Lecture Notes in Artificial Intelligence 7101

Subseries of Lecture Notes in Computer Science

LNAI Series Editors

Randy GoebelUniversity of Alberta, Edmonton, Canada

Yuzuru TanakaHokkaido University, Sapporo, Japan

Wolfgang WahlsterDFKI and Saarland University, Saarbrücken, Germany

LNAI Founding Series Editor

Joerg SiekmannDFKI and Saarland University, Saarbrücken, Germany

Sabina Jeschke Honghai LiuDaniel Schilberg (Eds.)

Intelligent Roboticsand Applications4th International Conference, ICIRA 2011Aachen, Germany, December 6-8, 2011Proceedings, Part I

13

Series Editors

Randy Goebel, University of Alberta, Edmonton, CanadaJörg Siekmann, University of Saarland, Saarbrücken, GermanyWolfgang Wahlster, DFKI and University of Saarland, Saarbrücken, Germany

Volume Editors

Sabina JeschkeRWTH Aachen University, IMA/ZLW & IFUDennewartstraße 27, 52068 Aachen, GermanyE-mail: jeschke.office@ima-zlw-ifu.rwth-aachen.de

Honghai LiuUniversity of Portsmouth, School of Creative TechnologiesIntelligent Systems and Biomedical Robotics GroupEldon Building, Winston Churchill Avenue, Portsmouth, PO1 2DJ, UKE-mail: honghai.liu@port.ac.uk

Daniel SchilbergRWTH Aachen University, IMA/ZLW & IFUDennewartstraße 27, 52068 Aachen, GermanyE-mail: daniel.schilberg@ima-zlw-ifu.rwth-aachen.de

ISSN 0302-9743 e-ISSN 1611-3349ISBN 978-3-642-25485-7 e-ISBN 978-3-642-25486-4DOI 10.1007/978-3-642-25486-4Springer Heidelberg Dordrecht London New York

Library of Congress Control Number: 2011941364

CR Subject Classification (1998): I.4, I.5, I.2, I.2.10, H.4, C.2

LNCS Sublibrary: SL 7 – Artificial Intelligence

© Springer-Verlag Berlin Heidelberg 2011This work is subject to copyright. All rights are reserved, whether the whole or part of the material isconcerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting,reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publicationor 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 liableto 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 lawsand regulations and therefore free for general use.

Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India

Printed on acid-free paper

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

Preface

Robots are increasingly being used for service duties, exploring inaccessible areasand for emergency and security tasks, besides their conventional application inindustrial environments. The trend toward intelligent and autonomous systemsis uninterrupted and poses new challenges for the interaction between humansand robots. Controlling robots is far beyond conventional programming specifictasks and cooperation between humans and robots becomes crucially important.As a result, the behavior of modern robots needs to be optimized toward thesenew challenges.

Against this background, the 4th International Conference on IntelligentRobotics and Applications picked“Improving Robot Behavior”as its central sub-ject. Building on the success of the previous ICIRA conference series in Wuhan,China, Singapore and Shanghai, China, the renowned conference left Asia forthe first time and took place between December 6–8, 2011 in Aachen, Germany.On the one hand, ICIRA 2011 aimed to strengthen the link between differentdisciplines developing and/or using robotics and its applications. On the otherhand, it improved the connection between different perspectives on the field ofrobotics - from fundamental research to the industrial usage of robotics.

The response from the scientific community was great and after an extensivereview 122 papers were selected for oral presentation at the conference. These high-quality papers from international authors cover a broad variety of topics, resem-bling the state of the art in robotic research. The papers accepted for the conferenceare presented in this volume of Springer’s Lecture Notes in Artificial Intelligence.The volume is organized according to the conference sessions. The sessions covera wide field of robotic research including topics such as “Robotics in Education”,“Human–Robot-Interaction”and“Bio-inspired Robotics”as well as“Robotics As-sembly Applications”,“Parallel Kinematics”or“Multi-Robot Systems”.

We would like to thank all authors and contributors who supported ICIRA2011 and the organization team under the direction of Max Haberstroh andRalph Kunze. Our special gratitude goes to the International Advisory Commit-tee and Program Chairs for their help and guidance, as well as the many externalreviewers who helped to maintain the high quality the conference demonstratedin the past three years. Our particular thanks goes to the keynote speakers Rudi-ger Dillmann (KIT, Germany), Dennis Hong (Virginia Tech, USA) and BradleyNelson (ETH Zurich, Switzerland) for their inspiring talks.

December 2011 Sabina JeschkeHonghai Liu

Daniel Schilberg

Conference Organization

Conference Chair

Sabina Jeschke RWTH Aachen University, Germany

Conference Co-chair

Xiangyang Zhu Shanghai Jiao Tong University, China

Program Chairs

Ulrich Epple RWTH Aachen University, AachenStefan Kowalewski RWTH Aachen University, Aachen

Program Co-chairs

Honghai Liu University of Portsmouth, UKJangmyung Lee Pusan National University, Republic of KoreaChun-Yi Su Concordia University, Canada

International Advisory Committee

Tamio Arai University of Tokyo, JapanHegao Cai Harbin Institute of Technology, ChinaToshio Fukuda Nagoya University, JapanKlaus Henning RWTH Aachen University, GermanyHuosheng Hu Essex University, UKOussama Khatib Stanford University, USAJurgen Leopold Huazhong University of Science and

Technology, ChinaMing Li National Natural Science Foundation of China,

ChinaPeter Luh Connecticut University, USAJun Ni University of Michigan, USANikhil R. Pal Indian Statistical Institute, IndiaGrigory Panovko Russian Academy of Science, RussiaMohammad Siddique Fayetteville State University, USAXinyu Shao Huazhong University of Science and

Technology, ChinaShigeki Sugano Waseda University, JapanMichael Wang Chinese University of Hong Kong, China

VIII Conference Organization

Kevin Warwick University of Reading, UKBogdan M. Wilamowski Auburn University, USAMing Xie Nanyang Technological University, SingaporeYoulun Xiong Huazhong University of Science and

