unnecessary. With knowledge about the explo-
ration velocity and its direction the information can be put into relation to each other.
For texture analysis multi-component force sensors should be used, as especially the
combined forces in the direction of movement and normal to the surface contribute
to haptic perception [184]. This dynamic measurement principle is comparable with
the intuitive exploration made by humans: To gain the texture of an object humans
gently touch and stroke over its surface. The surface structure excites the fingerprint
to oscillate and the vibrotactile sensors acquire the frequency image. The absolute
value of normal forces reached during such explorations are in a range of 0.3 to
4.5 N [35]. The force resolution is defined - as stated earlier - by the JND. The hap-
tic information about the texture is included into the high-frequency components
of the signal. For haptic applications the maximum frequency should be located at
1,000 Hz. The absolute value of nominal force should be chosen depending on the
elastic compliance of the object. If the object is softer, the absolute value can be
chosen lower as surface structures will deform and cannot be detected anymore. To
be able to measure equally good at soft and rigid objects, the nominal force should
take values ≤ 4.5 N. C ALDWELL [35] for examples decided to use F = 0.3 N.

10.1.5 Selection of Design Criteria

Following the description of the most relevant constrains, the limiting factors for a
sensor design in haptic applications can be found in physiological values. Nominal
force, force resolution, covered frequency range and measurement uncertainty can
be derived from humans’ haptic perception. For a quantitative analysis of these re-
quirements the contact between measuring object and force sensor is to bring into
focus. Measurement range and number of detectable force components are defined
by the application and the structure of the device. The geometrical dimensions and
other mechanical requirements are given depending on the point of integration into
the haptic system. The diagram displayed in figure 10.5 visualizes the procedure
how to identify the most important requirements for the sensor design
320 10 Force Sensor Design

structure of
haptic system

tele-
display
conditions of use

manipulation

closed loop open loop

control control
measurement
point of

end-effector and sensorless

end-effektor control configuration
control
detectable proper-
ties of the object

solid-
texture type of contact
mechanics

physiological parameters in consideration of

number of detactable force components

requirements

nominal load
resolution/measurement uncertainty
spatial resolution
nominal frequency range

Fig. 10.5 Tree diagram to identify the principle requirements on haptic force sensors. Beside me-
chanical characteristics of the object also physiological parameters of human haptic perception
have to be considered.
10.2 Sensing Principles 321

10.2 Sensing Principles

Within the previous section the most important criteria for the design and develop-
ment of a haptic sensor were named and introduced. Section 10.3 will summarize the
major requirements once again in tabular form. To help with the choice of a suitable
sensor principle, the variants according to figure 10.6 are presented in this section.
Beside the established measurement elements, such as resistive, capacitive, optic or
piezoelectric ones, other “exotic” sensor designs based on electro-luminescence or
active moving coils are discussed too.

resistive piezoelectrical

strain S prinziple of resonance

surface acoustic wave
conventional strain gauge
thinfilm strain gauge principle of ultrasound
thickfilm strain gauge
PZT-foil
stress T PVDT-foil
piezoresistive
contact area A

density r
optical

capacitive strain S
fibre bragg gratings
area A
distance d
distance d interferometry
intensity modulation

Fig. 10.6 Overview of established measurement principles for detecting forces in haptic systems.
Furthermore active sensor systems are also discussed.

Most sensor principles are active transformers using the displacement principle
for force measurement, which means that elasto-mechanic values such as stress or
strain are detected and the corresponding force is calculated. Sensors which belong
to the group of active transducers are resistive, capacitive, optic and magnetic ones
working according to the displacement-principle too. Piezoelectric, electro-dynamic
or electrostatic sensors belong to the group of passive transducers. After a short in-
troduction in elasto-mechanics each sensing principle will be discussed according
to its operating mode and several applications will be presented. All sensor prin-
ciples will be estimated according to their applicability for kinaesthetic and tactile
force measurement, and put into relation to the requirements known from chapter
6. At the end of this chapter a ranking method for the selection of suitable sensor
principles will be given.
322 10 Force Sensor Design

10.2.1 Basics of Elasto-mechanics

As aforementioned, a reasonable number of sensor principles base upon elasto-

mechanics. This section will summarize the fundamental knowledge which is nec-
essary for sensor design. A force is exerted to an elastic body. Depending on the
value of force the body deforms elastically. Internal stress T occurs resulting in a
shape change of the body - the strain S. Stress and strain are correlated by specific
material parameters, the so called elastic moduli si j . For a better comprehension a
short gedankenexperiment will be performed [220]. If a volume element Δ V is cut
from an object under load (fig. 10.7), the substitutive forces Δ F will act upon the
surfaces of the cuboid to keep the state of deformation. Due to the required state of
equilibrium the sum of all forces and torques acting upon Δ V must equal zero.

3 3
voxel dV DF T33
DA3
T32
T31
DF3 T23
T13
DF DF2 T22
DF1 T21
T12
T11
2 2
DA2
1 DA1 1

Fig. 10.7 Voxel dV of an elastic object. Due to deformation internal stress occurs which an be
described by element Ti j of the stress tensor [220].

Subdividing the force Δ F in its three components Δ F1 , Δ F2 and Δ F3 , just those

components remain orthogonal to the surface elements Δ A j . The quotient of the
acting force component Δ Fi and the corresponding surface element Δ A j results in a
mechanical stress Ti j . Following the balance condition Ti j = T ji six independent ten-
sion components remain, resulting in the stress tensor. The tensor elements can be
factorized into normal (stress parallel to the surface normal) and shear stress com-
ponents (stress orthogonal to the surface normal). Analyzing the volume element
Δ V before and after load, a displacement of the element Δ V with relation to the co-
ordinate system 123 such as a deformation happens. The sides of the cube change
their lengths and are not orthogonal to each other anymore (fig. 10.8).
For the description of this shape change a strain Si j is introduced. This strain is a
tensor too, consisting of nine elements (equ. 10.1)
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
d ξ1 S11 S12 S13 Δ x1
⎝ d ξ2 ⎠ = ⎝ S21 S22 S23 ⎠ · ⎝ Δ x2 ⎠ (10.1)
d ξ3 S31 S32 S33 Δ x3
10.2 Sensing Principles 323

position of point P in unstressed case:

Dx = (x1 x2 x3)
P’ translational displacement of P in stressed case:
P
dx Dx = Dx + dx
Dx
2

Fig. 10.8 The displacement of point P to P due to application of force visualizes the state of strain
[220].

Due to the volume constancy the following correlation can be defined

 
1 δ ξi δ ξ j
Si j = S ji = · + (10.2)
2 δ x j δ xi

and the matrix can be reduced to six linear independent elements. The normal strain
components act parallel to the corresponding normal to the surface. This is resulting
in a volume change. The shear components, acting normal to the surface, describe
the change of the angle between the borders of the volume element.
In case of isotropic materials, such as e.g. metals or Al2 O3 ceramics, the inter-
connection between the shape change mentioned before and the mechanical strains
can be formulated as follows:

⎛ ⎞ ⎛ ⎞ ⎛ ⎞
S1 s11 s12 s12 0 0 0 T1
⎜ S2 ⎟ ⎜ s12 s11 s12 0 0 0 ⎟ ⎜ T2 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ S3 ⎟ ⎜ s12 s12 s11 0 0 0 ⎟ ⎜ T3 ⎟
⎜ ⎟=⎜ ⎟ · ⎜ ⎟ (10.3)
⎜ S4 ⎟ ⎜ 0 0 0 2(s11 − s12 ) 0 0 ⎟ ⎜ T4 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝ S5 ⎠ ⎝ 0 0 0 0 2(s11 − s12 ) 0 ⎠ ⎝ T5 ⎠
S6 0 0 0 0 0 2(s11 − s12 ) T6

The six independent strain resp. stress components are summarized as a vector for
simplification. The components with the index 1, 2 and 3 mark the normal compo-
nents, those with indices 4, 5 and 6 the shear components [284]. The parameter si j
are regardless of direction. From them the known elasticity modulus E and shear
modulus G can be derived:
1 ν 1 2
s11 = , s12 = , = 2(s11 − s12 ) = (1 + 2ν ) (10.4)
E E G E
ν marks the so called P OISSON ratio, which is material dependent. With metals ν
achieves values between 0.25 and 0.35. For homogeneous materials equation 10.3
324 10 Force Sensor Design

can be reduced to a linear correlation T = E · S as a good approximation. For

anisotropic materials such as silicon or quartz the elastomechanic properties are
dependent on the orientation of the coordinate system (comp. section 10.2.3), re-
sulting in a matrix of elastic coefficients with up to 21 elements. For further reading
on elasto-mechanics e.g. [220] can be recommended.

10.2.1.1 Example “Beam Bending”

If a force vector is exerted to the tip of a beam bender made of isotropic materials
and clamped on one side (fig. 10.9), a bending moment MB acts upon the beam. The

y
x y
S2 b T(z)
S1 h z
z

l
F bending moment:
bending line w(z) MB = F(l-z)

Fig. 10.9 Behavior of a bending beam, the right-hand detail shows the stress distribution along the
profile.

mechanical stress components T (y) are linear distributed on the cross section and
take values of T (y) = c · y, whereas c is a proportional factor. The bending moment
equals the integral of the stress T (y) distributed on the cross section.
 
MB = y · T (y)dA = c · y2 dA (10.5)
A A
 2
As the integral A y dA equals the axial moment of inertia I, c is calculated as

MB
c= . (10.6)
I
The resulting strain components S1 and S2 act transversally to the beam’s surface.
Within the elastic region strain component S1 and the stress component T2 are cor-
related via the elasticity modulus E

T2 MB F · (l − z)
S2 = = = (10.7)
E I ·E I ·E
and therefore depending on the geometry of the cross section A of the beam, the
position z at the beam’s surface and the acting force F. For calculations of the strain
component S1 the transversal contraction has to be considered as follows
10.2 Sensing Principles 325

S1 = −ν · S2 . (10.8)
Further readings of elasto-mechanics, for example the calculations of defor-
mation of fiber-reinforced composites, the works of S CHNELL [220], W ERTH -
SCHÜTZKY [284] and BALLAS [8] are recommended.

10.2.1.2 Detection of Force

According to figure 10.9 acting forces can be measured evaluating both a resulting
strain distribution on the surface and a displacement of the beam. Following the
example above, the detection of strain S2 can be derived using B ERNOULLIs the-
ory. On this strain components acting transversal to the surface can be neglected for
slender and long beam-geometries. The stress or strain sensitive elements should
be placed in such a way, that a maximum surface strain change can be detected.
The correlations described above are examples for a clamped cantilever. Being able
to measure more than one force component, a suitable deformation element has to
be designed considering the elasto-mechanic correlations. Works of B RAY [28] can
help designing such an element. The primary objective is to generate a strain dis-
tribution in loading case, which enables to deduce the force components. Typically,
the interconnection between the force Fi and the electric signal vi of the sensor ele-
ment is given by a linear system of equations (e.g. [274]). Equation 10.9 shows an
example for a three-axial sensor:
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
v1 a11 a12 a13 F1
⎝ v2 ⎠ = ⎝ a21 a22 a23 ⎠ · ⎝ F2 ⎠ (10.9)
v3 a31 a32 a33 F3

It can be assumed that all force components contribute to each individual voltage
signal vi . The elements ai of the matrix can be found by calibrating the sensor. Dur-
ing the calibration process only one independent force component for each direction
is applied to the sensor and the resulting voltage components are measured. After
inverting the matrix A to A−1 the force vector can be calculated easily.

10.2.2 Resistive Strain Measurement

One of the most commonly used sensing principles for force sensing is based on the
resistive detection of strain components occurring in a (measuring) object. With the
resistive stress measurement a resistor pattern is applied on the bending elements
surface. The resistors have to be located in the areas of maximum strain. For your
information: The electrical resistance is defined via
326 10 Force Sensor Design

l l
R0 = ρ · =ρ· (10.10)
A b·h
ρ marks the specific resistance, l, b, h (length, width, height) define the volume of
the resistor itself. The total differential shown in equation 10.11 gives the relative
resistivity change resulting from the deformation.

dR dl db dh dρ
= − − + (10.11)
R0 l b
 h ρ

rel. volume changing piezoresistive part

The deformation causes on the one hand the change of the geometrical part Al , and on
the other hand the chance of the specific resistivity ρ . The material specific changes
will be discussed within section 10.2.3. Using equation 10.12 the correlation be-
tween strain and relative resistivity change is formulated:
 
dR d (N · μ )
= 2− ·S (10.12)
R0 S·N ·μ
 
:=k, k− f actor

whereas μ represents the electron mobility and N the number density of molecules.
The change of the resistivity can be measured using a so called W HEATSTONE
bridge circuit. This circuit is built of one to four active resistors connected in a
bridge circuit and fed by a constant voltage or constant current (fig. 10.10). Equa-
tion 10.13 calculates the bridge 10.10(c) with the assumption, that the basic resis-
tances R0i equal the resistance R0 . The values of R0 such as the k-factors are specific
to the material and listed in table 10.1 (further informations e.g. [215]).
! "
Vcc 1 r1 r2 r3 r4
Δu = = · − + − (10.13)
R0 · I0 4 R01 R02 R03 R04

The supply with a constant current has the advantage of the compensation of a
temperature dependent drift of the measurement signal. More advanced information
can be found in [44]. In case of metals a k-factor of approximately two occurs.
The material specific component of metals is less important and affects the first
decimal place only. In case of semiconductors and ceramic materials the material
specific component is dominant. In case of semiconductor-strain gauges k-factor
takes values up to 150. Using resistor pastes, applied in thick film technology on
substrates2 , and poly-silicon layers, sputtered in thin film technology the material
specific component is dominant. On this, k-factors achieve values of up to 18 in case
of thick-film resistors and up to approximately 30 for thin-film resistors. Table 10.1
lists the k-factor for several materials usually used in strain measurement.

2 For substrate material mainly (layer-) ceramics are used. Less frequent is the use of metals, as

isolating layers have to be provided then

10.2 Sensing Principles 327

RL RL RL

r1 R4 r1 R4 r1
r4
Dv Dv Dv
VCC VCC VCC
R2 R3 R2 r3 r2 r3

RL RL RL
(a) quater-bridge (b) half-bridge (c) full-bridge
configuration configuration configuration
Ri external precision resistor, ri strain gauge, RL lead resistance

Fig. 10.10 W HEATSTONE bridge configurations for evaluating one up to four resistors.

Table 10.1 k-factors of important resistive materials [284]

technology material k-factor basic resistance R0 in
Ω
foil strain gauge CuNi 2.1 120, 350, 700
thickfilm Bi2 Ru2 O7 18,3 ≈10k
TiON 4 ...5 ≈5k
thinfilm
Poly − Si 20 ...30 ≈5k
silicon micromachining p − Si 80.4 3k ... 5k

As mentioned earlier, strain gauges are manufactured in different technologies.

The most commonly used variants are foil-strain gauges; thick- and thin-film man-
ufactured measurement elements are found mainly in OEM-sensors and for specific
solutions in automation industry due to the necessary periphery and the manufactur-
ing process. Relevant literature can be found in the publications of PARTSCH [195]
and C RANNY [42].
Foil strain gauges are multilayer systems made of metallic measurement grids
and organic substrates. It is applied (fig. 10.11) and fixated on bending elements
via cold hardening cyanoacrylate adhesive (strain analysis) or via hot hardening ad-
hesives such as epoxy resin (transducer manufacture). These gauges are long-term
stable, robust, and especially used for high-precision tasks in wind-tunnel-scales
and balance sensors. Achievable dynamics, resolution and measurement range are
solely depending on the deformation element. The minimum size of the individ-
ual strain gauges taken of the shelf is in the area of 3 mm width and 6 mm length.
The measurement pattern itself is smaller in its dimensions. On this, it is possible
to shorten the organic substrate to finally achieve 1.5 mm width and 5 mm length
as a typical minimum size. If foil strain gauges are considered, the surface strains
328 10 Force Sensor Design

terminals organic substrate

(30...50 mm)
metal foil
(sensitive pattern, 5 mm) adhesive layer (5 mm)

substrate (steel, aluminium)

Fig. 10.11 Assembly of conventional strain gauges: the measuring grid is usually made of a tex-
tured metal foil. In case of special applications metal wires are applied.

resulting from the nominal load should be 1,000 μ m/m for an optimum usage of
the strain gauge. Many measurement pattern are applied for force and torque sen-
sors. Figure 10.12 shows a selection of commercialized measuring grids ready for
application on deformation elements.

two-element strain rosettes for detecting plain strain

(unknown principle directions)
90° 90° 90°
45° 45°
0°
0°
0°

three-element strain rosettes for detecting plain strain

(unknown principle directions)
90° 90°
45° 45°
90° 90° 45°0°
45°
0° 0°
0°
120° 60°
120° 60° 120° 60°

0° 0°
0°

Fig. 10.12 Compilation of possible grid configurations of strain gauges. See also [284].

Beside resistive foil strain gauges, semiconductor strain gauges can be purchased.
Their general design is comparable to conventional strain gauges, as the semicon-
10.2 Sensing Principles 329

ducting elements are assembled with organic substrates3 . The measurement ele-
ments are used identical to foil strain gauges and are available in different geo-
metrical configurations such as T-rosettes. Using measuring elements with a higher
k-factor (table 10.1) the deformation elements can be designed stiffer, allowing
smaller nominal strains. Such elements are especially relevant for the design of
miniaturized sensors for haptic systems, as small dimensions and high cut-off fre-
quencies have to be achieved. A commercially available example is the OEM-sensor
nano 17 from ATI (fig. 10.13). The strain elements are piezoresistive ones and their
k-factor takes values of approximately 150. Due to the high potential for minia-
turization and the manifold application in haptic systems, piezoresistive sensors -
especially silicon sensors - will be discussed in an independent subsection.

Fig. 10.13 Miniaturized force/torque sensor [6]. The resonance frequency of the sensor takes a
value of about 7,2 kHz.

10.2.3 Piezoresistive Silicon Sensoren

Published by Charles S. S MITH in 1954 for the first time [237], semiconducting
materials with a symmetric crystal structure such as silicon or germanium possess a
change in their conductivity σ due to an applied force or pressure. In the following
paragraphs this effect is discussed more deeply for monocrystalline silicon.

10.2.3.1 The Piezoresistive Effect

If a semiconducting material is deformed due to a load, stress components Ti are

generated inside the material. For your information: Due to the anisotropic prop-
erties of the material the elasto-mechanic properties are depending on the position
of the coordinate system, and consequently on the orientation of the crystal lattice.
3 Also single semiconducting elements without organic substrate are available. They are highly

miniaturized (width of about 230 μ m, length of about 400 μ m), but has to be insulated from the
deformation element.
330 10 Force Sensor Design

These stress components affect the electron mobility μ and - as a consequence - the
specific resistivity ρ . ρ is a material specific value, characterized via the parame-
ters electron mobility μ and number of charge carriers N (comp. section 10.2.1). By
considering these parameters the correlation between relative resistivity change and
the resulting strain tensor can be expressed to:

dρ dV d(N · μ ) V
= − , with ρ = , (10.14)
ρ V N·μ N · μ · |q|

whereas V is the volume of the resistive area and |q| is the charge of the particles.
Following the O HM’s law the specific resistance ρ is connected by the vector
E = (E1 ; E2 ; E3 )T of the electrical field and the current density J = (J1 ; J2 ; J3 )T :
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
E1 ρ11 ρ12 ρ13 J1 ρ1 ρ6 ρ5 J1
⎝ E2 ⎠ = ⎝ ρ21 ρ22 ρ23 ⎠ · ⎝ J2 ⎠ = ⎝ ρ6 ρ2 ρ4 ⎠ · ⎝ J2 ⎠ (10.15)
E3 ρ31 ρ32 ρ33 J3 ρ5 ρ4 ρ3 J3

Due to the symmetric crystalline structure of silicon4 six independent resistive com-
ponents ρi result, which are symmetrical to the diagonal of tensor ρ . With the matrix
of piezoresistive coefficients π the influence of the six acting stress components Ti
can be formulated. The cubic symmetry results in a reduction of the number of
piezoresistive and direction dependent coefficients to three. By doping silicone with
impurity atoms such as boron or phosphor areas of higher resistivity are generated.
By influencing the type and the concentration of dopant the three π -coefficients can
be influenced. Further information on doping can be found e.g. in [9].

⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
ρ1 ρ0 π11 π12 π12 0 0 0 T1
⎜ ρ2 ⎟ ⎜ ρ0 ⎟ ⎜ π12 π11 π12 0 0 0 ⎟ ⎜ T2 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ ρ3 ⎟ ⎜ ρ0 ⎟ ⎜ π12 π12 π11 0 0 0 ⎟ ⎜ ⎟
⎜ ⎟ = ⎜ ⎟+⎜ ⎟ · ⎜ T3 ⎟ · ρ0 (10.16)
⎜ ρ4 ⎟ ⎜ 0 ⎟ ⎜ 0 0 0 π44 0 ⎟
0 ⎟ ⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎜ T4 ⎟
⎝ ρ5 ⎠ ⎝ 0 ⎠ ⎝ 0 0 0 0 π44 0 ⎠ ⎝ T5 ⎠
ρ6 0 0 0 0 0 0 π44 T6

For homogenous silicone with a small concentration of dopant the values in ta-
ble 10.2 can be used.

Table 10.2 Piezoresistive coefficients of homogeneously doped silicon [9].

mm2 mm2 mm2
doping N in 1
cm−3
ρ in Ω cm π11 in N π12 in N π44 in N

n-Si 6 · 1014 11.7 −102.2 · 10−5 +53.4 · 10−5 −13.6 · 10−5

p-Si 1.8 · 1014 7.8 +6.6 · 10−5 −1.1 · 10−5 +138.1 · 10−5

4 face centered cubic

10.2 Sensing Principles 331

Depending on the angle between current density vector J and stress component Ti
three effects can be distinguished. With the so called longitudinal effect the current
i is guided parallel to the normal component of stress, with the transversal effect i
is guided normal to the normal component of stress, and with the shear effect i is
guided parallel or normal to the shear component of stress. Figure 10.14 visualizes
the described correlations.

doped area
(resistor) silicon T1 = Tlong, T2 = Ttrans, T4 = Tshear
h u
i
i
l

32
T2 T2
T1 T1 T4
1

(a) longitudinal effect (b) transversal effect (c) shear effect

(piezores. Hall effect)

Fig. 10.14 Visualization of the the piezoresistive effects: longitudinal, transversal and shear ef-
fect in silicon [284]. Transversal and longitudinal effect is normally used for commercial silicon
sensors.

For the resistivity change, depending on the orientation of the resistive area from
figure 10.14, the following equation becomes valid:

dρ
= πl · Tl + πq · Tq (10.17)
ρ0
As a consequence longitudinal and transversal stress components are influencing the
calculation of the resistivity change. Depending on the crystallographic orientation
of the resistive areas the π -coefficient is built from the longitudinal- and transversal
coefficient (table 10.2.3.1).
More advanced information for the design of piezoresistive silicon-sensors can
be found in the publications of BAO [9], W ERTHSCHÜTZKY [284] and S TAVROULIS
[240].

10.2.3.2 Examples of Piezoresistive Silicon Sensors

Piezoresistive silicon sensors for physical variables like pressure and force are com-
momly integrated in silicon deformation elements. For pressure transducers this
manufacture is state of the art and for all pressure ranges sensor elements can be
purchased. For example the company Silicon Microstructures Inc. (SMI) sells chips
332 10 Force Sensor Design

Table 10.3 Compilation of πl - und πq -coefficients for selected resistor assemblies dependent on
the crystallographic orientation [240].

surface longitudinal πl transversal πq

orientation
[100] π11 [010] π12
(100) π11 +π12 +π44 π11 +π12 −π44
[110] 2 [110] 2
π11 +2π12 +2π44 π11 +2π12 −π44
[111] 3 [112] 3
(110) π11 +π12 +π44 π11 +5π12 −π44
[110] 2 [001] 6

with glass-counter body for absolute pressure measurement with an edge length of
650 μ m (fig. 10.15(a)). In case of a suitable packaging these sensors can be arranged
in an array, to measure the uniaxial force- or pressure-distribution on a surface.

pressure sensor chip piezoresistive areas Ri

silicon deformation element
smi

borosilicate glass
(a) unseperated chips, edge length
(b) sectional drawing of the sensor
of about 650 mm

Fig. 10.15 Example of piezoresistive silicon pressure sensors [235].

In case of force sensors the realization of miniaturized multi-component force

sensors is subject to actual research. The dimensions of single sensor elements range
from 200 μ m to 2 mm. The nominal force covers a range of 300 mN to 2 N. Due
to the batch-manufacture of measurement elements, the realization of both single
sensor elements and array-design5 is possible. The sensitivity of the sensors takes
values of 2 % relative resistivity change in loading case. Figure 10.16 shows four
examples of actual topics of research. Variants (a) [266], (b) [271] and (d) [171]
were designed for the force measurement in haptic systems. Variant (c) [27] was
built for tactile, dimensional measurement technology. The force transmission is
always realized by beam- or rod-like structures.
Since 2007 a Hungarian manufacturer is selling the Tactologic system. Up to 64
miniaturized sensor elements are connected in an array of 3x3 mm2 . The sensor el-
ements have a size of 0.3x0.3 mm2 and are able to measure shear forces up to 1 N
5 By isolating arrays instead of single sensors in the last processing step.
10.2 Sensing Principles 333

(a) single tri-axial sensor (b) array of tri-axial sensors

1 mm

(c) single tri-axial sensor (d) single tri-axial sensor

Fig. 10.16 Different realisations of piezoresistive silicon force sensors: [266] (a), [271] (b), [27]
(c), [171] (d).

and normal force up to 2.5 N at nominal load. The force transmission is realized by
soft silicone dots, applied to every individual sensor element (fig. 10.17 (a) and (b)).
Using this array, static and dynamic loads in the range of kilohertz are measure-
able. But the viscoelastic material properties of the force transmission influence the
dynamics due to creeping, especially the measurement of the normal forces [270],
[269].

piezoresistive areas
force transmission

100 µm 1 mm

(a) one taxel of the sensing (b) 2x2 array for triaxial-force
array (Tactologic) measurement (Tactologic)

Fig. 10.17 Tactile multi-component force sensor [269].

334 10 Force Sensor Design

10.2.4 Further Resistive Sensors

Besides the resistive transducers presented until now, other more “exotic” design
exists, which will be introduced with three examples. All sensors are suitable for ar-
ray assembly to measure position-dependent pressure and a single force component.
The used measurement principles are based on the change of geometrical parameters
of the force elements. The examples shown in figure 10.18 (a) [139] and (b) [212]
use the load-dependency of the constriction resistance. With increased pressure6 the
electrical contact area A increases and the resistance decreases. The companies In-
terlink and TekScan use this effect for their sensor arrays (also called Force Sensing
Resistors - FSR). Interlink distributes polymer-foils printed with resistor pastes in
thick-film technology. Their basic resistance takes values in the region of M Ω . The
sensor foils have a height of 0.25 mm and a working range of zero to one Newton
respectively 100 N. Beside the sensitivity to force or pressure, the sensors show a
temperature dependency of 0.5 % K. The sensor foils from TexScan are located in
the range of 4.4 N up to 440 N, the spatial distribution reaches up to 27.6 elements
per centimetre. The available array size reach from approximately 13x13 mm2 up
to 0.5x0.5 m2 . The height of the foils is around 0.1 mm. The measurement inaccu-
racy takes a value of 10 %. The frequency range reaches from static up to 100 Hz.
Beside the application in data gloves, as described by B URDEA [34], the foil sen-
sors are used in orthopedy to detect the pressure distribution in shoes and prosthe-
sis and within automotive industry for ergonomic studies. Another approach is the
variation of the distance between two electrodes (Variant (b) in figure 10.18). The
sensing element is made of flexible substrates. The electrodes are arranged in rows
and columns. The gaps in between are filled with an electrically conductive fluid.
In loading case the fluid is squeezed out and the distance of the electrodes varies.
A disadvantage of this principle is given by the necessity for very large distance
variations up to 10 mm to achieve usable output signals. Until today this principle is
still a topic of research.

row
electrodes
conductive conductive
row
pyramids fluid
electrodes

column column
electrodes electrodes
(a) micromachined tactile array developed (b) variation of the electrodes distance
by Fraunhofer Institute IBMT

Fig. 10.18 Selected examples of foil sensors using the effect of a load-dependent constriction
resistance [139], [212].

6 The force can be calculated taking the contact area into account.
10.2 Sensing Principles 335

10.2.5 Capacitive Sensors

Within every capacitive sensor at least two electrodes are parallel located to each
other. Figure 10.19 shows a design based on a single measurement capacity.

d
A
U C = e0er A whereas er = 8,85 pF/m
d
e0, er

Fig. 10.19 Assembly of a single capacitance.

In contrast to the resistive principle - measuring the mechanical variables stress

and strain - the capacitive principle measures the integral values displacement (or
elongation) directly.
Concerning the working principle, three classes can be identified, which shows
some similarities to electrostatic actuators discussed in chapter 9.5. The first class
uses the displacement principle. On this, the mechanical load changes the electrode
distance d or the active electrode area A. In the third class the relative dielectric
εr is influenced. The change of electrode distance is usually used for measuring
force, pressure, displacement, and acceleration. In these cases the mechanical load
is directly applied to the electrode and displaces it relatively to the other one. The
resulting capacitance change can be calculated:

ΔC 1 ξ
= ≈± . (10.18)
C0 1 ± ξ /d d

ξ marks the change of distance. Additionally, the electrode distance can be kept
constant, and only one electrode can be parallel displaced (fig. 10.20). The active
electrode area varies accordingly and the resulting capacitance change can be used
to measure angle, filling level, or displacement. It is calculated according to:

ΔC ΔA
= 1± . (10.19)
C0 A0
The third option for a capacitance change is the variation of the relative dielectric.
This principal is often used for measuring a filling level e.g. of liquids, or as a prox-
imity switch for layer thickness. This capacitance change is calculated according
to
ΔC Δ εr
= 1± . (10.20)
C0 εr0
336 10 Force Sensor Design

10.2.5.1 Characteristics of Capacitive Pressure and Force Sensors

The main principle used for capacitive force- respectively pressure transducers is
measuring displacements. Consequently, the following paragraph will concentrate
on this principle. As stated within equation 10.18 for the change of distance, the
interconnection between capacitance change and mechanical load is nonlinear for
single capacities. The displacement ξ lies in the range of 10 nm to 50 μ m [284]. For
linearization of the characteristic curves an operating point has to be found, e.g. by
arranging three electrodes as a differential capacitor. The displacements ξ typical
for the working range are ≤ 10% than the absolute electrode distance d. In this
range the characteristic curve can be approximated as linear (fig. 10.20(a)). With
the principle varying the electrode’s surface the capacitance changes proportional to
it, resulting in a linear capacitance change (fig. 10.20(b)).

C, C2 C1

C0

C C2 C1
l/x
1 2 -1 -0,1 0,1 1
{

rangeability
(a) variation of electrode distance in case of single (1) and differential setup (2)
C

C0

C l/x
-1 1
(b) variation of effective electrode area

Fig. 10.20 Schematic view of capacitive sensors and their characteristic curve of capacitance.

The evaluation of the capacitance change can be made by an open- or closed-loop

measuring method. Concerning open-loop method, the sensor is either integrated in
a capacitive bridge circuit, or it is put into a passive oscillating circuit with coil
10.2 Sensing Principles 337

and resistor. Alternatively, the impedance can be measured at a constant measure-

ment frequency. An alternative could be the application of a pulse-width-modulation
(also called: re-charging method). A closed-loop approach is characterized by the
compensation of the displacement by an additional energy. The main advantage of
the closed-loop signal conditioning is the high linearity achieved by very small dis-
placements. Additional information can be found in [284] and [285]. The advantage
of capacitive sensor in contrast to resistive sensors lies in the effect of little power
consumption and high sensitivity. Additionally, the simple structure enables a low-
cost realization of miniaturized structures in surface micro-machining (fig. 10.21).
In contrast to the resistive sensors - where positions and dimensions of the resistive
areas have a direct influence on transfer characteristics - the manufacture tolerances
for capacitive sensors are quite high. Mechanically induced stress due to packag-
ing and temperature influence has almost no influence on their performance. Even
miss-positioning of electrodes next to each other do not change the transfer charac-
teristics, only the basic capacitance. The manufacturing steps with silicon capaci-
tive sensors are compatible to CMOS-technology. This allows a direct integration of
the sensor electronics on the chip, to minimize parasitic capacities. Especially with
miniaturized sensors7 a good signal-to-noise ration can be achieved [201]. The prob-
lem of parasitic capacities or leakage fields is one of the major challenges for the
design of capacitive actuators, as it achieves easily a level comparable to the capaci-
tance used for measurement. An additional challenge can be found in the constancy
of the dielectric value, which is subject to changes in open air-gap actuators due to
humidity or other external influence factors.

10.2.5.2 Examples of Capacitive Sensors

Concerning the manufacturing technology capacitive sensors integrated in haptic

systems can be distinguished in three classes. The first class is represented by minia-
turized pressure sensors, being realized using silicon microtechnology. Due to their
small size of few millimetres the moving masses of the sensor are low and thus
cover a wide dynamic range (frequencies from static to several kilohertz). As shown
before, the micro-machined capacitive sensors may be combined to arrays for mea-
suring spatially distributed load. As an example S ERGIO [223] reports the realiza-
tion of a capacitive array in CMOS-technology. A challenge is given by the capacity
changes in the range of femto Farad, which is similar to the capacity of the wiring.
A relief is given by a parallel circuit of several capacities to a sensor element [284].
The frequency range of the shown examples range from static measurement up to
several MHz upper cut-off frequency. Consequently, it is suitable for haptic related
measurements of tactile information. Another example is given by an array made of
poly-silicon. It has an upper cut-off frequency of 1,000 Hz and a spatial resolution of
0.01 mm2 suitable for tactile measurements. It was originally designed for acquisi-
tions of finger-prints. R EY [207] reports of the use of such an array for intracorporeal

7 Due to the small electrodes a small basic capacitance is achieved, comp. equation in figure 10.19.
338 10 Force Sensor Design

pressure measurement at the tip of a gripper. Once again the leakage capacities are
a problem, as they are within the range of the measured capacity changes.

capacitive combs

(a) 2-component force sensor, nominal load 1,5 mN

pipette
(injektion)
force sensor cell F

pipette
(holding)

capacitive combs

(b) 6-component force/torque sensor, nominal load 500 µN

Fig. 10.21 Examples of capacitive silicon multi-component force sensors [19] and [245].

Two examples of multi-component force sensors built in surface micro-machining

are shown in figure 10.21 (a) [19] and (b) [245]. The two-axial sensor8 is designed
for atomic force microscopy. The nominal load of this application lie in a range of
μ N. The three-axial sensor was designed for micro-manipulation, e.g. in molecu-
lar biology, with similar nominal values of several μ N. Both sensors are using the
displacement change for measurement. The second class is represented by ceramic
pressure load cells. They are widely used in automotive industry and industrial pro-
cess measurement technology. Substrate and measurement diaphragm are typically
made of Al2 O3 ceramics. The electrodes are sputtered on ceramic substrates. Sub-
strate and measurement diaphragm are connected via solder applied in thick-film
technology. In contrast to silicon sensors, ceramic sensors are macroscopic and have
dimensions in the range of several centimetres. Based on this technology sensors in

8 with respect to “force” component

10.2 Sensing Principles 339

differential-, relative-, and absolute-designs with nominal pressures in the range of

zero to 200 mbar such as in zero to 60 bar are available (e.g. fig. 10.22, Fa. Endress
und Hauser). The frequency range of these sensors is low, upper cut-off frequencies
of approximately 10 Hz are achieved. The third class is built from foil sensors, dis-

diaphragm (Al2O3)
CR 5,1 mm
CP
p 0,3 … 3 mm
thinfilm electrode 30 µm
CR
CP
main body
(Al2O3)
32

Fig. 10.22 Schematic view of a ceramic pressure sensor fabricated by company Endress und
Hauser [284] .

tributed e.g. by the company Althen. These capacitive sensor elements are arranged
in a matrix with a spacial resolution of ≤ 2x2 mm2 . As substrate a flexible poly-
mer foil is used. The height of such an array is 1 mm. The frequency range ranges
from static to approx. 1,000 Hz. Nominal loads up to 200 kPa can be acquired with
a resolution of 0.07 kPa. Due to creeping (comp. section 10.1.3) of the substrate and
parasitic capacities a high measurement inaccuracy exists. Another polymeric foil
sensor in the field of investigation is that one shown in figure 10.23 [38]. In contrast
to prior examples this array is used for direct force measurement. Normal forces
are detected measuring the change of electrode distance, shear forces by detecting
the change of active electrode surface. Similar to the sensor of the company Althen
static and dynamic changes up to 1,000 Hz can be measured. The spatial resolu-
tion is given with 1x1 mm2 . A disadvantage of the design is the high measurement
inaccuracy through creeping of the polymer and leakage capacities.

top view
1 electrode in upper layer
polyimide foil
side view
C1 C2

C4 C3
4 electrodes flexible dielectric
in lower layer

Fig. 10.23 Schematic view of capacitive shear force sensors [38].

340 10 Force Sensor Design

10.2.6 Optic Sensors

In the area of optical measurement technology sensors based on freely propagating

beams and fiber-optics are available. For force and pressure sensing mainly fiber-
optic sensors are used, which will be introduced further within this subsection. All
fiber-optic sensors have in common, that the mechanical load influences the trans-
mission characteristics of the optical transmission network, resulting in an influence
of the parameters of a reflected or simply transmitted electromagnetic wave. The
electromagnetic wave is defined by the wave equation [172].

