129 Pages
English
Gain access to the library to view online
Learn more

Development and control of an anthropomorphic telerobotic system [Elektronische Ressource] / Bartłomiej Stańczyk

Gain access to the library to view online
Learn more
129 Pages
English

Description

Lehrstuhl fur Steuerungs- und RegelungstechnikTechnische Universit at Munc henProf. Dr.-Ing./Univ. Tokio Martin BussDevelopment and Control of anAnthropomorphic Telerobotic SystemBart lomiej StanczykVollst andiger Abdruck der von der Fakult at fur Elektrotechnik und Informationstechnikder Technischen Universit at Munc hen zur Erlangung des akademischen Grades einesDoktor-Ingenieurs (Dr.-Ing.)genehmigten Dissertation.Vorsitzender: Univ.-Prof. Dr.-Ing. Eckehard SteinbachPrufer der Dissertation:1. Univ.-Prof. Dr.-Ing./Univ. Tokio Martin Buss2. Univ.-Prof. Dr.-Ing. habil. Heinz UlbrichDie Dissertation wurde am 21.02.2006 bei der Technischen Universit at Munc hen einge-reicht und durch die Fakult at fur Elektrotechnik und Informationstechnik am 17.07.2006angenommen.2ForewordThis thesis is a result of four years of work in the research group of my thesis adviserProf. Martin Buss; at rst at the Technische Universit at Berlin, then at the TechnischeUniversit at Munc hen. I would like to thank him for giving me the opportunity to conductexciting research in one of the most honorable robotics and control groups in Germany, forproviding me with great hardware and a wonderful research environment.During my time spent with the group, many people gave me support and advice innumerous ways. I would like to thank Dr. Konstantin Kondak from TU Berlin for leadingmy rst steps in Robotics. His help was invaluable during the design phase.

Subjects

Informations

Published by
Published 01 January 2006
Reads 39
Language English
Document size 4 MB