Technology, ChinaLotfi Zadeh California University of Berkeley, USA

Conference Area Chairs

Andrew Adamatzky University of the West of England, UKShamsudin H.M. Amin Universiti Teknologi Malaysia, MalaysiaNikos A. Aspragathos University of Patras, GreecePhilippe Bidaud Universite Pierre and Marie Curie, FranceDarwin G. Caldwell Italian Institute of Technology, ItalyJan-Olof Eklundh Center for Autonomous Systems, SwedenAshraf M. Elnagar University of Sharjah, United Arab EmiratesHubert Gattringer Johannes Kepler University Linz, AustriaVladimir Golovko Brest State Technical University,

Republic of BelarusJwusheng Hu National Chiao Tung Universty, TaiwanKarel Jezernik University of Maribor, SloveniaPetko Kiriazov Bulgarian Academy of Sciences, BulgariaHeikki Koivo Helsinki University of Technology, FinlandKrzysztof Koz�lowski Poznan University of Technology, PolandMaarja Kruusmaa Tallinn University of Technology, EstoniaDirk Lefeber Vrije Universiteit Brussel, BelgiumYangmin Li University of Macau, MacauBruce MacDonald University of Auckland, New ZealandEric T. Matson Purdue University, USAIvan Petrovic University of Zagreb, CroatiaMiguel A. Salichs Universidad Carlos III de Madrid, SpainJim Torresen University of Oslo, NorwayLaszlo Vajta Budapest University of Technology and

Economics, HungaryHolger Voos University of Luxembourg, LuxembourgCees Witteveen Delft University of Technology,

The NetherlandsChangjiu Zhou Singapore Polytechnic, Republic of Singapore

Conference Special Session Chair

Naoyuki Kubota Tokyo Metropolitan University, Japan

Conference Organization IX

International Program Committee

Fakhreddine Ababsa, FranceEhsan Aboosaeedan, IranSadek Crisostomo Absi Alfaro, BrazilCihan Acar, JapanCarlos Antonio Acosta Calderon,

SingaporeNitin Afzulpurkar, ThailandMojtaba Ahmadi, CanadaAndika Aji Wijaya, MalaysiaOtar Akanyeti, ItalyBerkant Akin, TurkeyMohammad Al Janaideh, JordanMohamed Al Marzouqi, UAEAhmed Al-ArajiAmna AlDahak, UAEKhalid A.S. Al-Khateeb, MalaysiaKaspar Althoefer, UKErdinc Altug, TurkeyFarshid Amirabdollahian, UKCecilio Angulo, SpainSherine Antoun, AustraliaSilvia Appendino, ItalyPhilippe S. Archambault, CanadaKartik Ariyur, USAPanagiotis Artemiadis, USAJoonbum Bae, USAFeng Bai, ChinaSubhasis Banerji, SingaporeSven Behnke, GermanyNicola Bellotto, UKCindy Bethel, USARichard J. Black, USAMisel Brezak, CroatiaElizabeth Broadbent, New ZealandMagdalena Bugajska, USADarius Burschka, GermanyQiao Cai, USABerk Calli, The NetherlandsJiangtao Cao, ChinaZhiqiang Cao, ChinaDavid Capson, CanadaBarbara Caputo, SwitzerlandGuillaume Caron, FranceAuat Cheein, Argentina

Xiaopeng Chen, ChinaIan Chen, New ZealandZhaopeng Chen, GermanyWenjie Chen, SingaporeYouhua Chen, USADimitrios Chrysostomou, GreeceXavier Clady, FranceBurkhard Corves, GermanyDaniel Cox, USAJacob Crandall, UAERobert Cupec, CroatiaBoris Curk, SloveniaMarija Dakulovic, CroatiaKonstantinos Dalamagkidis, GermanyFadly Jashi Darsivan, MalaysiaKamen Delchev, BulgariaHua Deng, ChinaMing Ding, JapanHao Ding, GermanyCan Ulas Dogruer, TurkeyHaiwei Dong, JapanZhenchun Du, ChinaHadi ElDaou, EstoniaMartin Esser, GermanyAndres Faına, SpainYongchun Fang, ChinaFaezeh Farivar, IranEhsan Fazl-Ersi, CanadaYing Feng, CanadaLucia Fernandez Cossio, SpainManuel Fernandez-Carmona, SpainKevin Fite, USAAntonio Frisoli, ItalyZhuang Fu, ChinaVelappa Gounder Ganapathy, MalaysiaZhen Gao, CanadaAntonios Gasteratos, GreeceYiannis Georgilas, UKHu Gong, ChinaDongbing Gu, UKLiwen Guan, ChinaLei Guo, ChinaAlvaro Gutierrez, SpainNorihiro Hagita, Japan

X Conference Organization

Hassan Haleh, IranKenji Hashimoto, JapanMitsuhiro Hayashibe, FrancePatrick Henaff, FranceSophie Hennequin, FranceDominik Henrich, GermanyK.V. Hindriks, The NetherlandsVesa Holtta, FinlandMasaaki Honda, JapanTianjiang Hu, ChinaYong’an Huang, ChinaCong-Hui Huang, TaiwanMathias Husing, GermanyDetelina Ignatova, BulgariaAtsutoshi Ikeda, JapanAkira Imada, BelarusMircea Ivanescu, RomaniaEdouard Ivanjko, CroatiaYumi Iwashita, JapanPatric Jensfelt, SwedenSeonghee Jeong, JapanLi Jiang, ChinaBahram Jozi, AustraliaTakahiro Kagawa, JapanYasuhiro Kakinuma, JapanKaneko Kaneko, JapanPizzanu Kanongchaiyos, ThailandShigeyasu Kawaji, JapanEunyoung Kim, USAChyon Hae Kim, JapanBalint Kiss, HungaryAndreja Kitanov, CroatiaBin Kong, ChinaPetar Kormushev, ItalyAkio Kosaka, USAVolker Krueger, DenmarkNaoyuki Kubota, JapanChung-Hsien Kuo, TaiwanBela Lantos, HungaryKiju Lee, USAKristijan Lenac, CroatiaGang Li, ChinaKang Li, UKZhijun Li, ChinaQinchuan Li, ChinaBin Li, China

Feng-Li Lian, TaiwanGeng Liang, ChinaChyi-Yeu Lin, TaiwanWei Liu, ChinaJindong Liu, UKJia Liu, ChinaXin-Jun Liu, ChinaBingbing Liu, SingaporeBenny Lo, UKYunjiang Lou, MacaoLeena Lulu, UAEDominic MaestasElmar Mair, GermanyTakafumi Matsumaru, JapanJouni Kalevi Mattila, FinlandJohannes Mayr, AustriaAbdul Md Mazid, AustraliaEmanuele Menegatti, ItalyQinhao MengHuasong Min, ChinaLei Min, ChinaSeyed Mohamed Buhari Mohamed