δ 2Ψ δ 2Ψ δ 2Ψ
∇2Ψ = + + 2 (10.21)
δ x2 δ y2 δz
Ψ represents an arbitrary wave. A possible solution for this differential equation
is the propagation of a plane wave in open space. In this case, the electrical field
E and the magnetic field B oscillate orthogonal to each other. The electrical field
propagating in z-direction is described by equation 10.22.
1
E(z,t) = A(z,t) · e jω0 t−β0 z (10.22)
2
A marks the amplitude of the envelope, ω0 is the optical carrier frequency and the
propagation constant β0 . With the propagation group velocity vg (λ )9 the E- and B-
field are connected. Depending on the transmitting medium the group velocity can
be calculated via the refraction index ng [173].
c0
vg (λ ) = (10.23)
n(λ )

According to the wave length λ , n different values result. The waves are propagat-
ing differently depending on their frequency and wavelength. A pulse “spread out”.
For further information the sources [172], [173], [190], [298] and [175] are rec-
ommended. If only mechanical load such as force or pressure influences the trans-
mission network, the resulting deformation can influence the transmission in two
different ways:
1. Material specific: Change of the refraction index n (photoelastic effect)
2. Geometric: Change of beam guidance
The photo-elastic effect describes the anisotropy of the refraction index influenced
by mechanical stress. Figure 10.24 visualizes this effect. The resulting refraction
index change is dependent on the applied stress T and is given by the following
equation [98]:
Δ n = (n1 − n2 ) = C0 · (T1 − T2 ) (10.24)

9 In vacuum it is equal to the speed of light c0 = 2.99792458· 108 m/s.

10.2 Sensing Principles 341

C0 is a material specific, so called photo-elastic coefficient. Ti marks the resulting in-

ternal stress. Depending on the refraction index polarization, wave length and phase
of beam are changing. In the geometric case, the mechanical load changes the con-
ditions of the beam guidance. Using geometrical optics the influences of mechan-
ical loads on intensity and phase of radiation can be characterized. A disturbing

F, p E3
3 E2
d path difference of
voxel of field components
an optical fibre T3, n3
T2, n2
2

E3 1
E2

Fig. 10.24 Visualization of the photoelastic effect [98]. Due to the different refraction indices
perpendicularly to the propagation direction the propagation velocity of each field component is
different and an optical path difference δ occurs. The polarization is changing.

source for all fiber-optical sensors can not be neglected: the temperature. The re-
fraction index is depending on temperature changes, and consequently influences
the properties of the guided wave. Beside the thermo-elastic coefficients describing
the strain resulting from temperature changes within any material, the temperature
directly influences the refraction index itself (paragraph 10.2.6.3). For temperature
compensation a reference fiber has to be used, un-loaded and only influenced by
temperature change. An advantage of all fiber-optical sensors is given by their im-
munity to electromagnetic radiation. The following paragraphs introduce the most
important principles for optical force- and pressure measurement.

10.2.6.1 Changing Intensity

In principal, two transducer types varying the intensity can be distinguished. Both
have in common, that the mechanical load varies the condition of total reflection
(fig. 10.25). The angle αc is defined as the critical angle for total reflexion and
defined by S NELLIUS’ law:
n2
sin(αc ) = (10.25)
n1
The numerical aperture NA gives the appropriate critical angle θc for coupling the
radiation into a multimode fiber:
#
sin(θc ) = n21 − n22 (10.26)
342 10 Force Sensor Design

cladding (n2)
air (n0)
ac
qc
core (n1)

Fig. 10.25 Guidance of multimode fibers. Beams injected with angles above θc are not guided in
the core.

If the angle varies due to the mechanical load and takes values larger than θc resp.
smaller values than αc , the conditions for total reflections are violated. The beam
will not be guided within the core of the fiber. The total intensity of the transmitted
radiation will become less. Figure 10.26 show a schematic sketch of the design of
the very first variant. The sensor element is attached to the end of a multimode fiber.
In a first variant the light (e.g. emitted by a laser-diode λ = 1550 nm) is coupled

sensing element

fibre fibre
transmitter coupling
fibre

receiver displaceable reflector

Fig. 10.26 Schematic view of a fiber optic sensor with intensity modulation.

into a multimode fiber. A reflective element is attached to the end of the trans-
mission line. The element can be designed as a deformable object or a rigid one
mounted on a deformable substrate. The mechanical load acts on this object. Due to
the load the reflective element will be deformed (in case of a flexible surface) or dis-
placed (in case of a rigid surface). Varying the displacement the mode of operation
is comparable to a displacement sensor. The intensity is directly proportional to the
displacement (fig. 10.27). The load itself is a function of displacement and directly
proportional to the elastic compliance n of the sensor element:

F(z) = n · z. (10.27)
10.2 Sensing Principles 343

If the geometry of the area changes, a part of the beam - according to the laws of
geometrical optics - is decoupled into the cladding (dispersion) and an intensity loss
can be measured at the detector (fig. 10.27).

reflective object reflective object

F, p
n1 n1
n2 n3 n2 n3

(a) stiff element: intensity is depending on the displacement

reflective coating reflective coating

F, p
n1 n1
n2 n3 n2 n3

(b) flexible element: intensity is depending on the deformation

Fig. 10.27 Variation of intensity due to displacement of rigid and flexible elements.

In academically publications from P EIRS [197] and K ERN [129] such a mode
of operation is suggested for multi-component force measurement. In this case the
measurement range is directly proportional to the mechanical properties of fixation
of the reflective body. Using the calculation method known from section 10.2.1 this
fixation can be designed. A disadvantage of this principal is the use of polymers for
the coupling of the reflective object. This leads to creeping of the sensor signal. The
measurement inaccuracy of these sensors lies in a range of 10% [129]. Their diam-
eter takes a value of few millimetres. The length depends the application. Another
source of noise is the temperature. A temperature change leads to a dilatation (or
shrinkage) of the polymer itself and displaces the reflective element. The displace-
ment change results in a defective measurement signal. Due to the small size an ar-
ray assembly is possible. The second variant is a so called “micro-bending sensor”.
Its fundamental design is schematically given in figure 10.28. Like before a beam is
coupled into a multimode fiber. Force, pressure or strain is applied by a comb-like
structure results in micro-bending of the fiber (fig. 10.28(b)). In case of deformation
- similar to the first variant - a part of the light is decoupled into the cladding. The
intensity of the measured light diminishes10 . The gaps between the comb-like struc-
tures for micro-bending sensors are in the range of one millimeter. The height of
the structure is in the same dimension [243]. To applicate mechanical loads an area

10 Both versions are possible: Measuring the transmitted and the reflected radiation.
344 10 Force Sensor Design

microbender
multi-mode fibre F, p

S
deformation element
(a) assembly of microbending elements (b) beam guiding in loading case

Fig. 10.28 Variation of beam guidance in case of microbending.

of ∼
= 1 cm length and a width of ≥ 5 mm is used. The measurement range depends
on the displacement of the bending structure and the diameter of the fiber itself.
PANDEY [191] describes the realization of a pressure sensor for loads up to 30 bar.
If the bending diameter becomes smaller, lower nominal pressures and forces are
possible. Concerning the detection of force components, only one-component sen-
sors can be realized using this principle.
If spatially distributed mechanical load has to be measured, multiple micro-
bending structures can be located along one fiber. To evaluate the several measuring
points, for instance optical time domain reflectometry (ODTR) can be used. This
device sends a pulsed signal (light pulses of around 10 μ s length) guided in a fiber,
and measures the reflexion depending on time. Based on the propagation velocity
of the beam inside the fiber v, the time delay for each measuring point can be calcu-
lated by relating them. Additional information can be found in [166], [243] or [191].
The dynamics of these sensors is only limited by the sensor electronics and could
theoretically be applied to the whole range of haptic applications.

10.2.6.2 Changeing of Phase

The variation of the phase of light by mechanical load is used for interferometric
sensors. The most commonly used type is based on the Fabry-Pérot-interferometer,
discussed in the following paragraph. Other variants are Michelson- and Mach-
Zehnder-interferometers. The assembly is made of two plane-parallel, reflective and
semi-transparent objects e.g. at the end of a fiber, building an optical resonator (fig.
10.29). The beam is reflected several times within the resonator and interferes with
each reflection. The resonance condition of this assembly is given by the distance
d of the reflective elements and the refraction index n within the resonator. The so
called “free spectral range” marks the phase difference δ , generating a constructive
superposition of beams:
2π
δ= · 2 · n · d · cos(α ) (10.28)
λ
10.2 Sensing Principles 345

optical resonator
d
fibre fibre
transmitter coupling n
fibre

receiver

(a) schematic assembly of a Fabry-Pérot interferometer

Transmission

l
dl
(b) transmission spectrum (c) interferences in a resonator

Fig. 10.29 Assembly and operating mode of a Fabry-Pérot interferometer.

Figure 10.29(b) shows the typical characteristics of the transmission spectrum of a

Fabry-Pérot interferometer. According to the formula shown above the correspond-
ing wavelength yields a transmission peak; all other wavelengths are damped and
annihilated. Due to the mechanical load the distance d of the surfaces is varied,
changing the conditions for constructive interference. Sensors using this principle
are e.g. used by the company LaserComponents for uniaxial force- or pressure mea-
surement, and can be bought for nominal pressures up to 69 bar [155]. The influ-
ence of temperature would also appear to be problematic too and has to be com-
pensated by a reference configuration parallel working. Beside pressure transduc-
ers, single component forces and strains can be measured (fig. 10.30). The design
equals a Michelson-Interferometer. The sensor element is made of two multimode
fibers, whereas the strain acts upon only one fiber. Identical to the Fabry-Pérot-
configuration the sensor element is made of two plane-parallel reflective surfaces,
whose distance varies according to varying strain. Inside the measuring electronics
a reference design is included. To measure the mechanical load the phase of refer-
ence and measuring assembly is compared. This measurement principle enables to
measure frequencies in the range of several kilohertz. The geometrical dimension is
given by the diameter of the fiber including some protective coating ≤ 1 mm, and
the length of 2 to 20 mm depending on the application itself. For pressure sensors
the measuring error with respect to nominal load takes a value of about 0,5%, with
strain gauges at a factor of 15·10−6 .
346 10 Force Sensor Design

reading unit sensing element

coupling reflective area
fixed mirrow

moveable reference points for measurement of

mirrow length changes
PC LED

{ photodiode
AD conversion & amplification

Fig. 10.30 Temperature compensation in interferometric strain sensing elements [108].

10.2.6.3 Changing Wavelength

For optical detection of strain so called fiber B RAGG grating sensors (fbg sensor)
are widely used. To realize the sensing element the refractive index of the core in a
single mode fiber is varied due to the position (fig. 10.31) and a grating arise [221].
The refractive index modulation can be described by
  
2π
n(z) = n0 + δ ne f f ective (z) = n0 + δ ne f f ective · 1 + s · cos z + φ (z) (10.29)
Λ

whereas n0 is the refractive index within the core, δ ne f f ective is the average of the
index’s modulation and s a measure of the intensity of the index’s modulation. Λ
marks the grating period and the phase shift φ (z) resulting from the measured value.
In idle situation results φ (z) = 0. Figure 10.31 gives a schematic drawing of the
assembly.
If light is coupled into the fiber, only parts of it are reflected according to the
law of B RAGG. The reflective spectrum shows a peak at the so called B RAGG-
wavelength λb . This wavelength depends on the refractive index n(z) and the grating
period Λ :
λb = 2nΛ . (10.30)
In loading case both grating distance and refractive index varies. The maximum
of the spectrum is shifting from λ0 to another wavelength. According to the wave-
length shift mechanical load can be determined. The following condition is achieved:

Δλ δ n/n
= (1 −C0 ) ·(S + αV K · Δ ϑ ) + ·Δϑ, (10.31)
λ0   δϑ
k− f actor

whereas αV K is the coefficient of thermal expansion of the deformation body, and

C0 the photoelastic coefficient. Beside the change induced by mechanical strain S
the change of temperature ϑ influences the wavelength shift in the same dimen-
10.2 Sensing Principles 347

input
intensity

Bragg grating

L cladding
l core y
1250nm 1400nm x
z
reflexion
intensity

ncore(z)
n0+dneff(z)(1+s)
l n0+dneff(z)
1304nm 1305nm n0 n0+dneff(z)(1-s)
z

Fig. 10.31 Operational mode of fbg sensors [221] .

sion. Compensating the influence of temperature another FBG sensor has to be in-
stalled as reference at an un-loaded area. The temperature compensation is after-
wards achieved by comparison between both signals. Analog to resistive strain sen-
sors a k-factor of k ≈ 0.78 can be achieved with constant measurement temperature.
Extensions up to 10.000 μ m/m can be achieved. The width of the sensor lies in the
area of single mode fibers. The sensors length is defined by the grating, which has
to be three milimetres at least to provide a useable reflective spectrum [133, 26, 53].
The resolution takes a value of 0.1 μ m/m and is - such as its dynamics - defined by
the sensor electronics. Similar to strain gauges these sensors can be mechanically
applied on deformation elements, whose dimensions and shapes define the measure-
ment range. A challenge with the application of fiber-sensors in this context is the
differing coefficients of thermal expansion between deformable element, glue and
fiber. Additionally, the reproducibility of the glue-process for fibers is not as high as
typically required. Especially creeping of glue results in large measurement errors.
Comparable with the micro-bending principle, fbg sensors are applicable to several
spatially distributed measurement points. To distinguish the several positions grat-
ings with different periods Λi and thus different B RAGG-wavelengths λb are used.
Further information on the application of fbgs can be found in [221], [53], [26] and
[133] .

10.2.7 Piezoelectric Sensors

Piezoelectric sensors are widely used, especially for measurement of highly dy-
namic activities. The measurement principle is based on a measure induced charge
displacement within the piezoelectric material, the so called reciprocal piezoelectric
effect (section 9.4). The charge displacement leads to an additional polarization of
348 10 Force Sensor Design

the material resulting in a change of charge on the surface. This can be measured
using electrodes (fig. 10.32). Beside the measurement of force, it is for pressure- and
acceleration measurement especially. For force measurement the longitudinal effect
is primarily used. Detailed information about the piezoelectric effect and possible
designs are found in section 9.4. The materials uesd for sensing elements will be
introduced in the following paragraph.

electrode
F
+Q

3
2

E3=0 D3 T3 1
-Q
piezoelectric
material

Fig. 10.32 Visualization of piezoelectricity [284].

The technically relevant materials can be distinguished into three groups. The
first group is built of mono-crystals such as quartz, gallium-, and orthophosphate11 .
The polarisation change in case of mechanical load is direct proportional to the
stress. The transfer characteristic is very linear and does not have any relevant hys-
teresis. The piezoelectric coefficients are long-term stable. One disadvantage is the
small coupling factor k of about 0.1. For remembrance: k is defined as the quotient
of the transformed to the absorbed energy. The second group is formed by polycrys-
talline piezo-ceramics, such as barium titanate (BaTiO3 ) or lead zirconate titanate
(PZT, Pb(ZiTi)O3 ), being manufactured in a sintering process. The polarization is
artificially generated during the manufacturing process (section 9.4). An advantage
of this material is the coupling factor, which is seven times higher than that one of
quartz. A disadvantage is the nonlinear transfer characteristics with a noticeable hys-
teresis, and a reduced long-term stability. The materials tend to depolarize. The last
group is build from partial crystalline plastic foils made of polyvinylidene fluoride
(PVDF). Its coupling factor lies with 0.1 to 0.2 in the area of quartz. Advantageous
are the limit size (foil thickness of a few μ m) and the high elasticity of the material.
The first two sensor materials are used in conventional force sensors, as e.g. dis-
tributed by the company Kistler. Nominal forces take values of 50 N to 1.2 MN. The
sonsor typically has a diameter of 16 mm and a height of 8 mm. Alternations of load
up to 100 kHz are measureable. Single- as well as multiple-component sensors are

11 This crystal is especially applicable for high temperature requirements.

10.2 Sensing Principles 349

state of the art. 10.33 shows the general design of a three-component-force-sensor

from Kistler.

F F

isolator
cap electrode 1
cap

electrode 2 piezoelectric electrode 2

elements

(a) possible assemblies für piezoelectric force sensors

x
+ + + + + +
2
y - - - - - -
z - - - - - - x
+ + + + + +

+ + + + + +
3
- - - - - -
- - - - - - y
+ + + + + +

+ + + + + +
4
- - - - - -
- - - - - - z
1 + + + + + +

(b) assembly of a 3-component sensor:1 electrodes, 2 quartz plates shear effect - Fx,
3 quartz plates longitudinal effect - Fz, 4 quartz plates schear effect - Fy

Fig. 10.33 Possible assemblies of piezoelectric force sensors [284].

Piezoelectric force sensors are typically used for the analysis of the dynamic
forces occurring during drilling and milling or for stress analysis in automotive in-
dustry. In haptic system these sensor variants can hardly be found. Not exclusively
but mostly because they are not suitable to measure static loads. Sensors based on
PVDF-foils as piezoelectric material are increasingly used for the measurement of
tactile actions. The piezoelectric effect however is used for the generation of a dis-
placement and not for its measurement, making this variant being described in sub-
section 10.2.8.1.
350 10 Force Sensor Design

10.2.8 Exotic Ones

The sensor designs shown in this subsection are not force or pressure sensors
for conventional purposes. All of them have been designed for different research
projects in the context of haptic systems. The focus of these developments lies in
the spatially distributed measurement of tactile information.

10.2.8.1 Resonance Sensors

For measurement of vibrotactile information e.g. the so called resonance principle

could be used. Figure 10.34(a) shows the principal design of such a sensor. A piezo-
electric foil (PZT or even PVDF) is used as an actuator. Electrodes on both sides
of the foil apply an electrical oscillating signal, resulting in mechanical oscillations
of the material due to the direct piezoelectric effect. The structure oscillates at its
resonance frequency f0 calculated by the following formula

1 n
f0 = · (10.32)
2d ρ

whereas d is the thickness, n the elasticity and ρ the density of the used material.
The load, responsible for the deformation, is proportional to the frequency change
[200]. For spatially distributed measurement the sensors are connected as arrays of
elements with 3x3 and 15x15 sensors. The dimensions of the sensing arrays takes
values of 8x8 mm2 resp. 14x14 mm2 . The thickness of the foil is << 1 mm. A huge
disadvantage of this principle is the high temperature dependency of the resonance
frequency from the piezoelectric material used. The coefficient lies at 11.5 Hz per
1 ◦ C within a temperature range between 20 to 30 ◦ C [64], [139]. So called surface

piezoelectric material
row electrode
transmitter
PZT foil

column electrode
receiver

(a) principle of a surface acoustic (b) principle of a resonant sensor array

wave resonator

Fig. 10.34 Schematical view of reonance sensors. (a) [284], (b) [200] .
10.2 Sensing Principles 351

acoustic wave resonators, SAW sensors, make use of the change of their resonance
frequency too. The excitation occurs via an emitter called “Inter-digital structure”
(fig. 10.34(b)). The mechanical oscillations with frequencies in the range of MHz
distribute along the surface of the material. They are reflected on parallel metal
structures and detected by the receiving structure. Due to mechanical values ap-
plied the material is deformed, the runtime of the mechanical wave changes, and
consequently the sensor’s resonance frequency. With this design, the temperature
is one of the major disturbing values. SAW sensors are used for measurement of
force, torques, pressure and strain. The dynamic range reaches from static to highly
dynamic loads.

10.2.8.2 Electrodynamic Sensor Systems

Within the research project TAMIC an active sensor system for the analysis of or-
ganic tissue in minimally invasive surgery was developed [216]. The underlying
principle is based on an electrodynamic actuated plunger excited to oscillations
(comp. section 9.2). The plunger is magnetized in axial direction. The movements
of the plunger induce voltages within an additional coil included in the system. The
material to be measured is damping the movement, which can be detected and quan-
tified by the induced voltage. The maximum displacement of the plunger is set to
one milimetre. The system is able to measure dynamically from 10 Hz to 60 Hz. The
nominal force lies in the range of 200 mN. The geometrical dimensions of the sys-
tem are a diameter of ≤15 mm, and a length of ≤400 mm, which is near to typical
minimally invasive instruments. Detailed information can be found in [222].
Another example for a miniaturized sensor for the measurement of spatially dis-
tributed tactile information is presented by H ASEGAWA in [86]. Figure 10.35 shows
the schematic design of one element. The elements are arranged in an array struc-

diaphragm with piezoresistors

diffused into it
force transmission
permanent
magnet

substrate
flat coil

Fig. 10.35 Schematic view of an active element [86]. The dimensions are 6x6x1 mm3 .
352 10 Force Sensor Design

ture. In quasi-static operation mode the system is able to measure contact force
and the measurement object’s elasticity. The upper surface is made of a silicon-
diaphragm with a small cubical for force-application to the center of the plane. The
displacement of the plate is measured identical to a silicon-pressure or -force sensor
with piezoresistive areas on the substrate. By the displacement the applied contact
force can be derived. For measuring the elastic compliance of the object, a cur-
rent is applied to the flat coil (fig. 10.35). In the center of the diaphragm’s lower
side a permanent magnet is mounted. The electrically generated magnetic field is
oriented in the opposite direction of the permanent magnet. The plate is displaced
by this electromagnetic actuator and the cube is pressed back into the object. The
force necessary to deform the object is used in combination with the piezoresistive
sensors signal for calculation of the object’s elastic compliance. In the dynamic op-
eration mode the coil is supplied with an oscillating signal, operating the diaphragm
in resonance. Due to interaction with the measured object the resonance condition
changes. By the changing parameters, such as phase rotation, resonance frequency,
and amplitude, elastic coefficients such as damping coefficients of the material can
be identified. Due to the high degree of miniaturization highly dynamic actions up
to several kilohertz are possible to be measured. The nominal force lies in the area
of 2 N, the resolution of the system is unknown.

10.2.8.3 Electroluminescence Sensors

A high resolution touch-sensors is presented by S ARAF [213]. It is thought to be

used for the analysis of texture on organ surfaces. Figure 10.36 shows a schematic
sketch. On a transparent glass substrate a layer-compound of 10 μ m height made
of gold- and cadmium sulfite particles12 is applied. The single layers are separated
by dielectric barriers. The mechanical load is applied on the upper gold layer, re-
sulting in a break-through of the dielectric layer and a current flow. Additionally,
energy is released in form of small flashes. This optical signal is detected using a
CCD-camera. The signal is directly proportional to the strain distribution generated
by the load. The resulting current density is measured and interpreted. The spatial
resolution of the design is given with 50 μ m. Nominal pressures of around 0.8 bar
can be detected. The sensor area has a size of 2.5x2.5 mm2 , the thickness of the
sensor is ≤ 1 mm and thus very thin. Additional information can be found in [207].

10.3 Selection of a Suitable Sensor

In earlier sections first of all sources for the requirements identification have been
presented. Afterwards, the presentation and discussion of the most relevant sensor
principles to measure forces was made. This section is thought as an aid, to select

12 a semi-conducting material
10.3 Selection of a Suitable Sensor 353

I
F, p
device under test
sensing element
gold-plated V
contact area transparent
electrode

glass substrate
CCD collective lens

Fig. 10.36 Schematic view of an electroluminiscence sensor [213].

an appropriate force sensor. Depending on the identified requirements found using

section 10.1, a suitable sensor principle can be chosen.
To get a better overview, the basic requirements described in section 10.1.5 are
collected in table 10.4. The requirements are distinguished concerning the human
perception in kinaesthetic and tactile information. More detailed information con-
cerning force- and spatial-resolution can of course be found in section 10.1. The
properties of active and passive transformers - the force measurement is done via
a mechanical variable such as strain or stress detected with elastomechanics - are
strongly dependent on the design of the deformation element. Especially the nom-
inal force, number of components to be measured and the dynamics are directly
influenced by the deformation element’s design.

Table 10.4 Compilation of main requirements on haptic sensors

type of information requirements values
nominal load FN (5 ... 100) N
kinaesthetic
resolution Δ F 5 %FN
frequency range (0 ... 10) Hz
nominal load FN ≤ 0.3 N bzw. ≤ 4.5 N
tactile resolution Δ F 5 %FN
frequency range (0 ... 1,000) Hz
spatial resolution Δ x structural dependent, ≥ 1 mm

A comparison with other sensor principles can hardly be done. Consequently,

the methods will be compared separately from each other. As an evaluation crite-
rion transfer characteristics and geometrical dimensions are chosen. Figure 10.37
classifies the principles according to k-factor and geometry. If the strain sensing el-
ement becomes smaller, the smaller the total force sensor design can be. A direct
354 10 Force Sensor Design

result of the smaller size is the minimized mass, providing an increased upper cut-
off frequency. If the k-factor of the sensing element is higher, lower absolute value
of strain is necessary to get a high output signal. Additionally, the overall design
can be made stiffer. This enables to detect smaller nominal forces. Concerning the
lower cut-off frequency the strain sensing elements are suitable for measuring static
loads. Using piezoresistive and capacitive silicon-sensor an upper cut-off frequency
of 10 kHz or more can be measured with high resolution. The other sensor principles

area
2
[mm ]
a optic strain gauge, e.g. FBGS
10 f b metal foil strain gauge
b c thickfilm strain gauge
d thinfilm strain gauge
e silicon strain gauge
f surface acoustic wave
5
a

0,5 c
d
e
k
2 5 10 15 20 40 80

Fig. 10.37 Comparison of different strain measurement technologies due to dimensions and gauge
factor.

can be compared contingent on the nominal load and the dimensions. Figure 10.38
classifies the presented principals according to their nominal load and corresponding
construction space.
Except the piezoelectric sensors, all sensor principles can be used for measur-
ing static and dynamic loads. The upper cut-off frequency mainly depends on the
mass of the sensor which has to be moved. Consequently, the more miniaturized
the sensor the higher the upper cut-off frequency becomes. Figure 10.39 compares
the presented sensor principles according to the detectable nominal load and the
corresponding dynamic range.
With the aid of the shown diagrams a pre-selection of suitable sensor principles
for the intended application can be done. Additional sensor properties such as res-
olution, energy consumption, costs or impact of noise are strongly depending on
the individual realization and will not be taken into account here. Advanced de-
scriptions of sensor properties have to be taken from literature highlighted in the
corresponding subsections for the individual principle.
10.3 Selection of a Suitable Sensor 355

volume
3
[mm ]
capacitive
(ceramic)

100 piezoelektrisch*)

piezoresistive
piezoelectric**) (ATI)
10 (system)

fibre optic

capacitive capacitive
(silicon) resistive (FSR) (foil)
1

piezoresistive***)
FN [N]
1 10 100
*) commercial force/torque sensors
**) active sensor systems classified as exotic sensors
***) Tactologic system

Fig. 10.38 Comparison of different measurement technologies due to dimensions and nominal
load.
356 10 Force Sensor Design

FN [N]

100 capacitive piezoelectric*)

(ceramic)

capacitive (foil)

piezoresistive (ATI)
10
resistive (FSR)

piezoelectric**) (system)
1

fibre optic piezoresistive***)

capacitive (silicon)

f [Hz]
0 1 10 100 1000
*) commercial force/torque sensors
**) active sensor systems classified as exotic sensors
***) Tactologic system

Fig. 10.39 Comparison of different measurement technologies due to nominal load and frequency
range.
Chapter 11
Application of Positioning Sensors

T HORSTEN A. K ERN

To acquire the user’s reaction in haptic systems, a measurement of positions

respectively their time derivatives (velocities, accelerations) is necessary. Several
measurement principles are available to achieve this. A mechanical inlfuence of
the sensor on the system has to be avoided for haptic applications, especially ki-
naesthetic ones. Consequently this discussion focuses on principles which do not
affect the mechanical properties significantly. Beside the common optical measure-
ment principles the use of inductive or capacitive sensors is promising especially
in combination with actuator design. This chapter gives an overview about the most
frequently used principles, amended by hints for their advantages and disadvantages
when applied to haptic systems.

11.1 Basic Principles of Position Measurement

For position measurement two principle approaches can be distinguished: differen-

tial and absolute measuring systems.

11.1.1 Incremental Principle

Differential systems acquire the change in discrete steps together with the direction
of change, and protocol (typically: count) these events. This protocol has to be set
back to a reference position by an external signal. If no step loss happens during
movement, a prior initialized differential system is able to provide the absolute po-
sition as output. If this initializing reference position is set in point which is passed

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_11, 357

c Springer-Verlag Berlin Heidelberg 2009
358 11 Application of Positioning Sensors

often, a differential system will be referenced frequently during normal operation.

Potential step losses would then affect the time till the next initializing event only.
Measurement of the steps is done via a discrete periodic event, typically encoded
in a code disc with grooves or a magnetic rotor. This event is transformed by the
sensor in a digital signal, whose frequency is proportional to the velocity of the
movement (fig. 11.1a). Some additional directional information is required to be
able to measure the absolute position. A typical solution for this purpose is the use
of two identical event types with a phase shift (between 1◦ and 179◦ , typically 90◦ ).
By looking at the status (high/low) of these incremental signals (fig. 11.1b) at e.g.
the rising edges of the first incremental signal (A), a low encodes one movement di-
rection, and a high encodes the opposite movement direction. Accordingly the count
process either adds or subtracts the pulses generated - in this case - by the second
signal (B). State-of-the-art microcontrollers are equipped with counters for incre-
mental measurement already. They provide input pins for count-signal and count-
direction. Discrete counters are sold as “Quadrature-Encoder” ICs and frequently
include actuator drive electronics, which can be applied for positioning tasks. Latter
prevents them from being useful for typical haptic applications.

Direction Direction
U[V] U[V]

A A

t t
B B

a) t b) t

Fig. 11.1 Principle of direction detection with two digital signals with a 90◦ phase-lag.

11.1.2 Absolute Measurement Principle

Absolute measurement systems acquire a position- or angle-proportional value di-

rectly. They are usually analog. A reference position for these systems is not neces-
sary. They have advantages with reference to their measurement frequency, as they
are not required to measure with a dynamic defined by the maximum movement ve-
locity. The acquisition dynamics of incremental principles is given by the necessity
not to miss any events. In case of absolute measurement principles the measurement
frequency can be adjusted to the process-dynamics afterwards, which is usually less
demanding. However by the analog measurement technology the efforts are quite
high for the circuit, the compensation of disturbances, and the almost obligate digi-
11.2 Requirements in the Context of Haptics 359

talization of the analog signal.

An alternative for the pure absolute measurement with analog technology is given
by a discrete absolute measurement of defined states. In section 9.2.2.1, figure 9.15,
a commutation of EC-drives with a discrete, position coding of magnet-angles with
field plates was already shown. This approach is based on the assumption to achieve
with m measurement points n states in a discrete resolution Δ D.

nm = Δ D. (11.1)
In case of the commutated EC-drive m = 3 measurement points, which are able
to have n = 2 states, could encode 8 positions on the circumference, but only six
were actually used. But there are other more complex code discs with several lanes
for one sensor each. These sensors are usually able to code two states. However e.g.
by the use of different colors on the disc many more states would be imaginable. A
resolution of e.g. 1◦ (360 discrete steps) would need the number of

log(Δ D)
m= = 8.49 (11.2)
log(n)
at least nine lanes for encoding.

11.2 Requirements in the Context of Haptics

Position measurement systems are primarily characterized by their achievable res-

olution and dynamics. For haptic devices, in dependence on the measurement basis
for computer mice and scanners, position resolutions are frequently defined as dots-
per-inch Δ Rinch . Consequently the resolution Δ Rmm in metric millimeters is given
as:
25, 4 mm dpi
Δ Rmm = . (11.3)
Δ Rinch
A system with 300 dpi resolution achieves an actual resolution of 84 μ m. In de-
pendency on the measurement principle used, different actions have to be taken to
achieve this measurement quality. With incremental measurement systems the sen-
sors for the acquisition of single steps (e.g. holes on a mask) are frequently less
resolutive, requiring a transformation of the user’s movement to larger displace-
ments at the sensor. This is typically achieved by larger diameters of code discs and
measurement at their edge. These discs are mounted on an axis, e.g. of an actuator.
With analog absolute systems an option for improving the signal is conditioning. It
is aimed at reducing the noise component in the signal relative to the wanted signal.
This is usually done by a suppression of the noise source (e.g. ambient light), the
modulation and filtering of the signal (e.g. lock-in amplifier, compare section 11.6)
or the improvement of secondary electronics of the sensors (high resolutive A/D-
360 11 Application of Positioning Sensors

transformer, constant reference sources).

Beside the position measurement itself, its dynamic has to be considered during
the design process. This requirement is relevant for incremental measurement sys-
tems only. Absolute measurement systems need a bandwidth equal to the bandwidth
provided by the interface and the transmission chain (chapter 12) for positioning
information. Incremental measurement systems however have to be capable of de-
tecting any movement event, independent from the actual movement velocity. The
protocol format, usually given by counters part of the microcontrollers, has to be
dimensioned to cover the maximum incremental frequency. This requires some as-
sumptions for the maximum movement velocity vmax . If e.g. a system with 300 dpi
position resolution move at a maximum velocity of 100 mm/s, the dynamic fink for
detecting the increments is given as
1 Δ Rmm
= (11.4)
fink vmax
For the example the necessary measurement frequency is given with fink =
1190 Hz. The effective counting frequency is usually chosen with factor two to four
higher than that, to have a security margin for counting errors and direction detec-
tion.

11.3 Optical Sensors

Optical Sensors for position measuring are gladly and frequently used. They excel
by their mechanical robustness and good signal to noise ratios. They are cheap and
in case of direct position measurement quite simple to read out1 .

Code Discs

Code discs represent the most frequently used type of position measurement sys-
tems with haptic devices, especially within the class of low-cost devices. They are
based on transmission (fig. 11.2a) or reflection of an optical radiation, which is in-
terrupted in discrete events. The necessary baffle is located near to the receiver. It
is manufactured by stamping, or printed on a transparent substrate (glass, plastic
material) via thigh-film technology or laser printers. For high requirements on res-
olution they are made of metal, either self-supportive or on a substrate again. In
theses cases the openings are generated by a photolithographic etching process. The
receivers can be realized in different designs. Figure 11.2 shows a discrete design

1 The examples presented here are discussed either for translatory and rotatory applications. But
all principles may be applied to both, as a translation is just a rotation on a circle with infinite
diameter.
11.3 Optical Sensors 361

with two senders in form of diodes and two receivers (photodiode, phototransistor).
The placement of sender/receiver-units have to allow the phase shift for directional
detection (section 11.1.1). An alternative is given by a fork light barriers already in-
cluding a compact sender/receiver unit. Additionally opto-encoder (e.g. HLC2705)
exists including the signal conditioning for direction-detection from the two incre-
mental signals. The output pins of these elements provide a frequency and one signal
for the direction information.

Gray Scale Values

With similar components, but for absolute measurement a gray scale disc or gray
scale sensor can be built. Once again there are transmission and reflection (fig. 11.2b)
variants of this sensor. In any case the reflection/transmission of the radiation varies
dependent on the angle or position of a code disc. The amplitude of the reflec-
tion gives absolute position information of the disc. For measurement, once again,
either a discrete design or the usage of integrated circuits in form of so called reflec-
tion sensors is possible. Although such sensors are frequently used as pure distance
switches only, they show very interesting proportional characteristics between the
received numbers of photons and their output signal. They are composed of a light
emitting diode as sender and a phototransistor as receiver. In some limits the output
is typically given by a linear proportional photoelectric current.

transmission - discrete design

reflexion - discrete design

reflection sensor
optical
encoder0

a) b)
forked light barrier

Fig. 11.2 Incremental optical position measurement (a), and absolute position measurement via
gray-scale values (b).
362 11 Application of Positioning Sensors

Reflection Light Switches

Reflection light switches show useful characteristics for a direct position measure-
ment too. In the range of several millimeters they have a piecewise linear depen-
dency between photocurrent and the distance from the sensor to the reflecting sur-
face. Consequently they are useful as sensor for absolute position measurement of
translatory movements (fig. 11.3a). By this method e.g. with the SFH900 or its SMD
successor SFH3201 within a near field up to ≈1 mm measurement inaccuracies of
some micrometers can be achieved. In a more distant field up to 5 mm the sensor is
1
suitable for measurement inaccuracies of 10 mm still.

Mice-Sensor

The invention of optical mice without ball resulted in a new sensor-type interesting
for other applications too. The optical mice-sensors are based on an IC measuring
an illuminated surface through an optic element (fig. 11.3b). The resolution of the
CMOS sensors typically used range from 16x16 to 32x32 pixels. By the image ac-
quired the chip identifies the movement of contrast difference in their direction and
velocity. The interface of the calculated values varies from sensor to sensor. The
very early types provided an incremental signal for movements in X and Y-direction
identical to approaches with code discs described above. They additionally had a se-
rial protocol included to read the complete pixel information. Modern sensors (e.g.
ADNB-3532) provide serial protocols for a direct communication with a microcon-
troller only. This allowed a further miniaturization of the IC and a minimization of
the number of contact pins necessary. The resolution of state-of-the-art sensor is in
between 500 to 1000 dpi, and is usually sufficient for haptic applications. Only the
velocity of position output varies a lot with the sensor types available at the market,
and has to be considered carefully for the individual device design. The frequency
is usually below 50 Hz. Additionally early sensor designs had some problems with
drift and made counting errors, which could be compensated only by frequent ref-
erencing.

The sensors are usually sold for computer-mouse-similar applications and corre-
sponding optics. But beside that it is also possible to make measurements of moving
surfaces with an adapted optic design at a distance of several centimeters.

Triangulation

Optical triangulation is an additional principle for contactless distance measure-

ment; however it is seldom used for haptic devices. A radiation source, usually a
laser, illuminates the surface to be measured, and the reflected radiation is directed
on different positions along a sensor array (fig. 11.4). The sensor array may be
11.3 Optical Sensors 363

Reflection sensor

LED camera-chip

reflecting
surface lense
a
a) b)
reflecting
surface

Fig. 11.3 Distance-measurement with reflection light switches (a), and via the movement of an
reflective surface in two DOFs “mice-sensor” (b).

made of discrete photodiodes. Frequently it is a CCD or CMOS row with the corre-
sponding high resolution. By focal point identification weighting several detectors
a further reduction of measurement inaccuracy can be achieved. Compared to other
optical sensors, triangulation sensors are expensive as the detection row with a suf-
ficient resolution is a high cost factor. Their border frequency (»1 kHz) and their
measurement inaccuracy (<10 μ m) leave nothing to be desired. It is one of the very
few principles, which can hardly be used for measuring rotating systems.

Detector
array

Laser

Fig. 11.4 Triangulation of a distance with laser-diode and detector array.

364 11 Application of Positioning Sensors

11.4 Magnetic Sensors

Beside the optical measurement principles, especially the group of magnetic mea-
surement principles is relevant for haptic devices. This is a consequence from the
fact that electrodynamic and electromagnetic actuators already require magnetic
fields to generate forces. For systematization, sensor for static fields, field plates
and hall-sensors, and sensor for induced currents and time dependent fields can be
distinguished.