Exrait

Lehrstuhl fur Steuerungs- und Regelungstechnik
Technische Universit at Munc hen
Prof. Dr.-Ing./Univ. Tokio Martin Buss
Development and Control of an
Anthropomorphic Telerobotic System
Bart lomiej Stanczyk
Vollst andiger Abdruck der von der Fakult at fur Elektrotechnik und Informationstechnik
der Technischen Universit at Munc hen zur Erlangung des akademischen Grades eines
Doktor-Ingenieurs (Dr.-Ing.)
genehmigten Dissertation.
Vorsitzender: Univ.-Prof. Dr.-Ing. Eckehard Steinbach
Prufer der Dissertation:
1. Univ.-Prof. Dr.-Ing./Univ. Tokio Martin Buss
2. Univ.-Prof. Dr.-Ing. habil. Heinz Ulbrich
Die Dissertation wurde am 21.02.2006 bei der Technischen Universit at Munc hen einge-
reicht und durch die Fakult at fur Elektrotechnik und Informationstechnik am 17.07.2006
angenommen.2Foreword
This thesis is a result of four years of work in the research group of my thesis adviser
Prof. Martin Buss; at rst at the Technische Universit at Berlin, then at the Technische
Universit at Munc hen. I would like to thank him for giving me the opportunity to conduct
exciting research in one of the most honorable robotics and control groups in Germany, for
providing me with great hardware and a wonderful research environment.
During my time spent with the group, many people gave me support and advice in
numerous ways. I would like to thank Dr. Konstantin Kondak from TU Berlin for leading
my rst steps in Robotics. His help was invaluable during the design phase. Designing
and building of the 7 DoF manipulator would not have been possible without Mr. Franz
Bachmann (TU Berlin), who not only put the design on paper, but personally crafted all
the aluminum elements. I also would like to thank Mr. Josef Gradl, Mr. Horst Kubick
and Mr. Thomas Lowitz at the TU Munc hen for additional developments. Many thanks
for your professionalism, creativity and enthusiasm.
I would like to express my gratitude to all the colleagues at the Lehrstuhl fur Steuerungs-
und Regelungstechnik, especially Sandra Hirche, Angelika Peer and Ulrich Unterhin-
ninghofen, with whom I cooperated closely on the SFB453 Project: for the scienti c input,
advice, help and for the great working atmosphere. Special thanks go to Hasan Esen, who
always supported and encouraged me as a friend.
I am also indebted to my students: Kwang-Kyu Lee, Hongquing Hu, Jiduo Wu, Klaas
Klasing and Martin Ernst. Thank you very much for the contributions you brought to the
project.
Finally, it is necessary to express my gratitude to my parents, my sister and my wife.
They were continuously supporting me with their love and patience, especially when
I needed it most.
Munich, October 2005. Bart lomiej Stanczykto my family
...Abstract
This thesis presents and summarizes the research work carried out during the develop-
ment and experimental evaluation of a dual arm anthropomorphic manipulator designed
for teleoperation purposes. The results are generalizable to other types of manipulators,
especially for robots acting in human environments and/or physically interacting with
humans.
Such applications pose speci c requirements that are in general di eren t from those of
traditional robot design. To assure safety and stable operation, \soft" robotic manipula-
tors are absolutely essential. For this reason, in this work, compliance control methods
were extensively studied and evaluated experimentally. As a result, an impedance control
strategy with underlying sti ness controller in motion control loop is proposed. To achieve
proper workspace matching between the human arm and the developed manipulator, an
adequate singularity handling strategy must be applied. Experimental comparison of such
solutions is performed and an innovative damped inverse kinematics method is introduced.
This method allows for traversing through singularities with simultaneous weighing of the
task space tracking error. Because the kinematic structures of the master/slave manipu-
lators in a teleoperation system are in general dissimilar, there is a need for a universal
kinematic interface, independent of the device structure. The classical approaches using
Euler or Cardanian angles fail due to algebraic singularities introduced by the represen-
tation. To avoid this problem, the unit quaternion was used for describing orientation
and corresponding rotational displacement. In order to avoid the collisions between the
arms and the workspace limits a virtual forces concept was introduced. This application
forms a basis for local optimization strategies and an intuitive force display for the human
operator. Finally, the force - position teleoperation control architecture was analyzed and