Ismail, Brunei DarussalamHyungpil Moon, Republic of KoreaRainer Muller, GermanyHyun MyungHiroyuki Nakamoto, JapanLazaros Nalpantidis, GreeceJohn Nassour, FranceAndreas C. Nearchou, GreeceSamia Nefti-Meziani, UKDuc Dung Nguyen, Republic of KoreaHirotaka Osawa, JapanMohammadreza Asghari Oskoei, UKChee Khiang Pang, SingaporeChristopher Parlitz, GermanyFederica Pascucci, ItalyFernando Lobo Pereira, PortugalAnton Satria Prabuwono, MalaysiaFlavio Prieto, ColombiaHong Qiao, ChinaMd. Jayedur Rashid, AASS, SwedenSushil Raut, IndiaNilanjan Ray, CanadaRobert Richardson, UKRoland Riepl, Austria

Conference Organization XI

Jorge Rivera-Rovelo, MexicoFabrizio Rocchi, ItalyStephen Rock, USAAndreja Rojko, SloveniaJuha Roning, FinlandAnis Sahbani, FranceSebastien Saint-Aime, FranceElsayed Sallam, EgyptMarti Sanchez-Fibla, SpainIngrid Schjolberg, NorwayKosuke Sekiyama, JapanNaserodin Sepehry, IranXinjun Sheng, ChinaDesire Sidibe, FrancePonnambalam Sivalinga G., MalaysiaJorge Solis, JapanKai-Tai Song, TaiwanPeter Staufer, AustriaGiovanni Stellin, ItalyChun-Yi Su, CanadaAnan Suebsomran, ThailandJussi Suomela, FinlandYoshiyuki Takahashi, JapanYuegang Tan, ChinaLi Tan, USABo Tao, ChinaKalevi Tervo, FinlandChing-Hua Ting, TaiwanFederico Tombari, ItalyAksel Andreas Transeth, NorwayNikos Tsourveloudis, GreeceAkira Utsumi, JapanKalyana Veluvolu, Republic of KoreaIvanka Veneva, BulgariaAihui Wang, JapanXiangke Wang, ChinaHao Wang, ChinaShuxin Wang, China

Furui Wang, USAGuowu Wei, UKStephen Wood, USAHongtao WuXiaojun Wu, SingaporeXianbo Xiang, ChinaElias Xidias, GreeceRong Xiong, ChinaCaihua Xiong, ChinaPeter Xu, New ZealandXipeng Xu, ChinaKai Xu, ChinaJijie Xu, USAXin Xu, ChinaGuohua Xu, ChinaBing Xu, ChinaXinqing Yan, ChinaWenyu Yang, ChinaZhouping Yin, ChinaMasahiro Yokomichi, JapanKuu-Young Young, TaiwanHanafiah Yussof, MalaysiaMassimiliano Zecca, JapanJianguo Zhang, UKWenzeng Zhang, ChinaXianmin Zhang, ChinaXuguang Zhang, ChinaYingqian Zhang, The NetherlandsDingguo Zhang, ChinaYanzheng Zhao, ChinaXiaoguang Zhao, ChinaYi Zhou, SingaporeHuiyu Zhou, UKChi Zhu, JapanLimin Zhu, ChinaChun Zhu, USAChungang Zhuang, ChinaWei Zou, China

Organizing Committee

Max HaberstrohRalph KunzeChristian TummelAlicia DrogeClaudia Capellmann

Katrin OhmenRichar BosnicRobert GlashagenLarissa MullerKathrin Schoenefeld

Table of Contents – Part I

Progress in Indoor UAV

On the Way to a Real-Time On-Board Orthogonal SLAM for an IndoorUAV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Mirco Alpen, Klaus Frick, and Joachim Horn

Quadrocopter Localization Using RTK-GPS and Vision-BasedTrajectory Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Ulf Pilz, Willem Gropengießer, Florian Walder, Jonas Witt, andHerbert Werner

Five-Axis Milling Simulation Based on B-rep Model . . . . . . . . . . . . . . . . . . 22Yongzhi Cheng, Caihua Xiong, Tao Ye, and Hongkai Cheng

Robotics Intelligence

Exploration Strategies for Building Compact Maps in UnboundedEnvironments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Matthias Nieuwenhuisen, Dirk Schulz, and Sven Behnke

The Basic Component of Computational Intelligence for KUKA KR C3Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Tadeusz Szkodny

An Experimental Comparison of Model-Free Control Methods in aNonlinear Manipulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Mateusz Przybyla, Rafal Madonski, Marta Kordasz, andPrzemyslaw Herman

Industrial Robots

Research on Modular Design of Perpendicular Jointed IndustrialRobots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Lin Song and Suixian Yang

Online Path Planning for Industrial Robots in Varying EnvironmentsUsing the Curve Shortening Flow Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Marcel Huptych, Konrad Groh, and Sascha Rock

Parallel-Populations Genetic Algorithm for the Optimization of CubicPolynomial Joint Trajectories for Industrial Robots . . . . . . . . . . . . . . . . . . . 83

Fares J. Abu-Dakka, Iyad F. Assad, Francisco Valero, andVicente Mata

XIV Table of Contents – Part I

Robotics Assembly Applications

Integrative Path Planning and Motion Control for Handling LargeComponents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Rainer Muller, Martin Esser, and Markus Janssen

Automatic Configuration of Robot Systems – Upward and DownwardIntegration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Gunther Reinhart, Stefan Huttner, and Stefan Krug

Process and Human Safety in Human-Robot-Interaction – A HybridAssistance System for Welding Applications . . . . . . . . . . . . . . . . . . . . . . . . . 112

Carsten Thomas, Felix Busch, Bernd Kuhlenkoetter, andJochen Deuse

Operation Simulation of a Robot for Space Applications . . . . . . . . . . . . . . 122Hui Li, Giuseppe Carbone, Marco Ceccarelli, and Qiang Huang

Re-grasping: Improving Capability for Multi-Arm-Robot-System byDynamic Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