Field Plates or Magnetic Dependent Resistors

Field plates or magnetic dependent resistors (MDR) are two pole elements with a re-
sistance being controlled by the presence of a magnetic field. They make use of the
G AUSS-effect, which is based on charge carriers being displaced by the L ORENTZ-
force when crossing a magnetic field. The resulting increase of the path length [187]
requires an increase of the ohmic resistance of the material. The parameter charac-
terizing this dependency is dependent on the electron mobility and the path length
in the magnetic field. A frequently used material is InSb with very high electron
mobility. For an additional increase of the effect, the conductor is formed like in the
shape of a meander similar to strain gauges. MDRs are not sensitive to the polarity
of the magnetic field. They are detecting the absolute value only. The increase of
resistance is nonlinear and similar to a characteristic curve of a diode or transistor.
A magnetic pre-load is recommended to use the plates in a linear working point.

Hall-Sensors

Hall-sensors are based on the G AUSS-effect too. In contrast to field plates they are
not measuring the resistance increase of the current within the semiconductor, but
the voltage orthogonal to the current. This voltage is a direct result of the displace-
ment of the electrodes along the path within the material. The resulting signal is
linear and bipolar in dependency on the field-direction. ICs with an integrated am-
plifier electronics and digital or analog output signals can be bought of the shelf. A
frequent use can be found with sensors being located at a phase angle α with di-
ametral magnetized rotational magnets (fig. 11.5). In this application a rotation and
rotation-direction is measured.

Inductance Systems

An often forgotten alternative for position measurement is the measurement of

changing inductances. The inductance of a system is dependent on many param-
eters, for example the magnetic permeability of a material in a coil. Using a dif-
11.5 Other Displacement Sensors 365

a S

Fig. 11.5 Measurement of the rotation angle of a magnet via field plates or hall-sensors.

ferential measurement in between two coils (fig. 11.6b) a displacement measure-

ment can be made, if a ferromagnetic material moves in between both coils as a
position-depending core. As alternatives the geometry of the magnetic circuit may
be changed, or its saturation may influence the inductance of the coils. Latter ap-
proach is used in systems, where grooves on a ferromagnetic material trigger events
in a nearby coil (fig. 11.6a).
A simple electronic for measuring inductance is the use of a LR-serial circuit, which
- for example with a microcontroller - is triggered with a voltage step. The mea-
surement value is given by the time the voltage at the resistor needs to trigger a
comparator voltage. The duration encodes the inductance, assuming a constant re-
sistance. For the actual design it has to be considered, that the winded coil has an
own resistance which cannot be neglected. As an alternative a frequency nearby
the resonance RL of the LR-circuit can be applied. The voltage amplitude measured
varies dependent on the inductance’s detuned by the movement of the ferromagnetic
core.

11.5 Other Displacement Sensors

Beside the displacement measurement principles discussed above there are some
rarely used principles still worth to be mentioned here.

Ultrasonic Sensors

Ultrasonic sensors (fig. 11.7) are based on the running time measurement in be-
tween the emission of acoustic oscillations and the moment of the acquisition of
366 11 Application of Positioning Sensors

magnetically soft
i1 i2
ferromagnetic
u1 u2 i1

u1
i1

u1
a) a b)

Fig. 11.6 Incremental measurement of a movement via induced currents (a) and differential mea-
surement of the position of a ferromagnetic core (b).

their reflection. The frequency chosen is dependent on the requirements on mea-

surement inaccuracy and the medium for propagation of the wave. As a rough rule
of thumb, the denser a material is, the less the damping becomes for acoustic waves.
For measurement in tissue frequencies between 1 and 40 MHz are applied. In wa-
ter frequencies between 100 and 500 kHz and in the atmosphere frequencies well
below 30 kHz are used. Whereas with medical applications in tissues the medium
shows a damping quite linear in the range of 1dB/Mhz/cm, the measurement within
the atmosphere is strongly dependent on the frequency chosen and usually nonlin-
ear. Additionally the acoustic velocity is dependent on the acoustic density of the
medium. For the transversal direction - typically used for measurement - velocities
between 340 m/s for air and 1500 m/s for water can be achieved. According to the
wave theory the minimum measurement inaccuracy possible in transversal direction
is λ2 , which and coupled to both factors mentioned above. It is a natural border of
minimum resolution to be achieved.
The most frequently used source and receiver for the mechanical oscillation are
piezoelectric materials (fig. 11.7b), whose step response oscillations are sharpened
by a coupled mass.

Capacitive Sensors

In section 9.5 the equations for the calculation of capacities between plates of elec-
trostatic actuators (equ. 9.79) were introduced. Of course the measurement of a
variable capacity, especially with the linear effect of a transversal plate displace-
ment, can be used for position measurement. This is especially interesting if there
are conductive components in the mechanical design, which already move relative
to each other. As the capacity is very much dependent on the permittivity of the
medium between the plates, which can be strongly influenced by oil or humidity,
such a measurement can be done on insusceptible or other well housed actuators
11.6 Electronics for Absolute Positions Sensors 367

Ultrasonic sensor

Rückwärtiger Dämpfer
Damper
a) b) Piezoceramic disk
Piezokeramische Scheibe
Contact surface
Ankoppelschicht

Fig. 11.7 Distance measurement via ultrasonic sensors (a) and cross-section through a medical
ultrasonic head with fixed focus (b).

only. Additionally leakage fields of conductors or geometries nearby are usually

of the same size as the capacity to be measured. But capacitive sensors for haptic
devices can be found in the context of another interesting application. The measure-
ment of the capacity of the handle, even when isolated by a non conductive layer,
allows identifying a human touch very securely.

11.6 Electronics for Absolute Positions Sensors

The absolute measurement of a position requires, as mentioned earlier, some addi-

tional effort in the electronic design compared to discrete sensors. Two aspects shall
be discussed in the context of this chapter.

Constant-Current Supply and Voltage References

For the generation of a constant radiation or the measurement of a bridge circuit the
use of constant currents is necessary. There is always the possibility to wire an oper-
ational amplifier as a constant current source, or use transistor circuits. Nevertheless
for designs with low quantities there are ICs which can be used as current sources
directly. The LM234 for example is a voltage-controlled 3-pin IC, providing a cur-
rent with a maximum error of 0.1%/V change in the supply voltage. The maximum
current provided is 10 mA, which is usually sufficient for the supply of optical or
resistive sensors.
The change of the signal is usually measured with relation to a voltage in the sys-
tem. In these cases, it is necessary to provide a voltage which is very well known
and independent from temperature effects or changes of the supply voltage. Com-
mon voltage regulators as used for electronic supply are not precise enough to fulfil
368 11 Application of Positioning Sensors

these requirements. An alternative is given by Z ENER-diodes operated in reverse di-

rection. Such diodes however are not applicable to high loads and are of course only
available in the steps of Z ENER voltages. Alternatively reference voltage sources are
available in many voltage steps. The REF02 for example is a six-pin IC, providing
a temperature-stable voltage of 5V with an error of 0.3%. The drivable load of such
voltage sources is limited, in case of the REF02 it is 10 mA, but this is usually not a
relevant limit as they are not thought as a supply to a complex circuit but only as a
reference.

Compensation of Noise

The obvious solution for the compensation of noise in a measurement signal is given
by the usage of a carrier frequency for modulating the signal. A prerequisite of
course is given by the sensor showing no damping at the modulating frequency.
This is usually no problem for optical sensors in the range of several kilohertz. At
the receiver, the signal is bandpass-filtered and equalized or otherwise averaged.
This suppresses disturbing frequencies or otherwise superimposed offsets.
A simple but very effective circuit for noise compensation is the use of so called
“lock-in” amplifiers. On the side of the sender a signal is switched between the
states on and off at a frequency f . In the receiver the wanted signal such as e.g. the
offset and other disturbing frequencies are receipt. A following amplifier is switched
with the same frequency between +1 and -1 in such a way, that with the receipt of
the wanted signal including the disturbance the positive amplification happens. Dur-
ing the period without the wanted signal, when the receiver measures the disturbing
values only, the signal is inverted with -1. The resulting signal is low-pass filtered
afterward, resulting in a subtraction of the noise signal and providing a voltage pro-
portional to the wanted signal only.

11.7 Acceleration and Velocity Measurement

Beside the direct position measurement, haptic systems sometimes demand some
knowledge about the first or second derivative of position in form of velocity or
acceleration. Such a necessity may be given with stability issues for closed-loop
systems or impedance behavior of users or manipulated objects. The acquisition can
either be done by direct measurement or by differentiation of the position-signal
with digital or analog circuits. Additionally it can be imagined to e.g. measure a
velocity and calculate the position by integration. The capabilities of integration
and differentiation and their limits, such as typical direct measurement principles,
are sketched in this section.
11.7 Acceleration and Velocity Measurement 369

11.7.1 Integration and Differentiation of Signals

The integration and differentiation of signals can either be done analog or digital.
Both variants have different advantages and disadvantages.

Analog Differentiation

The basic circuit for an active analog integrator is shown in figure 11.8a. It is a
high-pass filter, which already gives hints on the challenges connected with dif-
ferentiation. The high-pass behavior is limited in its bandwidth. The upper border
frequency is given by the resonance frequency fR = 2π1RC and by the bandwidth of
the operational amplifier. As these components are sufficiently dynamic for haptic
applications this should be no problem in practical realization. Due to the negative
feedback however the natural bandwidth limit of the operational amplifier at high
frequencies has a phase of 90◦ , adding to the phase of 90◦ from the differentiation.
This makes the circuit sensitive to become electrically instable and oscillate [255].
This effect can be compensated by a serial resistance with a capacity C, which is
identical to a linear amplification with the operational amplifier. This diminishes the
phase for high frequencies by 45◦ , resulting in a phase margin to the instable border
condition. Analog differentiation is an adequate method for the derivation of veloci-
ties from positioning signals. A double analog differentiation needs a careful design
of the corresponding circuit, as a number of capacitive inputs are placed in series.
Additionally it should be considered that the amplitude of the operational amplifier
is limited by the supply voltage. Accordingly the amplitude’s dynamic has to be
adjusted to the maximum signal change expected.

Analog Integration

The basic circuit of an active analog integrator is given in figure 11.8b. Analog
integration is a reliable method from analog calculation technique, but has limited
use for haptic applications. The circuit has an upper border frequency given by the
resonance fR = 2π1RC , and for a non-ideal operational amplifier it has a lower border
frequency too. This is a result of the current Ib at the input of the OP-amplifier
charging the capacitor with Uin = 0V continuously. If C=10 μ F and Ib = 1 μ A, the
voltage increases by 0.1 V per second. Whereas in signal processing applications
this can be compensated by high-passes in series, for haptic applications covering a
bandwidth from several seconds to 10 kHz this behavior is usually not acceptable.
370 11 Application of Positioning Sensors

C
R

U- R Ib U-
C
- -

Uin + Uout Uin +

U+ U+

a) b)

Fig. 11.8 Analog differentiation- (a) resp. integration-circuits (b) [255].

Digital Differentiation

Digital differentiation is realized by a subtraction of two consecutive measurement

values. It is very applicable, especially when the signal is measured at high frequen-
cies. The quality of the signal is dependent on the noise on the input. Frequently the
least-significant bit of e.g. an AD-conversion is rejected before the differentiation
is performed, as it is oscillating with the noise of the AD-conversion (quantization-
noise).

Digital Integration

Digital integration is the summation of continuous measurement values and the di-
vision of the sum by the number of values. Alternatively it can be the sum of discrete
changes of a measurement value. The incremental measurement of a digital encoder
is also a form of integration on the basis of change information. The procedure is
robust at high frequencies beyond the actual upper border frequency of the signal.
Beside a sufficient dimension of the register size for the measurement values to
prevent an overflow, there is nothing else to worry about.

11.7.2 Induction as a Velocity Measure

The most frequent variant to gain information about velocity is given by the digi-
tal signal processing of a position-measurement. Nevertheless to be able to measure
velocity directly, the use of a velocity-proportional physical effect is mandatory. Be-
side D OPPLER-ultrasonic measurement, which is seldom applicable to haptic sys-
tems due to the wavelengths (compare section 11.5), the use of electrical induction
is the most frequently used direct effect. Accordingly an electrical induced voltage
U is generated in a conductor of the length l, moving orthogonal in a magnetic field
B with the velocity v:
11.8 Conclusion on Position Measurement 371

U = v B l. (11.5)
Especially geometrical designs as given with electrodynamic actuators (sec-
tion 9.2) can be used for velocity measurements with inducing voltages in their
coils. In contrast to electrodynamic actuators the design requires a maximization of
conductor length, to generate a pronounced voltage signal. The inductivity of the
winding generates a low-pass characteristic in combination with its own resistance.
This limits the dynamic of the signal. The biggest error made with these kinds of
sensors is given by a bad homogeneity of the winding due to dislocation of single
turns. This manufacturing error results in different winding lengths moving in the
B-field at different positions of the sensor, which is directly affecting the quality of
the measured signal.

11.7.3 Force Sensors as Acceleration Sensors

In contrast to velocity measurement, to measure accelerations a wide variety of sen-

sors exists. Ignoring some exceptions, most of them are based on the relation
F
a= . (11.6)
m
In fact the force-measurement principles given in chapter 10 are added by a
known mass m only, resulting in a mechanical strain of a bending element or gen-
erating another acceleration-proportional signal. In professional measurement tech-
nology especially piezoelectric sensors for high dynamic measurements, but also
piezoresistive sensors for low-frequency accelerations are established. In mecha-
tronic systems with high quantities micromechanic acceleration sensors with comb-
like structures in silicon according to the capacitive measurement principle are used.
The requirements of automotive industry for airbags and drive stability programs to
measure acceleration in many directions made low-price and robust ICs available at
the free market, e.g. the ADXL series of Analog Devices. The bandwidth of these
sensors ranges from 400 Hz to 2.5 kHz with maximum accelerations >100 g in up to
three spatial directions. Only a wide variance of their characteristic values, e.g. the
output voltage at 0 g, requires a calibration of the individual sensor.

11.8 Conclusion on Position Measurement

With haptic devices position measurement is a subordinated problem. In the range of

physiological perceived displacements resolutions, there are enough sensor princi-
ples which are sufficiently precise and dynamic for position measurement. The cal-
culation or measurement of accelerations or velocities is easily possible too. Without
372 11 Application of Positioning Sensors

doubt, the optical measurement technology is the most frequently used technical so-
lution. Nevertheless especially for the design of specific actuators it is indicated to
ask the questions, whether there are other sensor principles applicable for a direct
integration into the actuator. If there are specific requirements for measurement in
the range of a few μ m positioning resolution, the proposed principles should be
treated with reserve. Measurements in the range of μ m require specific optical or
capacitive measurement technology. With the exception of special psychophysical
questions it is unlikely that such requirements are formulated for haptic devices.
Chapter 12
Interface Selection

T HORSTEN A. K ERN

After the decision for the actuator (chapter 9) used to generate the haptic feed-
back, and after the measurement of forces (chapter 10) or positions (chapter 11), it
becomes necessary to focus on the IT-interface. This interface has to be capable of
providing data to the actuation unit and catch and transmit all data from the sensors.
Its requirements result - such as with any interface - from the amplitude resolution
of the information and the speed at which they have to be transmitted. The focus of
this chapter lies on the speed of transmission, as this aspect is the most relevant bot-
tleneck when designing haptic devices. Haptic applications are frequently located
on the borderline, may it be with regards on the delay acceptable in the transmis-
sion, or the maximum data rate in the sense of a border frequency.

With regards to the interface two typical situations may be distinguished: Spa-
tially distributed tactile displays with a reasonable number of actuators; and pri-
marily kinaesthetic systems with a smaller number of actuators. In case of tactile
systems, pin-arrays, vibrators, or tactors the challenge is given by the application of
bus-systems for the reduction of cable lengths, and the decentralization of control.
Although there are still some questions about timing left, for example to provide
tactile signals in the right order despite of a decentralized control, the data rates
transmitted are usually not a challenge for common bus systems. VAN E RP points
out [267], that a 30 ms time delay between impulses generated by two vibrators at
the limbs may not be distinguished any more. For the data interface this observa-
tion implies for this application, that any time delay below 30 ms may be uncritical
for transmitting information haptically. This is a requirement, which can be fulfilled
by serial automation technology network protocols like CAN, or the time triggered
version TTCAN, without any problems. Accordingly this section concentrates on
requirements of haptic kinaesthetic devices with a small number of actuators only,

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_12, 373

c Springer-Verlag Berlin Heidelberg 2009
374 12 Interface Selection

whereas these devices usually have to satisfy tactile requirements according to their
dynamic responses too.

12.1 Border Frequency of the Transmission Chain

Chapter 4 stated that it is necessary to distinguish two frequency areas when talking
about haptic systems. The lower frequency range up to ≈ 10 Hz includes a bidi-
rectional information flow, whereas the high frequency area >10 Hz transmits infor-
mation only unidirectional from the technical system to the user. Although the user
himself influences the quality of this transmission by altering the mechanical cou-
pling, this change itself happens at lower frequencies only, and is - from the perspec-
tive of bandwidth - not relevant for the transmission. If this knowledge is applied
to the typical structures of haptic devices from subsection 2.2.2, some fascinating
results can be found. For the following analysis it is assumed, that the transmission
and signal conditioning of information happens digital. According to N YQUIST, the
maximum signal frequency has to be sampled at least two times faster. In practical
application this factor two is a purely theoretical concept, and it is strongly rec-
ommend to sample and analog system around 10 times faster than its maximum
frequency. The values within figures and texts are based on this assumption.

12.1.1 Bandwidth in a Telemanipulation System

For a telemanipulation system (fig. 12.1) the knowledge about the differing asym-
metric dynamics during interaction gives the opportunity to benefit directly for the
technical design. In theory it is possible to transmit the haptic information measured
at the object within the bandwidth of 1 Hz to 10 kHz, and replay it as forces or posi-
tions to the user. The user’s reactions may in this case be measured at a bandwidth
from static to 5 or 15 Hz only, and be transmitted via controller and manipulator
to the object. Although this approach would be functional indeed, the simplicity of
position measurement and the necessity to process them for e.g. passivity control re-
sult in movements being sampled and transmitted similar dynamic as in the opposite
transmission direction for haptic feedback.

12.1.2 Bandwidth in a Simulator-System

For a simulation-system with haptic feedback the different dynamics results in

slightly different findings. Nevertheless it is still true, that the movement information
may be sampled at a lower rate. However the simulator (fig. 12.2) has to provide the
force output at a frequency of 1 to 10 kHz. Due to this simple reason, the simulator
12.1 Border Frequency of the Transmission Chain 375

Physical Signal Data Signal Physical

1-10 kHz 10-100 kHz 10-100 kHz 1-10 kHz

Haptic Haptic
Manipulator controller device User

Object
5-15 Hz 50-150 Hz 50-150 Hz 5..15 Hz

Fig. 12.1 Block diagram of a telemanipulator with haptic feedback.

has to be aware of the actual position data for every simulation step. Consequently
with simulators the haptic ouput and the measurement of user reaction has to happen
at high frequency (exceptions, see section 12.2).

There are two approaches to integrate the haptic controller in the simulator. In
many devices it is designed as an external hardware component (fig. 12.2), which
reduces the computing load for the main simulator, and helps reducing the data rate
significantly in special data processing concepts with parametrizable models (sec-
tion 12.2). As an alternative the controller may be realized in software as a driver
computed by the simulation main computing unit (fig. 12.3). This is a concept used
especially for high power permanently installed simulation machines, or which is
used in cost-effective haptic devices for gaming industry with little requirements in
dynamics and haptic output.

Data Signal Data Signal Physical

10-100 kHz 10-100 kHz 1-10 kHz

Simulation
Haptic Haptic
controller device User

Virtual object

10-100 kHz 10..100 kHz 5..15 Hz

Fig. 12.2 Block diagram of a simulator with haptic feedback and an external controller.
376 12 Interface Selection

Data Signal Physical

10-100 kHz 1-10 kHz

Simulation
Haptic Haptic
controller device User

Virtual object

10..100 kHz 5..15 Hz

Fig. 12.3 Block-diagram of a simulator with haptic feedback and a controller as part of the driver
software .

12.1.3 Data Rates and Latencies

Table 12.1 summarizes the data rates necessary for kinaesthetic applications in some
typical examples. The data rates range from 200 kbit/s for simple applications up
to 50 Mbit/s for more complex systems. Such rates for the information payload -
still excluding the overhead necessary for the protocol and the device control - are
achieved by several standard interface types today (section 12.3).

Table 12.1 Example calculating the required unidirectional data rates for typical haptic devices.
DOFs Resolution 0.1 kHz 1 kHz 10 kHz
1 DOF 8 bit 800 bit/s 8 kbit/s 80 kbit/s
16 bit 1600 bit/s 16 kbit/s 160 kbit/s
3 DOF 8 bit 24 kbit/s 240 kbit/s 2.4 Mbit/s
16 bit 48 kbit/s 480 kbit/s 4.8 Mbit/s
6 DOF 8 bit 48 kbit/s 480 kbit/s 4.8 Mbit/s
16 bit 96 kbit/s 960 kbit/s 9.6 Mbit/s

Beside the requirements for the data rate there is another requirement consider-
ing the smallest possible latency. Especially interfaces using packets for transmis-
sion, with an uncertainty about the exact time of the transmission (e.g. USB) have
to be analyzed critically concerning this effect. Variable latencies between several
packets are a problem in any case. If there are constant latencies the reference to
other senses with their transmission channel becomes important: A collision is not
allowed to happen significantly earlier or later haptically than e.g. visually or acous-
tically. The range possible for latency is largely dependent on the way to present the
other sensual impressions. This interdependencies are subject to current research
and are analyzed e.g. by the group around B USS at the TU-Munich.
12.2 Concepts for Bandwidth Reduction 377

12.2 Concepts for Bandwidth Reduction

Whoever ever tried to process a continuous data flow of several megabit with a
PC, and in parallel make this PC do some other tasks too, will have noticed that the
management of the data flow binds immense computing power. With this problem in
mind and as a result from the question about telemanipulation with remotely located
systems several solutions for bandwidth reduction of haptic data transmission have
been found.

12.2.1 Analysis of the Required Dynamics

The conscious analysis of the dynamics of the situation at hand should be ahead of
every method to reduce bandwidth, as already mentioned in chapter 6 and calculated
in section 3.3. The limiting cases to be analyzed are given by the initial contact or
collision with the objects. If the objects are soft, the border frequencies are in the
range <100 Hz. If there are stiff objects part of interaction and if there is the wish
to feed back these collisions too, the frequencies up to a border >1 kHz will have
to be transmitted. Additionally it has to be considered that the user is limited con-
cerning its own dynamics, or may even be further limited artificially. The DaVinci
System (fig. 1.8 on page 15) as an unidirectional telemanipulator filters e.g. the
high frequencies of the human movements to prevent a trembling of the surgical
instruments.

12.2.2 Local Haptic Model in the Controller

A frequently used strategy being part of many haptic libraries is the usage of lo-
cal haptic models. These models allow a much faster reaction on the user’s input
compared to the simulation of a complete object interaction (fig. 12.4). Such mod-
els are typically linearized functions dependent on one or more parameters. These
parameters are actualized by the simulation at a lower frequency. For example each
degree-of-freedom of the haptic system may be equipped with a model of spring,
mass and damper, whose stiffness-, mass- and friction-coefficient is updated to the
actual value at each simulation step, e.g. every ≈ 30
1
s. This approach does not permit
the simulation of nonlinear effects in this simple form. The most frequent nonlinear
effect when interacting with virtual worlds is the lift-off of a tool from a surface.
Dependent on the delay of the actualization of the local model, the lift-off will be
perceived as “sticking”, as the tools is held to the simulated surface by the local
model in one simulation step, whereas it is suddenly released within the next. Con-
cepts, which model nonlinear stiffnesses, compensate this effect satisfactory. By
making the additional calculations necessary for the local model, a significant data
378 12 Interface Selection

reduction between simulation and haptic controllers is achieved. Distantly related

concepts are used in automotive applications too, where CAN bus-systems are con-
figured in their haptic characteristics by a host, and report selection events in return
only.

Data Signal Data Signal Physical

50-100 Hz 10-100 kHz 1-10 kHz

Haptic controller
Simulation with local
model Haptic
F device User
m k d
v
Virtual object

50-100 Hz 10..100 kHz 5..15 Hz

Fig. 12.4 Block diagram of a simulator with haptic feedback and a local haptic model inside the
controller.

12.2.3 Event-Based Haptics

K UCHENBECKER presented in 2005 the concept of “Event-based haptics” [146]

and brought it into perfection since. It is based on the idea to split low frequency in-
teraction and high-frequency unidirectional presentation, especially of tactile infor-
mation (fig. 12.5). These tactile events are stored in the controller and are activated
by the simulation. They are combined with the low-frequency signal synthesized
from the simulation, and are presented to the user as a sum. In an improved ver-
sion, a monitoring of the coupling between haptic device and user is added, and the
events’ intensities are scaled accordingly. The design generates impressively real-
istic collisions with comparably soft haptic devices. As any other highly dynamic
system it nevertheless requires a specialized driver electronics and actuator selection
to achieve full performance.

12.2.4 Movement Extrapolation

Another very frequently used method for bandwidth reduction on the path to mea-
sure user reaction is given by extrapolation of the movement. Especially with simu-
lators using local models it is often necessary to have some information about steps
12.2 Concepts for Bandwidth Reduction 379

Data Signal Data Signal Physical

50 Hz 10-100 kHz 1-10 kHz

Haptic controller
Simulation
+ Haptic
device User

Virtual object

50-100 Hz 50..100 Hz 5..15 Hz

Fig. 12.5 Block diagram of a simulator with haptic feedback and with events of high dynamic
being held inside the controlling structure.

in between two complete measurement sets, as the duration of a single simulation

step varies strongly, and the available computing power has to be used most effi-
ciently. The extrapolation becomes a prediction with increased latency and a further
reduced transfer rate. Prediction is used for haptic interaction with extreme dead
times.

12.2.5 Compensation of Extreme Dead Times

The working group of N IEMEYER from Telerobotics Lab. at the Stanford University
works on the compensation of extreme dead-times of several seconds by prediction
[178]. The dead-time affects both paths: the user’s reaction and the information
to the user, such as the haptic feedback generated. The underlying principle is an
extension of the telemanipulation system, which is added with a controller of the
manipulator and a powerful controller for the haptic feedback (fig. 12.6). Latter can
be understood as an own simulator of the manipulated environment. During move-
ment, a model of the environment is generated in parallel. If a collision happens
in the real world, the collision is placed as a wall in the model, and its simulation
provides a haptic feedback. Due to the time lag the collision does not happen at the
position where it happened in reality. During the following simulation the collision
point is relocated slowly within the model back to its correct position. By succes-
sive exploration of the environment a more detailed haptic model is generated. The
method has the status of a research project.
380 12 Interface Selection

Physical Signal Data Signal Data Signal Physical

1-10 kHz 10-100 kHz variable variable 10-100 kHz 1-10 kHz
Haptic
controller
Manipulator with adjustable Haptic
Manipulator control device User
world model

Object
5-15 Hz 50-150 Hz variable variable 50-150 Hz 5..15 Hz

Fig. 12.6 Block diagram of a telemanipulator with compensation of long dead times by an adept-
able world-model.

12.2.6 Compression

As any data data stream, haptic data can be compressed for reducing their band-
width. This may happen based on numerical methods on each individual packet,
however it may also be possible to make use of the special properties of the haptic
human-machine interaction and haptic perception. The following list shall give a
short overview about common approaches:
• A first approach for compressing haptic data is given in the situational adaption
of digitalization on the path for measuring user reaction. S HAHABI [226] (Uni-
versity of Southern California, USA) compares different digitalization methods
adapting their time- and amplitude-discretization on the actual movement veloc-
ity.
• Since several years now the working group around B USS (TU-München, Ger-
many) does intensive research on the perceptional impact of loss of resolution
and bandwidth in haptic data streams [148]. They are coupling their research
with the analysis of user-reactions, and base their algorithm on the psychophysi-
cal perception and an benefit-effort-analysis.
• The working group around E L S ADDICK (University Ottawa, Canada) wants to
achieve data reduction by standardizing the haptic interaction in a descriptive
data format “HAML”. It models the environment in a comparably little set of
parameters, which gives advantages in tele-interaction applications with a larger
group of participants. Especially varying data transmission paths can be compen-
sated more easily on this abstract description, in comparison to classic telema-
nipulation approaches with a transmission of explicit forces and positions. As a
by-product of this work, concepts for the unidirectional replay of haptic data in
form of a “haptic player” are developed [47].
• Another obvious approach for compression is the usage of limitations given
by haptic perception. The working group of K UBICA (University of Waterloo,
Canada) demonstrates [296] an analysis of an interaction with a virtual environ-
ment at different velocities. The identified dependency of the force perception
threshold on the velocity was successfully used as basis for data reduction.
12.3 Technical Standard Interfaces 381

• The working group of W ERTHSCHÜTZKY (TU-Darmstadt, Germany) works on

the measurement of haptic perception in the range of milli-Newton and microme-
ter resolution up to border frequencies of 1 kHz. They aim on building a database
as a basis for compression with frequency dependent perception curves.

12.3 Technical Standard Interfaces

Most haptic devices are operated with personal computer or related systems. They
offer a high flexibility in configuration as well as for research projects as for gaming-
and design applications. Within this section different standard interfaces typical to
PC hardware architecture are highlighted and discussed with respect to haptic de-
vices applications.

Serial Port

The serial port is an interface, which is - dependent on the operating system used1
- quite simple to be addressed. The serial interface of a home computer is based on
the RS232 standard. This standard defines, beside several timing-aspects, the bits
being encoded in between ±3 V to ±15 V for low- and high-levels. Compellable a
connection to digital circuits with 5 V logic has to happen via a converter, such as
the MAX232. Usually only two lines (RxD and TxD) are necessary for data trans-
mission. The maximum specified data rate is 20.000 bit/s, whereas data rates far off
specifications with 56 kbit/s or 128 kbit/s are possible too. Both, the data rate and
the number of bytes carrying the actual data payload can be configured in a wide
range. Two systems communicating with each other have to be adjusted in these
parameters. A simple parity-control with one bit is also integrated. With respect to
its data rate the serial port is absolutely suitable for the control of simple haptic
devices. However due to its master-slave architecture a bidirectional data commu-
nication is connected with a large data-overhead as a results of the coordination of
both units. As modern PC as not necessarily equipped with serial ports only, sev-
eral USB-to-serial converters are offered at the market. They emulate a COM-port
in software and transmit the data through the USB-connector and the typical 9-pole
socket. The data rates achievable with these connectors are usually sufficient, but
they show some unpredictable delays and time-loss due to software-emulation of
the interface. This makes them hardly usable for time critical haptic applications.

1 As a rule of thumb: with Linux or DOS-Systems a direct communication using the serial interface

is a lot simpler than compared to Windows operating systems.

382 12 Interface Selection

Parallel Port

The parallel port is, if it is still part of modern PCs, a 25-pole double row SUB-
D-socket on the computer back plane. Similar to the serial interface its easiness of
use is largely dependent on the operating system. In an idea case (with several Linux
distributions and with DOS) the address of the port can be directly written with three
consecutive bytes. In this case the first byte represents the eight data lines, the two
following bytes are used for setting and reading of the control lines. The parallel
port is set to work with 5 V logic levels with a maximum source current of ≈ 5 mA
and sinking currents of maximum ≈ 20 mA. An overload current or leveling the
pins to wrong voltages should be strictly avoided, as typical PCs do not show any
protective circuitry. The data transmission of modern parallel ports is usually bidi-
rectional, allowing read and write operations on the same eight data lines. But as
the change between data transmission directions needs some time, the control lines
are frequently used as input, whereas the data lines are operates as output. Writing
to and reading from the port can be made with frequencies up to 100 kbit/s without
much effort. An extension of the parallel port is given by the “Enhanced Parallel
Port” (EPP) and the “Enhances Capabilities Port” (ECP). Whereas the ECP excels
by Plug&Play-functionalities mainly, the EPP had been designed for an increase in
bidirectional data transmission rates up to 2 Mbit/s. This increase was achieved by a
Hardware-implementation of the data protocol. For a slave using the EPP - such as a
haptic device - this additional hardware requires of course some more hardware on
the device’s end too, as the protocol has to be realized near to the interface. From the
perspective of data rates, the parallel-port is highly suitable for haptic devices, espe-
cially in EPP-mode. Especially low latencies of <100 μ s between writing command
and the availability of the data makes it very attractive. Only the dwindling avail-
ability of this port to standard PCs makes it necessary to use other type of interfaces.

By their flexibility in software drivers, the serial and the parallel port makes it
possible to be interfaced as debugging ports to microcontroller circuits or to bus-
systems such as I2C or CAN.

USB

The USB-port is a serial port with a predefined data transmission protocol. It con-
tains two data lines, an electrical ground and a 5 V supply, which can be drained with
up to 100 mA per device attached. According to the USB specifications this load can
be increased up to 1 A, if the host accepts it. An extension of the standard named
“Power USB” considers additional lines for higher currents and even larger volt-
ages. The USB-clients receive an identifier when being connected to the bus, which
marks the data packages, sent to or from them. Each USB-component has “Device
Descriptor” uniquely identifying the manufacturer and the product. Additionally
each device is classified in several standard classes. The “human-interface-device-
class” (HID) is reserved for input systems. Devices with active, haptic feedback are
12.3 Technical Standard Interfaces 383

grouped in an own class of “physical-interface-devices” (PID). Each manufacturer

of an USB driver circuit has to apply for a unique product id. Due to the require-
ments and the complexity connected to the implementation of an USB-conform
protocol, it is recommended to use USB-interface circuits of the shelf for product
designs with little quantities manufactured. Such interface circuits offer parallel or
serial data lines to be interfaced by an own microcontroller, which itself does not
have to bother about the USB interface any more. The USB interface can be oper-
ated in different modes. For the transmission of larger, time-crictical data volumes
the “isochrone” transfer is most suitable. Its theoretical limit is given by the data
packets, which are transmitted in micro-frames according to the USB 2.0 standard.
The duration of a micro-frame is given by 125 μ s. So called “full-speed-systems”
are able to transmit up to 1023 bytes with each micro-frame. High-speed devices
are able to transmit even three times more bytes per micro-frame. According to
the specifications transmission rates of up to 40 MByte/s are possible with devices
combining several isochronous endpoints in one unit. The data rate of isochronous
transfer is optimized for unidirectional transmission only. In case of bidirectional
communication the data rate is reduced accordingly. Nevertheless the speed of the
USB-port covers any requirements given by USB devices. However two special as-
pects have to be checked in the context of the individual application:
• A micro-frame of 125 μ s duration (8 kHz) is the upper limit of the available
bandwidth. Without compression and decoding this gives the natural bandwidth
limit according to N YQUIST of 4 kHz.
• The data rate has a tolerance of 0.1%.

FireWire - IEEE1394

FireWire, Apple’s brand name, according to the IEEE1394 standard is a serial trans-
mission format similar to USB. In fact it is a lot older than the USB specification.
The six-pole FireWire Connector includes a ground and a supply line too. The volt-
age is not controlled and may take any value between 8 to 33 V. FireWire 400 defines
up to 48 W power to be transmitted. The data rates are - dependent on the port design
- 100, 200, 400 or 1600 kbit/s. This is completely sufficient for any haptic applica-
tion. Even fiber optics-transmission over 100 m distance with up to 3200 kbit/s are
specified in the standard. The bus-hardware additionally includes a concept to share
memory areas between host and client, enabling very latency less transmissions.
Even networks without an explicit host can be established. The interface according
to IEEE1394 is the preferred design for applications with high data transmission
rates. Only the little propagation of this interface in personal computers hinders a
wide application.
384 12 Interface Selection

Ethernet

The capabilities of the Ethernet-interface available with any PC are enormous but
largely dependent on the protocol used. Whereas the naked interface enables trans-
mission rates of 10 Mbit or 100 Mbit or even Gigabit, the available data payload
within the transmission is largely dependent on the interlacing of the underlying
protocols. The very well known TCP/IP protocol has a header portion of 40 byte.
The Ethernet protocol adds another 18 bytes for the Ethernet frame, and some more
8 byte for the whole packet, resulting in an overhead per packet of 66 byte. This
packet may contain up to 1460 bytes of data. This is sufficient for typical haptic ap-
plications with respect to the available space per packet. Assuming a six-DOF kine-
matics with 16 bit (2 byte) resolution in their sensors and actuators, each packet has
to carry only 12 bytes of data, with one packet for force-output and one for position-
input. The number of bytes carried in each packet has a lower limit depending on the
physical design of the network. In the area of home networks it is 50 byte, making it
necessary to add arbitrary data on the example from above. A cycle of the haptic de-
vice example would transmit 232 byte, which is 1.856 kbit. With a 10 Mbit network
and theoretical bandwidth of 8 kHz would be available. Even when considering that
the data have to be extended with some additional overhead (address negotiations,
status-information), this is still sufficient for many haptic applications. A disadvan-
tage in using the Ethernet is given by the high efforts necessary for packet confection
and protocol formulation, which would usually overload the computation power of
standard microcontrollers. Additionally a high number of clients reduce the data rate
within a network significantly. Using switches compensates this reduction to some
extend.

Measurement Equipment and Multi-Functional Interface Cards

Measurement- and multi-function interface cards are a simple approach to interface

to hardware designs. They are available for internal and external standard interfaces,
such as PCI, AGB or PCMCIA. They are usually equipped with several standard
sw-drivers optimized for their hardware capabilities. When considering a prototype
design they should be considered in any case. Their biggest disadvantage is given
by the data processing happening inside the hosting PC and within the restrictions
of the operating system. Especially in combination with non-realtime operating sys-
tems like Windows the dynamics of controllers necessary for haptic applications
may become not fast enough.

HIL-Systems

“Hardware In the Loop” (HIL) systems were first used in control engineering and
compensate the disadvantages from multifunctional cards for rapid prototyping and
interfaces to haptic systems. HILs include a powerful controller with proprietary or
12.4 Final Remarks about Interface Technology 385

open real-time operating system. The programs operated on these controllers have
to be built on standard PCs and are transmitted as with any other microcontroller
system too. Frequently the compilers allow programming with graphical program-
ming language such as MatLab/Simlink or LabView. The processors of the HILs are
connected via specialized bus-systems with variable peripheral components. Rang-
ing from analogue and digital output over special bus- and actuator-interfaces a wide
range of components is covered. HIL-systems are predestined for the always time-
critical applications of haptics in design phase. But compared to other solutions they
have a high price too.