a tele-assembly experiment in 6 DoF was successfully performed. The experimental results
con rm the high performance of the developed hardware and control strategies.
Kurzfassung
In der vorliegenden Dissertationsschrift werden die Forschungsarbeiten, die w ahrend der
Entwicklung und experimentellen Bewertung eines speziell fur Telepr asenzanwendungen
konzipierten, zweiarmigen, anthropomorph gestalteten Manipulators durchgefuhrt wur-
den, vorgestellt und zusammengefasst. Die Ergebnisse lassen sich auf andere Arten von
Manipulatoren ub ertragen, insbesondere auf jene, die in fur Menschen gestalteten Umge-
bungen eingesetzt werden oder in direkten physikalischen Kontakt mit menschlichen Perso-
nen treten. Die besonderen Anforderungen der genannten Anwendungsgebiete unterschei-
den sich grunds atzlich von denen klassischer Robotikanwendungen. Um Sicherheit und
Stabilit at zu gew ahrleisten, sind "nachgiebige" Roboterarme unabdingbar. Aus diesem
Grund wurden Methoden nachgiebiger Positionsregelungen ausgiebig untersucht und ex-
perimentell bewertet. Als Ergebnis wird eine Impedanzregelungsstrategie mit unterlagerter
Stei gk eitsregelung vorgeschlagen. Um die Arbeitsbereiche des menschlichen Armes und
des entwickelten Manipulator aufeinander abzustimmen, muss eine angemessene Strate-
gie zur Behandlung von Singularit aten angewandt werden. Ein experimenteller Ver-
gleich derartiger L osungen wird durchgefuhrt, und eine besonders attraktive Methode
iiizur ged ampften inversen Kinematik wird vorgestellt, welche es erlaubt, Singularit aten
zu durchlaufen und dabei den kartesischen Regelfehler im Arbeitsraum abzusch atzen.
Da die kinematischen Strukturen auf Operator- und Teleoperatorseite eines Teleoper-
ationssystems ublic herweise verschieden sind, bedarf es einer allgemein verwendbaren,
ger ateunabh angigen Schnittstelle. Klassische, auf Euler- oder Kardan-Winkeln basierende
Ans atze sind aufgrund der m oglichen algebraischen Singularit aten ungeeignet. Um dieses
Problem zu umgehen, werden Einheitsquaternionen zur Beschreibung von Orientierung
und Orientierungsfehler verwendet. Zur Vermeidung von Kollisionen zwischen den bei-
den Armen und den Arbeitsraumgrenzen wird ein auf virtuellen Kr aften beruhendes
Konzept vorgestellt, welches als Grundlage fur lokale Optimierungsstrategien und eine
intuitive Kraftruc kkopplung an den menschlichen Operator dient. Schlie lic h wird die
Kraft-Positions-Teleoperationsstruktur untersucht und die erfolgreiche Durchfuhrung eines
Experiments zum telepr asenten Fugen gezeigt. Die experimentellen Ergebnisse besttigen
die hohe Performanz der entwickelten Manipulatoren und Steuerungsmethoden.
ivContents
1 Introduction 1
1.1 Teleoperation: De nition and Motivation . . . . . . . . . . . . . . . . . . . 2
1.2 Requirements for the Telemanipulator Development . . . . . . . . . . . . . 5
1.3 Main Contributions and Outline of the Dissertation . . . . . . . . . . . . . 6
2 Telemanipulator Kinematics 9
2.1 Telemanipulator Kinematic Design . . . . . . . . . . . . . . . . . . . . . . 10
2.1.1 Forward Kinematics for a 7 DoF Manipulator . . . . . . . . . . . . 10
2.1.2 De nition of the Elbow Angle: Self Motion Parametrization . . . . 12
2.2 Closed Form Inverse Kinematics Algorithm for 7 DoF Anthropomorphic Arm 13
2.2.1 Position-Based Inverse Kinematics Algorithm . . . . . . . . . . . . 13
2.2.2 Velocity-Based Inverse for a Spherical Wrist . . . . . . . 17
2.3 Singularity-Robust Inverse Kinematics . . . . . . . . . . . . . . . . . . . . 18
2.3.1 Jacobian Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.2 Singularity Robustness: State of the Art . . . . . . . . . . . . . . . 19
2.3.3 The Least Square (Pseudoinverse) Method . . . . . . . . . . . . . . 20
2.3.4 Damped Least Square Method . . . . . . . . . . . . . . . . . . . . . 20
2.3.5 Weighted DLS Method with Task Priority . . . . . . . . . . . . . . 21
2.3.6 Adjoint Jacobian Method . . . . . . . . . . . . . . . . . . . . . . . 22
2.3.7 Jacobian Transpose Method . . . . . . . . . . . . . . . . . . . . . . 23
2.4 Experimental Evaluation of the WDLS Method . . . . . . . . . . . . . . . 24
2.5 Redundancy Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.5.1 Pseudoinverse with Optimization Criteria . . . . . . . . . . . . . . . 28
2.5.2 Extended Kinematics Method . . . . . . . . . . . . . . . . . . . . . 29
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3 Task Space De nition and Orientation Error 31
3.1 Problem of Dissimilar Kinematics in Teleoperation . . . . . . . . . . . . . . 31
3.2 Orientation Representation and Orientation Error . . . . . . . . . . . . . . 32