Burkhard Corves, Tom Mannheim, and Martin Riedel

A Parallel Kinematic Concept Targeting at More Accurate Assembly ofAircraft Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Christian Lochte, Franz Dietrich, and Annika Raatz

Dimensional Synthesis of Parallel Manipulators Based on Direction-Dependent Jacobian Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

Marwene Nefzi, Clement Gosselin, Martin Riedel,Mathias Husing, and Burkhard Corves

Rehabilitation Robotics

EMG Classification for Application in Hierarchical FES System forLower Limb Movement Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

Dingguo Zhang, Ying Wang, Xinpu Chen, and Fei Xu

Situated Learning of Visual Robot Behaviors . . . . . . . . . . . . . . . . . . . . . . . . 172Krishna Kumar Narayanan, Luis-Felipe Posada,Frank Hoffmann, and Torsten Bertram

Humanoid Motion Planning in the Goal Reaching Movement ofAnthropomorphic Upper Limb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Wenbin Chen, Caihua Xiong, Ronglei Sun, and Xiaolin Huang

Human Sitting Posture Exposed to Horizontal Perturbation andImplications to Robotic Wheelchairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

Karim A. Tahboub and Essameddin Badreddin

Table of Contents – Part I XV

Automatic Circumference Measurement for Aiding in the Estimation ofMaximum Voluntary Contraction (MVC) in EMG Systems . . . . . . . . . . . . 202

James A.R. Cannan and Huosheng Hu

Classification of the Action Surface EMG Signals Based on the DirichletProcess Mixtures Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

Min Lei and Guang Meng

Displacement Estimation for Foot Rotation Axis Using aStewart-Platform-Type Assist Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

Ming Ding, Tomohiro Iida, Hiroshi Takemura, and Hiroshi Mizoguchi

Mechanisms and their Applications

Inverse Kinematics Solution of a Class of Hybrid Manipulators . . . . . . . . . 230Shahram Payandeh and Zhouming Tang

Stiffness Analysis of Clavel’s DELTA Robot . . . . . . . . . . . . . . . . . . . . . . . . . 240Martin Wahle and Burkhard Corves

Optimum Kinematic Design of a 3-DOF Parallel Kinematic Manipulatorwith Actuation Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

Fugui Xie, Xin-Jun Liu, Xiang Chen, and Jinsong Wang

Integrated Structure and Control Design for a Flexible PlanarManipulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

Yunjiang Lou, Yongsheng Zhang, Ruining Huang, and Zexiang Li

Effects of Clearance on Dynamics of Parallel Indexing CamMechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270

Zongyu Chang, Lixin Xu, Yuhu Yang, Zhongqiang Zheng, andTongqing Pan

Design and Compliance Experiment Study of the Forging Simulator . . . . 281Pu Zhang, Zhenqiang Yao, Zhengchun Du, Hao Wang, andHaidong Yu

Design of Compliant Bistable Mechanism for Rear Trunk Lid of Cars . . . 291Shouyin Zhang and Guimin Chen

Multi Robot Systems

DynaMOC: A Dynamic Overlapping Coalition-Based MultiagentSystem for Coordination of Mobile Ad Hoc Devices . . . . . . . . . . . . . . . . . . 300

Vitor A. Santos, Giovanni C. Barroso, Mario F. Aguilar,Antonio de B. Serra, and Jose M. Soares

XVI Table of Contents – Part I

Design of a High Performance Quad-Rotor Robot Based on a LayeredReal-Time System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312

Jonas Witt, Bjorn Annighofer, Ole Falkenberg, and Uwe Weltin

Simple Low Cost Autopilot System for UAVs . . . . . . . . . . . . . . . . . . . . . . . . 324S. Veera Ragavan, Velappa Ganapathy, and Chee Aiying

A Marsupial Relationship in Robotics: A Survey . . . . . . . . . . . . . . . . . . . . . 335Hamido Hourani, Philipp Wolters, Eckart Hauck, and Sabina Jeschke

Multi-objective Robot Coalition Formation for Non-additiveEnvironments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

Manoj Agarwal, Lovekesh Vig, and Naveen Kumar

Development of a Networked Multi-agent System Based on Real-TimeEthernet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356

Xiong Xu, Zhenhua Xiong, Jianhua Wu, and Xiangyang Zhu

A Conceptual Agent-Based Planning Algorithm for the Production ofCarbon Fiber Reinforced Plastic Aircrafts by Using Mobile ProductionUnits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366

Hamido Hourani, Philipp Wolters, Eckart Hauck,Annika Raatz, and Sabina Jeschke

Robot Mechanism and Design

Trajectory Tracking and Vibration Control of Two Planar RigidManipulators Moving a Flexible Object . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376

Balasubramanian Esakki, Rama B. Bhat, and Chun-Yi Su

Concept and Design of the Modular Actuator System for the HumanoidRobot MYON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388

Torsten Siedel, Manfred Hild, and Mario Weidner

Design of a Passive, Bidirectional Overrunning Clutch for Rotary Jointsof Autonomous Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397

Manfred Hild, Torsten Siedel, and Tim Geppert

DeWaLoP-Monolithic Multi-module In-Pipe Robot System . . . . . . . . . . . . 406Luis A. Mateos and Markus Vincze

Design and Control of a Novel Visco-elastic Braking Mechanism UsingHMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416

Keith Gunura, Juanjo Bocanegra, and Fumiya Iida

Table of Contents – Part I XVII

Parallel Kinematics, Parallel Kinematics Machinesand Parallel Robotics

Topological Design of Weakly-Coupled 3-Translation Parallel RobotsBased on Hybrid-Chain Limbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426

Huiping Shen, Tingli Yang, Lvzhong Ma, and Shaobin Tao

Working Space and Motion Analysis on a Novel Planar ParallelManipulator with Three Driving Sliders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436

Huiping Shen, Wei Wang, Changyu Xue, Jiaming Deng, andZhenghua Ma

Optimal Kinematic Design of a 2-DoF Translational ParallelManipulator with High Speed and High Precision . . . . . . . . . . . . . . . . . . . . 445

Gang Zhang, PinKuan Liu, and Han Ding

Modeling and Control of Cable Driven Parallel Manipulators withElastic Cables: Singular Perturbation Theory . . . . . . . . . . . . . . . . . . . . . . . . 455

Alaleh Vafaei, Mohammad A. Khosravi, and Hamid D. Taghirad

CAD-2-SIM – Kinematic Modeling of Mechanisms Based on theSheth-Uicker Convention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465