12.4 Final Remarks about Interface Technology

The interface subordinates to the requirements of the system. Any realistic appli-
cation and its required data rate can be covered with today’s standard components.
Only commercial- or company-interests may prevent the choice of a suitable inter-
face for a haptic device. This is a complete difference to the situation at the begin-
ning of the 21st century. At this time highly specialized interfaces were designed
for haptic devices, to cover the high requirements on data transmission rates. Ac-
cordingly even today commercial products with own ISA or PCI interface cards can
be found on the market. Other solutions require an EPP parallel port still. Never-
theless the design of controller circuit suitable for the USB protocol should not be
underestimated. Especially its layout and programming for the still high data rates
of haptic devices offers enough room for errors. Although the technical specifica-
tions are sufficient to fulfill the requirements, the first design and operation is far
from being trivial.
“This page left intentionally blank.”
Chapter 13
Software Design

A LEXANDER R ETTIG
The ultimate display would, of course, be a room within which the computer can
control the existence of matter. A chair displayed in such a room would be good
enough to sit in. Handcuffs displayed in such a room would be confining, and a
bullet displayed in such a room would be fatal. With appropriate programming
such a display could literally be the Wonderland into which Alice walked.
I VAN E. S UTHERLAND, 1965 [246]
A central application area for haptic systems is the so called “Virtual Reality”.
This names a concept of human-computer-interaction, which has developed rapidly
during the last 20 to 25 years. The vision itself has been formulated already at the
beginning of the development of computer graphics in the sixties of the last century
by S UTHERLAND.
The focus of this chapter about software design is on this very application area
of computer haptics in virtual worlds. Other application areas of haptic systems as
the modification of haptic properties of manual control elements or telemanipulation
respectively need a system structure, which is mainly defined by control engineering
aspects as discussed in other chapters of this book (chapter 7 and 12).
Following a short overview to motivate the subject and provide some terminology
and concepts different topics will be discussed from a software developer’s point
of view. According to the goal of this book the chapter at hand does not provide
an exhaustive depiction but gives a basic understanding with creating interest for
further activity with the topic in mind.

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_13, 387

c Springer-Verlag Berlin Heidelberg 2009
388 13 Software Design

13.1 Overview About the Subject “Virtual Reality”

“Virtual reality”1 , or VR, terms a technology, which among others should fulfill
the demand to recreate the natural environment as close to real as possible inside a
computer simulation or to let unreal things become real and present them to the user
quasi as if they were true. The displayed virtual environment should behave in the
way the user would expect because of his experiences in the natural world. Ideally
VR allows a perfectly intuitive handling with the computer which needs not to be
learned.
Three main criterions have been identified during the research about virtual
worlds which should be fulfilled as possible to achieve this goal: a quality of the
display and form of presentation, which enables immersion, natural or intuitive in-
teraction and realistic or at least plausible behavior of the displayed environment.

Immersion

Immersion, “diving” into the virtual world, happens when the user quasi forgets that
he is interacting with a virtual world and the real world largely steps into the back-
ground. To achieve this it is crucial to activate as many senses as possible with the
best quality as possible: mainly the dominant visual sense as well as the auditory
sense, but also — as soon as manipulative interaction becomes relevant for the ap-
plication — the haptic perception and finally the olfactory or even the sense of taste
may be stimulated. The latter ones typically are only dealt with in a kind of exotic
scenarios though. At least a variety of sense modalities should be incorporated, this
is called multimodal presentation.

Natural Interaction

Experienced computer users are easily apt to consider operating the computer by
a mouse to be natural and intuitive. However, actually so-called more or less ab-
stract interaction metaphors find use which must be understood and trained first:
is it “natural” to “grab” a virtual object by pressing a mouse button after an arrow
symbol was moved onto the visual representation of the object on the screen by
means of a movement spatially separate of it of a small plastic box (the mouse)?
Really natural grabbing is certainly something else: the hand is moved to where the
object is seen, although in a fraction of seconds contact is taken up with the sense
of touch before finally the fingers are closed around the object. In addition, that the
meaning associated to an interaction metaphor often is only partially obvious — one
opens a document by dropping its icon onto the icon of a text processor program?

1 There are concepts “Augmented Reality” (AR), “Mixed Reality” (MR) or even “Augmented Vir-

tuality” as well, which name different kinds of mixture and embedding of real and virtual objects
into a real or virtual environment. This differentiation is not needed in our context though.
13.1 Overview About the Subject “Virtual Reality” 389

Those doubtless are useful and reasonable concepts, but their semantics have to be
learnt and understood explicitly. By using different input devices as data gloves up
to full body tracking systems, which track the movement of all extremities, the VR
research tries to provide actually natural interaction with complex data worlds.

Natural Object Behavior

The third criterion, plausible behavior, inter alia includes, that dropped things fall,
liquids swash and flow, that you can not go through walls and objects in the virtual
environment can not penetrate each other. These behavioral characteristics have to
be provided by different simulations.

Especially the latter aspect of natural object behavior, the impenetrability of solid
objects, is of importance: although it could be implemented using collision detection
and more or less physically correct simulation in a purely graphical presentation, but
then there is a possible discrepancy between the real movement of the user and the
visual echo of this movement. It occurs, for example, if a virtual hammer stops on
the surface of the virtual wall, while the real hand of the user has already deeply
banged into the virtual brickwork. Haptic rendering2 with force-reflecting robots
makes it possible to overcome this shortcoming.
This reflection is not only of academic significance but motivated by a practical
benefit for the user (or his employer). One benefit which by no means should be
underestimated in this respect is the fun factor. Apparently this is directly commer-
cially relevant in the computer games industry. In the wake of the rapid development
of computer technology, which is very much driven by computer gaming, the per-
sonal computers available on the mainstream market by now are capable to display
VR scenes in a complexity and quality, which was achievable solely with special-
ized graphics workstations from SGI, SUN and IBM still a few years ago. Thus VR
technologies are about to enter the mass market: a lot of computer games in particu-
lar from the genre of so-called first-person shooters or motorsport simulations with
their excellent graphics and realistic physics simulations can — albeit with some
conceptual concessions — be called “Virtual Reality”. Today the wealth of detail of
these virtual worlds is enormous, immersion happens very quickly, you forget that
you play, almost everything behaves plausibly. The interaction, however, is hardly
natural or intuitive as long as the game is controlled by keyboard and mouse. Suit-
able input devices fill this gap partly: the motorsport simulation makes much more
fun immediately when you have a vibrating steering wheel in your hands and throt-
tles below your feet — and the driving performance gets better.

2The term rendering generally denotes the presentation process done by a computer system. With-
out further definition often the image generation process for the graphical representation is meant,
but analogously the production of structured, information-carrying stimuli for other modalities is
called acoustic, olfactory, or just haptic rendering.
390 13 Software Design

This factor, better user performance, is also a driving force for the use of virtual
reality in industrial or medical applications. Another major aspect is the possibil-
ity to save costs while increasing the quality of products at the same time by using
virtual prototypes. The automotive industry as a pioneer in this area demands, for
example, that vehicles could be presented ahead of the construction of any physi-
cal prototype in life-size in photo quality and accurate, verifiable and interactively
adjustable lighting, enabling designers to decide on form details, the car paint and
exact color of the interior. In the best case the software should be that intuitive, that
nobody needs an explanation on how to use it.
In the same way virtual prototypes are used to examine, whether the vehicle can
be assembled as planned. Since the worker himself is subject to this analysis also it
suggests itself to let a real person do the virtual assembly. This approach is superior
in many aspects to a non-interactive assembly simulation relying only on a human
model.
To get applicable results from this kind of investigation, the virtual assembly
scenario has to be as realistic as possible. Especially in situations with narrow con-
struction spaces which are difficult to access, haptic feedback for collisions must
be provided, because the mechanic needs and uses it more or less consciously for
finding his way and accurately placing either the tool or the component.
Quite similarly haptic feedback vastly improves interactive simulations of medi-
cal interventions for the training of certain surgical procedures. Obviously the ma-
nipulative fine motor skills of the surgeon are of central importance during opera-
tions. Particularly in minimal invasive surgery much practice is needed for achieving
these skills, since the physician can not interact in the operation field with his fin-
gers, but has to do it in an entirely unfamiliar manner: the interaction is indirect
using long thin instruments, which are inserted into the situs through small inci-
sions. These punctures act as pivot points so that some of the movement directions
are mirrored relatively to the movement of the hands. Additionally also the view
is not direct, but via a screen showing an image taken by an endoscopic camera,
which up to now still usually is monoscopic. The spatial orientation and hand eye
coordination in this situation have to be learned and trained intensively. Tradition-
ally models or animal cadavers are used, but more and more VR training simulators
are establishing in this domain. Providing realistic visualization, simulation of soft
tissue deformation, haptic feedback and tools for training analysis and evaluation
they are much more flexible concerning different scenarios and much better suited
for tracing the learning progress.
13.2 Design and Architecture of VR-Systems 391

13.2 Design and Architecture of VR-Systems

13.2.1 Hardware Components

Generally the central hardware component of virtual reality systems consists of

high-performance computers or clusters of computers which communicate via net-
work. Connected to them are input devices and display systems using different in-
terfaces.
The input devices mouse and keyboard as known from desktop computers are
only of minor importance in VR applications. To fulfill the claim of natural in-
teraction many other input apparatuses have been developed. Central are so called
tracking systems, which offer absolute three-dimensional measuring of positions
and mostly also of orientations in space, i.e. these provide 6 degrees of freedom
(DoF). Using a tracking system it is possible to locate an interaction device exactly
as well as to measure the (head) position and viewing direction of the user. The latter
is amongst others necessary to achieve a perspectively correct visual rendering of
the virtual environment or an appropriate spatial acoustics simulation. Several tech-
nologies are used for 6 DoF-tracking: magnetic tracking systems consist of sensors
measuring the magnetic field, which is emitted by a field generator. Analogously
ultrasound tracking systems have ultrasound emitters and sensors. Another class are
camera based systems, where image processing methods are used to detect selected
features of the real world (so called “marker”) in video streams. Their relative loca-
tion is then calculated by taking the camera parameters into account. Usually such
tracking systems operate with an update rate of 30 to 50 Hz. Next to the update rate
also the latency between measuring and provision of the data in the VR system is
important: if it gets too large, noticeable delays between user action and the visual
result arise which destroy immersion: when for example the image follows with
some time offset when the head is moved, immediately the impression gets lost,
that the perceived visual stimuli come from objects with fixed location in space.
On the side of output devices several technical approaches are used for generating
specific stimuli for the different sense modalities. For the auditory perception it’s the
task of speaker systems, from headphones to surround sound systems and finally to
speaker arrays by which precisely located sound events could be generated.
However, by far the most attention is given to the most dominant distance sense
of man, the visual sense. Based on the physiological model of the trichromatic vision
almost all optical display systems generate image frames at frequencies between 50
to 120 Hz, which are built of square picture elements (pixels) mixing the additive
primary colors red, green and blue. The impression of continuity of motion is cre-
ated by means of that the change of images is fast enough to trick the human vision
system, which commonly happens from about 15 Hz [74].
To generate the impression of spatial depth stereoscopic displays are of exceed-
ing importance in VR, because the presentation of a separate image with different
perspective for each eye is the most effective way to stimulate the depth impression
at least for people with normal seeing capabilities.
392 13 Software Design

The sizes of graphical displays used in VR applications range from small optical
units in data glasses or head mounted displays (HMD) to large tiled projections
beamed by a bunch of projectors or projections onto the sides of a big cube which
one enters into, a so called CAVE (recursive acronym for Cave Automatic Virtual
Environment). Standard monitors or autostereoscopic monitors are also used, but
they don’t meet the demand for immersion very well.
Force reflecting haptic devices take an exceptional position amongst the hard-
ware components of VR-systems: they are both input and output devices for they
provide the input functionality of (mechanical) tracking systems in their passive de-
grees of freedom and also the (stimulus) output functionality by force transmission
in their active degrees of freedom.

13.2.2 Device Integration and Device Abstraction

Many modern VR-systems implement the concept of device servers and logical
devices for the integration of various input and output devices. A server is a com-
puter or a running software process respectively, which offers a service to other
programs. The service a device server provides is to supply input data on request
or receive output data following a specified protocol3 . The device server abstracts
from the special characteristics of the physical device as far as reasonable. For exam-
ple all 6 DoF tracking devices provide position and orientation data, but they differ
from manufacturer to manufacturer with respect to the control sequences which are
used to initialize and calibrate them and to trigger the data retrieval. These techni-
cal details could be hidden by the device server, which is adapted to the particular
tracking system. It then as a service provides the logical device “position and ori-
entation” comprising a standardized protocol for the transfer of control and payload
data. This service furthermore can occur network transparently: the input or output
device may be connected to another computer in the network, if the communication
protocol between VR system and device server is designed accordingly and suitable
configuration methods enable the VR-systems to gain access to the remote server.
The concept of logical devices disburdens the application developer who designs
a virtual scenario from thinking about the exact type e.g. of the tracking system
to be used. Later on it is even possible to replace an absolute by a relative 6 DoF
input device like the so called SpaceMouse4 without the need to change anything
of the scenario, as long as the device driver of the SpaceMouse also implements the
logical device “position and orientation”. By the way this concept is quite common
in the graphical “desktop” user interfaces of modern PCs: computer mice which are

3 The term protocol refers to a scheme which describes the semantics of an otherwise abstract data
stream and determines the order of its elements.
4 The SpaceMouse consists of a knob roughly of the size of the palm of the hand. It includes

sensors, which measure forces and torques the user exerts on it. The internal software converts
these to relative movements.
13.2 Design and Architecture of VR-Systems 393

connected via the USB port register themselves as human interface device (HID).
Included device drivers take care, that the operating system can use the device as a
“pointing device”.
Quite similarly at least on a single computer the VR-system developer5 needs not
to worry about servers for the output of audio and image data as the operating system
and standard libraries provide software interfaces to the respective device drivers,
which control the hardware, i.e., the sound and graphics card. The latter on their
own are responsible for sending appropriate signals to loudspeakers or monitors via
their hardware interfaces. As soon as more complex output systems on distributed
systems get involved though, the standard driver interfaces are not sufficient any
more and VR-systems have to implement the needed functionality itself.
This holds e.g. for graphical display systems as tiled projection screens or
CAVEs, which often are driven by clusters of computers, each one of it control-
ling one or more projectors. Server processes running on all of these computers
collectively provide a service for the output of the composed image6 . Similar con-
cepts are used, when the acoustic output should be carried out by a separate audio
workstation.
The device technologies in the area of VR are utterly varied, the development of
concepts is still very dynamic and thus no common standards have been established
yet. Every VR-system defines its own protocols for device connections7 . It would be
desirable, if there were cross-system classes for haptic devices similar to the USB-
device class HID in the future. They should be based on popular interfaces which
are capable of real time data transmission as e.g. IEEE 13948 (section 12.3).
Whereas small latencies in the data transmission via networks are mostly unprob-
lematic for pure input devices, they become critical if they occur in the control loop
of haptic devices. Therefore a network transparent device abstraction is not suitable
for haptics without further ado. We go into that in more detail for the special case of
the software renderer in section 13.2.5 and relating to the latencies in chapter 12.
Another aspect of device abstraction is of major relevance for haptic devices: the
application developer is used to model the virtual environment in a Cartesian coor-
dinate system and in the same way simulations are typically performed in Cartesian
coordinates. The necessary conversion of the input data from device coordinates to
Cartesian coordinates should be “hidden” from the application by the device driver
or server as well as the transformation of the Cartesian output forces and moments
into motor forces and moments and the appropriate control currents. Both conver-

5 Very often one distinguishes between the system developer, who creates the software infrastruc-

ture (in our case the VR-system libraries and executables), and the application developer, who
uses this infrastructure in order to model a virtual environment with object data and behavioral
descriptions.
6 On UNIX-like systems with graphical user interface based on the X-Windows system this concept

is well established. Any computer with a running X-Server can display graphical output for all
computers in the network which use the X protocol.
7 There are some efforts to develop software libraries which define a standard server interface for

various input devices, e.g. the free open source VRPN (Virtual Reality Peripheral Network) system.
8 Apple’s FireWire, SONYs i.Link
394 13 Software Design

sions need internal knowledge about the kinematics of the device and the mechanical
and electronic design, which almost without exception are irrelevant for the design
of the virtual scenario.

13.2.3 Software Components

The central software instance of a VR-system is often called object manager or VR

kernel. It manages the scene, i.e. all objects of the virtual environment, and controls
the data flow between all hardware and software components.
In many cases the central data structure is a so called scene graph. It structures
objects into groups in a transformation hierarchy and defines their relations. The
whole virtual environment shares a global world coordinate system. Nodes which
hold transformations define the relative location of their subordinate objects in the
virtual world and parent child relations in the graph represent spatial and kinematics
dependencies.
For example the four geometry objects, which represent the wheels of a vehicle,
are connected via transform nodes to the node, which contains the car body, which
itself is linked to the world coordinate system via a transformation node (fig. 13.1).
By changing the parameters of the transform nodes over time it’s easily possible to
describe the dependent movements of the individual parts of the car, i.e. the rotation
of the wheels around their axes relative to the car body as well as the superimposition
of the forward movement of the whole vehicle.
The object manager furthermore manages the behavioral descriptions of the
scene. These define the effects of an “event” on the objects and which subsequent
events are initiated by it.
According to the implementation concepts of the particular VR-system these de-
pendencies are coded directly in scripts or parameterizable modules or they are ex-
pressed in so-called behavior graphs or a combination of both is used.
After loading the scene description and setup of all data structures the VR kernel
starts a cycle repeating upon system termination, which can be phased into 3 steps9 :
1. Collecting all events
2. Propagation of the events through the behavior graph or triggering the event pro-
cessing in the appropriate modules respectively
3. Display of the final state of the cycle on all output devices

The first phase amongst others comprises of requesting the data of the input de-
vice’s via the device’s servers depicted above. In the second phase the event data
is passed to the various simulations and these calculate the next step. Since simula-
tions may run concurrently as described below this phase ends with a synchroniza-

9Basically this process is the same for every interactive software system including all widely used
programs that have graphical user interfaces — from word processors up to computer games.
13.2 Design and Architecture of VR-Systems 395

tion leading to a consistent state, which is interpreted by the different renderers for
the individual sense modalities to create the respective display data.

13.2.3.1 Events, Event Propagation and Behavior Graph

There are several sources for events: obviously user actions via input devices are
events, which either occur at discrete points in time (finally you only click a mouse
button now and then) or e.g. when using a tracking system at every time step, since
tracking systems usually send data continuously.
Less obvious are those events, which originate from internal changes as the step-
ping forward of the simulation time, which triggers the calculation of a new state of
animations or simulations. Moreover there are events, which result indirectly from
user input or simulations as for example the collision of two moved objects.
On the implementation side events are represented by state changes of objects,
i.e. by changes of the values of internal variables of them, which define their current
properties.
The task of the object manager is to propagate event changes according to the
scene description. This means, that it has to send messages to all concerned objects
to inform them about relevant events, which they should react onto. To illustrate the
underlying principles we will utilize the terminology of the ISO standard X3D [1].
As a rule even VR-systems which don’t implement this standard work with similar
concepts.
Besides the software object instances which correlate to those objects of the vir-
tual environment, which are arranged in the scene graph and rendered by a renderer
(visible objects by the graphic renderer, sound sources by the acoustic renderer,
tangible objects by the haptic renderer), there are software object instances which
exclusively provide behavioral functionality. Included amongst others these are sim-
ulations or animations, which for example store a predefined motion to be played
back, controlled by a timer.
Both object types — scene and behavioral objects — are stored as nodes in
the behavior graph. Therefore the scene graph could be considered as a subgraph
embedded into the behavior graph. In addition to the parent-child-relations of the
transformation hierarchy the behavior graph contains directed edges which model
reaction chains of events. These edges are called routes (in X3D).
Routes deliver messages between nodes. In the simplest case a message is the
transfer of a new value: an animation may for example generate a new position
which should be transferred to the transform node belonging to the object to be ani-
mated. Within other concepts the message transfer correlates to the call of a function
of the receiving node which may take several parameters. The state change initiated
by the message in the receiving node very often results into the generation of new
events, which again have to be transferred to linked nodes. This process of distribut-
ing or propagating messages is called event cascade.
We will give a simple scenario to exemplify this concept: for the replay of a time
controlled animation a timer is instantiated in the scene. It generates a time signal
396 13 Software Design

for each run of the VR application loop (the event is “a new loop has been started”).
The event is transferred to an animation node by a message which contains the
new time stamp. The animation node evaluates a stored function which maps the
time value to a new position value (this evaluation is the subsequent event). The
event is propagated to a transform node, which sets the new position value in the
transformation it holds internally. When the next rendering is done (section 13.2.5)
all objects which are descendants to the transform node in the hierarchy will appear
at the new position.

T A

A
G

T T T T =
G G G G

Fig. 13.1 Simple scene and behavior graph of a driving vehicle: S: root node of the scene,
T: transformation node, G: geometry node, A: animation node, clock symbol: timer. Routes are
dotted.

This composition of behavioral and scene objects make it possible to construct

networks of dependencies which model very complex behavior inside a virtual
world. Messages about events commonly are not only sent to one but many ob-
jects (by creating several outgoing routes). In the same way an object may receive
many types of messages and react differently to them.
The control of this data flow is the responsibility of the object manager. In some
constellations of the event cascade cycles could happen which have to be broken to
avoid infinite loops, which otherwise would lead to a seemingly frozen system. The
events which could not be processed for that reason in one time step but should not
get lost are fed into the event cascade of the next time step.

13.2.4 Simulation

Despite of the diversity of behavior which could be created by the composition of

simple behavioral elements it is not possible to describe everything what eventu-
ally is of interest in a virtual environment. Many desired behavioral properties of a
13.2 Design and Architecture of VR-Systems 397

scenario cannot be defined by simple timer or event controlled processes but need
more complex simulations. For example it is not feasible to predefine animations for
every single activity during a surgical operation — taking hold of tissue, clamping,
cutting at arbitrary spots, suturing — which would properly produce local deforma-
tions or even topological changes due to incisions. Rather this behavior of organs
must be simulated in real–time and depending upon the use case different levels of
physical correctness are necessary: if e.g. the focus is to train dexterity (motor skills)
the reactions of the virtual tissue only need to be plausible and a kind of close to
reality. If on the other hand a lesson is about improving diagnostic abilities (sensor
skills and interpretation), where specific properties of anatomical structures should
be examined for pathologic alterations, these have to be simulated with much higher
accuracy and realism.
To implement the relations which have been identified during simulation model-
ing usually data structures and calculation methods are needed, which are not part
of a general VR-system. Very often moreover there already exists an implementa-
tion of the physical model in a specialized library or system for physical simulation,
which at the best could be reused and integrated into a given VR-system. To do so a
concept is needed how the application loop of the VR-system exchanges data with
the simulation component and synchronizes with it.
Simple simulations possibly may be implemented directly as nodes of the be-
havior graph to be processed within the application loop. The node encapsulates the
interface of the simulation library, delivers the events it receives as input data to the
simulation and feeds the results of it as subsequent events into the event cascade.
However this approach is only suitable if the simulation does not need too much
computing power. Inside the application loop only a restricted amount of computa-
tion time can be allocated for performing a simulation step, since the various output
systems need to be updated at regular intervals in order to guarantee a minimal
graphical frame rate for example. Otherwise the virtual environment may feel slug-
gish or start to judder. Even if the simulation itself fulfills real time demands it may
happen that the necessary computing resources are not available in the context of
the complete system which also has to manage input and output devices and process
the event cascade. In this case it is inevitable to decouple the simulation from the
cycle of the application main loop.
To support the clarification of the problem we will at first discuss the term
“real-time”, which is used more sloppy in the domain of computer graphics as e.g.
in robotics: one essential temporal requirement of a graphical display is obviously
to nearly always achieve update rates above a frequency of 15 Hz to create the vi-
sual impression of motion and avoid perceivable delays between user interaction
and the reaction of the system. This often is already taken to be enough to refer to
it as real-time computer graphics. More precisely though it would be to say that an
application provides interactive computer graphics to emphasis, that the system is
fast enough for fluid interaction.
When speaking of real-time in the simulation domain it is meant, that the sim-
ulated time or model time proceeds as fast as the real world time. Simulations can
stretch or compress time (slow motion or fast motion) for to, for example, show pro-
398 13 Software Design

cesses, which are too fast or too slow in reality to observe them. Simulations of this
kind do not run in real-time. The underlying interpretation of “real-time” however
does not define any constraint on the time needed to calculate one single simulation
step.
On the other hand real-time capability of a control system especially in robotics
denotes that the result of a calculation is available within a given period for use as
control signal. If it is enough to fulfill this requirement on average and exceptions
are largely harmless as long as they are not too frequent this is called “soft” real-
time. Within “hard” real-time constraints the system must complete each calculation
before a deadline in any case. The time to deadline may differ largely though: some
systems only require a guaranteed result in a period on the time scale of seconds
whereas others need to be clocked in microseconds.
To meet the requirement of hard real-time with multi-tasking (see below) control
computers have to run under a real-time operating system like LynxOS, QNX or
VxWorks. These take care, that processes with according privileges get the proces-
sor within a guaranteed time span for a defined period. Based upon this guarantee
the processes running on the system can give real-time guarantees themselves.
The problem touched above — that not enough resources can be allocated for a
simulation — now presents itself as follows: a simulation may, for example, not be
able to perform a simulation step during the duration of one cycle of the application
main loop, which may have to run with 15 Hz. Thus the simulation does not satisfy
the soft real-time requirements of the VR-system. But it possibly meets the real-
time condition for simulations if it for example can perform a step of 0.2 seconds
of model time with a frequency of 5 Hz and a processor load of 50 %. Now the
integration of the simulation into the VR-system becomes possible, if it is separated
from the main loop into its own loop: about every three main application loop cycles
the simulation can feed new data into the behavior graph via the node associated
with it. Then the next simulation step can be started with new input data coming
from other parts of the virtual world.
The concurrency of the VR main loop and the simulation is realized by outsourc-
ing the simulation into its own so-called thread. Threads are a technology provided
by the operating system, which allows it to let parts of a program run temporally
independent from the rest of the program. This is similar to the concept of processes
which is the basis of multitasking. Simply speaking the processor switches the pro-
gram it processes in very short time slices. For the user it appears as if the programs
would run at the same time. On modern processors with multiple processor cores or
computer systems with multiple processors the execution of some of the programs
(in case of processes) or program parts (in case of threads) actually takes place at
the same time. For complex scenarios this concept is developed even further: it is
also possible to outsource the simulation onto a separate computer or a computer
cluster10 as well. This is called a distributed system.
The next issue is the correct synchronization between the simulation and the
rest of the system: each time data has to be exchanged between the two loops both

10 Cluster in this context denotes a network of computers which collectively perform a task.
13.2 Design and Architecture of VR-Systems 399

systems have to be in a consistent state. If a rotation matrix was only filled with new
data to the half when it is transferred this would result into data corruption. Most
likely the inconsistent copy of the rotation matrix wouldn’t have certain essential
mathematical properties. Problems of this kind are avoided by defining dedicated
spots in the program where the involved threads wait for each other and thus only
“approved” data is exchanged. We don’t go into the problems of the synchronization
of processes and threads in more detail here but as to be expected the effort to
implement it becomes the more, the more complex the distribution of the system is:
if threads run on one computer common memory in the RAM can be used for data
exchange whereas in the case of cluster systems network communication including
inherent latency problems gets involved.

13.2.5 Subsystems for Rendering

The main purpose of a VR-system is of course to display the simulated virtual en-
vironment. To do so various rendering subsystems have to interpret the current state
of the scenario regularly and create an appropriate stimulus (a rendering) for every
intended sense modality.
For the most obvious one, the visual or graphical presentation, this means, that
about every 10 to 50 milliseconds the graphical rendering has to be triggered. Since
it is only sensible to create a new image when a new state of the virtual scene is
available usually the graphical rendering is directly integrated into the main appli-
cation loop of the VR-system. After the event cascade has been processed and, if
necessary, data from concurrent simulations has been synchronized into the main
thread, the system starts a so-called traversal of the scene graph: the renderer recur-
sively marches through the transformation hierarchy and thus by and by all nodes
which contain information relevant for visualization are visited. The respective data
is transferred to the basic graphics library11 : affine transformation matrixes are accu-
mulated from data for translation, rotation and scaling in transform nodes, material
nodes provide color settings and textures, light nodes direction, color, brightness and
attenuation of light sources and last but not least the triangles and other polygons
stored in geometry nodes are “pumped”12 to the graphics library. The library is re-
sponsible for transferring the data efficiently onto the graphics card via appropriate
drivers. The graphics card finally performs the computation of a color for each pixel
autonomously and creates the output signal for the connected monitor or projector.
It’s worth to notice, that the graphics card controls the output device with a con-
stant frequency (e.g. 60 Hz for TFT flat screens and preferably higher for CRT dis-
plays) which is completely asynchronous to and independent of the rate in which
images are computed. For CRT displays a frequency of 85 Hz and above is recom-

11 This basic graphics library usually is DirectX on Windows-Systems, on others the platform
independent OpenGL (Open Graphics Library)
12 It is spoken of “polygon pumps” in this context as well.
400 13 Software Design

mendable for the reason that this avoids the perception of flicker by the receptors in
the periphery of the retina, which are specialized in detection of movement: due to
the functional principle of a cathode ray tube (CRT) the brightness of each pixel al-
ternates, because the phosphor on the screen gets activated only for a short moment
when the electron beam passes it on its zigzag way and luminesces decreasingly
until activated again. On the one hand the afterglow duration should not be too long
to enable the sharp display of fast movements without smear artifacts, but on the
other hand a short afterglow duration abets the distracting perception of flicker. Flat
screens or LCD or DLP projectors don’t have this problem, because their backlight
provides quasi constant brightness and pixel colors are created by filtering.
Hence for the graphical display one has to distinguish between image update rate
and display refresh rate. The first one is determined by the cycle time of the VR
application loop and has to be high enough, that single images fuse into motions,
whereas the latter, higher one is determined by the graphics card and drivers and ad-
justed to the frequency, at which flicker gets imperceptible dependent on properties
of the display hardware. Obviously there is no benefit of an image update rate of
the VR-system above the refresh rate of the monitor or projection system and only
would wastefully occupy computing power.

Some of the concepts of acoustic rendering are similar to those of graphic render-
ing. To take the location of a virtual sound source into account for sound synthesis
the transformation hierarchy is evaluated and the orientation and position of the
sound source as well as of the user determined. These parameters for the generation
of stereo effects13 are input into the audio library14 together with audio chunks —
the audio data fragmented into short pieces — from each sound source. The audio
library forwards this data to the audio card, which generates the final output signal
for the loudspeakers. The VR-system has to take care, that enough audio data is
available for the audio card until the next update cycle. Otherwise distracting drop
outs become audible.
Humans perceive frequencies above more than 10 kHz, to achieve a clear audio
experience free of artifacts dynamics of about 20 kHz are preferable. Due to the
enormous computing power it would need the generation of sounds by simulating
the vibration behavior of the objects in the virtual scenario in realtime is impracti-
cable. Most VR-systems therefore use the trick to combine the sounds from a set of
prepared samples, even if that doesn’t allow for the full diversity of sounds which is
possible when interacting with real oscillatory objects. To compensate this the audio
renderer may also interpolate between samples associated to different locations of
the object the user interacts with [205].

Thus the requirements for the graphical and acoustic rendering are determined
essentially by the biological properties of the senses. Neither the visual nor the

13For realistic surround sound simulation geometry information has to be included additionally.
14 For example platform independent OpenAL (Open Audio Library) or specific for Microsoft
DirectSound3D.
13.2 Design and Architecture of VR-Systems 401

auditive sense retroact to the technical device, there is no coupling loop15 which
possibly could cause instabilities.
In contrast the haptic rendering has besides the physiological a control theoreti-
cal component too, because within haptic systems there is bilateral energy exchange
between haptic device and user. Therefore as in general for controlling coupled sys-
tems hard real-time requirements have to be fulfilled when driving haptic devices,
in particular force feedback devices: too large latencies in the control loop — i.e.
control information is not updated for too long — destabilizes the system. Because
typical haptic devices are tuned for low damping this could happen in a resonance
frequency and carries the danger of self destruction of the system or hurting the user.
This leads to the postulation, that high control frequencies of above 1000 Hz (chap-
ter 12) should be achieved for the most frequently used devices with low impedance
(chapter 5) or when high force gradients are wanted. Besides this technical reason
the high frequency has implications for the quality and fidelity of the haptic impres-
sion. Since the haptic sense perceives vibrations of above one kilohertz (section 4.3)
people find haptic systems which are able to display information in this range to be
more realistic. This especially applies for the extreme cases, free space movement
and stiff contacts.

13.2.6 Decoupling of the Haptic Renderer from other Sense

Modalities

The visual presentation needs update rates, which are oriented to the temporal reso-
lution of the visual sense. Even for the perception in the peripheral field of view and
for extra fast movements 70 Hz is quite enough. For professional applications even
lower frequencies of about 20 Hz suffice, because there virtual objects as well as the
users usually move much slower than in computer games. Thus the requirements to
a graphical renderer with respect to the update rate are about the factor 10 to 100
lower than to a haptic renderer. It follows, that the haptic renderer has to be decou-

15 Strictly speaking even in this case there is a feedback loop which is closed via the interaction of
the user with the virtual world. The instabilities, which occur in practice, usually don’t result from
specific properties of the graphical or acoustical rendering, but from unstable physics simulations.
Though resonances occur in the presentation, but not because of the representation. An object
may jitter on the monitor, but the luminous flux from the monitor doesn’t get chaotic thereby or
overdrives and endangers the user either.
This may be different with acoustical rendering. There it indeed could happen, that the user’s
interaction and a slow synthesis of the sounds lead to resonance effects, which — if the sound sys-
tem is accordingly powerful — may cause acoustic overdrive. But this problem is mostly uncritical
due to the limited capacity of the sound hardware.
402 13 Software Design

pled from the cycle of the main application loop of the VR-system and the other,
slower renderers, for which such high update rates are neither feasible nor fruitful16 .
The decoupling of the haptic rendering loop is implemented using the same
principles as the concurrency of simulations (see sections 12.1 and 12.2): follow-
ing the initialization phase of the VR application the system starts a thread, which
concurrently to the application main loop executes the haptic rendering loop with
high priority and frequency. It consists of the query of position data from the haptic
device, collision detection, contact classification with force simulation and finally
actuating the device driver with the calculated forces and moments. At the begin-
ning of its cycle the main thread reads position data of the haptic device from the
haptics thread to feed it into the event cascade. The data flow in the opposite direc-
tion is more intricate:
In complex scenes the collision detection is the time critical part of haptic ren-
dering. Therefore it is an efficient approach to reduce the amount of data on which
collision detection has to be performed in the haptics thread: if the main thread only
supplies the haptics thread with a local extract of the scene, the global collision
detection could run with the lower frequency of the main loop whereas the high-
frequency collision detection only “sees” object parts, which currently are close to
the haptic probe. After processing the event cascade the main thread estimates an
area from the current position and movement of the haptic probe17 , which includes
the haptic probe and the predicted path of it at least up to the next collision cycle of
the main loop. The estimated area should have the smallest volume possible, but it is
better to overestimate a little than to underestimate. All polygons of the scene which
intersect the area are collected and pumped to the haptics thread and rendered by it.
Collecting the relevant polygons itself can be implemented using collision detec-
tion algorithms by approximating the area by a polygonal hull object, which then is
tested against the polygons. Obviously the collision detection in the main thread has
to be done regularly and in not too long intervals, because otherwise the prediction
of the motion of the haptic probe gets too inaccurate and the portion of the scene
which has to be provided to the haptics thread gets too large. This again would make
the local collision detection in the haptics thread more costly.
The data flow between main application thread and haptics thread has to be de-
signed carefully, thus that the haptics thread never has to be stopped longer to wait
for new data from the main thread. During the data exchange between threads it is
inevitable to protect the memory areas which currently are written by one thread
from any, even reading access by other threads, because the data is not consistent
during the write operation. To achieve this the memory area is locked by the writing
thread using operation system functionality. Any other thread running into a state-
ment during the program execution which needs access to the said memory area, has

16 Although the limit frequencies for acoustic rendering are above those of haptic rendering it could
stay in the main application loop, when the sample-based approach is used. Only a physically based
sound simulation would have to be decoupled, but this should not be discussed further here.
17 The velocity of the haptic probe can be calculated from the position and duration of the last

cycle in the main thread of course, but it’s more accurate, if this is done in the haptics thread also.
13.2 Design and Architecture of VR-Systems 403

to be stopped until the writing thread unlocks it again. To prevent that the haptics
thread gets thwarted by the main thread when it writes the local scene portion the
locking phase has to be kept short.
A good approach for this is to use double buffering: while the haptics thread
is rendering a local copy of the scene in one memory area, the main thread fills
a second memory buffer with the next local scene copy. Not before this copy is
transferred completely to the buffer, the haptics thread is notified, that it should
start using the data in the second buffer for rendering and release the first one to be
refilled by the main thread. Only for the transfer of the notification and the buffer
switch both threads must be synchronized. This should be done that way, that the
main thread has to wait for the haptics thread if necessary but not vice versa — on
the one hand because of the higher criticality of the haptic rendering, on the other
hand because the haptic loop reaches the synchronization point in the program more
often due to the higher loop frequency anyways and therefore the average waiting
period of the main thread is much shorter than it would be for the haptics thread
conversely.
The concept of the local scene copy is also used with success, if the haptic ren-
dering loop runs on another computer than the main application loop. For example
there are haptic devices like the FCS HapticMASTER (fig. 5.9 on page 102), which
are sold with a separate driver computer on which the haptic loop is executed. In this
case it is not necessary any more to have a dedicated high-frequency haptics thread
in the VR-system itself. Apart from the differences in the inter process communica-
tion on operation system level the same principles are used for data exchange and
synchronization.