3.2.1 Euler Angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2.2 Orientation Error with Nonminimal Representation . . . . . . . . . 35
3.2.3 Euler Angles vs. Unit Quaternion: a Simulation Example . . . . . . 37
3.3 Extended Task Space Formulation . . . . . . . . . . . . . . . . . . . . . . . 38
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4 Manipulator Control 41
4.1 Manipulator Dynamic Modeling . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 Compliant Control Methods: State of the Art . . . . . . . . . . . . . . . . 42
4.2.1 Sti ness Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
vContents
4.2.2 Impedance Control in Task Space . . . . . . . . . . . . . . . . . . . 48
4.2.3 Mixed Sti ness-Imp edance Approach . . . . . . . . . . . . . . . . . 50
4.3 Experimental Comparison of the Control Algorithms . . . . . . . . . . . . 50
4.3.1 Criteria for Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.3.2 Trajectory Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.3.3 Impedance Display Fidelity . . . . . . . . . . . . . . . . . . . . . . 59
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5 Telemanipulation Control Loop 65
5.1 T Control Architectures . . . . . . . . . . . . . . . . . . . . 65
5.1.1 Two-port Model of the Bilateral Teleoperation System . . . . . . . 66
5.1.2 Four Channel Bilateral Control Architecture . . . . . . . . . . . . . 66
5.1.3 Stability and Performance Analysis of a Two-port System . . . . . 68
5.1.4 Two Channel Bilateral Control Architectures . . . . . . . . . . . . . 69
5.2 Analysis of the Teleoperation Control Loop . . . . . . . . . . . . . . . . . . 70
5.3 Telerobotic Response Requirement . . . . . . . . . . . . . . . . . . . . . . 73
5.4 Controller Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.5 Teleoperation Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6 Collision Avoidance with Virtual Forces in Dual Arm Teleoperation 81
6.1 Collisions in Dual Arm Con guration . . . . . . . . . . . . . . . . . . . . . 82
6.2 Robot Modeling and Collision Detection . . . . . . . . . . . . . . . . . . . 84
6.3 Virtual Forces Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.4 Collision Avoidance with Redundancy Utilization . . . . . . . . . . . . . . 85
6.5 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7 Conclusions and Future Work 91
7.1 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
A Appendix 95
A.1 Forward Kinematics for the 7 DoF Arm . . . . . . . . . . . . . . . . . . . . 95
A.2 Analysis of the Elbow Motion . . . . . . . . . . . . . . . . . . . . . . . . . 97
A.3 Hardware and Implementation Details . . . . . . . . . . . . . . . . . . . . 99
A.4 Mass Properties of the Manipulator Links . . . . . . . . . . . . . . . . . . 100
viNotations
Abbreviations
DoF Degree of Freedom
FK Forward Kinematics
IK Inverse
LS Least Square
DLS Damped Least Square
WDLS Weighted DLS
SJA Sti ness control with inverse dynamics in joint space
SJB Modi ed sti ness control with inverse dynamics
STT Sti ness control in the task space
RAC Resolved acceleration control
PDJ Joint space PD control
TO Teleoperator
Conventions
Throughout this thesis the term \force" stands for both linear force and torque, while
\motion" in terms of a generalization of position, velocity and acceleration refers to both
translational and angular motion quantities.
Scalars, Vectors, and Matrices
Scalars are italicized in both upper and lower cases. Vectors are denoted by lower case
letters in boldface style, e.g., the vector x is composed of elements x . Matrices are denotedi
by upper case letters in boldface type, e.g., the matrix M is composed of elements Mij
(i-th row, j-th column).
x scalar
x vector
X matrix
2d dx_ , x are equivalent to x and x2dt dt
One exception is reserved for the Cartesian wrench F , which is a vector written in
capital for consistence with the literature, e.g., [84].
viiNotations
Simpli ed Trigonometric Notation
s sin()
c cos()
s sin(q )i i
c cos(q )i i
Subscripts and Superscripts
x apparent value of xa
x desired value of xd
x value x associated with the environmente
x value x asso with the forcef
x value x associated with the extended kinematics formulationE
x value x asso with the human operatorh
x value x associated with joint coordinatesj
x value x asso with Cartesian coordinatesk
x value x associated with the left arml
x value x asso with the master manipulatorm
x value x associated with rotational coordinateso
x value x asso with translational cop
x value x associated with the right armr
x value x asso with the slave manipulators
x value x associated with the elbow coordinate
f ;f ;f components of vector f in x ;y ;z directionx y z
viii