Bertold Bongardt

Handling and Manipulation

Non-rigid Object Trajectory Generation for Autonomous RobotHandling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478

Honghai Liu and Hua Lin

Robotized Sewing of Fabrics Based on a Force Neural NetworkController . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486

Panagiotis N. Koustoumpardis and Nikos A. Aspragathos

Dynamic Insertion of Bendable Flat Cables with Variation Based onShape Returning Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496

Yuuki Kataoka and Shinichi Hirai

A Vision System for the Unfolding of Highly Non-rigid Objects on aTable by One Manipulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509

Dimitra Triantafyllou and Nikos A. Aspragathos

Tangibility in Human-Machine Interaction

Optimizing Motion of Robotic Manipulators in Interaction with HumanOperators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520

Hao Ding, Kurniawan Wijaya, Gunther Reißig, and Olaf Stursberg

XVIII Table of Contents – Part I

Haptic Display of Rigid Body Contact Using Generalized PenetrationDepth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532

Jun Wu, Dangxiao Wang, and Yuru Zhang

Assistive Robots in Eldercare and Daily Living: Automation ofIndividual Services for Senior Citizens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542

Alexander Mertens, Ulrich Reiser, Benedikt Brenken,Mathias Ludtke, Martin Hagele, Alexander Verl,Christopher Brandl, and Christopher Schlick

Key Factors for Freshmen Education Using MATLAB and LEGOMindstorms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553

Alexander Behrens, Linus Atorf, Dorian Schneider, and Til Aach

Navigation and Localization of Mobile Robot

Adaptive Dynamic Path Following Control of an Unicycle-Like MobileRobot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563

Victor H. Andaluz, Flavio Roberti, Juan Marcos Toibero,Ricardo Carelli, and Bernardo Wagner

A Study on Localization of the Mobile Robot Using Inertial Sensorsand Wheel Revolutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575

Bong-Su Cho, Woosung Moon, Woo-Jin Seo, and Kwang-Ryul Baek

Robust and Accurate Genetic Scan Matching Algorithm for RoboticNavigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584

Kristijan Lenac, Enzo Mumolo, and Massimiliano Nolich

Beacon Scheduling Algorithm for Localization of a Mobile Robot . . . . . . . 594Jaehyun Park, Sunghee Choi, and Jangmyung Lee

Position Estimation Using Time Difference of Flight of the Multi-codedUltrasonic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 604

Woo-Jin Seo, Bong-Su Cho, Woo-Sung Moon, and Kwang-Ryul Baek

Detecting Free Space and Obstacles in Omnidirectional Images . . . . . . . . 610Luis Felipe Posada, Krishna Kumar Narayanan,Frank Hoffmann, and Torsten Bertram

A Composite Random Walk for Facing Environmental Uncertainty andReduced Perceptual Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 620

C.A. Pina-Garcia, Dongbing Gu, and Huosheng Hu

Motion Design for Service Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 630Elias Xidias, Nikos A. Aspragathos, and Philip Azariadis

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639

Modeling and Control of Cable Driven Parallel

Manipulators with Elastic Cables:Singular Perturbation Theory

Alaleh Vafaei1, Mohammad A. Khosravi2, and Hamid D. Taghirad2

1 Electrical and Computer Engineering Department,2 Advanced Robotics and Automated Systems (ARAS),

Faculty of Electrical and Computer Engineering, University of Tehran,K.N. Toosi University of Technology

Abstract. This paper presents a new approach to the modeling and con-trol of cable driven parallel manipulators and particularly KNTU CDRPM.First, dynamical model of the cable driven parallel manipulator is derivedconsidering the elasticity of the cables, and then this model is rewrittenin the standard form of singular perturbation theory. This theory usedhere as an effective tool for modeling the cable driven manipulators. Next,the integrated controller, applied for control of the rigid model of KNTUCDRPM in previous researches, is improved and a composite controller isdesigned for the elastic model of the robot. Asymptotic stability analysisof the proposed rigid controller is studied in detail. Finally, a simulationstudy performed on the KNTU CDRPM verifies the closed-loop perfor-mance compared to the rigid model controller.

1 Introduction

Cable driven parallel robots are a special kind of parallel robots in which rigidlinks are replaced by cables. This has produced some advantages for cable drivenones that has attracted the attention of researches [1,2,3]. High acceleration dueto the reduced mobile mass, larger workspace, transportability and ease of as-sembly/disassembly, economical structure and maintenance are among these ad-vantages. The most important limitation of cable driven robots is that, the cablessuffer from unidirectional constraints that can only be pulled and not pushed. Inthis class of robots, the cables must be in tension in the whole workspace. Cablesare sagged under compression forces, and therefore, to enable tension forces inthe cables throughout the whole workspace, the mechanism must be designedover-constrained [4]. KNTU CDRPM is an over-constrained parallel manipula-tor that uses a novel design to achieve high stiffness, accurate positioning forhigh-speed maneuvers [5]. Controller must ensure that the cables are always inpositive tension by using an appropriate redundancy resolution scheme, [5].

The major challenge in the controller design of these robots is deformationof the cables under tension. Elongation is one kind of these deformations thatcauses position and orientation errors. Moreover, the flexibility of the cables may

S. Jeschke, H. Liu, and D. Schilberg (Eds.): ICIRA 2011, Part I, LNAI 7101, pp. 455–464, 2011.c© Springer-Verlag Berlin Heidelberg 2011

456 A. Vafaei, M.A. Khosravi, and H.D. Taghirad

Fig. 1. The KNTU CDRPM, a perspective view

lead the system to vibration, and cause the whole system to be uncontrollable [6].Although cable behavior has been the subject of researches in civil engineeringbut different use of them in parallel robots requires new studies. Cables in parallelrobots are much lighter than one used in civil engineering and usually we havelarge changes in cable length and the tension exerted to them. Reported studieson the effect of cable flexibility on modeling, optimal design and control of suchmanipulators are very limited and usually neglected.