13.2.7 Haptic Interaction Metaphors

As aforementioned shortly, the term interaction metaphor calls a concept in the

context of “human-computer interaction” (HCI), which defines how a user action is
linked to the reaction of the computer system. The way people work with a computer
today by moving the mouse to control a little arrow and to click on mouse buttons to
trigger actions is based on a complex interaction metaphor, which consists of many
smaller interaction metaphors. One of them is e.g. the metaphor “drag and drop”:
pointing to an object, click, hold the button moving the mouse and thereby dragging
the object onto another one as the icon of the tray, let it go. This interaction metaphor
is patterned after the natural grasping, moving and releasing of an item.
For the hapticly enriched interaction with a virtual environment on a very basic
level two interaction metaphors can be discerned.
Ideally if a perfect haptic device was available, the user could directly interact
with all objects of the virtual environment including the sense of touch: when grasp-
ing in the very moment of the contact between fingers and hand to the virtual object
a haptic perception would occur. The fingers of the hand as well as any other part
of the body would be totally free in movement without resistances except for those,
404 13 Software Design

which are defined by the virtual world. Everywhere at the body haptic stimuli could
be provided. This shall be called direct haptic interaction.
Partially it is possible: the end-effector of the first versions of the haptic device
PHANTOM for example is a small plastics cone like a thimble where the user puts
his fingertip into to move the TCP. The force feedback of the system leads to the
impression that you (almost) directly touch the surface of virtual objects.
To extend this to the whole hand it would be necessary to have many effectors,
one for each finger or even better one for each phalanx, several for the palm and
the back of the hand as well. Some developments of exosceletons lead into this
direction, but there are no devices yet which could generate an even only roughly
realistic impression. The software side seems to be the less demanding part of this
challenge: the hand can be modeled in sufficient granularity using todays software
technologies. The collision detection and force simulation between hand model and
virtual environment is an ambitious, but not an fundamentally impossible task. On
the side of the hardware development even the exact tracking of all possible hand
movements is extremely difficult: each finger can be moved in 4 degrees of freedom,
the thumb in 3 and additionally multiple motions of the metacarpus are possible.
Conversely for each of these degrees of freedom actuators of best quality have to be
plugged together in a very confined space. Even if only the fingertips are considered
the technical complexity is enormous.
Because of that another interaction metaphor comes in handy which is used with
all widespread haptic devices: a kind of indirect interaction with the virtual environ-
ment called tool-handle metaphor. Within this concept one does not directly touch
the virtual environment with the fingers, but indirectly via the item one holds in the
hand. It is as one would touch the environment with the tip of a pen instead of the
fingertip.
This admittedly is a severe restriction but it still allows utterly valuable haptic
interaction while requiring much less effort with regard to the device technology.
First of all in many applications in reality the haptic perception is indirect also.
That’s obvious for minimal invasive surgery but also eating with a knife and fork,
feeling out the notch of a screw with the screwdriver, writing or drawing with a pen
or brush, modeling clay with tools and doing woodwork with a chisel are examples.
When simulating these applications in virtual reality it is feasible to equate the end-
effector with the tool. If the end-effector of the haptic device could be exchanged
then it is possible the effector can be a clone of the grip of the real tool. It’s no-
ticeable, that the applications mentioned above are of less explorative nature but are
more related to the modification of objects.
But the tool-handle metaphor is also useful in different problems: whenever it’s
suitable to equate the end-effector with a (substantially stiff) virtual object which
should be moved around in the virtual environment it’s a good choice to use a real
representative of it as end-effector. In combination with rapid prototyping this ap-
proach is of special interest for assembly simulation as aforementioned in the intro-
duction. The part which should be tested for assemblability in VR can be produced
by stereolithography and mounted as end-effector. The mechanic simply takes and
moves it as in the real assembly situation.
13.3 Algorithms 405

The choice of the interaction metaphor plays an essential role for the design of
the device as well as for the planning of the software application and the algorithms
to be used.

13.3 Algorithms

Depending on the purpose of a visual presentation of virtual objects, different prop-

erties can be used, others can be ignored. Many aspects like contour, color, reflective
properties, transparency, local color-variations (texture) contribute to a realistic vi-
sual impression. The most important properties however are those, which allow to
recognize the shape of the object. Consequently first efforts were aimed at contours
and shape-defining edges during the historical development of computer graphics.
With increasing computing power of the hardware and improved capabilities of the
software the generation of brightness variations, color and other visual object prop-
erties could be realized.
A very similar approach can be found with haptic rendering, enabling the user to
perceive shapes first and material properties second. Such impressions are generated
when body parts of the user — usually finger or hand — hit a mechanical resistance
during explorative movements. This resistance has to show certain properties: with
regards to haptic rendering a spatial and temporal coherence has to exist, it has to
depend on the exerted force, and there must be a spatial relation between user and
the resistance. Consequently a haptic device has to be able to generate varying forces
at different positions in space.
To get a three-dimensional representation of the involved objects they must be
described in a mathematical model first. In general the same representations can be
used as in 3D computer graphics applications. Instead of those attributes necessary
for a visual representation, properties like stiffness, elasticity (hardness), roughness
or stickiness stand in.
The dominating representation to define spatial structures within the so called
“real-time”-computer-graphics are sets of triangles approximating the surface of
an object. Even CAD-systems providing surface representations of higher order18
for modeling generate a triangle-based intermediate representation to render on the
screen. This is a concession to the fact, that today’s graphic hardware is highly
optimized for the handling of triangles. In special application areas (e.g. image
acquisition technology in medical context like computer- or magnet-resonance-
tomography) direct volume models are visualized via Direct Volume Rendering
(DVR), making use of regular 3D-grids of scalar- or vector-data (voxels). Both
representations may be used for haptic simulations too, however additional data
structures for collision detection (section 13.4) have to be generated.
Beside the virtual objects, the interacting user has to be represented within the
virtual environment too. In the graphical rendering the corresponding counterpart of

18 e.g. non-uniform-rational-b-spline tensor surfaces, NURBS

406 13 Software Design

the eye would be the camera, characterized by position, orientation, aperture angle
and other parameters. For the haptic rendering at least the hand — for the most
frequent type of interaction in virtual worlds, which is manual interaction — is
represented by a point in space. Its coordinates are derived from the position of the
handle of the haptic device, the Tool Center Point (TCP). However this point in
the virtual environment has to be distinguished from the TCP in the real world. In
general coordinate transformations are necessary to match the position data from
the device’s TCP to the position of the point in the virtual environment. A simple
example would be a molecular construction kit with haptic feedback, making it
necessary to scale movements by several orders of magnitude. The representation of
the TCP in the virtual scene is called haptic-interaction-point, or short HIP, in the
following.
For some haptic rendering algorithms another object beside the HIP exists to
represent the user-interaction. This object may be point-like or more complex. It
is called haptic probe. Whereas the HIP always represents the position of the TCP
within the virtual scene, the haptic probe is the object used to figure out the contact
situation. In so called “penalty”-based approaches (section 13.3.2) the position of
HIP and haptic probe are identical, whereas they may differ in other methods.

The most basic component of a haptic-rendering algorithm is a collision detec-

tion tuned to the object representation. It calculates the localizaton of the contacts
between user and virtual object. The general concepts of collision detection are sub-
ject to section 13.4.
The collision detection is strongly integrated within the total structure of haptic-
rendering algorithm. Put in a scheme, the cycle of haptic rendering can be system-
ized as follows:
• Position-/orientation detection of the user’s interaction
• Collision-detection and evaluation
• Simulation
• Force or position output via the haptic interaction device
The following sections present several methods to render virtual objects hapti-
cally.

13.3.1 Virtual Wall

The simplest haptic model enables the virtual touch of a smooth plain surface with
a dot-like probe, as it would happen with e.g. a hard pencil’s tip. For this purpose a
haptic device is actuated that way, that no resistance is excerted as long as the user
moves in free space. For simplification we assume the free space to be the half-space
with positive x-coordinate. Then any movement of the HIP through the yz-plane
would get immediately slowed down by a counter force F. The user perceives this
constraint as contact with a virtual wall.
13.3 Algorithms 407

The calculation of the counter force in the simplest format is:

!
K · (x, 0, 0) for x < 0
F= (13.1)
(0, 0, 0) else

with K being the proportional factor which controls the stiffness of the wall. Concep-
tional a spring according to H OOK’s law is simulated, being effective in the normal
direction of the yz-plane only (fig. 13.2). Movement components are not constrained
in parallel to the plane. This makes the plane feel perfectly smooth. A large spring
constant K results in a fast force increase, making the virtual wall feel stiff, whereas
a small K generates the impression of a soft wall, feeling like made of foam plastic.

HIP
x

Fig. 13.2 Virtual Wall

This model of the virtual wall also could be interpreted as force field. In front of
the wall (the yz-plane) the force-field is zero. Behind the wall surface the potential
of the field increases proportionally to the distance to the surface.
This model can be easily generalized to a plane of arbitrary orientation and po-
sition. If a plane is defined in the H ESSE’s normal form with a normalized normal-
vector n and a distance d of the plane to the origin according to
⎛ ⎞⎛ ⎞
x0 n0
d = xn = ⎝ x1 ⎠ ⎝ n1 ⎠ (13.2)
x2 n2

the calculation of the force F for a position p of the HIP is given as:
!
K(d − pn)n for pn < d
F= (13.3)
0 else

The formulation of the control loop of the haptic rendering for the virtual wall is
given as:
repeat:
read HIP position p
408 13 Software Design

s = d - DotProduct(p, n)
if s > 0:
F = ScalarMultiplication(K * s, n)
else:
F = 0
actuate HIP with F
If K increases, the control loop has to be processed with an increased frequency
to guarantee stability of the whole system. As a result of dead-times from discretiza-
tion and other system components, the system tends to generate more energy with
increasing product of K and δ t. This makes the virtual wall to feel not as passive
anymore as it is “expected” from a real wall. If the damping is not sufficient, the
resulting energy makes the system oscillate in its resonance which can be perceived
easily (section 7.3.3).
If an additional damping hardware or some closed-loop control in an external unit
is not possible or reasonable (section 7), some additional damping in the control-
loop of the haptic rendering may increase the stability of the overall system. One
option is to add some damping force to the spring-force, actuating against the move-
ment of the HIP and being proportional to its velocity v with a proportional factor
D, the damping constant. Consequently a spring-damper system
$
Kq − Dv, for s > 0
F= (13.4)
0 else
with s = d − pn and q = s · n results. The vector q describes the distance between
the HIP p from the plane in the reverse normal direction, being the penetration
depth in the virtual wall. To still render an ideally smooth surface it has to be made
sure to use the normal component of the velocity for the damping only: v = q̇. If
otherwise v = ṗ is used, the impression of friction on the surface is generated. This
impression however is not realistic, as it is independent of the contact force and
shows no transition from static to dynamic friction.
In the implementation19 of the control loop the velocity v is approximated by
q −qi−1
the differential quotient tii −ti−1 of the values from the actual time-step (i) and the
time-step before (i − 1):
initialize q_old, t_old
repeat:
read HIP position p

19 In the implementation of time critical systems such as haptic control loops care should be taken
not to make repetitive calculations of similar or unnecessary values. Modern compilers are able to
optimize such calculations out of the machine code. Some manual “optimizations” may even hin-
der the compiler to generate an optimized code. This makes it necessary to optimize the algorithm
first by some complexity analysis. Afterwards it is implemented and only after careful considera-
tions detail-optimizations should be made to the code. Anyway tests should be made to prove that
the optimized code is performing better than the non-optimized version. Only very experienced
programmers are able to compete against good compilers in this discipline.
13.3 Algorithms 409

read current time t

s = d - DotProduct(p, n)
q = ScalarMultiplication(s, n)
if s > 0:
v = ScalarMultiplication(1/(t_old - t),(q_old - q))
F = ScalarMultiplication(K, q) -
ScalarMultiplication(D, v)
else:
F = 0
display F at HIP
q_old = q
t_old = t
It is obvious that even with damping included the duration of a control cycle
should not get too long, as the quality of the velocity approximation degrades. Nev-
ertheless a clever tuning of the parameters K and D may achieve some good haptic
quality even with slow sampling rates. If oneself is in the lucky situation to cooperate
with the hardware-engineers and to be able to influence the hardware architecture, it
is recommended to examine the whole signal processing chain including hardware
and control architecture according to chapters 5, 7 and 12. Usually this makes it pos-
sible to find a much more optimized solution than tuning the rendering algorithm
only.

13.3.2 “Penalty” Methods

The haptic-rendering method of a virtual wall in the previous section follows a so

called “penalty” based approach. These methods are characterized by a cycle start-
ing with a collision-detection to check, whether the HIP is within the boundaries of
the object — whether it is “in contact”20 Afterwards the contact is classified, and a
repulsion- or “penalty”-force is calculated. This force is directed against the impact
direction and tries to push the HIP out of the boundaries of the object.
The “classification of contact” characterizes the penetration depth and direction
as a basis for the calculation of the penalty-force. This step is trivial for the virtual
wall and resembles the calculation of the vector q. The function of the collision
detection in this simple model is covered by the query whether s > 0. Collision
detection and classification with complex objects requires much more efforts. In
fact the most time-critical part of real applications is given by the implementation
of the collision detection.
As written before, penalty based methods with dot-like haptic probes do not have
to distinguish between their own position and the position of the HIP. An obvious
definition of the penetration depth and direction is given by the minimum distance
20The exceptional case of the HIP being exactly on the virtual surface at the moment of measure is
very unlikey due to the time discretization. Nevertheless formally it is part of the contact-situation.
410 13 Software Design

of the HIP to the surface of the virtual object (fig. 13.3). The repulsive force shall
try to move the HIP to the nearest surface point. This allows the calculation of the
penetration depth and direction for every point within the object in advance, as it
does not change during interaction. In this case the classification of the contact can
be done very efficient in a preprocess, which will be described for the case of the
polygonal representation of virtual objects as polygonal surfaces in the following.

Fig. 13.3 Minimum distance between several points and object surface

Beside some negligible exceptions, every point of the object’s interior has a mini-
mum distance to exactly one point of the object’s surface. Polygonal surface models
have the features vertices, edges, and surfaces. All points of the object’s interior
which have an equal distance to the same feature — the same edge, surface or ver-
tex - can be summarized in a common region, called VORONOI region [154] of
the attribute. The VORONOI regions of all features partition the interior of the ob-
ject (fig. 13.4). The partition is computed via so called medial-axes-transformation
(MAT), which will not be detailed further in the context of the book. The term ori-
gins from the name of the boundary surfaces of the said regions which are called
medial surfaces (and in the two-dimensional case medial axes). They consist of the
set of points which have minimum distance to more than one single surface point.
Each Voroni-region of a feature of the object’s surface is bounded by the fea-
ture and medial surfaces. Convex polyhedrons have flat medial surfaces, whereas
in general VORONOI regions of concave edges or vertices show curved medial sur-
faces. To represent VORONOI regions they usually are approximated by polyhedrons
themselves.
During runtime of the haptic rendering algorithm the VORONOI region has to be
identified which contains the HIP. If the HIP is exactly on a medial-surface during
one cycle (a usually purely theoretical situation), the above definition of the pene-
tration direction is not unique. In this case one neighboring area has to be chosen,
e.g. that one lying into the direction of the actual movement. The search for the
VORONOI region can be done with a classic hierarchical bounding volume algo-
rithm (section 13.4).
13.3 Algorithms 411

Fig. 13.4 Inner VORONOI regions of an object. Areas in light grey belong to dotted features,
medium grey to dashed features and dark grey to drawn through features (and the thick dot repre-
senting a convex vertex).

Once the VORONOI region has been identified, the rest of the classification algo-
rithm is simple: the vector from the HIP to the closest point of the adjacent feature
is calculated, which is a standard procedure in any geometry calculation.
Penalty-based approaches have some fundamental difficulties though. The clas-
sification of the contact and the identification of the penetration depth and direction
consider the actual situation only. This results in instabilities at the borders of the
VORONOI regions with perceivable and frequently annoying changes in the force-
direction and intensity. Especially with thin structures and relatively small spring-
constants this effect becomes very dominant, when the HIP switches into the region
corresponding to an object surface opposite to the original penetration point. As a
result an unexpected change in force direction happens; the original resistance is
replaced by a force acting in the direction of movement pushing the HIP out of the
object on the opposite surface. Figure 13.5 visualizes this situation.

HIP
F

Fig. 13.5 “Tunneling through” thin structures: sudden inversion of the force direction when cross-
ing the medial axis (M)
412 13 Software Design

This conceptional disadvantage — not to cover the “history” of the contact — and
its consequence leads to the constraint-based methods presented in the following.

13.3.3 Constraint-based Methods

The principle of constraint-based methods is based on the approach to model the

virtual objects of the scene as borders of the free space impenetrable for the haptic
probe. Whereas the user-guided HIP is able to penetrate an object, the haptic probe
is kept on the surface of the object and additionally constrained to keep the minimal
distance to the HIP. From the relative position between HIP and haptic probe a force
is calculated pulling the HIP in direction to the haptic probe. This force is made
perceivable for the user when displayed by the haptic device.
In 1995 Z ILLES and S ALISBURY [299] proposed one representative of this al-
gorithmic class called god-object21 algorithm (also named surface-contact-point al-
gorithm). This algorithm is one of the most central algorithms for haptic rendering
and is implemented in many haptic-rendering libraries. It solves the major problems
of the penalty-based methods described before, and enables a one-point-interaction
with polygonal surface models. The algorithm is suitable for the control of haptic
devices with three active translatory degrees of freedom.
The method can be described best when following the trajectory of the HIP and
a virtual probe in contact with a virtual object (fig. 13.6). Initially the HIP is in free
space. The location of the haptic probe p0 is identical to the HIP. Now the user moves
the handle of the haptic device. The algorithm checks whether the haptic probe is
able to follow the HIP by performing a ray intersection test against the scene. If
the movement does not cross an object’s surface, the haptic probe is set to the new
HIP position and the next cycle begins. Otherwise the crossed polygon’s supporting
plane22 E1 is set as a constraint to the haptic probe and limits its movement: the
point p1 on the plane has to be identified, which has the minimum distance to the
HIP. This is the new candidate for the position of the haptic probe. But before setting
this point it has to be checked whether the passage-way between p0 and p1 is not
blocked. With non convex objects it may happen that a second polygon has to be
considered, resulting in its supporting plane E2 becoming another constraint to the
position of the haptic probe. In a three-dimensional case the position of the probe
is limited to the intersection line of both supporting planes. The new candidate p2
for the position of the haptic probe is the point nearest to the HIP on this line. A
third time the passage-way between the points p0 and p2 has to be checked. It may
be that the supporting plane of another polygon limits the possible positions of the

21 The haptic probe is called “god-object” by the authors, another typical term is virtual proxy.
22 The supporting plane is the infinitely extended plane the points of a flat polygon lie on. If the
vertices of a polygon are not in one plane, it has to be subdivided into flat polygons before, usually
into triangles. This step should go into a preprocess and will not be discussed in further detail in
this book.
13.3 Algorithms 413

haptic probe on the collective intersection point of all planes. Otherwise p2 is the
new location of the haptic probe. As three supporting planes constrain the possible
position of the haptic probe to one point already, another test in the same cycle is
not necessary.
After the new position of the haptic probe was found, a spring-damper system
(equ. 13.4) between HIP and haptic probe is used to calculate the force to be dis-
played haptically.

p0 p2
p1
E1
E2
HIP

Fig. 13.6 Procedure of the god-object algorithm (diagram for 2D case). The optimum position of
the haptic probe is dotted.

It should be noted that the algorithm does not guarantee, that the haptic probe
reaches a local minimum distance to the HIP within one time step. Such a situation
is displayed in figure 13.6, where the optimum point lies on the second supporting
plane and not on the first one. Nevertheless the algorithm limits the haptic probe
to the intersection lines of both supporting planes. This is not critical, as already
within the next time-step the haptic probe moves into the direction of the optimum
point. Practically due to the high update rates no artifacts can be perceived, as the
method converges in a few time-steps to the optimum position of the haptic probe.
The theoretical disadvantage that the algorithm in some rare cases may need several
time-steps to converge is compensated by the advantage that it terminates for one
time-step after three tests at the latest: a result is guranteed to be computed within a
reliable time-frame.
In this algorithm the largest part of the runtime is usually consumed by the ray
intersection tests for the identification of constraint planes. This makes it necessary
to use sophisticated collision detection methods. On the contrary the minimization
of the distance while satisfying the plane-constraints is less demanding. It is done as
follows: the distance between haptic probe p and HIP q is minimized by searching
the minimum of the energy function (equ. 13.5) with one to three constraints given
by the plane equations (equ. 13.6).
1
f (p) = (p − q)2 (13.5)
2

Ei : ni p − di = 0 (13.6)
By the method of L AGRANGE-multipliers the function to be minimized can be
set up as
414 13 Software Design

3
1
h(p, λ1 , λ2 , λ3 ) = (p − q)2 + ∑ λi (ni p − di ) (13.7)
2 i=1

The minimum is found by setting the partial derivatives δδphi and δh

δ pi to zero and
solving the resulting system of equations:
⎛ ⎞⎛ ⎞ ⎛ ⎞
1 0 0 n10 n20 n30 p0 q0
⎜ 0 1 0 n11 n21 n31 ⎟ ⎜ p1 ⎟ ⎜ q1 ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ 0 0 1 n12 n22 n32 ⎟ ⎜ p2 ⎟ ⎜ q2 ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ n10 n11 n12 0 0 0 ⎟ ⎜ λ1 ⎟ = ⎜ d1 ⎟ (13.8)
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝ n20 n21 n22 0 0 0 ⎠ ⎝ λ2 ⎠ ⎝ d2 ⎠
n30 n31 n32 0 0 0 λ3 d3

The matrix is symmetric and always contains the identity matrix within the upper
left 3 × 3 submatrix and the zero matrix within the lower right corner. In the first two
passes per time step only one respectively two constraints are defined. Consequently
the system of equations is reduced to the upper left 4 × 4 respectively 5 × 5 matrices.
In the third pass it is sufficient to solve the system of equations according to
⎛ ⎞⎛ ⎞ ⎛ ⎞
n10 n11 n12 p0 d1
⎝ n20 n21 n22 ⎠ ⎝ p1 ⎠ = ⎝ d2 ⎠ (13.9)
n30 n31 n32 p2 d3
which represents the calculation of the intersection points of the three supporting
planes, as the solution for the Lagrange-multipliers is of no importance.
Due to the very simple structure all systems of equations can be inverted ana-
lytically in all three cases so that just the coefficients have to be set in at runtime.
It is recommended to derive manually optimized program code out of this method,
minimizing the number of necessary mathematical operations. As an example, the
solution of the 3 × 3 system of equations (equ. 13.9) can be implemented most effi-
ciently with the C RAMER’s rule23 .
An extension of the god-object algorithm from a one-point interaction to an in-
teraction with a spherical probe was suggested by RUSPINI, KOLAROV and K HATIB
[211] in 1997. It solves one problem of the god-object algorithm that due to numer-
ical errors the haptic probe may drop at edges of polygons into the inner volume
of the object. By replacing the dimensionless point by an object with volume this
effect of numerical gaps can be prevented.
Additionally the algorithm, which in some haptic-libraries implementing it is
called RUSPINI algorithm, contains a concept for force-shading. It is used for
smoothing the sensibly edged surface of polygonal objects resulting from the dis-
continuities of the normal directions at vertices and edges.
In graphical rendering e.g. PHONG-shading is used to cover the edges between
polygons. For lighting calculation for each surface point a normal vector is used
which is interpolated from the normal vectors stored at each vertex of the polygon.

23 For larger systems of equations this approach leads to inefficient code though.
13.3 Algorithms 415

This idea is adapted for force-shading, but has to be extended by an additional step.
As the direction of the force is defined by the direction of the spring-damper sys-
tem between the haptic probe and the HIP, a new position for the haptic probe has
to be derived from the interpolated normal first. The authors of the algorithm pro-
pose an adaptation in two phases: after calculating a first position p according to
the constraint-method of the god-object algorithm, a plane is put through p whose
normal vector is collinear to the interpolated normal. Onto this force-shading plane
the HIP q is projected. In the second phase the final position p of the haptic probe
is calculated by using the newly found point q in a second pass of the constraint-
method. By this means the direction of the spring-damper system approximates the
the direction of the interpolated normals.
An interesting fact about this approach for smoothing the direction of forces is the
modification of the displayed force, which is achieved by altering the position of the
haptic probe, and not by directly adding forces to the HIP. Avoiding a conceptional
break in doing so is highly valuable for the stability of the method. Additionally
this method is suitable for the integration of friction, the display of locally different
stiffness of the same object, and the generation of fine surface structures by haptic
textures.

13.3.4 6 DoF-interaction: Voxmap-PointShell-Algorithm

Another fundamental algorithm was developed by M C N EELY, P UTER -BAUGH and

T ROY [169] from Boeing in 1997 to solve questions in the area of virtual proto-
typing. The VPS algorithm combines a penalty-based approach including hybrid
collision-detection with the simulation of rigid-body dynamics and the concept of a
virtual coupling as suggested in 1993 by C OLGATE et al. [40] (see below).
In contrast to the methods discussed above the algorithm allows the output of a
force vector and additionally of torques in three degrees of freedom. This is a basic
requirement for the interactive planning of assembly pathes with virtual prototypes.
It makes use of the tool-handle-metaphor and equates the virtual part with the handle
of the haptic device.
One challenge in the development of the algorithm was to enable haptic rendering
for highly complex scenes with huge polygon counts as they are generated from
airplane CAD data. A clever representation of the scene was the right solution: the
collision detection is not performed on the polygonal data but in form of hybrid
collision detection between a so called Voxmap representing the static part of the
scene and a so called PointShell which represents the “dynamic object”, i.e. the part
to be assembled.
416 13 Software Design

Fig. 13.7 Collision between dynamic object (light grey) and static part of scene (dark grey) with
anti-parallel surface normals at the contact points.

Hybrid Collision Detection

The Voxmap is a global data structure representing the whole static scene. It consists
of a three-dimensional grid of cubes and is generated in a preprocess in several steps.
In each grid element or Voxel (abbreviation for volume-element in analogy to pixel)
the information is stored, whether it is part of an object, whether it contains surfaces
or whether it represents empty space of the scene (fig. 13.8). In a following step the
voxels are classified in more detail, by e.g. storing the distance to the next surface
point in each empty space voxel24 . Afterwards the Voxmap is available as three-
dimensional array of scalar data in a coherent memory area.
The rasterization process of the first generation step for the volume generation
can be performed highly efficient on the graphics card by simply rendering the scene
layer for layer. The volume model is the stack of the generated layers. The resolu-
tion of the layers should be chosen according to the detail level necessary for the
application. As an example: a resolution of 1 mm is usually sufficient for assembly
in automotive applications.
The dynamic part, the one to be moved by the user through the scene, is rep-
resented by the PointShell (fig. 13.9), which can be calculated in advance too. It
consists of a multitude of points located at the surface of the dynamic object compa-
rable to the data a laser-scanner would generate. Also the digital generation process
of the point cloud is similar to a scan. The points are sampled by ray-tests in a reg-
ular raster, which can be realized with the help of the graphics card too. Once again
the density of the sampling should be chosen to resemble the main details of the
object and has to be tuned to the resolution of the Voxmap. For each surface point

24 The calculation of this distance-field could be done via an Euclidian-distance-transformation

(EDT)
13.3 Algorithms 417

Fig. 13.8 Voxmap of the static part of the scene: surface-voxels are dark grey, inner voxels light
grey and free-space voxels are white.

the corresponding surface-normal is calculated and stored negated. The reason for
this is given a little further below.

Fig. 13.9 PointShell of the dynamic object with negated surface normals

Within each cycle of the haptic rendering a simple collision detection algorithm
is applied on the Voxmap-PointShell-representation of the scene: according to the
actual transformation of the dynamic object, each point of the PointShell is trans-
formed into the coordinate system of the Voxmap. Afterwards it is checked, in which
voxel the point comes to rest: if the hidden voxel represents empty space, no local
collision happens. Otherwise the point must be taken into account in the subsequent
418 13 Software Design

classification and force simulation. The advantage of this hybrid representation in

the collision detection is given by the fact, that for each PointShell-point only one
access to the Voxmap has to be done. In addition, this access could be calculated
highly efficient directly from the transformed coordinates of the PointShell-point:
the three-dimensional index into the Voxmap can be achieved by rounding the point-
coordinates to integer values.
As a result the runtime behaviour of the collision detection is completely inde-
pendent of the complexity of the static model part of the scene. This advantageous
runtime complexity however is paid with relatively high memory footprint25 , as the
memory requirements scale cubically with the resolution of the Voxmap.
The runtime complexity of the total collision-detection step with one test for each
PointShell-point scales linearly with the number of points in the point cloud. The
latter in itself depends quadratically on the size of the dynamic object in relation to
the sampling density. Consequently, dynamic objects with a large surface may be
problematical26 .

Local Penalty Method

After collecting all colliding points of the PointShell local penalty-forces Fi are
calculated and summed up. For the identification of the individual forces the sur-
face normals ni stored in the PointShell will be used: during contact of two real
objects their surface normals at the contact point are anti-parallel (fig. 13.7). On a
microscopic size this holds also for edges and vertices. As a consequence the con-
tact normal for each PointShell-point can be used as surface normal of the dynamic
object too.
The next challenge is to identify the penetration depth of the PointShell-point pi .
With a clever trick this problem is transferred to the model of the virtual wall. A
plane is put into the center m j of the surface-voxel containing pi with the normal
direction ni . Then, a distance vector pi to this plane is calculated. Multiplied with a
spring constant the local penalty force Fi (fig. 13.10 and equ. 13.3) is calculated:
!
K(m j ni − pi ni )ni for pi ni < m j ni
Fi = (13.10)
0 else
If the local forces Fi of all n detected contacts have been calculated, the total
collision force FK and the total collision torque MK is summed up:

25 A quite common correlation: the more memory can be used, the better the runtime performance
will be.
26 In theses cases hierarchic bounding volume methods (section 13.4) provide some room for

optimization, which however are not discussed further in the context of this book.
13.3 Algorithms 419

ni
mj
pi

Fig. 13.10 Determination of the local penalty force

1
n∑
FK = Fi
i
(13.11)
1
n∑
MK = Fi × li
i

with li = pi − S and S being the actual position of the reference point (preferably the
center of gravity) of the dynamic object.
Scaling by the factor 1/n avoids the emergence of too strong forces and smoothes
the effects of contact-points being added and falling away between the cycles.
Now an obvious idea would be to actuate the 6 DoF haptic device directly with
Fk and Mk . This would be identical to a hard coupling between HIP and the dynamic
object resembling the haptic probe. Practically this approach frequently results into
instable system behavior and a bad quality of the haptic impression. The dynamics
of the haptic device is usually not sufficient to display the simulated forces.

Virtual Coupling

Similar to the god-object algorithm, which derives the force to be displayed by the
device from the force of the spring-damper system between haptic probe and HIP, a
so called virtual coupling couples the dynamic object with the handle of the haptic
device. The position of the handle in the virtual scene is given by the HIP. In the VPS
algorithm the coupling is designed as a six DoF spring-damper system: translatory
springs with corresponding dampers represent three translatory DoF and rotatory
springs with corresponding dampers the other three rotatory DoF. The simulation
420 13 Software Design

of the torques generated by the virtual coupling is implemented analogously to the

force-simulation of the translatory spring-damper system:

FV = Kt q − Dt v
(13.12)
MV = Kr θ − Dr ω

with Kt and Kr beeing translatory and rotatory spring constants and Dt and Dr rep-
resenting the corresponding damping constants. The values q and v are the vector
distance between dynamic object and virtual handle (representing the displacement
of the virtual spring in idle position 0) and their relative velocity. The vector θ rep-
resents the relative rotation between the dynamic object and the virtual handle, ω is
the relative angular velocity.

Rigid-Body Dynamics

Up to now it is open, how the position and motion of the dynamic object is identified
within the virtual scenario. In the steps described till here each haptic rendering
cycle calculates forces and torques of the virtual collisions and inside the virtual
coupling. Now still their impact on the dynamic object has to be simulated.
One approach is given by the integration of motion equations according to N EW-
TON and E ULER27 :

F = FK + FV = ma
(13.13)
M = MK + MV = Iα + ω × Iω

with the translatory acceleration a = v̇ and the angular acceleration α = ω̇ . As ad-

ditional necessary values the mass m of the dynamic object and its inertial tensor I
describing the mass distribution of the object have to be provided.
Both values can be derived from the surface model of the object with the methods
already implemented: a volume model of he dynamic object is generated with the
same voxelization-method known from the generation of Voxmaps. To each center
wk of an occupied voxel a point mass mk is assigned. The total mass m is given by
the sum of all mk and the components Iik of the inertial tensor
⎛ ⎞
I00 I01 I02
I = ⎝ I10 I11 I12 ⎠
I20 I21 I22

are calculated with

27For a very helpful introduction into the implementation of rigid-body dynamics simulation based
on N EWTON-E ULER equations [13] is recommended.
13.4 Collision Detection 421

n
Ii j = ∑ mk (r2k δi j − rki rk j ) (13.14)
k=1

with the K RONECKER-symbol δi j and r = v − S. S denotes the point of action of

the virtual coupling at the object. Optimally the object is modelled so that S coin-
cides with the center of mass of the object. The latter is important to prevent the
emergence of torques during translatory movements in collision-free space, which
otherwise would disturb the haptic impression.

Including these components the VPS algorithm is complete: in the virtual scene
the collision detection and identification of collision-forces is done. Together with
the forces and torques from the virtual coupling they act upon the dynamic object
and result in its movement. Finally the torques and forces of the virtual coupling
are displayed by the haptic device. The coupling therefore builds the bridge be-
tween real handle and virtual part: it is like the user would move the dynamic object
through the world by pulling a rubber band attached to it. According to the applica-
tion the haptic impression can be controlled by configuring the mass of the dynamic
object adapted to the spring stiffness and particularly to the damping of the virtual
coupling.
This overview about the VPS algorithm shows how object representations, col-
lision detection, physical modelling and finally the simulation of the forces to be
displayed are interconnected and have to be harmonized with each other.
At the same time the limits of a certain modelling concept for a usage beyond
the intended application become obvious. A volumetric description of the static part
of the scene cannot be used without major efforts to represent deformable models.
Movements of single objects of the static scene or interaction between objects ex-
cept for the dynamic object cannot be realized directly with the hybrid collision
detection. Nevertheless it should be noted, that the algorithm proved very valuable
for the interactive verification of assembly steps of rigid parts in rigid installation
spaces. For this purpose it is one of the best methods at hand.

13.4 Collision Detection

The core element of a haptic-rendering algorithm is efficient collision detection. At

a frequency of close to 1000 Hz it must be checked whether there is contact be-
tween the haptic probe and the virtual object. Collision detection is also relevant for
applications in the area of virtual-reality which not necessarily include haptics. Fun-
damentally any kind of physics simulation, where virtual objects interact with each
other, should be mentioned. In scientific as well as in industrial applications and
in computer games the collision detection is of central importance. Consequently it
was focused on this topic since the very beginning of virtual reality.
Beside the obvious usages of collision detection algorithms — to find out
whether virtual objects collide — there are other less obvious ones: within so called
422 13 Software Design

raytracers the paths of single light rays are simulated physically to create a photo-
realistic image. For one image millions of collision tests have to be made between
rays and the virtual scene (using ray intersection tests).
Based on the following considerations the most fundamental questions and con-
cepts of collision detection methods as well as principles for their optimization shall
be discussed. This will provide a good basic understanding of the challenges related
to it. For a more comprehensive discussion specialized literature for collision detec-
tion is recommended [295, 50].
For all problems where large amounts of data have to be processed efficiently,
algorithms with low complexity have to be designed. Complexity is a measure how
processing-time, memory requirements, and the need for other resources for the
algorithm depend on the problem size, which is usually measured by the number of
elements to be processed.
If for example a collision between two objects shall be tested, this can be done
“brute force” by testing all triangles of the one object against all triangles of the other
object. The effort increases quadratically with the size of the problem: if the number
of triangles of both objects doubles, the runtime of the algorithm quadruples. Thus
the algorithm has quadratic order, the (runtime-)complexity is O(n2 ).
The size of the models which can be displayed graphically on state-of-the-art
VR-systems and even on off-the-shelf PCs in real time (section 13.2.4) is several
hundred-thousand or even more than a million triangles. Even in such complex
scenarios efficient collision detection is desired, quadratic complexity however is
completely impracticable.
The total complexity of a collision detection method can be improved with a
plausible idea: in several steps the size of the problem is reduced by performing
simple and more “economic” tests, excluding as many elements as possible from
the more complex tests. The necessary steps are usually arranged in a collision-
pipeline, which is processed step by step.
For a fast exclusion of many elements their spatial coherence is analyzed and ex-
ploited. For example neighbouring triangles can be collected into a bounding object.
If a test of the bounding object shows no intersection, all contained elements can be
excluded from all further analysis.
Different geometry types are suitable as bounding objects: spheres, cubes, convex
polyhedrons with defined orientation of surfaces (so called discrete oriented poly-
topes, DOP), to name the most frequently used ones. Usually only one bounding
geometry type is used within the same collision detection step. A “good” bounding
geometry can be tested efficiently for intersections and approximates the included
elements closely, because this reduces the number of false positive tests. The qual-
ity of the approximation achievable with a bounding geometry depends highly on
the shape of the bounded object. Figure 13.11 shows this effect for different bound-
ing geometry types: spheres for example approximate elongated objects only badly.
Consequently flexible bounding geometry types are advantageous for the approx-
imation quality. However they require more effort for intersection tests. A good
compromise are DOPs which reach a good approximation quality in combination
with relatively simple intersection tests. But there is no universal optimum choice
13.4 Collision Detection 423

of a bounding geometry type for every application, as the runtime performance at

collision detection is dependent on many factors.

Fig. 13.11 Types of hull geometries in 2D (from left to right): minimal sphere, axis aligned bound-
ing box (AABB), object oriented bounding box (OOBB), discrete oriented polytope (DOP)

The test for the intersection of two spheres is simple: may m0 and m1 be the
centers of two spheres and r0 and r1 their radii. They do not intersect, if |m0 −
m1 | > r0 + r1 . For an efficient implementation the monotony of the square in the
first quadrant should be used to avoid the extraction of the root for the calculation of
the distance between m0 and m1 . The test condition then is (m0 − m1 )2 > (r0 + r1 )2 .
This saves one root for the costs of an additional multiplication for squaring the sum
of the radii, which still requires considerable less processor cycles. The optimization
of these tests is relevant, as they are frequently used within the collision detection
process as depicted below. The calculation of the bounding sphere (which exists and
is unique) is not trivial, but can be done using the algorithm of H OPP [96].
In average fewer effort is required by the test for intersection of two axis aligned
boxes. One axis aligned box Q can be characterized by two vertices qmin and qmax .
The intersection of both boxes is empty — no collision happens — if at least one
coordinate of the minimum of one box is larger than the maximum of the other box
(fig. 13.12).
This makes it possible to formulate a test, which has to be processed in the av-
erage of all collision-free cases only to the half of its code, as it already excludes a
collision before:
if qMin[0] > pMax[0]:
return "no collision"
else if qMax[0] < pMin[0]:
return "no collision"
else if qMin[1] > pMax[1]:
return "no collision"
else if qMax[1] < pMin[1]:
return "no collision"
else if qMin[2] > pMax[2]:
return "no collision"
424 13 Software Design

y
p max,1
P
p min,1
q max,1

q min,1
x
q min,0 p min,0 q max,0 p max,0

Fig. 13.12 Intersection test of boxes: no intersection of the intervals on thed y-axis guarantess
absence of collisions.

else if qMax[2] < pMin[2]:

return "no collision"
else:
return "boxes intersect"
The full potential of tests via bounding volumes is used in hierarchic collision
detection algorithms working with trees of nested bounding volumes28 .
The root of such a tree resembles the total object, which is nothing but “all tri-
angles”. During generation of the bounding-volume hierarchy (BVH) all triangles
of a node are distributed onto its child nodes recursively until the leaves of the tree
contain exactly one triangle (fig. 13.13).
An advantageous split is essentially for the efficiency of the following collision
detection. Quality criterions are

• The tree should be balanced according to the number of its child-nodes and there-
fore be optimized in its depth. This is identical to the requirement, that the geo-
metric primitives of a node should be distributed equally among its child nodes.
• The distribution of the geometrical primitives should be chosen to minimize the
volumes of the bounding geometries and reduce overlapping on the same level
of the tree.