It should be noticed that a complete dynamic model of cable robots is verycomplicated. Furthermore, such complicated models are useless for controllerdesign strategies, although they can accurately describe dynamic intrinsic char-acteristics of cables. Thus, in practice it is proposed to include only the dominanteffects in the dynamics analysis. For this reason in many robotics applications,cables mass have been neglected and cable has been considered as a rigid element[7,8]. With those assumptions the dynamics of cable driven robot is reduced tothe end-effector dynamics, that will lead to some inaccuracies in tracking errorand especially the stability of the manipulator. In this paper a more precisemodel of the cable driven robot considering cable flexibility is derived and beingused in the controller design and stability analysis. Using natural frequencies ofsystem, Diao and Ma have shown in [9] that in fully constrained cable drivenrobots the vibration of cable manipulator due to the transversal vibration ofcables can be ignored in comparison to that of cable axial flexibility. By thismeans, this model can describe the dominant dynamic characteristics of cableand can be used in the dynamic model of cable robot. Based on this observation,in this paper axial spring is used to model cable dynamics.

In this paper, considering axial flexibility in cables, a new dynamical model forcable driven robots is presented. This model is formulated in the standard formof singular perturbation theory. The most contribution of this theory in solvingthe control problems of the systems is in the modeling part [10]. By using theobtained model, the control of the system is studied. Next, the stability of the

Modeling and Control of Cable Driven Parallel Manipulators 457

system is analyzed through Lyapunov second method and it is proven that theclosed–loop system with the proposed control algorithm is stable. Finally theperformance of the proposed algorithm is examined through simulation.

2 Singular Perturbation Standard Model

The singular perturbation model of a dynamical system is a state space modelwhere the derivatives of some of the states are multiplied by a small positivescalar ε, that is [11]

x = f(x, z, ε, t) x ∈ Rn (1)

εz = g(x, z, ε, t) z ∈ Rm (2)

It is assumed that f , g have continuous derivatives along ( t , x , z , ε ) ∈ [0, t1]×D1 ×D2 × [0, ε0], on their domains D1 ⊂ Rn and D2 ⊂ Rm. Putting ε = 0, thedimension of the standard model reduces from m+ n to n, since the differentialequation (2) changes to

g(x, z, ε, t) = 0 (3)

The model (1) and (2) is an standard model, if and only if, the equation (3), hask ≥ 1 distinct real solutions:

z = hi(t, x) ∀[t, x] ∈ [0, t1] , i = 1, 2, 3, . . . (4)

This assumption ensures that the reduced model with appropriate order of n isrelated to the roots of equation (3). For achieving the i-th reduced order model,substitute (4) in (1) and assume ε = 0, then:

x = f(t, x, h(t, x), 0) (5)

This approximation is a wise simplification of the dynamic system in which thehigh frequency dynamics is neglected, which is sometimes called a quasi-steadymodel. Since the velocity of variable z i.e. z = g/ε can be a large number whileε is small and g �= 0, therefore, variable z converges rapidly to the roots ofequation g = 0, the quasi-steady form of (2). The equation (5) is often calledslow model.

3 Dynamics

Due to redundancy characteristic of KNTU CDRPM and other over–constrainedcable driven parallel manipulators, the sagging of the cables is neglected. Asimple model that can hold elastic characteristic of the cable and also can beused in controller design procedure, is to model the cable as a spring. This simplemodel can be well included in singular perturbation theory in order to derivea dynamic model for KNTU CDRPM considering elasticity of the cables. Inwhat follows, we will first describe the dynamics of rigid robot briefly and thendynamic equations of the elastic system are derived using rigid ones. In the nextstep the dynamics equations are formulated in the standard form of singularperturbation theory.

458 A. Vafaei, M.A. Khosravi, and H.D. Taghirad

3.1 Dynamics with Ideal Cables

The rigid model of parallel robots can be formulated into the general form of[12]:

M(x)x + C(x, x)x+G(x) = JT τ (6)

in which, x is the vector of generalized coordinates showing the position andorientation of the end-effector, M(x) is a 6×6 matrix called mass matrix, C(x, x)is a 6× 6 matrix representing the Coriolis and centrifugal forces, G(x) is a 6× 1vector of gravitational forces, J n×6 denotes the Jacobian matrix, τ n×1 is thecable tension vector. n is equal to the number of cables and for KNTU CDRPMit is equal to 8. The actuator dynamics can be represented as

MmL+DL+ τ = u (7)

in which, L is the n×1 cable length vector, Mm is a diagonal n×n inertia matrixof actuators, D a diagonal n×n matrix including viscous friction coefficients foractuators (pulleys), τ n×1 cable tension vector, u : n×1 actuator input vector.Use equations (6) and (7) to derive

Meq(x)x + Ceq(x, x)x+Geq(x) = JTu (8)

in which,Meq(x) = M(x) + JTMmJ

Ceq(x, x) = C(x, x) + JTMmJ + JTDJGeq(x) = G(x)

(9)

3.2 Dynamics with Real Cables

In parallel manipulators with elastic cables, actuator position is not directlyrelated to end-effector position, and therefore, both the actuator and the end-effector positions must be taken into state vector. In other words both the cablelength in the unloaded state and the cable length under tension are taken as statevector. For modeling a parallel manipulator with n cables, we assume L1i : i =1, 2, ..., n indicate the length of i-th cable under tension and L2i : i = 1, 2, ..., nindicate the i-th cable without tension. In the case of rigid system, we have:L1i = L2i(∀ i). In vector representation

L = (L11, ..., L1n, L21, ..., L2n)T = (LT1 |LT

2 ) (10)

The kinetic energy of the system is

T =12xTM(x)xT +

12LT

2MmL2 (11)

The sum of total potential energy of the system is

P = P1 + P2(L1 − L2) (12)

Modeling and Control of Cable Driven Parallel Manipulators 459

In which P1 is the potential energy of the rigid robot and the second term, thepotential energy of the i-th cable which its elasticity is approximated with alinear spring, is as follows

P2 =12

(L1 − L2)TK(L1 − L2) (13)

and K is the matrix of the stiffness coefficients of cables. Now the Lagrangianof the system is derived by L = T − P , as

L =12xTM(x)x+

12LT

2MmL2 − P1 − 12

(L1 − L2)TK(L1 − L2) (14)

The total dynamic equations of the system is derived simply by applying theLagrange equations{

M(x)x + C(x, x) x+G(x) = JTK(L2 − L1)MmL2 + K(L2 − L1) +DL2 = u

(15)

in which, the relation between x and L1 is obtained by L1 = Jx. Furthermore,in eq. (15), K is the n × n diagonal stiffness matrix of the cables, M(x) the6× 6 inertia matrix, C(x, x) a 6× 6 matrix with Coriolis and centrifugal terms,G(x) the 6 × 1 vector of gravitational forces, J the n× 6 Jacobian matrix, Mm

the diagonal n × n inertia matrix of actuators(pulleys), D the diagonal n × nmatrix including viscous friction coefficients for actuators, and n = 8 for KNTUCDRPM.