28 A tree is a cycle-free interconnected graph, consisting of nodes and connecting edges. Directed
trees have one root. Leaves are nodes which do not have any “child-nodes” and no outgoing
branches. Usually hierarchic collision detection algorithms use binary trees, i.e. trees where all
nodes except for the leaves have two child-nodes. Octrees, trees with eight child-nodes, are an-
other frequent variation.
13.4 Collision Detection 425

Fig. 13.13 Schematic representation of the split of a geometry inside a bounding-volume hierarchy

How good both criterions can be fulfilled at the same time depends on the structure
of the individual models. For models made of triangles with large variations in their
size, no optimum distribution can be found covering both criterions equally good.
After the generation of the bounding volume hierarchy in a preprocess the data
structures are available, which enable the application to quickly filter out large parts
of the objects during the collision tests. If the bounding volumes do not intersect
in a node of the tree, all child nodes and the included triangles can be neglected
for the following tests, whereas nodes, which could not be excluded, are recursively
processed in more detail during the subsequent tests. Very fast only branches close
to the actual contact point remain.
If the question is about the existence of collision only, the collision detection can
be interrupted as soon as a collision was found for one branch. For this question
therefore a near-contact leads to the worst average performance: the hierarchy has
to be traversed deeply up to the demanding triangle tests to just return “no collision”
as result finally.
However for haptic applications exact information about the contact is necessary
to be able to classify it. This requires all contact points to be identified, i.e. all
triangles which intersect another object. In this exact collision detection the near-
collision doesn’t differ from collision concerning its runtime performance.
In many applications it is not sufficient to regard the collision between a geomet-
rically simple probe (e.g. a point or sphere) with objects of the virtual scene but the
contact between complex geometrical models. In this situation both tested objects
are represented by bounding volume hierarchies, which are processed in parallel
recursive traversal.
426 13 Software Design

function CollisionTest( node0, node1 ):

if bounding volumes of node0 and node1 are disjoint:
return
else if node0 and node1 are leafs:
if TriangleTest( node0, node1 ) is positive:
add triangle pair to collision list
else if node0 "larger than" node1:
for each child node cn of node0:
CollisionTest( cn, node1 )
else:
for each child node cn of node1:
CollisionTest( node0, cn )
The function CollisionTest of the pseudo-code-listing above is called with
the root nodes of both object’s bounding-volume hierarchies. During the recursive
processing the list is filled with all pairs of triangles of the one or the other object
which collide (it is an algorithm of the class of exact collision detections). If the list
is empty at its end, in the current time-step there are no collisions. As metric (for
the test “larger than”) for the choice of the path to descend the number of children
or the size of the bounding volume may be chosen.
Beside spatial coherence also temporal coherence is exploited often: in a system
simulated with small time-steps usually only small changes happen from step to
step. Information from the last collision detection cycle can be recycled as a start-
ing point for the search of a collision in the current time-step. Collision detection
algorithms according to this principle are called incremental.
A good example for this is an algorithm which often very quickly proves that
no collision exists and thus used in early phases of collision pipelines. The convex
hulls of two objects given it tries to identify a plane, which lies in between them.
If so, the objects are in different of the both disjunctive half spaces generated by
the plane. The existence of such a separating plane guarantees that no collision
happens.
The first test within each cycle of the collision detection checks whether a sepa-
rating plane existed within the prior cycle. This plane is tested for being valid in the
current cycle too. If this is the case, the further processing of the collision pipeline
for the object pair can be stopped. If this is not the case, a new separating plane
is searched nearby the old plane until one is found or a maximum number of trials
was made. In the latter case the processing of the collision pipeline is continued, as
collision can’t be excluded (fig. 13.14).
Another incremental procedure is given by the Closest-Feature-Tracking algo-
rithm of L IN and C ANNY [163]. It searches for the pair of surface features (points,
edges, faces) of two polygonal objects which constitute the minimum distance be-
tween both (fig. 13.15). The method starts with an arbitrary feature pair and analyzes
specificly29 pairs of neighbouring features whether they are closer to each other. The

29 For the choice of the next feature pair the VORONOI regions of the objects are analyzed.
13.4 Collision Detection 427

Fig. 13.14 Separating plane E

search gives a fast result if a good starting pair was chosen. Such a starting pair is
often available in form of the feature pair of the last collision test.

Fig. 13.15 Pairwise closest features of a set of objects

For the haptic rendering with a point-like probe the question for the collision
detection has to be formulated different from the collision with complex models: it
has to be identified whether the point is within the object, or whether the movement
of the point passed a surface of an object within the last time-step.
A point is within a closed polyhedron30 if a ray casted to infinity starting from
this point in an arbitrary direction hits an odd number of surfaces. With convex
polyhedrons the ray hits exactly once. With concave objects numerous hits may
occur (fig. 13.16). When implementing this test some exceptional cases have to be
considered: the ray may hit an edge or point within the numerical precision. In this
case the corresponding surfaces have to be counted very carefully. Grazing touches
of the corresponding surfaces (only relevant for the concave case) should not be
counted. A penetration should be counted only once even when there are two or
more surfaces touching the ray.
30 An object modelled of triangles representing a solid body is such a closed polyhedron. For

non-closed polyhedrons with a “hole” in their surface an inner region can not be defined so easily.
428 13 Software Design

Fig. 13.16 Ray intersection tests with odd (drawn through ray) and even (dashed ray) number of
intersection points for an inner and an outer point respectively.

The question about a point shape probe having crossed a surface within the last
time-step is answered by a ray intersection test too, but the test is performed with
the line between the last and the current position of the probe instead with the ray.
In both cases the ray intersection test can be optimized using the efficient colli-
sion detection methods described above: the ray resp. the line may be tested recur-
sively against the bounding volume hierarchy of the object in question.
The analysis of the two collision queries for the one-point-interaction leads to an-
other basic distinction. The collision tests between complex objects as described so
far only consider static states of the objects within one time-step. As a consequence
of the time discretization collisions can be missed if a relative small or fast object
is in collision only in between two simulation steps. These cases are addressed by
the methods of dynamic collision detection (in contrast to static collision detection),
which additionally are suitable to calculate the moment of collision back in time and
interpolate the corresponding position of the involved objects.
The basic concept is to approximate the volume occupied by the moving object
since the last simulation step and to test this volume for collisions with other objects.
It is important to avoid the calculation of the envelopes of both objects. If the relative
movement of the objects is expressed within the coordinate system of one of the
objects, it is sufficient to calculate the envelope of the other “dynamic” object.
The methods for hierarchic collision detection can be applied to dynamic col-
lision detection too. We start from the assumption that for both involved objects
boundary hierarchies exists. During the recursive collision detection algorithm an
envelope for the bounding volume of the current node in the bounding volume hi-
erarchy of the dynamic object is approximated and tested for intersection. Once
again the principle is to exclude large parts of the geometry as effortless and early
as possible from the more performance-critical collision tests.
The approximation of the envelope of the bounding geometry as well as of the
individual triangles is simple for translations. In these cases the movement volume
itself can be characterized by a polygonal object. The bounding surfaces of rotating
objects become so called ruled surfaces though which are curved and need a more
complex representation. Therefore once again a rough outer approximation of the
envelope by an bounding object of the chosen bounding geometry type (sphere, box,
13.5 Conclusion 429

DOP) is used. Only in the leaves of the hierarchy exact tests are performed for the
triangles, which are necessary when the contact moments are calculated back in
time.

13.5 Conclusion

For the integration of haptic rendering in existing virtual-reality systems a modifi-

cation of their architecture is usually necessary. In case of a new development of a
multimodal system which should include the haptic sense an appropriate architec-
ture should be chosen: it should provide functionalities for multithreading and/or
distribution including mechanisms for synchronization. This is needed to decouple
the higher-frequent haptic rendering of the application behavior and graphic render-
ing cycle.
Additionally the VR-system has to include data-structures and algorithms for ex-
act collision detection between haptic probes and virtual objects as well as physical
models for the calculation of collision- or frictional forces. If the forces calculated
inside the haptic-simulation are intended to affect the virtual scene, instances for
physical simulation, rigid-body dynamics, deformable objects etc. have to be im-
plemented in the scenario and linked to the objects of the scene. The resulting data
flow requires an open eye on the synchronization between the different modules.
Due to the hybrid functionality of haptic devices — they are input- and output-
devices at the same time — an adequate concept of device abstraction has to be
realized too. At least the communication between high-frequency haptic rendering
thread and haptic device has to be designed with little latency. With “dangerous”
devices, which are extremely strong or fast, it is strongly recommended to give hard
realtime-guarantees to the haptic rendering loop. Among other things this requires to
build on top of realtime operating systems. Softer realtime-requirements should be
fulfilled for smaller haptic devices in any case, as the fundamental system stability
and an acceptable quality of the haptic feedback can be achieved easier.
“This page left intentionally blank.”
Chapter 14
Final Remarks on the Design of Haptic Systems

Like any other design process the design of haptic systems is largely influenced
by the optimization of a technical system based on the balancing of a plurality of
decisions on separate components, which, as a rule, influence each other. In the
beginning, the requirements of the customer respectively of the project have to be
defined. The methods presented in chapter 6 are intended to systematically identify
the most important aspects of these requirements. However, the engineer should be
conscious of the fact that for the design of a sense-related interface less precise and
definite terms are available than he may be used to. Additionally, the knowledge on
the part of the customer may result in considerable confusion, as especially hap-
tic terms like, e.g., resolution or dynamics may be used in the wrong context or
understood in a wrong way. A better definition of the requirements without major
misunderstandings is achieved by e.g. giving the customer aids, “shows-and-tells”
of haptics. It is necessary for the customer and the engineer to come to a common
understanding based on references known to both.

For this reason, an understanding of the specialties of haptic perception on the

part of the engineer is necessary. It should not be limited to the technical charac-
teristics described in chapters 2 and 3, but also include some knowledge about the
“soft”, i.e. psychological and social aspects of haptics (chapter 1). In chapter 4 an at-
tempt is made to bridge the gap between these aspects and the technical description
of haptic perception. However, the application of the methods presented requires in-
creased attention, as the data basis of the models is not yet sufficient. Nevertheless,
they show a path suitable for further optimization of haptic devices.

Based on the requirements discussed above, the technical design process may
begin. One of the very first decisions is the choice of the haptic system’s structure
(chapter 5). Although this decision is at the very beginning of the design process, a
rough sketch of the favored structure of the device to be developed has necessarily

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_14, 431

c Springer-Verlag Berlin Heidelberg 2009
432 14 Final Remarks on the Design of Haptic Systems

to be made. This demands a considerable previous knowledge of all the branches

of haptic device design, which later will be needed again during the actual de-
sign phase. Besides the already mentioned decision on the general structure (chap-
ter 7) , the basis of the design of kinaesthetic and tactile systems is its kinematic
structure (chapter 8). After the considerations made for kinematics, concerning the
transmission- and gearing-proportions, the working volume, and the resolution to
be achieved, suitable actuators are chosen or even designed. In chapter 9 the basis
for this is provided by comparing the different actuation principles. Examples of
their realizations, even of unusual solutions for haptic applications, provide a useful
collection for any engineer to combine kinematic requirements of maximum forces
and translations with impedances and resolutions.

As closed-loop admittance controlled systems with kinaesthetic and tactile ap-

plications are gaining in importance, force sensors have to be considered as another
component of haptic devices. In chapter 10 this technology is introduced, providing
the tools, as well as conveying the chances but also the challenges connected with
their application.
A frequent application of haptic devices is to be found in the human-machine in-
terface of simulators, be it for games ranging from action to adventure games, or
for more serious applications for training surgeons or in the military, respectively in
industrial design. In addition to the output of haptic information an input of user-
movements is required. The measurement principles typically used are discussed
in chapter 11. The design steps presented so far will enable the haptic device to
provide a tactile or kinaesthetic output to the user, often measuring a reaction, too.
Especially with today’s computer technology, the data will be almost always inter-
faced with a standard PC. The requirements derived from this interface are subject
to a presentation of standard interface technology given in chapter 12, whereby the
interfaces’ performances are compared with each other.

Due to the rather frequent application of haptic devices in simulators, an inter-

face with a simulation engine is required. An insight into the requirements and
challenges of suitable haptic algorithms is helpful for the hardware engineer to
improve the communication with software-engineers and the interfacing with their
VR-environments. An appropriate introduction is given in chapter 13.

The cross-section given in this book is meant to improve and further speed up the
design of haptic devices, and to avoid the most critical errors typically made during
the design process. The research on the area of haptic devices is making impressive
progress. Every few months adapted control-engineering concepts appear; the us-
age of haptic perception for the design is subject to current research. Actuators are
being continuously improved; even new principles with haptically interesting prop-
erties regularly appear on the market. Closed-loop controlled systems become more
and more interesting, due to the slowly increasing availability of highly dynamic
high-resolution force sensors, This dynamics of a still young discipline commits
the developing engineers to monitor current research attentively. For this purpose,
14 Final Remarks on the Design of Haptic Systems 433

finally, a list enumerating teams active in the haptic area has been compiled in the
appendix-chapter 15.
“This page left intentionally blank.”
 Part III
Appendix
“This page left intentionally blank.”
Chapter 15
URLs

The following tables are a collection of all names and groups familiar to the authors.
Naturally the lists are not compelling, but they provide a starting point for the re-
search about relevant sources. Additions to this list for future editions of this book
may be sent to the editor via email.

Table 15.1: URLs of labs and individuals working in the area of haptics, taken and
extended from HapticsSymposium.org

Institute Head URL Type

ALAB, Tokyo Univ. S HINODA www.alab.t.u-tokyo.ac.jp/ Lab
A Labaratory for Teleop- F IORINI http://metropolis.sci.univr.it/r_d/r_d.php Lab
eration and Autonomous
Intelligent Robots (AL-
TAIR)
Artificial Intelligence K HATIB http://robotics.stanford.edu/ Lab
Lab, Robotics
Bionengineering, De- B URDET www.bg.ic.ac.uk/staff/burdet/Home.html Person
partment of
Bioinstrumentation Lab., J ONES http://bioinstrumentation.mit.edu/jones/ Person
Haptic group
Biorobotics Labaratory H ANNAFORD http://brl.ee.washington.edu Lab
BioRobotics Labaratory RYU http://robot.kut.ac.kr/ Lab
Biomimetics and Dex- C UTKOSKY http://bdml.stanford.edu/DML Lab
terous Manipulation
Labaratory
Computer Graphics G ROSS http://graphics.ethz.ch Lab
Labaratory ETH Zürich
Delft Haptics Lab H ELM and oth- www.tudelft.nl Lab
ers
see next page ...

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_15, 437

c Springer-Verlag Berlin Heidelberg 2009
438 15 URLs

... continuation
Institute Head URL Type
EduHaptics P ROVANCHER www.eduhaptics.org Edu.
Fujimoto Lab F UJIMOTO http://drei.mech.nitech.ac.jp/ fujimoto/ Lab
Group of Robots and In- F ERRE and oth- http://138.100.76.36/en/default.asp Group
telligence Machines ers
Haptics and Embedded P ROVANCHER http://heml.eng.utah.edu/pmwiki.php/Main/HomePageLab
Mechatronics Lab
Haptics Grasp Lab K UCHENBECKERhttp://haptics.grasp.upenn.edu/ Lab
Haptic Exploration Lab O KAMURA www.haptics.me.jhu.edu/ Lab
Haptic Interface Re- TAN www.ecn.purdue.edu/HIRL/index Lab
search Lab
Haptics Research Group PANCHANATAN http://haptics.asu.edu/index.php Lab
Haptics and Virtual Real- C HOI http://hvr.postech.ac.kr/wiki/wiki.php Lab
ity Lab
Haptics Labaratory H AYWARD www.cim.mcgill.ca/%7Ehaptic/grouphome.html Lab
Haptiklabor H EGEL www.haptiklabor.de/ Lab
Haptix Labaratory G ILLESPIE http://www-personal.umich.edu/%7Ebrentg/Web Lab
Harvard Biorobotics H OWE http://biorobotics.harvard.edu/ Lab
Labaratory
Human Factors Engineer A NDREW [email protected] Person
Human machine inter- B URDEA www.caip.rutgers.edu/vrlab/ Lab
face Lab
Human sciences group, G RIFFIN www.isvr.soton.ac.uk/HSG Inst.
Insitute of sound and vi-
bration reserach
Institute of Automatic B USS , H IRCHE , www.lsr.ei.tum.de/ Inst.
Controle Engineering S TURSBERG ,
S CHMIDT
Institute of Robotics and H IRZINGER www.dlr.de/rm/en/ Inst.
Mechatronics
Institute for Elektrome- S CHLAAK, www.institute-emk.de/ Inst.
chanical Design W ERTH -
SCHUETZKY
Interactive Systems Re- H ARWIN http://www.reading.ac.uk/isrg/isrg- Group
search Group (ISRG) haptics.asp
Interdisciplinary In- H OWE www.oucom.ohiou.edu/IINR/ Lab
stitute for Biomedical
Research (IINR)
Ishibashi and Sugawara I SHIBASHI, http://nma.elcom.nitech.ac.jp/index_e.html Lab
Lab S UGAWARA
Laboratoire de Systèmes B LEULER http://lsro.epfl.ch Lab
Robotiques (LSRO)
Laboratoire Systemes K HEDDAR http://lsc.univ-evry.fr/ Lab
Complexes (Fre-CNRS)
Mechatronics and Hap- O’M ALLEY http://mems.rice.edu/ %7Emahil Lab
tics Interfaces (MAHI)
Mechatronics Lab YOKOKOHJI www.mechatronics.me.kyoto-u.ac.jp Lab
see next page ...
15 URLs 439

... continuation
Institute Head URL Type
Microdynamic System H OLLIS www.msl.ri.cmu.edu/ Lab
Labaratory
MIT Touch Lab S RINIVASAN http://touchlab.mit.edu/ Lab
Multimedia Commu- S ADDIK and http://www.mcrlab.uottawa.ca/ Lab
nication and Research others
(MCR) Lab
Multimodal Interaction B REWSTER www.dcs.gla.ac.uk/ stephen/ Group
Group
Multisensory Perception E RNST www.kyb.tuebingen.mpg.de/ bu/people/- Lab
and Action marc
PERCRO Bergamasco www.percro.org/ Lab
Physics of man, depart- K APPERS, www.phys.uu.nl/ wwwpm/HumPerc Inst.
ment KOENDERING
Precision and Intelli- S ATO http://kenwww.pi.titech.ac.jp/ Lab
gence Laboratory
Psychology at Hamilton G ESCHEIDER , http://academics.hamilton.edu/psychology/home/ Inst.
B URR , Y EE ,
W ELDON ,
VAUGHAN
Psychology at Carnegie K LATZKY www.psy.cmu.edu/faculty/klatzky/lab/ Lab
Mellon, spacial and hap-
tic
Labaratories for Intelli- C OLGATE , http://lims.mech.northwestern.edu Labs
gent Machine Systems P ESHKIN ,
(LIMS) LYNCH
Robotics Group, B ICCHI www.piaggio.ccii.unipi.it/newrobotics/ roboticre- Group
“E.Piaggio” search.html
Salisbury Research S ALISBURY http://jks-folks.stanford.edu/home.html Lab
Group
Sensory Motor Neuro- W ING www.symon.bham.ac.uk/labs.htm Lab
science
Sensory Perception and M AC L EAN www.cs.ubc.ca/labs/spin Lab
Research Group
Skripte zur Psychologie Z WISLER www.zwisler.de/ Person
Tachi Lab TACHI http://tachilab.org Lab
Telerobotics and Control K WON http://robot.kaist.ac.kr/ Lab
Lab (TCL)
Telerobotics Lab N IEMEYER http://telerobotics.stanford.edu Lab
The Senses of Touch PATTERSON www.ggy.bris.ac.uk/postgraduates/ggmp Person
TNO, Haptics E RP, VAN www.tno.nl Inst.
Touch Laboratory L EDERMAN http://psyc.queensu.ca/%7Echeryl/labpage.html Lab
Haptic group Utah H OLLERBACH , www.cs.utah.edu/ jmh/Haptics.html Group
J OHNSON ,
M ASCARO ,
P ROVANCHER
Virgina Touch Labora- G ERLING www.sys.virginia.edu/ggerling/ Lab
tory
see next page ...
440 15 URLs

... continuation
Institute Head URL Type
Virtual Reality and Ac- BAUR http://vrai-group.epfl.ch/ Group
tive Interfaces (VRAI)
Group
Virtual Reality in H ARDERS http://www.vision.ee.ethz.ch/research/projects _med.cgi?topic=6 Lab
Medicine
Virtual Reality Lab K ESAVADAS www.haptics.buffalo.edu Lab
Visualisation and Image S TETTEN www.vialab.org Lab
Analysis Labaratory
(VIALAB)
VR-Lab. Univ. of I WATA , YANO http://intron.kz.tsukuba.ac.jp/index_e.html Lab
Tsukuba

Table 15.2 Established conferences, workgroups, journals with reasonable parts of haptic hard-
ware
Title URL Type
Eurohaptics Conference www.eurohaptics.com Conf.
Haptics-e www.haptics-e.org/ Journ.
HapticsSymposium www.hapticssymposium.org Conf.
Haptics technical comittee (IEEE) www.worldhaptics.com Soc.
Transactions on Haptics (IEEE) www.computer.org/th Journ.
Worldhaptics Conference www.worldhaptics.com Conf.
15 URLs 441

Table 15.3 Commercial manufacturers of haptic-related products

Manufacturer URL Type
Artificial Muscle www.artificalmuscle.com HW
Butterfly Haptics www.butterflyhaptics.com HW
Chai3D www.chai3d.org SW
ForceDimension www.forcedimension.com HW
Guitammer Company Inc. www.ButtKicker.com HW
haptiklibary www.haptiklibrary.org SW
Haption www.haption.com HW
Haptx www.haptx.com SW
Immersion www.immersion.com HW & SW
Moog www.fcs-cs.com/robotics HW
MPB Technologies Inc. www.mpb-technologies.ca HW
Novint www.novint.com HW
Quanser www.quanser.com HW
reachin www.reachin.se SW
SensAble www.sensable.com HW & SW
Sensegraphics www.sensegraphics.com SW & HW
Xitact www.xitact.com HW
“This page left intentionally blank.”
Chapter 16
Mechanical Impedances and Admittances for
Translatory and Rotatory Systems

T HORSTEN A. K ERN

For the description of technical systems a model of the real physical properties
is mandatory. A central toolset in the area of electromechanics is the use of lin-
ear, time-invariant, and concentrated elements. The following section gives a short
introduction into this topic. Further details can be found especially in [158].
Concentrated elements represent real physical object properties. They simplify
these properties in three fundamental aspects:

• The objects represented by the elements do not have any spatial size - they are
concentrated on a single spot.
• The object properties are linear.
• The object properties are time-independent.
Accepting above limits three types of elements are sufficient to model all relevant
mechanical effects.:
• Mass/inertia m resp. moment of inertia Θ
• Linear or rotary spring with elasticity n or spring-constant/stiffness k
• Viscous friction resp. rotatory viscou friction with damping d resp. frictional
coefficient r
They are represented by the symbols according to figure 16.1.
Using these elements modelling can be made in time- and frequency domain. For
the analysis about stability and frequency plots a description in L APLACE-domain
with complex numbers is advantageous. As there is a limited operation time ex-
pected in all cases the complex frequency is given as s = j ω + ϕ (mit j · j = −1).
For calculating the complex system behaviour two quotients are distinguished:
F M
• mechanical impedance Z transl = v resp. Z rot = Ω
v Ω
• mechanical admittance Y tranl = F bzw. Y rot = M

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_16, 443

c Springer-Verlag Berlin Heidelberg 2009
444 16 Mechanical Impedances and Admittances for Translatory and Rotatory Systems

Fig. 16.1 Symbols for concentrated mechanical elements for translatory (a) and rotatory (b) sys-
tems according to the 2nd analogy.

For the elements this gives us:

translatory mass: sm = Z
1 k
translatory spring: sn = s = Z
translatory friction: r=d=Z
rotatory mass (moment of inertia): sΘ = Z
1 k
rotatory spring: sn = s = Z
rotatory friction: r=s=Z

The elements can be combined to networks. Their usage happens analog to elec-
trical networks. Depending whether velocity or force (torque or rotational velocity)
is used as differential value, two analogies can be formulated. They differ accord-
ing to the handling of the node- and mesh-equations in network-theory. This book
makes use of the 2nd analogy, whereas F is the flux- and v the differential-value.
A transition of the mechanical networks in their electrical equivalent circuit does
not happen in the context of this book. However learning and using this method is
recommend to any electrical engineer to improve the understanding of mechanical
systems. The limitations of concentrated elements can be avoided by a transition
of the network to finite elements and a solution of the resulting wave-equations.
However in the context of this book this advanced method is not used.
Chapter 17
Details about Gyrators and Transformers

S TEPHANIE S INDLINGER

For the design of electromechanical systems the interdependencies between elec-

trical and mechanical networks are formulated by the use of transducers. These
transducers are linear, lossless and time-invariant quadrupoles.
The electromechanical converters are distinguished in active and passive ones.
They show the following properties [158]:

Table 17.1 Properties of passiv and active electromechanical transducers [158].

PASSIVE TRANSDUCER ACTIVE TRANSDUCER

reversible coupling between mechanical and mechanical values control the power-flow be-
electrical or magnetical field values tween the inner energy source and the output
no inner energy source required inner energy source required
signal flow is possible in both directions signal flow is possible in one direction only
only an input of either electrical or mechani- even with missing mechanical power at the in-
cal power generates the corresponding output put and electric output signal exists
power

Examples for passive transducers of electromechanical networks are:

• electrostatic transducer,
• piezoelectric transducer,
• magnetic transducer,
• electrodynamic transducer,
• piezomagnetic transduver.
They are either represented by a transformatory or a gyratory coupling.

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_17, 445

c Springer-Verlag Berlin Heidelberg 2009
446 17 Details about Gyrators and Transformers

     
μL X 0 μK
transformatory coupling = ·
λL 0 1/X λK
     
μL 0 Y μK
gyratory coulping = ·
λL 1/Y 0 λK
In a transformatory network the flow-values μ and differential values λ are cou-
pled to values of the same class. Gyratory networks couple flow- with differential
values and vice versa. The indices L and K mark idle mode resp. short circuit opera-
tion. The transducer-coefficients X (transformer) and Y (gyrator) have to be derived
from the pyhsical interdependencies. Further information about the identification of
transducer-coefficients and a method for the solution of electromechanical networks
can be found in [158].

Example:

The basic model of an electrodynamic transducer (fig. 17.1) is identical to a me-

chanical oscillator. A cylindric disc held by the elasticity n with the mass m and the
coil-length l with the turn-number w is able to move along its axis in an air gap. The
circular air-gap is filled with the magnetic flux density B0 of a permanent magnet.
A current in the coil generates a L ORENTZ-force along the coil’s axial direction.

n ,r m v w ,l

B0 hS h
r b

i
u

Fig. 17.1 Sketch of an electrodynamic actuator [158]

This L ORENTZ-force may be described by the following two equations:

F = B0 · l · i und u = B0 · l · v (17.1)

Consequently the electrodynamic transducer is a transformatory coupling of an

electrical and a mechanical networt, as the flow-values i and F, such as the differ-
ential values v and u are coupled with each other.
The electrodynamic transducer from figure 17.1 is represented by its mechanical
equivalent circuit in figure 17.2:
The transducer coefficients are given by the following equations:
17 Details about Gyrators and Transformers 447

i R Lb FW

u uW
( ) ( )( )
uW
i
=
1/X
0
0
X
v
FW
nI n m 1
r
v

{
nL

Fig. 17.2 Electromechanical transformer visualized in the electromechanical schematic diagram

[158]

1
X= transformatory coupling constant
B0 · l
 
μ 0 · w 2 · 2π r · h 2 hs
Lb = 1− inductivity of the coil, in blocked position
b 4 h
ρl
R= , l = 2π r · w resistance of the coil
A
nL = n elasticity
m = ρ ·VS mass of the coil
1 1
r= , ω2 =
ω0 · Q · nL 0 nL · m
For transformation of the elements to the other side of the electromechanical
network the following transformation laws are given:
• A series circuit stays a series circuit
• A parallel circuit stays a parallel circuit
The transformation from the mechanical to the electrical side gives:
m
Cm =
X2
n
Ln = 2
X
1
Rr =
r · X2
h 1
Z= 2=
X z · X2

The transformation from the electrical to the mechanical side results in:
448 17 Details about Gyrators and Transformers

C
mC =
X2
nI = L · X2
1
rR =
R · X2
h = Z · X2
1
z =
Z · X2
Chapter 18
Impedance Values of Grasps

The following tables provide the parameter for the model given by figure 4.8 and
equation 4.7 in chapter 4. They parametrize the different grasping situations dis-
cussed in section 4.2.4.

Table 18.1: Mean values of the mechanical impedance model according

to figure 4.8 for different grasping siutations.

Grasp/touch k1 m1 d1 d2 k2 k3 d3 m2
[N/m] [kg] [Ns/m] [Ns/m] [N/m] [N/m] [Ns/m] [kg]
Power grasps
Cylinder 412.61 1.577 43.43 33.06 31271 15007 182.77 0.13
Sphere 2500.7 4.32 45.72 31.35 21033 9743 150.60 0.098
Ring 17.71 10.0 0.0032 31.35 5843.7 2906.7 34.54 0.016

Precision grasps
Pen 45◦ 1357.1 1.7376 23.38 3.269 36672 3544.6 12.22 0.029
Pen vertical 44.73 5.44 4.55 17.92 17794 1782.7 12.92 0.029
Pen horizontal 212.49 3.26 7.56 8.15 22092 3672.7 13.73 0.043

Finger
normal 2ṁm 203.21 75.02 1.0854 3.1672 6656.0 478.73 8.3689 0.0114
normal 15ṁm 0.091 37.28 3.79 3.18 9273.5 839.92 12.22 0.018
shear lateral 54.5 10.0 0.323 4.88 12935 191.62 4.4342 0.0178
shear distal 77.56 9.892 9.443 3.003 22874 2004 4.0377 0.0195
shear 45◦ 1053.0 90.44 5.47 7.16 26854 1090.2 15.26 0.006

T.A. Kern (ed.), Engineering Haptic Devices, DOI 10.1007/978-3-540-88248-0_18, 449

c Springer-Verlag Berlin Heidelberg 2009
450 18 Impedance Values of Grasps

Table 18.2: Linear interpolated dependencies of the model’s parameters

from grasp- and touch-forces according to figure 4.8 for different grasp-
ing situations. Interpolation according to c = a · F + b

Grasp/touch k1 m1 d1 d2 k2 k3 d3 m2
[N/m] [kg] [Ns/m] [Ns/m] [N/m] [N/m] [Ns/m] [kg]
Power grasps
Cylinder a = -62.4 -0.216 1.46 -0.409 -365 1330.0 3.27 0.0043
b = 1360 4.88 21.20 39.3 36800 5300.0 133 0.065
Sphere a = -49.0 -0.111 -.0359 -0.788 13.3 109.0 4.94 .00015
b = 3250 6.01 46.3 43.4 20800 8090.0 75.2 0.096
Ring a = 2.26 −2.26 10−16 -.0054 0.143 304.0 150.0 1.72 -.0003
b = -14.0 10.0 0.107 5.47 1590.0 811.0 10.5 0.206

Precision grasps
Pen 45◦ a = -74.3 0.0616 -0.776 0.0247 -134.0 363.0 0.551 0.00372
b = 1840.0 1.34 28.4 3.11 37500 1190.0 8.64 0.00447
Pen vertical a = -21.4 -1.26 1.69 0.428 1460.0 204.0 0.759 0.00465
b = 584.0 13.6 6.46 15.1 82.80 454.0 7.99 -.0087
Pen horizontal a = -27.5 0.56 1.16 -0.619 380.0 229.0 0.409 .00883
b = 391 -0.371 0.193 12.2 19600 2190 11.1 -.0144

Finger
normal 2ṁm a = -124.0 15.2 -0.088 -0.106 -1350 -36.3 1.84 -0.002
b = 606.0 25.5 2.09 3.2 11000 361.0 2.39 0.0180
normal 15ṁm a = 0.187 -5.95 0.861 -0.233 -1940 374.0 1.48 .000675
b = 0.0311 56.6 0.993 3.94 15600 375 4.75 0.0159
shear lateral a = -14.0 −106−10 0.177 0.509 1250.0 63.3 0.363 0.00141
b = 100.0 10.0 -0.558 3.23 8860 14.1 3.25 0.0133
shear distal a = -291.0 0.0571 1.61 -0.711 -8590 367.0 0.266 0.00405
b = 1720 9.71 4.22 5.31 508000 811.0 3.17 0.00636
shear 45◦ a = -54.7 -1.94 0.469 1.65 -299.0 74.5 -.0776 .000295
b = 1230.0 96.8 3.95 1.73 27800 848.0 15.5 .00538
References 451

References

1. X3D, H-Anim, and VRML 97 Specifications.

web|3D: http://www.web3d.org/x3d/specifications/.
2. A DAMS , R.; H ANNAFORD , B. A . Control law design for haptic interfaces to virtual reality.
Control Systems Technology, IEEE Transactions on 10, 1 (Jan. 2002), 3 – 13.
3. A DAMS , R. J. Stable Haptic Interaction with Virtual Environments. PhD thesis, University
of Washington, Washington, 1999.
4. AG, E. R. A. X15G Datasheet. Datenblatt.
5. A LLAM , H. Single element based tactile display. Mechatronics, 2004. ICM ’04. Proceedings
of the IEEE International Conference on (2004), 261 – 267.
6. ATI. Industrial Automation - Multi Axis Force/Torque Sensors - http://www.ati-
ia.com/products/ft/sensors.asp. 4. Juni 2008.
7. BALLAS , R. G. Statisches und dynamisches Verhlante piezoelektischer Multilayer-
Biegeaktoren und Beitrag zur Sensorintegration. PhD thesis, Technische Universität Darm-
stadt, 2006.
8. BALLAS , R. G. Piezoelectric multilayer beam bending actuators: static and dynamic behav-
ior and aspects of sensor integration. Microtechnology and MEMS. Springer, Berlin [u.a.],
2007.
9. BAO , M.-H. Micro mechanical transducers: pressure sensors, accelerometers and gyro-
scopes, 2 ed., vol. 8 of Handbook of sensors and actuators. Elsevier, Amsterdam [u.a.],
2004.
10. BAR -C OHEN , Y. Electroactive polymer (EAP) actuators as artificial muscles - reality, po-
tential, and challenges, vol. 98 of SPIE Press monograph. SPIE Press, Bellingham, Wash,
2001.
11. BAR -C OHEN , Y. Biomimetics - biologically inspired technologies. CRC, Taylor und Francis,
Boca Raton [u.a.], 2006.
12. BAR -C OHEN , Y.; M AVROIDIS , C.; B OUZIT, M.; D OLGIN , B.; H ARM , D. L.; KOPCHOK ,
G. E.; W HITE , R. Virtual reality robotic telesurgery simulations using MEMICA haptic
system. In 8th Annual International Symposium on Smart Structures and Material - EAPAD
(2001).
13. BARAFF , D.; W ITKIN , A.; K ASS , M. Physically based modelling, course notes 36. In ACM
SIGGRAPH ’99 (August 1999).
14. BAUMANN , K.; L ANZ , H. Mensch-Maschine-Schnittstelle elektronischer Geräte. Springer-
Verlag, Berlin, 1998.
15. BAUMANN , R. Haptic Interface for Virtual Reality Based Laparoscopic Surgery Training
Environment. PhD thesis, Département de Microtechnique, Polytechnique Fédérale de Lau-
sanne, 1997.
16. B ELFORD , J. D. The stepped horn - Technical Publication TP-214. Morgan Electro Ceram-
ics, ...
17. B ERKELMAN , P. J.; B UTLER , Z. J.; H OLLIS , R. L. Design of a hemispherical magnetic
levitation haptic interface device. In ASME International Mechanical Engineering Congress
and Exposition (Atlanta, Nov. 1996 1996), ASME, Ed., pp. 483–488.
18. B ERTONYMUS. Wikipedia - Sinn (Wahrnehmung). 19.09.2005 2005.
19. B EYELER , F.; M UNTWYLER , S.; NAGY, Z.; G RAETZEL , C.; M OSER , M.; N ELSON , B.
Design and calibration of a MEMS sensor for measuring the force and torque acting on a
magnetic microrobot. Journal of Micromechanics and Microengineering, IOP 18, 025004
(Dezember 2007).
20. B ICCHI , A. The sense of touch and its rendering - progress in haptics research, vol. Vol. 45
of Springer tracts in advanced robotics. Springer, Berlin ;Heidelberg, 2008.
21. B ICCHI , A.; R AUGI , M.; R IZZO , R.; S GAMBELLURI , N. Analysis and design of an electro-
magnetic system for the characterization of magneto-rheological fluids for haptic interfaces.
Magnetics, IEEE Transactions on 41, 5 (2005), 1876 – 1879.
452 18 Impedance Values of Grasps

22. B IET, M.; G IRAUD , F.; S EMAIL , B. New tactile devices using piezoelectric actuators. In
ACTUATOR 2006, 10th International Conference on New Actuators (2006), pp. 989 – 992.
23. B ÉKÉSY, G. V. Physiologie der Sinneshemmung. Goldmann, München, 1970.
24. B LEICHER , F. Parallelkinematische Werkzeugmaschinen. Neuer Wissenschaftlicher Verlag,
Graz, 2003.
25. B LUME , H.-J.; B OELCKE , R. Mechanokutane Sprachvermittlung, als ms ed., vol. 137 of
Reihe 10. VDI-Verl., Düsseldorf, 1990.
26. B OTSIS , J.; H UMBERT, L.; C OLPO , F.; G IACCARI , P. Embedded fiber Bragg grating sensor
for internal strain measurements in polymeric materials. Optics and Lasers in Engineering
43 (2004).
27. B RAND , U.; B ÜTTGENBACH , S. Taktile dimensionelle Messtechnik für Komponenten der
Mikrosystemtechnik. Technisches Messen 12 (2002).
28. B RAY, A.; BARBATO , G.; L EVI , R. Theory and practice of force measurement. Monographs
in Physical Measurement. Academic Press, London [u.a.], 1990.
29. B RISSAUD , M. Characterization of piezoceramics. Ultrasonics, Ferroelectrics and Fre-
quency Control, IEEE Transactions on 38, 6 (Nov. 1991), 603 – 617.
30. B RONSTEIN , I. N.; S EMENDJAJEW, K. A.; M USIOL , G.; M ÜHLIG , H. Taschenbuch der
Mathematik, 3 ed. Harri Deutsch, Frankfurt am Main, 1997.
31. B ROOKS , T. L. Telerobotic Response Requirements. Proceedings of the IEEE International
Conference on Systems, Man and Cybernetics (1990), 113–120.
32. B ÖSE , H.; M ONKMANN , G. J.; F REIMUTH , H.; T UNAYAR , A.; K HALED , W.; BAUMANN ,
M. A new haptic sensor-actuator system based on electrorheological fluids. In Actuator
2004: 9th International Conference on New Actuators (2004), HVG Hanseatische Veranst.-
GmbH, Div. Messe Bremen, pp. 300 – 303.
33. B ÖSE , H.; T RENDLER , A. Smart fluids - properties and benefit for new electromechanical
devices. In AMAS workshop SMART’03 (2003), pp. 329 – 336.
34. B URDEA , G. Force and touch feedback for virtual reality. A Wiley-Interscience Publication.
Wiley, New York [u.a.], 1996.
35. C ALDWELL , D.; L AWTHER , S.; WARDLE , A. Multi-modal cutaneous tactile feedback. In
Intelligent Robots and Systems ’96, IROS 96, Proceedings of the 1996 IEEE/RSJ Interna-
tional Conference on (Osaka, 4. - 8. November 1996), vol. 2, pp. 465 – 472 vol.2.
36. C ARLSON , J. D.; S TANWAY, R.; J OHNSON , A. R. Electro-rheological and magneto-
rheological fluids: a state of the art report. In Actuator 2004: 9th International Conference
on New Actuators (Bremen, 2004), HVG Hanseatische Veranst.-GmbH, Div. Messe Bremen,
pp. 283 – 288.
37. C HAPUIS , D.; C HAPUIS , D.; M ICHEL , X. A Haptic Knob with a Hybrid Ultrasonic Motor
and Powder Clutch Actuator. In EuroHaptics Conference, 2007 and Symposium on Haptic
Interfaces for Virtual Environment and Teleoperator Systems. World Haptics 2007. Second
Joint (2007), X. Michel R. Gassert, Eds., pp. 200 – 205.
38. C HASE , T.; L UO , R. A thin-film flexible capacitive tactile normal/shear force array sensor.
In Industrial Electronics, Control, and Instrumentation, 1995., Proceedings of the 1995 IEEE
IECON 21st International Conference on (Orlando, FL ISBN: 0-7803-3026-9, 6101995),
vol. 2, pp. 1196 – 1201 vol.2.
39. C OLGATE , J.; B ROWN , J. Factors affecting the Z-Width of a haptic display. Robotics and
Automation, 1994. Proceedings., 1994 IEEE International Conference on (8-13 May 1994),
3205 – 3210 vol.4.
40. C OLGATE , J. E.; G RAFING , P. E.; S TANLEY, M. C.; S CHENKEL , G. Implementation of
stiff virtual walls in force-reflecting interfaces. In Proceedings of the IEEE Virtual Realtiy
Annual International Symposium(VRAIS) (Seattle, September 1993), pp. 202–2008.
41. C ONRAD , H.; F ISHER , M.; S PRECHER , A. F. Characterization of the structure of a model
electrorheological fluid employing stereology. In Proc 2nd Int Conf on ER Fluids (1990).
42. C RANNY, A.; C OTTON , D. P. J. .; C HAPPELL , P. H.; B EEBY, S. P.; W HITE , N. M. Thick-
film force, slip and temperature sensors for a prosthetic hand. Measurement Science and
Technology, IOP 16 (März 2005), 931–941.
References 453

43. DAI , Z.; G IRAUD , F.; L EMAIRE -S EMAIL , B.; M ARTINOT, F. A force feedback device
actuated by piezoelectric travelling wave ultrasonic motors. In ACTUATOR 2006, 10th In-
ternational Conference on New Actuators (2006), pp. 600 – 603.
44. D EAN , M. H.; D OUGLAS , R. D. H. Semiconductor and conventional strain gages. Aca-
demic Press Inc., New York, 1962.
45. D ENG , K.; E NIKOV, E.; Z HANG , H. Development of a pulsed electromagnetic micro-
actuator for 3D tactile displays. Advanced intelligent mechatronics, 2007 ieee/asme interna-
tional conference on (2007), 1 – 5.
46. D OERRER , C. Entwurf eines elektromechanischen Systems für flexibel konfigurierbare
Eingabefelder mit haptischer Rückmeldung, vol. 3 of Mikro- und Sensortechnik. Inst. für
Elektromechan. Konstruktionen, Darmstadt, 2003.
47. E ID , M.; M ANSOUR , M.; I GLESIAS , R.; E L S ADDIK , A. A Device Independent Haptic
Player. In Virtual Environments, Human-Computer Interfaces and Measurement Systems,
2007. VECIMS 2007. IEEE Symposium on (Ostuni ISBN: 978-1-4244-0820-7, 2007), pp. 7
– 12.
48. E ISNER , E. Complete Solutions of the ’Webster’ Horn Equation. The Journal of the Acous-
tical Societey of America 41, 4 (2) (1966), 1126 – 1146.
49. E L WAHED , A. K.; S PROSTON , J. L.; S TANWAY, R.; W ILLIAMS , E. W. An improved
model of ER fluids in squeeze-flow through model updating of the estimated yield stress.
Journal of Sound and Vibration 268, 3 (2003), 581 – 599.
50. E RICSON , C. Real-Time Collision Detection (The Morgan Kaufmann Series in Interactive
3-D Technology) (The Morgan Kaufmann Series in Interactive 3D Technology). Morgan
Kaufmann, December 2004.
51. E RP, J. V.; JANSEN , C.; D OBBINS , T.; V EEN , H. V. Vibrotactile Waypoint Navigation at
Sea and in the Air: two Case Studies. 166–173.
52. FALLER , A.; [B EARB .], M. S. Der Körper des Menschen: Einführung in Bau und Funktion,
13 ed. Thieme, Stuttgart [u.a.], 1999.
53. F ERDINAND , P.; M AGNE , S.; D EWYNTER -M ARTY, V.; M ARTINEZ , C.; OUGEAULT, S.;
B UGAUD , M. Applications of Bragg grating sensors in Europe. In International Conference
on Optical Fiber Sensors OFS (Wiliamsburg, Virginia, 28. - 31. Oktober 1997).
54. F ERNANDEZ , J. M.; P ERRIARD , Y. Comparative analysis and modeling of both standing
and travelling wave ultrasonic linear motor. IEEE Ultrasonic Symposium (2003), 1770 –
1773.
55. F LEISCHER , M.; S TEIN , D.; M EIXNER , H. Ultrasonic piezomotor with longitudinally os-
cillating amplitude-transforming resonator. Ultrasonics, Ferroelectrics and Frequency Con-
trol, IEEE Transactions on 36, 6 (Nov. 1989), 607 – 613.
56. F ÖLLINGER , O. Nichtlineare Regelungen. 2. Anwendung der Zustandsebene. R. Oldenbourg
Verlag GmbH, 1970.
57. F ÖLLINGER , O. Nichtlineare Regelungen. 3. Ljapunow-Theorie und Popow-Kriterium. R.
Oldenbourg Verlag GmbH, 1970.
58. F ÖLLINGER , O. Nichtlineare Regelungen 1. Grundlagen und Harmonische Balance. R.
Oldenbourg Verlag GmbH, 1991.
59. F LUECKIGER , M.; F LUECKIGER , M.; B ULLO , M. fMRI compatible haptic interface ac-
tuated with traveling wave ultrasonic motor. In Industry Applications Conference, 2005.
Fourtieth IAS Annual Meeting. Conference Record of the 2005 (2005), M. Bullo D. Chapuis,
Eds., vol. 3, pp. 2075 – 2082 Vol. 3.
60. F RANKLIN , G.; P OWELL , J.; E MAMI -NAEINI , A. Feedback Control of Dynamic Systems.
Pearson, 2002.
61. F U , C.-Y.; O LIVER , M. Direct measurement of index finger mechanical impedance at low
force. In Eurohaptics Conference, 2005 and Symposium on Haptic Interfaces for Virtual En-
vironment and Teleoperator Systems, 2005. World Haptics 2005. First Joint (2005), pp. 657
– 659.
62. F UNG , R.-F.; YAO , C.-M.; C HANG , D.-G. Dynamic and contact analysis of a bimodal
ultrasonic motor. Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on
46, 1 (Jan. 1999), 47 – 60.
454 18 Impedance Values of Grasps

63. F UNG , Y.- C . Biomechanics: mechanical properties of living tissues, 2 ed. Springer, New
York [u.a.], 1993.
64. G EHIN , C., B. C.; T EISSYRE , Y. Design and characterization of new force resonant sensor.
Sensor and Actuators A Physical 84 (2000), 65–69.
65. G ERLING , G.; T HOMAS , G. The effect of fingertip microstructures on tactile edge percep-
tion. In Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2005. WHC
2005. First Joint Eurohaptics Conference and Symposium on (2005), pp. 63 – 72.
66. G ESCHEIDER , G. A. Psychophysics: the fundamentals, 3 ed. Erlbaum, Mahwah, NJ [u.a.],
1997.
67. G ESCHEIDER , G. A.; B OLANOWSKI , S. J.; H ARDICK , K. R. The frequency selectivity of
information-processing channels in the tactile sensory system. Somatosens Mot Res 18, 3
(2001), 191–201. Hamilton College, Department of Psychology, Clinton, NY 13323, USA.
[email protected].
68. G ESCHEIDER , G. A.; B OLANOWSKI , S. J. J.; V ERRILLO , R. T.; A RPAJIAN , D. J.; RYAN ,
T. F. Vibrotactile intensity discrimination measured by three methods. J Acoust Soc Am 87,
1 (1990), 330–8. Hamilton College, Clinton, New York 13323.
69. G ESCHEIDER , G. A.; S KLAR , B. F.; VAN D OREN , C. L.; V ERRILLO , R. T. Vibrotactile
forward masking: psychophysical evidence for a triplex theory of cutaneous mechanorecep-
tion. J Acoust Soc Am 78, 2 (1985), 534–43.
70. G ETZINGER , G. Haptik - Rekonstruktion eines Verlustes. Profil, Wien, 2006.
71. G HODDSI , R.; B EEBE , D.; W HITE , V.; D ENTON , D. Development of a tangential tac-
tor using a LIGA/MEMS linear microactuator technology Microelectromechanical System
(MEMS). In Microelectromechanical System (MEMS), DSC-Vol. 59 (Atlanta, 1996), pp. 379
– 386.
72. G IRAUD , F.; S EMAIL , B. Analysis and phase control of a piezoelectric traveling-wave
ultrasonic motor for haptic stick application. Industry Applications, IEEE Transactions on
40, 6 (2004), 1541 – 1549.
73. G OETHALS , P.; S ETTE , M. M.; R EYNAERTS , D.; B RUSSEL , E. V. Flexible Elastore-
sistive Tactile Sensor for Minimally Invasive Surgery - Eurohaptics Conference. Springer,
Berlin/Heidelberg, 2008, pp. 573–579.
74. G OLDSTEIN , E. B. Sensation and Perception, 7 ed. Wadsworth Publishing Co., Belmont,
California, January 2006.
75. G ONG , X.; W U , J.; H UANG , X.; W EN , W.; S HENG , P. Influence of liquid phase on
nanoparticle-based giant electrorheological fluid. Nanotechnology 19, 16 (2008), 165602
(7pp).
76. G OSLINE , A. H.; C AMPION , G.; H AYWARD , V. On The Use of Eddy Current Brakes as
Tunable, Fast Turn-On Viscous Dampers for Haptic Rendering. In Eurohaptics Conference
(Paris, 2006), pp. 229–234.
77. G RIMALDI , G.; L AMMERTSE , P.; VAN D EN B RABER , N.; M EULEMAN , J OS , M. M. A
New Myohaptic Device to Asses Wrist Function in the Lab and in the Clinic - The Wrist-
alyzer - Eurohaptics Conference. In Haptics: Perception, Devices and Scenarios, M. Ferre,
Ed. Springer, Berlin/Heidelberg, 2008.
78. G RUNWALD , M. Der bewegte Sinn: Grundlagen und Anwendungen zur haptischen
Wahrnehmung. Birkhäuser, Berlin, 2001.
79. G RUNWALD , M. Human haptic perception: basics and applications. Birkhäuser, Berlin,
2008.
80. H AGEDORN , P.; S ATTEL , T.; S PEZIARI , D.; S CHMIDT, J.; D IANA , G. The importance
of rotor flexibility in ultrasonic traveling wave motors. Smart Materials and Structures 7
(1998), 352 – 368.
81. H AGOOD , N. W.; M C FARLAND , A. J. Modeling of a piezoelectric rotary ultrasonic mo-
tor. Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on 42, 2 (March
1995), 210 – 224.
82. H AJIAN , A. Z.; H OWE , R. D. Identification of the Mechanical Impedance at the Human
Finger Tip. Journal of Biomechanical Engineering 119 (Feb. 1997 1997), 109–114.
References 455

83. H ALE , K.; S TANNEY, K. Deriving haptic design guidelines from human physiological,
psychophysical, and neurological foundations. Computer Graphics and Applications, IEEE
24, 2 (March-April 2004), 33 – 39.
84. H ANNAFORD , B.; A NDERSON , R. Experimental and simulation studies of hard contact
in force reflecting teleoperation. Robotics and Automation, 1988. Proceedings., 1988 IEEE
International Conference on (24-29 April 1988), 584 – 589 vol.1.
85. H ANNAFORD , B.; RYU , J.-H. Time-domain passivity control of haptic interfaces. Robotics
and Automation, IEEE Transactions on 18, 1 (Feb. 2002), 1 – 10.
86. H ASEGAWA , Y.; S HIKIDA , M.; ET AL . An active tactile sensor for detecting mechanical
characteristics of contacted objects. Journal of Micromechanics and Microengineering, IOP
16 (2006), 1625–1632.
87. H ASSER , C. J. Force-Reflecting Anthropomorphic Hand Masters. Report AL/CF-TR-1995-
0110, Armstrong Laboratory, Wright-Patterson Air Force Base, Ohio (1995).
88. H ASSER , C. J.; DANIELS , M. W. Tactile Feedback with Adaptive Controller for a Force-
Reflecting Haptic Display - Part I: Design. In Proceedings of the Fifteenth Southern Biomed-
ical Engineering Conference (1996), pp. 526–529.
89. H AYWARD , V. A Brief Taxonomy of Tactile Illusions and Demonstrations That Can Be
Done In a Hardware Store. Brain Research Bulletin 756, 6 (2008), 742–752.
90. H AYWARD , V.; A STLEY, O. R. Performance Measures for Haptic Interfaces. Robotics
Research: The 7th Int. Symposium 1 (1996), 195–207.
91. H AYWARD , V.; C RUZ -H ERNANDEZ , M. Tactile display device using distributed lateral
skin stretch. In Symposium on Haptic Interfaces for Virtual Environment and Teleoperator
Systems, IMECE 2000 Conference (2000), J. Wikander, Ed.
92. H E , S.; C HEN , W.; TAO , X.; C HEN , Z. Standing wave bi-directional linearly moving ultra-
sonic motor. Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on 45,
5 (Sept. 1998), 1133 – 1139.
93. H ELIN , P.; S ADAUNE , V.; D RUON , C.; T RITSCH , J.-B. Linear ultrasonic motors using
surface acoustic waves mechanical model for energy transfer, vol. 2 of Solid State Sen-
sors and Actuators, 1997. TRANSDUCERS ’97 Chicago., 1997 International Conference on.
Chicago, IL, 16-19 June 1997.
94. H EYDT, R.; C HHOKAR , S. Refreshable Braille display based on electroactive polymers. In
23rd International Display Research Conference (2003), pp. 111 – 114.
95. H IRATA , H.; U EHA , S. Design of a traveling wave type ultrasonic motor. Ultrasonics,
Ferroelectrics and Frequency Control, IEEE Transactions on 42, 2 (March 1995), 225 – 231.
96. H OPP, T. H.; R EEVE , C. P. “an algorithm for computing the minimum covering sphere in
any dimension”.
97. H OPPACH , E.; S CHLAAK , H. F.; W EISSMANTEL , H. Elektrische Kleinantriebe - Skriptum.
Tech. rep., Darmstadt, 2006.
98. H OU , L. Erfassung und Kompensation von Fehlereffekten bei der statischen Kraftmessung
mit monolithischen Nd:YAG-Laserkristallen. PhD thesis, Universität Kassel, 1999.
99. H U , M.; X U , Z.; J IN , L.; WANG , X. Performance Simulation of Traveling Wave Type Ul-
trasonic Motor, vol. 3 of Electrical Machines and Systems, 2005. ICEMS 2005. Proceedings
of the Eighth International Conference on. 27-29 Sept. 2005.
100. H UANG , X.; S HENG , P.; W EN , W.; YANG , S. Formation of polarized contact layers and
the giant electrorheological effect. International Journal of Modern Physics B (IJMPB) 21,
28/29 (2007), 4907–4913.
101. H UGONY. Über die Empfindung von Schwingungen mittels des Tastsinnes. Zeitschrift Bi-
ologie 96 (1935), 548–553.
102. H USTY, M.; K ARGER , A.; S ACHS , H.; S TEINHILPER , W. Kinematik und Robotik.
Springer, Tokio, 1997.
103. H USTY, M. L. An algorithm for solving the direct kinematics of general stewart-gough
platforms. Mechanism and Machine Theory 31, 4 (1996), 365–379.
104. IEEE. Standard on Piezoelectrity. 1988.
105. I KEDA , T. Fundamentals of piezoelectricity. Oxford Univ. Press, Oxford [u.a.], 1990.
456 18 Impedance Values of Grasps

106. I KEI , Y.; I KEI , Y.; S HIRATORI , M. TextureExplorer: a tactile and force display for vir-
tual textures. In Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2002.
HAPTICS 2002. Proceedings. 10th Symposium on (2002), M. Shiratori, Ed., pp. 327 – 334.
107. I KEI , Y.; YAMADA , M.; F UKUDA , S. Tactile Texture Presentation by Vibratory Pin Arrays
Based on Surface Height Maps. In ASME Dynamic Systems and Control Division, N. Olgac,
Ed., vol. 67. American Society of Mechanical Engineers, New York, NY, 1999, pp. 97 –102.
108. I NAUDI , D. SOFO Sensors for Static and Dynamic Measurements. In 1st FIG Interna-
tional Symposium on Engineering Surveys for Construction Works and Structural Engineer-
ing (Nottingham, 28. Juni -1. Juli 2004).
109. I SHII , H. Bottles: A Transparent Interface as a Tribute to Mark Weiser. IEICE: Transactions
on Infomration and Systems E87-D, 6 (June 2004 2004), 1299–1311299–1311.
110. I SRAR , A.; C HOI , S.; TAN , H. Z. Detection Threshold and Mechanical Impedance of the
Hand in a Pen-Hold Posture. In Intelligent Robots and Systems, 2006 IEEE/RSJ International
Conference on (Beijing, 2006), pp. 472 – 477.
111. I SRAR , A.; C HOI , S.; TAN , H. Z. Mechanical Impedance of the Hand Holding a Spheri-
cal Tool at Threshold and Suprathreshold Stimulation Levels. In EuroHaptics Conference,
2007 and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Sys-
tems. World Haptics 2007. Second Joint (Tsukaba ISBN: 0-7695-2738-8, 2007), pp. 56 –
60.
112. I WAMOTO , T.; S HINODA , H. Ultrasound tactile display for stress field reproduction - ex-
amination of non-vibratory tactile apparent movement. In Haptic Interfaces for Virtual En-
vironment and Teleoperator Systems, 2005. WHC 2005. First Joint Eurohaptics Conference
and Symposium on (2005), pp. 220 – 228.
113. I WAMOTO , T.; S HINODA , H. Two-dimensional Scanning Tactile Display using Ultrasound
Radiation Pressure. In Haptic Interfaces for Virtual Environment and Teleoperator Systems,
2006 14th Symposium on (2006), pp. 57 – 61 ISBN: 1–4244–0226–3.
114. JANDURA , L.; S RINIVASAN , M. A. Experiments on Human Performance in Torque Dis-
crimination and Control. Proceedings of the ASME Winter Annual Meeting, Dynamic Sys-
tems and Control, DSC-Vol. 55-1 (1994), 369–375.
115. JANSEN , C.; W ENNEMERS , A.; VOS , W.; G ROEN , E. FlyTact: A Tactile Display Im-
proves a Helicopter Pilot’s Landing Performance in Degraded Visual Environments - Euro-
haptics Conference. In Haptics: Perception, Devices and Scenarios, M. Ferre, Ed. Springer,
Berlin/Heidelberg, 2008, pp. 867–875.
116. J ENDRITZA , D. J. Technischer Einsatz neuer Aktoren: Grundlagen, Werkstoffe, De-
signregeln und Anwendungsbeispiele, 2 ed., vol. 484 of Kontakt Studium. expert-Verl.,
Renningen-Malmsheim, 1998.
117. J OLLY, M. R.; C ARLSON , J. D. Controllable squeeze film damping using magnetorheolog-
ical fluid. In Actuator 96, 5th Int. Conf. on New Actuators (Bremen, 1996).
118. J ONES , L. Using the Haptic System as a Means of Communication, a Collection of Infor-
mation - http://bioinstrumentation.mit.edu/jones/pages/Actual/mealani. 2003.
119. J UNGMANN , M. Entwicklung elektrostatischer Festkörperaktoren mit elastischen Dielek-
trika für den Einsatz in taktilen Anzeigefeldern. PhD thesis, TU Darmstadt, 2004.
120. J UNGMANN , M.; S CHLAAK , H. F. Taktiles Display mit elektrostatischen Polymeraktoren.
Konferenzband des 47. Internationalen Wissenschaftlichen Kolloquiums, Technische Univer-
sität Ilmenau (2002).
121. J UNGNICKEL , U. Miniaturisierte Positioniersysteme mit mehreren Freiheitsgraden auf der
Basis monolithischer Strukturen. PhD thesis, Darmstadt, 2004.
122. K ACZMAREK , K.; W EBSTER , J.; BACH - Y R ITA , P.; T OMPKINS , W. Electrotactile and
vibrotactile displays for sensory substitution systems. Biomedical Engineering, IEEE Trans-
actions on 38, 1 (1991), 1 – 16.
123. K ACZMAREK , K. A.; BACH -Y-R ITA , P. Tactile Displays (Chapter 9). In Virtual Environe-
ments and Advanced Interface Design, W. Barfield T. Furness, Eds. Oxford University Press,
New-York, 1995, pp. 349–414.
References 457

124. K AJIMOTO , H.; K ANNO , Y.; TACHI , S. Forehead Electro-tactile Display for Vision Substi-
tution. In Eurohaptics Conference (Paris, 2006).
125. K ALLENBACH , E.; E ICK , R.; Q UENDT, P.; S TRÖHLA , T.; F EINDT, K.; K ALLENBACH ,
M. Elektromagnete - Grundlagen, Berechnung, Entwurf und Anwendung, 3., bearb. u. erg.
aufl ed. Vieweg + Teubner, Wiesbaden, 2008.
126. K ANT, I. Anthropologie in pragmatischer Hinsicht. Reclam, Stuttgart, 1983.
127. K ENALEY, G. L.; C UTKOSKY, M. R. Electrorheological fluid-based robotic fingers with
tactile sensing. In Robotics and Automation, 1989. Proceedings., 1989 IEEE International
Conference on (1989), pp. 132 – 136 vol.1.
128. K ERDOK , A. E. Characterizing the Nonlinear Mechanical Response of Liver to Surgical
Manipulation. PhD thesis, Harvard University, Cambridge, May 2006.
129. K ERN , T. A. Haptisches Assistenzsystem für diagnostische und therapeutische Katheter-
isierungen: “Mit dem Blindenstock durch Adern tasten!”. PhD thesis, 2006.
130. K ERN , T. A.; K LAGES , S.; M EISS , T.; W ERTHSCHÜTZKY, R. Study of the Influence of
Varying Diameter and Grasp-Forces on Mechanical Impedance for the Grasp of Cylindrical
Objects. Eurohaptics Conference 1 (3.- 6. Juni 2006 2006), 113–118.
131. K ERN , T. A.; RÖRUP, H.; W ERTHSCHUTZKY, R.; TAMMER , R. A remotely controlled
lightweight MRI compatible ultrasonic actuator for micrometer positioning of electrodes
during neuroethological primate research / Telemetrisch gesteuerter, MRT-kompatibler Ul-
traschallaktor zur Mikrometerpositionierung von Elektrode. Biomed Tech (Berl) 53, 6 (2008),
292–9. 1Institute for Electromechanical Design, Darmstadt University of Technology, Darm-
stadt, Germany.
132. K ERN , T. A.; S CHAEFFER , A.; W ERTHSCHÜTZKY, R. An Interaction Model for the Quan-
tification of Haptic Impressions. In Haptics: Perception, Devices and Scenarios, M. Ferre,
Ed. Sppringer, Berlin Heidelberg, 2008, pp. 139–145.
133. K ERSEY, A. D.; DAVIS , M. A.; PATRICK , H. J., E . A . Fiber Grating Sensors. Journal of
Lightwave Technology 15, 8 (1997).
134. KG, P. I. P. G. . C. Piezo linear driving mechanism for converting electrical energy into
motion, has a group of stacking actuators for driving a rotor in a guiding mechanism, 2003.
135. K HALIL , H. K. Nonlinear Systems. Prentice Hall, 2002.
136. K IKUUWE , R.; S ANO , A.; M OCHIYAMA , H.; TAKESUE , N.; T SUNEKAWA , K.; S UZUKI ,
S.; F UJIMOTO , H. The tactile contact lens. In Sensors, 2004. Proceedings of IEEE (2004),
pp. 535 – 538 vol.2.
137. K IM , K. J.; TADOKORO , S. Electroactive Polymers for Robotic Applications - Artificial
Muscles and Sensors. Springer eBook Collection, Engineering [Dig. Serial], Springer-11647
[Dig. Serial]. Springer-Verlag London Limited, London, 2007.
138. K IRCHNER , J. Mehrkriterielle optimierung von parallelkinematiken. Verlag Wis-
senschaftliche Scripten, Zwickau (2001).
139. K LAGES , S. Neue Sensorkonzepte zur Zungendruckmessung. Diplomarbeit, Technische
Universität Darmstadt, Darmstadt, August 2004.
140. K LOSE , S. Abschätzung der Miniaturisierbarkeit eines Ultraschallaktors. PhD thesis, Tech-
nische Universität Darmstadt, 2006.
141. KONTARINIS , D. A.; H OWE , R. D. Tactile Display of Contact Shape in Dextrous Telema-
nipulation. ASME Winter Annual Meeting: Advances in Robotics, Mechatronics and Haptic
Interfaces, New Orleans, DSC-Vol. 49 (1993), 81–88.
142. KONYO , M.; TADOKORO , S.; YOSHIDA , A.; S AIWAKI , N. A tactile synthesis method
using multiple frequency vibrations for representing virtual touch. In Intelligent Robots and
Systems, 2005. (IROS 2005). 2005 IEEE/RSJ International Conference on (2005), pp. 3965
– 3971.
143. KORNBLUH , R.; P ELRINE , R. Electrostrictive polymer artificial muscle actuators. In
Robotics and Automation, 1998. Proceedings. 1998 IEEE International Conference on
(1998), vol. 3, pp. 2147 – 2154 vol.3.
144. KORNBLUH , R.; P ELRINE , R.; J OSEPH , J.; H EYDT, R.; P EI , Q.; C HIBA , S. High-Field
electrostriction of elastomeric polymer dielectrics for actuation. In 6th Annual International
Symposium on Smart Structures and Material - EAPAD (1999), pp. 149 – 161.
458 18 Impedance Values of Grasps

145. KOYAMA , T.; KOYAMA , T.; TAKEMURA , K. Development of an ultrasonic clutch for multi-
fingered exoskeleton haptic device using passive force feedback for dexterous teleoperation.
In Intelligent Robots and Systems, 2003. (IROS 2003). Proceedings. 2003 IEEE/RSJ Inter-
national Conference on (2003), K. Takemura T. Maeno, Eds., vol. 3, pp. 2229 – 2234 vol.3.
146. K UCHENBECKER , K.; F IENE , J.; N IEMEYER , G. Event-based haptics and acceleration
matching: portraying and assessing the realism of contact. 18-20 March 2005.
147. K UNSTMANN , C. Handhabungssystem mit optimierter Mensch-Maschine-Schnittstelle für
die Mikromontage, als ms ed., vol. 751 of Reihe 8, Mess-, Steuerungs- und Regelungstechnik.
VDI-Verl., Düsseldorf, 1999.
148. K USCHEL , M.; K REMER , P.; H IRCHE , S.; B USS , M. Lossy data reduction methods for
haptic telepresence systems. Robotics and Automation, 2006. ICRA 2006. Proceedings 2006
IEEE International Conference on (2006), 2933 – 2938.
149. K YUNG , K.-U.; A HN , M.; K WON , D.-S.; S RINIVASAN , M. A compact broadband tactile
display and its effectiveness in the display of tactile form. In Haptic Interfaces for Virtual En-
vironment and Teleoperator Systems, 2005. WHC 2005. First Joint Eurohaptics Conference
and Symposium on (2005), pp. 600 – 601.
150. K YUNG , K.-U.; K YUNG , K.-U.; L EE , J. Y. Haptic Stylus and Empirical Studies on Braille,
Button, and Texture Display. Journal of Biomedicine and Biotechnology vol. 2008, Article
ID 369651 (2007), 11.
151. K YUNG , K.-U.; K YUNG , K.-U.; L EE , J. Y. Pen-like Haptic Interface and Its Application
on Touch Screen. In Robot and Human interactive Communication, 2007. RO-MAN 2007.
The 16th IEEE International Symposium on (2007), J. Y. Lee J.-S. Park, Eds., pp. 9 – 13.
152. K YUNG , K.-U.; K YUNG , K.-U.; PARK , J.-S. Ubi-Pen: Development of a Compact Tactile
Display Module and Its Application to a Haptic Stylus. In EuroHaptics Conference, 2007 and
Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems. World
Haptics 2007. Second Joint (2007), J.-S. Park, Ed., pp. 109 – 114.
153. L ANDAU , K. Die Arbeit im Dienstleistungsbetrieb: Grundzüge einer Arbeitswissenschaft
der personenbezogenen Dienstleistung. Ulmer, Stuttgart, 1996.
154. L ANGETEPE , E.; Z ACHMANN , G. Geometric Data Structures for Computer Graphics, 1 ed.
AK Peters, 2005.
155. L ASER C OMPONENTS. Laser Components Group - Faseroptische Sensoren -
http://www.lasercomponents.com/de/1134.html. 4. Juni 2008.
156. L AWRENCE , D. Stability and transparency in bilateral teleoperation. Robotics and Automa-
tion, IEEE Transactions on 9, 5 (Oct. 1993), 624 – 637.
157. L AWRENCE , D.; PAO , L.; A PHANUPHONG , S. Bow spring/tendon actuation for low cost
haptic interfaces. Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2005.
WHC 2005. First Joint Eurohaptics Conference and Symposium on (18-20 March 2005), 157
– 166.
158. L ENK , A.; P FEIFER , G.; W ERTHSCHÜTZKY, R. Elektromechanische Systeme: mechanische
und akustische Netzwerke, deren Wechselwirkungen und Anwendungen. Springer, Berlin
[u.a.], 2001.
159. L EONDES , C. T. MEMS/NEMS - handbook techniques and applications. Springer, New
York, NY, 2006.
160. L EUSCHKE , R.; K URIHARA , E.; D OSHER , J.; H ANNAFORD , B. High fidelity multi fin-
ger haptic display. Eurohaptics Conference, 2005 and Symposium on Haptic Interfaces
for Virtual Environment and Teleoperator Systems, 2005. World Haptics 2005. First Joint
(18202005), 606 – 608.
161. L EVESQUE , V.; L EVESQUE , V.; PASQUERO , J. Braille Display by Lateral Skin Deformation
with the STReSS2 Tactile Transducer. In EuroHaptics Conference, 2007 and Symposium on
Haptic Interfaces for Virtual Environment and Teleoperator Systems. World Haptics 2007.
Second Joint (2007), J. Pasquero V. Hayward, Eds., pp. 115 – 120.
162. L I , W. H.; D U , H.; G UO , N. Q.; KOSASIH , P. B. Magnetorheological fluids based haptic
device. Sensor Review 24, 1 (2004), 68 – 73.
References 459

163. L IN , M. C.; C ANNY, J. F. Efficient algorithms for incremental distance computation. In

IEEE Conference on Robotics and Automation (1991), pp. 1008–1014.
164. L UNZE , J. Regelungstechnik 2. Springer, Berlin, 2005.
165. L UNZE , J. Regelungstechnik 1. Springer, Berlin, 2006.
166. L UO , F.; L IU , J.; M A , N.; M ORSE , T. A fiber optic microbend sensor for distributed sensing
application in the structural strain monitoring. Sensors and Actuators A Physical 75 (1999),
41–44.
167. L UTZ , H.; W ENDT, W. Taschenbuch der Regelungstechnik, 5 ed. Deutsch, Frankfurt am
Main, 2003.
168. M AGNUSSEN , B JÖRN ; H AGEMANN , B. S. D. D. E. W. K. V. P. Piezomotor and methods
for the production and exitement thereof, 30. Okt.. 2003.
169. M C N EELY, W. A.; P UTERBAUGH , K. D.; T ROY, J. J. Six degree-of-freedom haptic ren-
dering using voxel sampling. In SIGGRAPH (1999), pp. 401–408.
170. M EHLING , J.; C OLGATE , J.; P ESHKIN , M. Increasing the impedance range of a haptic
display by adding electrical damping. In Eurohaptics Conference, 2005 and Symposium on
Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2005. World Haptics
2005. First Joint (2005), pp. 257 – 262.
171. M EISS , T.; K ERN , T. A.; K LAGES , S.; W ERTHSCHUETZKY, R. Fertigung eines
Miniaturkraftsensors mit asymmetrischem Grundkörper zur Anwendung bei Katheterisierun-
gen. In MikroSystemTechnik Kongress (Dresden, 2007), VDI-VDE, Ed.
172. M EISSNER , P. Seminar zu speziellen Themen der optischen Nachrichtentechnik - Simula-
torische und experimentelle Untersuchungen optischer WDM Übertragungssysteme. Skrip-
tum, Technische Universität Darmstadt, Juni 2006.
173. M EISSNER , P. Optische Nachrichtentechnik I - Skriptum zur Vorlesung. Skriptum, Technis-
che Universität Darmstadt, 2007.
174. M ERLET, J. P. Parallel Robots, 2. ed ed. Solid mechanics and its applications. Springer,
Dordrecht, 2006.
175. M ESCHEDE , D. Optics, light and lasers: the practical approach to modern aspects of pho-
tonics and laser physics, 2 ed. Optik, Licht und Laser engl. Wiley-VCH-Verl., Weinheim,
2007.
176. M ILLER , B.; C OLGATE , J.; F REEMAN , R. Environment delay in haptic systems. Robotics
and Automation, 2000. Proceedings. ICRA ’00. IEEE International Conference on 3 (24-28
April 2000), 2434 – 2439 vol.3.
177. M ILNER , T. E.; F RANKLIN , D. W. Characterization of Multijoint Finger Stiffness: Depen-
dence on Finger Posture and Force Direction. IEEE Transactions on Biomedical Engineer-
ing, Vol. 43, Nr. 11 (1998), 1363–1375.
178. M ITRA , P.; G ENTRY, D.; N IEMEYER , G. User Perception and Preference in Model Medi-
ated Telemanipulation. In EuroHaptics Conference, 2007 and Symposium on Haptic Inter-
faces for Virtual Environment and Teleoperator Systems. World Haptics 2007. Second Joint
(Tsukaba, 2007), pp. 268 – 273.
179. M ÜLLER , G.; G ROTH , C. FEM für Praktiker - Grundlagen: Basiswissen und Arbeits-
beispiele zur Finite-Element-Methode mit dem FE-Programm ANSYS Rev. 5.5 : mit CD-
ROM, Installationsanleitung und zahlreichen Beispielen von: FEM für Praktiker, 7 ed.,
vol. 23 of Edition expertsoft. expert-Verl., Renningen-Malmsheim, 2002.
180. M OAL , P. L.; J OSEPH , E.; F ERNIOT, J.-C. Mechanical energy transductions in standing
wave ultrasonic motors: Analytical modelling and experimental investigations. European
Journal of Mechanics - A/Solids 19, 5 (2000), 849–871.
181. M URAYAMA , J.; B OURGILA , L.; L UO , Y.; A KAHANE , K.; H ASEGAWA , S.; H IRSBRUN -
NER , B.; S ATO , M. SPIDAR G&G: A Two-Handed Haptic Interface for Bimanual VR
Interaction. In Eurohaptics, vol. 1. Universität München, München, 2004, pp. 138–146.
182. NAKASHIGE , M.; H IROTA , K.; H IROSE , M. Linear actuator for high-resolution tactile
display. Robot and Human Interactive Communication, 2004. ROMAN 2004. 13th IEEE
International Workshop on (2004), 587 – 590.
460 18 Impedance Values of Grasps