3.3 Singular Perturbation Model

The spring stiffness matrix K which connects two equations in (15) enablesus to formulate these equations in singular perturbation form. /without loss ofgenerality, assume that all of the cables stiffness are equal. Then write the elasticforces in the cables in the form z = k(L1 − L2) , K = kI. Since the singularperturbation theory is defined usually for small terms, define ε = 1/k, thereforeε→ 0 as k → ∞. Multiplying two sides of the first line of equation (15) by M−1

and consider z = k(L1 − L2), we have{x = −M−1(x)JT z −M−1(x)(C(x, x)x+G(x))−εz = M−1

m z −M−1m DL2 +M−1

m u− L1(16)

Considering the following equations,

L2 = L1 − εz

L1 = Jx

L1 = Jx+ J x

(17)

We can summarize equation (16), which is in the standard form of singularperturbation theory in the form{

x = a1(x, x) +A1(x)zεz = a2(x, x, εz) +A2(x)z +B2u

(18)

460 A. Vafaei, M.A. Khosravi, and H.D. Taghirad

InverseKinematics

PD TJe

pwKx

e

Trajectory RedundancyK

dx

dx e wF xF F uLxTrajectory

PlanningCDRPM

RedundancyResolution

vwKdx

dxw x x

x

IDCIDCF

IDC

Fig. 2. The cascade control scheme

In which

A1 = −M−1(x)JT

a1 = −M−1(x)(C(x, x)x+G(x))a2 = −εM−1

m Dz +M−1m DJx− JM−1(x)(C(x, x) +G(x)) + J x

A2 = −(J(x)M−1(x)JT (x) +M−1m ) ,

B2 = −M−1m

Note that the rigid model is the marginal mode of the elastic model of eq. (6),when the stiffness of the cables tends to infinity or ε→ 0.

4 Control

4.1 Control Law for the Rigid Model

The controller applied to the rigid model is a combination of two control loopswith an inverse-dynamic controller. The first control loop is a PD controller injoint-space and the second one in work space (Fig. 2). It is shown that thiscontroller can improve the performance of the control system up to 80% com-pared to conventional single loop controllers [5]. The structure of this controlleris illustrated in Fig. 2 and the control law is defined as:

F = Fj + Fx

Fj = JT (Kpj(Ld − L) +Kvj(Ld − L))Fx = Kpw(xd − x) +Kvw(xd − x) +Meqxd +Geq + Ceqxd

u = P + Pn = (JT )†F + (I − JT †JT )ke

(19)

in which, (·)† denotes the pseudo inverse and (·)d denote the desired values. Pand Pn are defined as

F = JTP0 = JTPn

and ke is an n dimensional vector which is optimized through redundancy resolu-tion scheme, [5]. Kpj,Kvj ,Kpw and Kvw are diagonal positive definite matrices.

Modeling and Control of Cable Driven Parallel Manipulators 461

Stability Analysis of the Closed-loop System. First, let us derive the errordynamics to prove the stability of the closed-loop system using the controller inequation (19). According to the robot dynamic equations (8) and control law wecan write

Meqx+ Ceq x+Geq = Kpw(xd − x) +Kvw(xd − x) +Meqxd+Geq + Ceqxd + JT (Kpj(Ld − L) +Kvj(Ld − L))

(20)

Or,Meqe+ (Kvw + JTKvjJ)e+Kpwe+ JTKpjeL + Ceq e = 0 (21)

in which, eL = Ld −L and e = xd − x. Now, introduce a Lyapunov candidate toprove the stability of the system under control.

V =12eTMeqe+

12eTKpwe+

12eT

LKpjeL (22)

in which, Meq, Kpw and Kpj matrices are positive definite, therefore V is positivedefinite. The derivative of Lyapunov function is:

V = eTMeqe +12eT Meq e+ eTKpwe+ eT

LKpj eL (23)

Substitute the term Meq e from the dynamic equations of the system.

V = eT (−(Kvw + JTKvjJ)e−Kpwe− JTKpjeL − Ceq e)+ 1

2 eT Meq e+ eTKpw e+ eT

LKpj eL(24)

Hence,

V = −eT (Kvw + JTKvjJ)e +12eT (Meq − 2Ceq)e

= −eT (Kvw + JTKvjJ + 2JTDJ)e ≤ 0 (25)

note that JTKvjJ is a positive semi-definite (PSD) matrix, because Kvj is PDand

yT (JTKvjJ)y = yT (JTK1/2vj K

1/2vj J)y = zT z ≥ 0. (26)

Therefore, Kvw +JTKvjJ+2JTDJ which is sum of two PSD matrices and a PDmatrix, is a PD matrix. Then we can conclude V ≤ 0. Therefore, we know thatthe motion of the robot will converge to the largest invariant set that satisfiesV = 0. In this case, V = 0 results in e = 0. Therefore, from equation (21) thelargest invariant set is

Kpwe+ JTKpjeL = 0 (27)

It is shown in Appendix that J.e has the same sign of eL , hence, we can writeeL = αJe, α > 0 and then we can rewrite equation (27) in this form:

(Kpw + αJTKpjJ).e = 0, α > 0 (28)

According to the above equation and positive definiteness of (Kpw + αJTKpjJ)it is concluded that e = 0. Therefore, as time tends to infinity we have x = xd

and this means the end-effector position converges to the desired trajectory.