183. NAKATANI , M.; H OWE , R. D.; TACHI , S. The Fishbone Tactile Illusion. In Eurohaptics
Conference (Paris, 3.- 6. Juni 2006 2006), Academie de Technologies.
184. NAKATANI , M.; H OWE , R. D.; TACJI , S. The Fishbone Tactile Illusion. In Eurohaptics
(Paris, 2006), IEEE, Ed.
185. N EUGEBAUER , R. Parallelkinematische Maschinen: Entwurf, Konstruktion, Anwendung.
VDI-Buch. Springer, Berlin [u.a.], 2006.
186. N ÉRET, G. Erotik in der Kunst des 20. Jahrhunderts. Taschen, Tokyo, 1998.
187. N ÜHRMANN , D. Das Groß e Werkbuch Elektronik, 7. ed. Franzis’ Verlag, Poing, 1998.
188. O GUZTORELI , M. N.; S TEIN , R. B. Optimal task performance of antagonistic muscles. Biol
Cybern 64, 2 (1990), 87–94. Department of Mathematics, University of Alberta, Edmonton,
Canada.
189. PAHL , G.; B EITZ , W.; F ELDHUSEN , J. Konstruktionslehre: Grundlagen erfolgreicher Pro-
duktentwicklung; Methoden und Anwendung, 7 ed. Springer, Berlin [u.a.], 2007.
190. PALAIS , J. C. Fiber optic communications, 5 ed. Pearson Prentice-Hall, Upper Saddle River,
NJ, 2005.
191. PANDEY, N.; YADAV, B. C. Embedded fibre optic microbend sensor for measurement of
high pressure and crack detection. Sensors and Actuators A Physical 128 (2006), 33–36.
192. PANG , X. D.; TAN , H. Z.; D URLACH , N. I. Manual Discrimination of Force Using Active
Finger Motion. Perception Psychophysics, Vol. 49, Nr. 6 (1991), 531–540.
193. PARE , M.; E LDE , R.; M AZURKIEWICZ , J. E.; S MITH , A. M.; R ICE , F. L. The Meiss-
ner corpuscle revised: a multiafferented mechanoreceptor with nociceptor immunochemical
properties. J Neurosci 21, 18 (2001), 7236–46. Departement de Physiologie, Universite de
Montreal, Montreal, Quebec, Canada H3C 3J7.
194. PARTHASARATHY, M.; K LINGENBERG , D. J. Electrorheology: Mechanisms and models.
Materials Science and Engineering: R: Reports 17, 2 (1996), 57 – 103.
195. PARTSCH , U. LTCC-kompatible Sensorschichten und deren Applikation in LTCC-
Drucksensoren, 1 ed., vol. 9 of Elektronik-Technologie in Forschung und Praxis. Detert,
Templin/Uckermark, 2002.
196. P EI , Q.; ROSENTHAL , M. A.; P ELRINE , R.; S TANFORD , S.; KORNBLUH , R. Multifunc-
tional electroelastomer roll actuators and their application for biomimetic walking robots. In
Proceedings of SPIE, 5051 [31] (2003), pp. 281 – 290.
197. P EIRS , J.; C LIJNEN , J.; R EYNAERTS , D.; VAN B RUSSEL , H.; H ERIJGERS , P.;
C ORTEVILLE , B.; B OONE , S. A micro optical force sensor for force feedback during mini-
mally invasive robotic surgery. Sensors and Actuators A: Physical 115, 2-3 (2004), 447–455.
198. P ELRINE , R.; KORNBLUH , R. Electromechanical transduction effects in dielectric elas-
tomers: actuation, sensing, stiffness modulation and electric energy generation. In Dielec-
tric Elastomers as Electromechanical Transducers, F. Carpi, D. De Rossi, R. Kornbluh,
P. Sommer-Larsen, Eds., 1 ed. Elsevier, Amsterdam, 2008, p. 344.
199. P HILLIPS , R. W. Engineering applications of fluids with a variable yield stress. PhD thesis,
University of California, Berkeley, 1969.
200. P ING , L.; Y UMEI , W. An arbitrarily distributed tactile sensor array using piezoelectric
resonator. In Instrumentation and Measurement Technology Conference, 1996. IMTC-96.
Conference Proceedings. ’Quality Measurements: The Indispensable Bridge between Theory
and Reality’., IEEE (Brussels ISBN: 0-7803-3312-8, 1996), vol. 1, pp. 502 – 505 vol.1.
201. P UERS , R. Capacitive sensors: when and how to use them. Sensors and Actuators A: Physi-
cal 37-38 (1993), 93–105.
202. R AJAGOPAL , K. R.; RUZICKA , M. Mathematical modeling of electrorheological materials.
Continuum Mechanics and Thermodynamics 13, 1 (2001), 59 – 78.
203. R ASKIN , J. The Humane Interface. Addison-Wesley, Reading, Massachusetts, 2000.
204. R AUSCH , J. Analyse der mechanischen Eigenschaften von Lebergewebe bei intrakorporaler
Interaktion. Diplomarbeit, Technische Universität Darmstadt, Darmstadt, 2005.
205. R EISSELL , L.-M.; PAI , D. K. High Resolution Analysis of Impact Sounds and Forces. In
EuroHaptics Conference, 2007 and Symposium on Haptic Interfaces for Virtual Environment
and Teleoperator Systems. World Haptics 2007. Second Joint (Tsukaba, 2007), pp. 255 – 260.
References 461

206. R EN , K.; L IU , S.; L IN , M.; WANG , Y.; Z HANG , Q. M. A compact electroactive polymer
actuator suitable for refreshable Braille display. Sensors and Actuators A: Physical 143, 2
(2008), 335 – 342.
207. R EY, P.; C HARVET, P.; ET AL . A high density capacitive pressure sensor array for finger-
print sensor application. In International Conference on Solid-State Sensors and Actuators
(Chicago, 1997), IEEE, Ed.
208. R IENER , R.; V ILLGRATTNER , T.; K LEISER , R.; N EF, T.; KOLLIAS , S. fMRI-Compatible
Electromagnetic Haptic Interface. In Engineering in Medicine and Biology Society, 2005.
IEEE-EMBS 2005. 27th Annual International Conference of the (Shanghai ISBN: 0-7803-
8741-4, 2005), pp. 7024 – 7027.
209. RÖSE , A.; K ERN , T. A.; E ICHER , D.; S CHEMMER , B.; S CHLAAK , H. F. INKOMAN An
intracorporal manipulator for minimally invasive surgery. In Proceedings Biomedizinische
Technik (Zürich, 06.-09.09. 2006).
210. ROSEN , J.; S OLAZZO , M.; H ANNAFORD , M. D. B. Task Decomposition of Laparo-
scopic Surgery for Objective Evaluation of Surgical Residents Learning Curve Using Hidden
Markov Model. Computer Aided Surgery 7 (2002).
211. RUSPINI , D. C.; KOLAROV, K.; K HATIB , O. The haptic display of complex graphical en-
vironments. In SIGGRAPH ’97: Proceedings of the 24th annual conference on Computer
graphics and interactive techniques (New York, NY, USA, 1997), ACM Press/Addison-
Wesley Publishing Co., pp. 345–352.
212. RUSSELL , R. A. A tactile sensory skin for measuring surface contours. In Tencon 1992
Region 10 Conference (Melbourne, November 1992), IEEE, Ed.
213. S ARAF, R.; M AHESHWARI , V. High-Resolution Thin-Film Device to Sense Texture by
Touch. Tech. Rep. 5779, 9. Juni 2006.
214. S ATTEL , T. Dynamics of ultrasonic motors. IV, 167.
215. S CHAUMBURG , H. Sensoren, vol. 3 of Werkstoffe und Bauelemente der Elektrotechnik.
Teubner, Stuttgart, 1992.
216. S CHIMKAT, J.; K IESEWETTER , L.; G EVATTER , H.; A RNDT, F.; S TECKENBORN , A.;
S CHLAAK , H. F. Moving Wedge Actuator: An Electrostatic Actuator for Use in a Microre-
lay. In MICRO SYSTEM Technologies ’94, 4th International Conference and Exhibition on
Micro, Electro, Opto, Mechanical Systems and Components (1994), VDE-Verlag, pp. 989 –
996.
217. S CHIMMEL , P.; N OEVER , P. Out of actions: zwischen Performance und Objekt 1949 - 1979.
Cantz, Ostfildern, 1998. Kämpfen mit Schlamm.
218. S CHLAAK , H.; H OPPACH , E.; W EISSMANTEL , H.; E ICHER , D. A . Elektrische
Kleinantriebe - Skript. Vorlesungsmaterialien. Eigenverlag, Darmstadt, 2007.
219. S CHMIDT, R. F. Neuro- und Sinnesphysiologie, 4 ed. Springer, Berlin [u.a.], 2000.
220. S CHNELL , W.; G ROSS , D.; H AUGER , W. Technische Mechanik 2: Elastostatik, 7 ed.
Springer, Berlin [u.a.], 2002.
221. S CHREIER -A LT, T. Polymerverkapselung mechatronischer Systeme: Charakterisierung
durch eingebettete Faser-Bragg-Gitter-Sensoren. PhD thesis, TU Berlin, 2006.
222. S EKTION FÜR M INIMAL I NVASIVE C HIRURGIE T ÜBINGEN , U. T. Verbundprojekt TAMIC
- Entwicklung eines taktilen Mikrosensors für die Minimal Invasive Chirurgie - Schluss-
bericht. Tech. rep., 1999.
223. S ERGIO , M.; M ANARESI , N.; C AMPI , F.; ET AL . A dynamically reconfigurable monolithic
CMOS pressure sensor for smart fabric. IEEE Journal of solid-state circuits 38, 6 (2003),
966–968.
224. S ERINA , E LAINE , R.; M OTE , C.; R EMPEL , D. Force Response of the Fingertip Pulp to Re-
peated Compression - Effects of Loading Rate, Loading Angle and Anthropometry. Journal
of Biomechanics 30, 10 (1997 1997), 1035–1040.
225. S ESSLER , G. M. Akkustik I - Skriptum zur Vorlesung. 2000.
226. S HAHABI , C.; O RTEGA , A.; KOLAHDOUZAN , M. R. A comparison of different haptic
compression techniques. Multimedia and Expo, 2002. ICME’02. Proceedings. 2002 IEEE
International Conference on 1 (2002).
462 18 Impedance Values of Grasps

227. S HENG , P. Mechanism of the giant electrorheological effect. International Journal of Mod-
ern Physics B (IJMPB) 19, 7/9 (2005), 1157–1162.
228. S HERRIT, S.; D OLGIN , B. P.; BAR -C OHEN , Y.; PAL , D.; K ROH , J.; P ETERSON , T. Mod-
eling of horns for sonic/ultrasonic applications. Ultrasonics Symposium, 1999. Proceedings.
1999 IEEE 1 (1999), 647 – 651 vol.1.
229. S HIMOGA , K. B. A Survey of Perceptual Feedback Issues in Dextrous Telemanipulation:
Part I. Finger Force Feedback. Proceedings of the IEEE Virtual Reality Annual International
Symposium (VRAIS), Seattle, Washington (1993), 263–270.
230. S HIMOGA , K. B. A Survey of Percepual Feedback Issues in Dextrous Telemanipulation: Part
II. Finger Touch Feedback. Proceedings of the IEEE Virtual Reality Annual International
Symposium (VRAIS), Seattle, Washington (1993), 271–279.
231. S HUAI , L.- G .; K UANG , Y.- H .; Z HOU , Z.- T. Research on robot tactile display based on
water jet technology. In Robotics and Biomimetics (ROBIO). 2005 IEEE International Con-
ference on (Shatin ISBN: 0-7803-9315-5, 2005), pp. 676 – 680.
232. S IEGEL , M. Tactile display development: the driving-force for tactile sensor development.
Haptic Virtual Environments and Their Applications, IEEE International Workshop 2002
HAVE (2002), 115 – 118.
233. S IMONE , C. Modelling of neddle insertion forces of percutaneous therapies. Diplomarbeit,
Johns Hopkins University, Baltimore, 2002.
234. S LOTINE , J.-J. E.; L I , W. Applied Nonlinear Control. Prentice Hall, 1991.
235. SMI. Silicon Microstructures Incooperated - Pressure Sensors Products - http://www.si-
micro.com/pressure-sensor-products.html. 4. Juli 2008.
236. S MITH , C. U. The biology of sensory systems. Wiley, Chichester [u.a.], 2000.
237. S MITH , C., S. Piezoresistance Effect in Germanium and Silicon. Phys. Rev., American
Physical Society 94, 1 (April 1954), 42–49.
238. Newsclae Technologies - Ingenioulsy small motion systems.
239. S RINIVASAN , M. A.; C HEN , J.-S. Human Performance in Controlling Normal Forces of
Contact with Rigid Objects. ASME Winter Annual Meeting: Advances in Robotics, Mecha-
tronics and Haptic Interfaces, New Orleans, DSC-Vol. 49 (1993), 119–125.
240. S TAVROULIS , S. Rechnergestützter Entwurf von piezoresistiven Silizium-Drucksensoren mit
realem mechanischem Wandler. PhD thesis, 2004.
241. S TELZMANN , U.; G ROTH , C.; M ÜLLER , G. FEM für Praktiker - Strukturdynamik: Ba-
siswissen und Arbeitsbeispiele zu FEM-Anwendungen der Strukturdynamik von: FEM für
Praktiker, 3 ed., vol. 44 of Edition expertsoft. expert-Verl., Renningen, 2002.
242. S TEVENS , S. S. Psychophysics: introduction to its perceptual, neural, and social prospects.
A Wiley-Interscience publication. Wiley, New York [u.a.], 1975.
243. S U , L.; C HIANG , K. S.; L U , C. Microbend-induced mode coupling in a graded-index
multimode fiber. APPLIED OPTICS 44, 34 (Dezember 2005).
244. S UMMERS , I. R.; C HANTER , C. M. A broadband tactile array on the fingertip. The Journal
of the Acoustical Society of America 112, 5 (2002/11/00/ 2002), 2118 – 2126.
245. S UN , Y.; N ELSON , B.; P OTASEK , D.; E NIKOV, E. A bulkmicrofabricated multi-axis ca-
pacitive cellular force sensor using transverse comb drives. Journal of Micromechanics and
Microengineering, IOP 12 (2002), 832–840.
246. S UTHERLAND , I. E. The Ultimate Display. In Proceedings of IFIP Congress (1965), pp. 506
– 508.
247. TAKAMATSU , R.; TANIGUCHI , T.; S ATO , M. Space browsing interface based on head
posture information. In Computer Human Interaction, 1998. Proceedings. 3rd Asia Pacific
(Shonan Village Center ISBN: 0-8186-8347-3, 1998), pp. 298 – 303.
248. TAKASAKI , M.; K URIBAYASHI K UROSAWA , M.; H IGUCHI , T. Optimum contact condi-
tions for miniaturized surface acoustic wave linear motor. Ultrasonics 38, 1-8 (2000/3 2000),
51 – 53.
249. TALBOT, W. H.; DARIAN -S MITH , I.; KORNHUBER , H. H.; M OUNTCASTLE , V. B. The
sense of flutter-vibration: comparison of the human capacity with response patterns of
mechanoreceptive afferents from the monkey hand. J Neurophysiol 31, 2 (1968), 301–34.
References 463

250. TAMMER , R.; E HRENREICH , L.; J URGENS , U. Telemetrically recorded neuronal activity
in the inferior colliculus and bordering tegmentum during vocal communication in squirrel
monkeys (Saimiri sciureus). Behav Brain Res 151, 1-2 (2004), 331–6. German Primate
Centre, Kellnerweg 4, 37077 Gottingen, Germany.
251. TAN , H. Z.; PANG , X. D.; D URLACH , N. I. Manual Resolution of Length, Force, and
Compliance. ASME Advances in Robotics, DSC-Vol. 42 (1992), 13–18.
252. TAN , H. Z.; S RINIVASAN , M. A.; E BERMANN , B.; C HENG , B. Human Factors for the De-
sign of Force-Reflecting Haptic Interfaces. Proceedings of the ASME Winter Annual Meet-
ing: Dynamic Systems and Control, DSC-Vol. 55-1 (1994), 353–359.
253. TANG , H.; B EEBE , D. J. A microfabricated electrostatic haptic display for persons with
visual impairments. Rehabilitation Engineering, IEEE Transactions on Neural Systems and
Rehabilitation 6, 3 (1998), 241 – 248.
254. TAYLOR , P. M.; P OLLET, D. M.; H OSSEINI -S IANAKI , A.; VARLEY, C. J. Advances in an
electrorheological fluid based tactile array. Displays 18, 3 (1998), 135 – 141.
255. T IETZE , U.; S CHENK , C. Halbleiter-Schaltungstechnik: / U. Tietze, 12 ed. Springer, Berlin
[u.a.], 2002.
256. T SAGARAKIS , N.; H ORNE , T.; C ALDWELL , D. SLIP AESTHEASIS: a portable 2D
slip/skin stretch display for the fingertip. In Haptic Interfaces for Virtual Environment and
Teleoperator Systems, 2005. WHC 2005. First Joint Eurohaptics Conference and Symposium
on (2005), pp. 214 – 219.
257. T SAI , L.- W. Robot analysis - the mechanics of serial and parallel manipulators. A Wiley-
Interscience publication. Wiley, New York [u.a.], 1999.
258. T SAI , L. W.; J OSHI , S. Kinematics and optimization of a spatial 3-upu parallel manipulator.
Journal of Mechanical Design 122 (2000), 439.
259. T SAI , L. W.; WALSH , G. C.; S TAMPER , R. E. Kinematics of a novel three dof trans-
lational platform. Robotics and Automation, 1996. Proceedings., 1996 IEEE International
Conference on 4 (1996).
260. U CHINO , K. Piezoelectric actuators and ultrasonic motors. Kluwer, Boston [u.a.], 1997.
261. U CHINO , K.; G INIEWICZ , J. R. Micromechatronics. No. 22 in Materials engineering.
Dekker, New York [u.a], 2003.
262. U HEA , S. Recent development of ultrasonic actuators, vol. 1 of Ultrasonics Symposium,
2001 IEEE. Atlanta, GA, 7-10 Oct. 2001.
263. U HEA , S.; T OMIKAWA , Y.; K UROSAWA , M.; NAKAMURA , N. Ultrasonic motors - Theory
and application. Oxford University Press, 1993.
264. U NBEHAUEN , H. Regelungstechnik I. Vieweg und Teubner, 2007.
265. U NBEHAUEN , H. Regelungstechnik II: Zustandsregelungen, digitale und nichtlineare Regel-
systeme. Vieweg und Teubner, 2007.
266. VALDASTRI , P.; ROCCELLA , S.; B ECCAI , L.; C ATTIN , E.; M ENCIASSI , A.; C ARROZZA ,
M.; DARIO , P. Characterization of a novel hybrid silicon three-axial force sensor. Sensors
and Actuators A Physical 123-124 (2005), 249–257.
267. VAN E RP, J. Vibrotactile spatial acuity on the torso: effects of location and timing parameters.
Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2005. WHC 2005. First
Joint Eurohaptics Conference and Symposium on (2005), 80 – 85.
268. VARADI , P. Piezomotor with a guide, 3. April 2003.
269. VASARHELYI , G.; A DAMA , M.; ET AL . Effects of the elastic cover on tactile sensor arrays.
Sensors and Actuators A Physical 132 (2006), 245–251.
270. VASARHELYI , G.; F ODOR , B.; ROSKA , T. Tactile sensing-processing - Interface-cover
geometry and the inverse-elastic problem. Sensors and Actuators A Physical 140 (2007),
8–18.
271. VAZSONYI , E.; A DAM , M.; D USCO , C.; V IZVARY, Z.; T OTH , A. L.; BARSONY, I. Three-
dimensional force sensor by novel alkaline etching technique. Sensors and Actuators A
Physical 123-124 (2005), 620–626.
272. V IDAL V ERDÚ , F.; NAVAS G ONZALES , R. Thermopneumatic Approach for Tactile Dis-
plays. In Mechatronics & Robotics (Aachen, Germany, 2004), pp. 394–399.
464 18 Impedance Values of Grasps

273. V ITRANI , M. A.; N IKITCZUK , J.; M OREL , G.; M AVROIDIS , C.; W EINBERG , B. Torque
Control of Electrorheological Fluidic Resistive Actuators for Haptic Vehicular Instrument
Controls. Journal of Dynamic Systems, Measurement, and Control 128, 2 (2006), 216–226.
274. VOYLES , R.; M ORROW, J. D.; KOHSLA , P. K. The shape from motion approach to rapid
and precise force/torque sensor calibration. Journal of Dynamic Systems, Measurement and
Control 119, 2 (1997), 229–235.
275. V SRHELYI , G.; DM , M.; V ZSONYI , .; B RSONY, I. A . Effects of the elastic cover on tactile
sensor arrays. Sensors and actuators. A, Physical 132, 1 (2006), 245–251.
276. WAGNER , C.; L EDERMAN , S.; H OWE , R. A tactile shape display using RC servomotors. In
Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2002. HAPTICS 2002.
Proceedings. 10th Symposium on (Orlando, FL ISBN: 0-7695-1489-8, 24252002), pp. 354 –
355.
277. WAHED , A. K. The characteristics of a homogeneous electrorheological fluid in dynamic
squeeze. In Actuator 2004: 9th International Conference on New Actuators (Bremen, 2004),
HVG Hanseatische Veranst.-GmbH, Div. Messe Bremen, pp. 605 – 608.
278. WALLASCHEK , J. A . Contact mechanics of piezoelectric ultrasonic motors. Smart Materials
and Structures, 3 (1998), 369 – 381.
279. Piezoaktoren in ihrer Anwendung.
280. W EAVER , W.; T IMOSENKO , S. P.; YOUNG , D. H. Vibration problems in engineering, 5.
ed ed. Wiley, New York [u.a.], 1990.
281. W EINBERG , B.; N IKITCZUK , J.; F ISCH , A.; M AVROIDIS , C. Development of electro-
rheological fluidic resistive actuators for haptic vehicular instrument controls. Smart Mate-
rials and Structures 14, 6 (2005), 1107 – 1119.
282. W EISSMANTEL , H.; H OPPACH , E. Elektrische Kleinantriebe - Skriptum zur Vorlesung.
2002.
283. W EN , W.; H UANG , X.; S HENG , P. Particle size scaling of the giant electrorheological
effect. Applied Physics Letters 85, 2 (2004), 299–301.
284. W ERTHSCHUETZKY, R. Mess- und Sensortechnik - Band II: Wirkprinzipien von Primärsen-
soren. Skriptum, Technische Universität Darmstadt, 2007.
285. W ERTHSCHUETZKY, R. Mess- und Sensortechnik - Band III: Analoge Sensorelektronik.
Skriptum, Technische Universität Darmstadt, Darmstadt, 2007.
286. W IKER , S. F.; H ERSHKOWITZ , E.; Z IK , J. Teleoperator Comfort and Psychometric Sta-
bility: Criteria for Limiting Master-Controller Forces of Operation and Feedback During
Telemanipulation. Proceedings of NASA Conference on Space Telerobotics, Vol. 1 (1989),
99–107.
287. W ILKINSON , R. S.; F UKAMI , Y. Responses of isolated Golgi tendon organs of cat to sinu-
soidal stretch. J Neurophysiol 49, 4 (1983), 976–88.
288. W INFIELD , L.; G LASSMIRE , J.; C OLGATE , J. E.; P ESHKIN , M. T-PaD: Tactile Pattern
Display through Variable Friction Reduction. In EuroHaptics Conference, 2007 and Sympo-
sium on Haptic Interfaces for Virtual Environment and Teleoperator Systems. World Haptics
2007. Second Joint (Tsukaba, 2007), pp. 421 – 426.
289. W OLF, U. Aristoteles’ “Nikomachische Ethik”, 2 ed. Wiss. Buchges., [Abt. Verl.], Darm-
stadt, 2007.
290. YAMAMOTO , A.; I SHII , T.; H IGUCHI , T. Electrostatic tactile display for presenting sur-
face roughness sensation. Industrial Technology, 2003 IEEE International Conference on 2
(2003), 680 – 684 Vol.2.
291. YAMAOKA , M.; YAMAMOTO , A KIO , H. T. Basic Analysis of Stickiness Sensation for
Tactile Displays. In Eurohaptics Conference (Madrid, 2008), Springer.
292. Y EONGMI , K.; OAKLEY, I.; J EHA , R. Design and Psychophysical Evaluation of Pneumatic
Tactile Display. In SICE-ICASE, 2006. International Joint Conference (Busan ISBN: 89-
950038-5-5, 2006), pp. 1933 – 1938.
293. YOSHIKAWA , T.; I CHINOO , Y. Impedance identification of human fingers using virtual
task environment. Intelligent Robots and Systems, 2003. (IROS 2003). Proceedings. 2003
IEEE/RSJ International Conference on 4 (2003), 3094 – 3099 vol.3.
References 465

294. Y U , O.; Z HARRI. An exact mathematical model of a travelling wave ultrasonic motor, vol. 1
of Ultrasonics Symposium, 1994. Proceedings., 1994 IEEE. Cannes, 1-4 Nov. 1994.
295. Z ACHMANN , G. Virtual Reality in Assembly Simulation — Collision Detection. PhD thesis,
Darmstadt University of Technology, Germany, Department of Computer Science, may 2000.
296. Z ADEH , M. H.; WANG , D.; K UBICA , E. Perception-based lossy haptic compression con-
siderations for velocity-based interactions. Multimedia Systems 13, 4 (2008), 275–282.
297. Z HU , M. Contact analysis and mathematical modeling of traveling wave ultrasonic motors,
vol. 51 of Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on. June
2004.
298. Z IEMANN , O. POF-Handbuch: optische Kurzstrecken-Übertragungssysteme, 2 ed. Springer,
Berlin [u.a.], 2007.
299. Z ILLES , C. B.; S ALISBURY, J. K. A constraint-based god-object method for haptic display.
In IROS ’95: Proceedings of the International Conference on Intelligent Robots and Systems
(Washington, DC, USA, 1995), vol. 03, IEEE Computer Society, p. 3146.
“This page left intentionally blank.”
Index

A capacitive actuator 194

capacitive actuators 279
absolute threshold 40 capacitive principle 192
active 10, 27 capacitive sensors 366
active haptic interaction 60, 63 charge constant 246
actuation principle 192 closed-loop admittance controlled 95
actuator design 191 closed-loop impedance controlled 95
actuator designs, piezoelectric 248 cluster “always” 122
actuator, electrostatic 277
cluster “kinaesthetic” 118
actuator, piezoelectric 243
cluster “omni-dimensional” 121
addressability 23
cluster “surface-tactile” 119
adhesion 28
cluster “vibro-directional” 121
admittance controlled, closed-loop 100
cluster “vibro-tactile” 120
admittance controlled, open-loop 98
code disc 360
admittance, mechanical 27, 443
π -coefficients 330
ampere turns 203
anthropometric 11 coefficient of strain, piezoelectric 245
artificial muscle 292 coefficient of tension, piezoelectric 245
assistive system, haptic 25 coefficients, piezoelectric 245
collision detection
B dynamic 428
static 428
bar-shaped resonator 250 commutated, electronic 216
basic equations, piezoelectric 244 commutation 216
basic piezoelectric actuator designs 248 compression 380
being, physical 6 conductor 201
bending moment 324 constraint 183
bending-actuator 194, 266 contact grasp 65
Bernoulli 325 control design
bimorph, piezoelectric 263 for Haptic Systems 152
block-commutation 218 control law design 156
Bragg 346 control quality 132
Braille 266 controller, haptic 22
brake, electromagnetic 237
Coulomb’s law 277
C coupling factor, piezoelectric 247
Curie temperature 247
C 82, material property 247 current source, analog 223

467
468 Index

customer, experiments with 122 electromechanical network 261

electronic-commutated 216
D electrostatic actuators 277
energy density, magnetic 208
D/A converter 221 energy, magnetic 229
DC-drive 215 enhanced parallel port 382
DC-motor 193 EPP 382
dead-man-switch 64 equations traveling wave motor 255
Delta.3 189 Ethernet 384
Delta-Robot 170 event 118
Denavit-Hartenberg-parameter 169, 177, Event-based haptics 378
178, 181 evolutionary algorithm 174
design piezoelectric actuators 258 experiments 122
design process 11 external supply rate 150
designs of DEA 289 extreme dead times 379
determinant 172
dielectric elastomer actuator (DEA) 285,
287 F
difference threshold 40
Difference-Limen 40 FA-I 36
differentiation of signals 369 FA-II 37
digital to analog converter 221 Fabry-Pèrot 344
direct current magnetic field 207 Falcon 167
direct kinematics 169, 190 fast-adapting 36
direct method by LYAPUNOV 148 Fechner’s law 44
discretization 87 feedforward 159
disturbance compensation 158 feedforward filter function 159
DL 40 FFB 24
DoF 391 fiber bragg grating 346
dots-per-inch 359 field driven actuator 277
DPI 359
field plates 364
driver electronics, electrodynamic 219
field response 212
dynamic 37
field strength, magnetic 203
dynamic behavior 132
filling-factor 203
dynamic collision detection 428
FIP 82
dynamics 46
FireWire 383
E flow mode 295
flux 203
EC-drive 216 flux density, magnetic 203
EC-motor 193 flux, magnetic 203
ECP 382 foil sensor 334, 339
effect, piezoelectric 243 force sensing resistor 334
elastic constant 245 force, magnetomotive 203
elasto-mechanics 322 force-feedback 24
electric field 277 force-impression 82
electric motor 193 forward kinematics 169
electrical time constant 212 forward kinematics for serial mechanisms
electro-rheological fluid 292 177
electrochemical principle 192 frequency domain analysis 89
electrodynamic principle 192 frequency-dependency 46, 48
electromagnetic actuators 229 friction 28
electromagnetic brake 237 fully parallel 183
electromagnetic principle 192 functional requirements 124
Index 469

G input strictly passive 150

integral criterion 157
Gaussian distribution 45 integration of signals 369
gear-DOF 183 intensity 341
gears 197 interaction metaphor 388, 403
gearwheel 198 interaction, haptic 23
Grübler formula 182, 186 interface 17
gradient approach 174 interface, haptic 24
Grasps 64 inverse kinematics 169, 187, 190
gray scale sensor 361 iron-less rotor 213

H J

H-bridge 221 Jacobian-matrix 169

hall sensors 364 Jacobian matrix, direct 172
Hammerstein-model 140 Jacobian matrix, inverse 172
HapKeys 226 Jacobian matrix, rank 173
haptic assistive system 25 Jacobian matrix, singular values 174
haptic controller 22 JND 40, 319
haptic device 22 joint, driven 166
haptic device, active 27 joint, linear 175
haptic device, passive 27 joint, passive 166
haptic display 24 joint, rotatory 176
haptic interaction 23, 117
joint, un-driven 166
haptic interface 24
Just-Noteable-Difference 40
haptic loop 127
Just-Noticeable-Difference 40
haptic manipulator 24
haptic marker 24
K
haptic rendering 389
haptic shape 28
haptic simulator 25 k-factor 326, 353
haptic surface properties 28 kinaesthetic 321
haptic texture 28 kinaesthetic perception 10
haptics 21 kinaesthetic sensors 37
hardware in the loop 384 kinaesthetics 21
harmonic balance 145 kinematic dependencies 170
hexapod 166 kinematic design, complete process 188
HID 382, 393 kinematic scheme 184
high-grade parallel 183 kinematics 166
HIL 384
human-interface-device 382 L
hybrid kinematic 166, 188
hydraulic 194 level of parallelism 183
linear SISO-systems 132
I Lissajusfigure 251
LM 43
identical condition 183 local haptic model 377
IEEE1394 383 longitudinal actuator 246
impedance 63 longitudinal effect, magnetic 231
impedance controlled, closed loop 97 longitudinal effect, piezoelectric 244
impedance controlled, open-loop 96 Lorentz-force 199
impedance coupling 62, 64 Lossless 150
impedance, mechanical 27, 443 Lyapunov 145
Induction 211 Lyapunov functions 148
470 Index

M open-loop admittance controlled 95

open-loop impedance controlled 95
magn. resistance 203 optical position sensors 360
magnetic circuits 203, 232 output equation 137
magnetic cross section 232 output strictly passive 150
magnetic dependent resistors 364 overshoot 154
magnetic energy 229, 234
magnetic field strength 203 P
magnetic field, direc current 207
magnetic flux 203
Pacinian 37
magnetic flux density 203, 205
parallel kinematic 165, 166, 188
magnetic resistance 203
parallel port 382
magneto-rheological-fluid 297
parallel-plate capacitor 277
magnetomotive force 203
partly parallel 183
magnetorheological principle 192
passive 10, 27
manipulator, haptic 24
passive haptic device 90
marker, haptic 24
passive haptic interaction 60
masking 41
passivity 145
material properties, piezoelectric 247
passivity, control engineering 63
materials, piezoelectric 247
percentile 45
mechanical admittance 27
percentiles 12
mechanical commutation 215
Perception 80
mechanical impedance 27
perception, kinaesthetic 84
mechanically commutated 217
perception, tactile 85
mechanism 166, 170
mechanism, parallel 182, 186 permanent magnet 206
mechanism, serial 175 permanent magnets 234
Meissner 36 permeability 203
Merkel 37 permeability number 234
mice-sensor 362 permittivity 203
micro frame 382 PHANTOM Omni haptic device 176
micro-bending sensor 343 phase plane analysis 145
model time 397 photo-elastic effect 340
motor cortex 36 photodiode 361
moving coils, electrodynamic 213 phototransistor 361
moving magnet 216 physical-interface-device 382
multi-input-multi-output-systems 136 PID 382
multimodality 40 PID-Control 156
piezoelectric actuators 243
N piezoelectric actuators, design 258
piezoelectric basic equations 244
natural behavior 389 piezoelectric bimorph 263
natural interaction 388 piezoelectric coefficient of strain 245
neurobiology 19 piezoelectric coefficient of tension 245
nociceptive 21 piezoelectric coefficients 245
nominal load 314, 354 piezoelectric coupling factor 247
nonlinear stability analysis 145 piezoelectric effect 243
nonlinear system description 139 piezoelectric equation 246
Nyquist´s stability criterion 142 piezoelectric longitudinal effect 244
piezoelectric material properties 247
O piezoelectric materials 247
piezoelectric motor 194
observer based piezoelectric principle 192
state space control 160 piezoelectric sensors 347
Index 471

piezoelectric shear effect 244 resolution 23

piezoelectric special designs 250 resonance principle 350
piezoelectric stack 194 resonance-actuator 193
piezoelectric stepper motors 257 resonator, bar-shaped 250
piezoelectric transversal effect 244 resonator, circular 254
piezoelectrical Bimorph 266 reverse kinematics 169
piezoelelectric shear effect 257 robot 166
Pixel 391 robot arm 166
plunger type magnet 238 root locus method 142, 157
plunger-type magnet 194 roughness 29
pneumatic 194 Routh Hurwitz criterion 142
Popov inequality 146 Ruffini 37
Popov plot 146
Popov-criterion 145 S
positioning system 166
power grasp 65
SA-I 37
power grasps 70
SA-II 37
power law 44
Sashida 254
power loss, electrodynamic 201
SAW sensors 351
precision grasp 65
SCARA-robot 167
precision grasps 71
self-supportive 215
propagation 395
semiconductor 328
psychophysics 19, 40
senses 5
PT1 -System 135
sensor-less commutation 218
PT2 -System 137
sensors, kinaesthetic 37
puls-width-modulation 221
sensors, tactile 36
PVDF, material property 247
serial kinematic 165, 166, 188
PWM 221
serial port 381
PZT-4, material property 247
servo-drive 216
PZT-5a, material property 247
shaker 193
Q shape 117
shape, haptic 28
quadrant controllers 219 shape-memory alloy 192
Quarz, Kristallstruktur 243 shape-memory wire 194
Quarz, material property 247 shear effect, piezoelectric 244, 257
shear mode 294
R silicon sensors 331
simulator system 374
RA 36 simulator, haptic 25
rapid-adapting 36 singular position 172
rare earth 206 singularity 172
real-time 397 sinus-commutation 218
reflection light switches 362 slow-adapting 37
relative resistivity change 326 somatotrope 36
reluctance 203 spatial masking 41
reluctance drives 237 special designs, piezoelectric 250
reluctance effect 229 specifications 124
reluctance effect, magnetic 231 squeeze mode 295
remanence flux density 208 stability analysis 141
rendering 389 standing wave 250
requirements 124, 319, 352 state feedback control 159
resistance, magnetic 203 state space control 159
resistivity change 330 state space representation 136
472 Index

state space vector 138 transfer characteristics, optimization 174

state strictly passive 150 transmission chain 374
static collision detection 428 transmission-ratio 197
stepper motor 236 transparency 27
stepper motors, piezoelectric 257 transversal actuator 246
stepper-motor 194 transversal effect 229
Stevens 44 transversal effect, electromagnetic 230
Steward-Gough-platform 166 transversal effect, piezoelectric 244
strain gauge 327 traveling wave motor, equations 255
stress 324 traveling wave motor, linear 254
strictly passive 150 travelling wave 254
Successivenes Limen 43 travelling wave motor 273
surface adhesion 28 triangulation 362
surface friction 28 two-point threshold 52
surface micro-machining 338 two-point-threshold 42
surface properties, haptic 28
surface-wave actuators 195 U
surgery 14
system description 132 Ubi-Pen 268
System Engineering 95 Uchinomotor, Uchino 255
system equation 136 Ultraschallaktoren 250
system specification 124 ultrasonic sensors 365
system stability 131 ultrasonic-actuator 193
system strukturing 152 universal serial bus 382
system transfer function 141 upper cut-off frequency 354
USB 382
T user 22, 29
user-impedance 27
tactile 21, 317, 319, 332, 337, 350
tactile perception 10
V
tactile sensor 36
tactile systems 290, 298
tacton 16, 29 vibrotactile display 267
tactor 24 virtual work 173
telemanipulation 14 viscoelastic material performance 316
telemanipulation system 25, 374 voice-coil-actuator 193
telemanipulator 25 voigtsche notation 245
temperature, monitor 224
temporal masking 41 W
texture 117, 317
Texture Explorer 267 wave, standing 250
texture, haptic 28 wave, traveling 254
thermal principle 192 Weber’s law 43
thermoceptive 21 Wheatstone 326
threshold 40 Wiener-Model 139
threshold, difference 40 wire 201
time constant, electrical 212
Tool Center Point, TCP 168 Z
topology 166, 189
total reflection 341 Z 27
transfer characteristics 170, 173 Z-width 27