462 A. Vafaei, M.A. Khosravi, and H.D. Taghirad

Table 1. Geometric and Inertial Parameters of the KNTU CDRPM

Description Quantity

K: Spring stiffness matrix 100I8×8

Mm: Inertia matrix of actuators 0.006I8×8

D: Viscous friction coefficients for actuators 0.244I8×8

The parameters of controllers:

Kp = 13500, Kv = 700Kpj = 105I8×8, Kdj = 104I8×8

Kpw = 107diag(80, 50, 1000, 77.5, 14, 19.5)Kdw = 107diag(24, 9, 600, 16.5, 1.14, 5.7)

4.2 Control Law for the Elastic Model

Control of the systems with real cables can be done using a composite controlscheme that is a well-known technique in the control of singularly perturbedsystems [10]. In this framework the control effort utot consists of two main parts,i.e. u the control effort for slow subsystem, the model in eq. (8), and uf thecontrol effort for fast subsystem. Here we use a control law that is combinationof rigid model control and a PD controller for the fast dynamics

ut = u+ Kp(L1 − L2) + Kv(L1 − L2) (29)

As a practical point of view, it must be said that L1 can be measured by anencoder and L2 by a string pot. In next section, it is shown through simulationthat this controller can stabilize the closed-loop system with real cables and reachto a desired tracking error. Stability analysis of the system with this compositecontroller will be discussed in later researches.

4.3 Simulation Study

In this section, the performance of the proposed controller is demonstratedthrough simulating the KNTU CDRPM. The dynamic equations of the CDRPM

−0.4−0.2

00.2

−0.4−0.2

00.2

0

0.5

1

x(m)y(m)

z(m

)

−20

0

20

−20

0

200

0.5

1

θx(deg)θy(deg)

θ z(deg

)

Fig. 3. Desired path in the workspace

Modeling and Control of Cable Driven Parallel Manipulators 463

0 5 10−2

−1.5

−1

−0.5

0

0.5

1

1.5x 10

−4

time(sec)

Error

of Po

sition

(m)

eX

eY

eZ

0 5 10−2

−1.5

−1

−0.5

0

0.5

1

1.5

2x 10

−3

time(sec)

Error

of O

rienta

tion(d

eg)

eθX

eθY

eθZ

Fig. 4. The tracking error of the controller for elastic model

0 0.5 1 1.5

x 10−3

−0.2

0

0.2

0.4

0.6

0.8

time(sec)

Error

of P

ositio

n(m)

eX

eY

eZ

0 0.5 1 1.5

x 10−3

−50

0

50

100

150

200

250

300

time(sec)

Error

of O

rienta

tion(d

eg)

eθX

eθY

eθZ

Fig. 5. The tracking error of the rigid model controller on the elastic model

considering the elasticity of the cables are shown in eq. (15).These equations inthe standard form of singular perturbation theory are shown in eq. (18). Table1 shows robot and controller specifications, other parameters are the same aswhat is given in [5]. The desired path of the manipulator in 3D is cylindrical andis shown in Fig. 3. The tracking performance of the CDRPM using the proposedcontroller is shown in Fig. 4. As seen in this figure, the proposed control topologyis capable of reducing the tracking errors less than 0.15 millimeters in positionand less than 2 × 10−3 degrees in orientation. The tracking error of a singlecontroller for the rigid model i.e.u in eq. (19) is shown in Fig. 5 for comparison.It is obvious that this controller cannot stabilize the cable driven manipulator.

5 Conclusions

A dynamical model for cable driven manipulators considering the flexibility ofthe cables is proposed using cable model as a linear axial spring. The modelis formulated in standard form of singular perturbation theory. A compositecontrol is employed for control of cable driven manipulators, which is compositionof the controller for the rigid model and a PD controller for controlling thefast dynamics. It is shown that the rigid control law can stabilize the systemwith ideal and inflexible cables asymptotically. The efficiency of the proposedcontroller is verified through simulations on KNTU CDRPM.

464 A. Vafaei, M.A. Khosravi, and H.D. Taghirad

A Appendix

Here, we will show that Jex = J(xd − x) has the same sign of el = (�d − �),the proof will be done by reduction to the absurd ( or contradiction). Therefore,assume that they have different sign:

ld − l = αJ(xd − x) , α < 0 (30)

Therefore, ∃M � 1ε ⇒ Δl

M = αM JΔx.

ΔlM = dl and we know that dl Jdx, so from equation (30) we have:

Jdx = dl α

MJΔx (31)

dx α

MΔx (32)

Which is a wrong expression when α < 0. Thus by contradiction, we can concludethat α > 0 , i.e. J(xd−x) and (ld−l) have the same sign.

��

References

1. Albus, J., Bostelman, R., Dagalakis, N.: The nist robocrane. J. of Robotic Sys-tems 10, 709–724 (1993)

2. Kawamura, S., Kino, H., Won, C.: High-speed manipulation by using parallel wire-driven robots. Robotica, 13–21 (2000)

3. Merlet, J.-P., Daney, D.: A new design for wire driven parallel robot. In: 2nd Int.Congress, Design and Modeling of mechanical systems, Monastir (March 2007)

4. Pham, C., Yeo, S., Yang, G., Kurbanhusen, M., Chen, I.: Force-closure workspaceanalysis of cable-driven parallel mechanisms. Mechanism and Machine Theory 41,53–69 (2006)

5. Vafaei, A., Aref, M.M., Taghirad, H.D.: Integrated controller for an over-constrained cable driven parallel manipulator: Kntu cdrpm. In: IEEE Int. Conf. onRobotics and Automation (ICRA), pp. 650–655 (May 2010)

6. Khosravi, M.A., Taghirad, H.D.: Dynamics analysis and control of cable drivenrobots considering elasticity in cables. Presented at CCToMM 2011 Symposium(2011)

7. Gorman, J., Jablokow, K., Cannon, D.: The cable array robot: Theory and exper-iment. In: IEEE International Conference on Robotics and Automation (2001)

8. Alp, A.B., Agrawal, S.K.: Cable suspended robots: feedback controllers with posi-tive inputs. In: American Control Conference, pp. 815–820 (2002)

9. Diao, X., Ma, O.: Vibration analysis of cable-driven parallel manipulators. Multi-body Syst. Dyn. 21, 347–360 (2009)

10. O’Reilly, J., Kokotovic, P.V., Khalil, H.: Singular Perturbation Methods In Control:Analysis and Design. Academic Press (1986)

11. Khalil, H.K.: Nonlinear Systems, 3rd edn. Prentice-Hall (2002)12. Aref, M.M., Gholami, P., Taghirad, H.D.: Dynamic analysis of the KNTU

CDRPM: a cable driven redundant manipulator. In: IEEE/ASME InternationalConference on Mechtronic and Embedded Systems and Applications (MESA), pp.528–533 (2008)