ICINCO 2012 Final Program and Book of Abstracts · ICINCO 2012 Final Program and Book of Abstracts 9th International Conference on Informatics in Control, Automation and Robotics

ICINCO 2012 Final Program and Book of Abstracts

9th International Conference on

Informatics in Control, Automation and Robotics

Rome, Italy 28 – 31 July, 2012

Sponsored by INSTICC – Institute for Systems and Technologies of Information, Control and Communication

Co-sponsored by IFAC – International Federation of Automatic Control

In Cooperation with AAAI – Association for the Advancement of Artificial Intelligence EUROMICRO ACM – Association for Computing Machinery ACM SIGART – ACM Special Interest Group on Artificial Intelligence

Copyright © 2012 SciTePress – Science and Technology Publications All rights reserved Printed in Portugal http://www.icinco.org [email protected]

TABLE OF CONTENTS ICINCO 2012

9th International Conference on Informatics in Control, Automation and Robotics 3

Table of Contents Table of Contents .............................................................................................................................................. 3 Foreword ............................................................................................................................................................ 5 Organizing and Steering Committees ................................................................................................................ 6 Program Committee ........................................................................................................................................... 7 Auxiliary Reviewers ......................................................................................................................................... 11 Panel ................................................................................................................................................................ 12 Keynote Lectures ............................................................................................................................................. 13

Resilience of Dynamical Transportation Networks ..................................................................................... 13 Space Robotics - Guidance, Navigation and Control Challenges .............................................................. 14 Recent Advances in Physical Human-Robot Interaction ............................................................................ 15 Optimization in Design of Automated Machining Systems ......................................................................... 16

Special Sessions ............................................................................................................................................. 17 Special Session on Artificial Neural Networks and Intelligent Information Processing – ANNIIP 2012 ...... 17 Special Session on Intelligent Vehicle Controls & Intelligent Transportation Systems – IVC&ITS 2012 .... 19 Special Session on Operations Management and Decision Making in Today’s Competitive Environment – OMDM 2012 ..................................................................................................... 20

Awards ............................................................................................................................................................. 21 Selected Papers Book ..................................................................................................................................... 21 Social Event and Banquet ............................................................................................................................... 22 General Information ......................................................................................................................................... 24 Rooms Layout .................................................................................................................................................. 25 Program Layout ............................................................................................................................................... 26 Final Program and Book of Abstracts ............................................................................................................... 27

Contents ....................................................................................................................................................... 29 Saturday Sessions ....................................................................................................................................... 39 Sunday Sessions ......................................................................................................................................... 47 Monday Sessions ......................................................................................................................................... 69 Tuesday Sessions ........................................................................................................................................ 89

Notes............................................................................................................................................................... 109

FOREWORD ICINCO 2012


Foreword

This book contains the abstracts of the 9th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2012) which was sponsored by the Institute for Systems and Technologies of Information, Control and Communication (INSTICC) and held in Rome, Italy. ICINCO 2012 was co-sponsored by the International Federation of Automatic Control (IFAC) and held in cooperation with the Association for the Advancement of Artificial Intelligence (AAAI), Euromicro and the ACM Special Interest Group on Artificial Intelligence (ACM SIGART).

The ICINCO Conference Series has now consolidated as a major forum to debate technical and scientific advances presented by researchers and developers both from academia and industry, working in areas related to Control, Automation and Robotics that benefit from Information Technology.

In the Conference Program we have included oral presentations (full papers and short papers) and posters, organized in four simultaneous tracks: “Intelligent Control Systems and Optimization” , “Robotics and Automation” , “Signal Processing, Sensors, Systems Modelling and Control” and “Industrial Engineering, Production and Management” . We have included in the program three plenary keynote lectures, given by internationally recognized researchers, namely – Alessandro De Luca ( Università di Roma "La Sapienza", Italy), Munther A. Dahleh (MIT, United States), Alexandre Dolgui (Ecole des Mines de Saint-Etienne, France) and Jurek Sasiadek (Carleton University, Canada).

The meeting is complemented with three special sessions: Special Session on Artificial Neural Networks and Intelligent Information Processing (ANNIIP), Special Session on Intelligent Vehicle Controls & Intelligent Transportation Systems (IVC&ITS) and Special Session on Operations Management and Decision Making in Today’s Competitive Environment (OMDM).

ICINCO received 360 paper submissions, including special sessions, from 58 countries, in all continents. To evaluate each submission, a double blind paper review was performed by the Program Committee. Finally, only 197 papers are published in these proceedings and presented at the conference. Of these, 144 papers were selected for oral presentation (40 full papers and 104 short papers) and 53 papers were selected for poster presentation. The full paper acceptance ratio was 11%, and the oral acceptance ratio (including full papers and short papers) was 40%. As in previous editions of the Conference, based on the reviewer’s evaluations and the presentations, a short list of authors will be invited to submit extended versions of their papers for a book that will be published by Springer with the best papers of ICINCO 2012.

Conferences are also meeting places where collaboration projects can emerge from social contacts amongst the participants. Therefore, in order to promote the development of research and professional networks the Conference includes in its social program a Conference Social Event & Banquet in the evening of July 30 (Monday).

We would like to express our thanks to all participants. First of all to the authors, whose quality work is the essence of this Conference. Next, to all the members of the Program Committee and auxiliary reviewers, who helped us with their expertise and valuable time. We would also like to deeply thank the invited speakers for their excellent contribution in sharing their knowledge and vision. Finally, a word of appreciation for the hard work of the INSTICC team; organizing a conference of this level is a task that can only be achieved by the collaborative effort of a dedicated and highly capable team.

Commitment to high quality standards is a major aspect of ICINCO that we will strive to maintain and reinforce next year, including the quality of the keynote lectures, of the workshops, of the papers, of the organization and other aspects of the conference. We look forward to seeing more results of R&D work in Informatics, Control, Automation and Robotics at ICINCO 2013. Jean-Louis Ferrier, University of Angers, France

Alain Bernard, Ecole Centrale de Nantes, France

Oleg Gusikhin, Ford Research & Adv. Engineering, U.S.A.

Kurosh Madani, University of PARIS-EST Créteil (UPEC), France

ICINCO 2012 ORGANIZING AND STEERING COMMITTEES

6 9th International Conference on Informatics in Control, Automation and Robotics

Organizing and Steering Committees

Conference Chair Jean-Louis Ferrier, University of Angers, France

Program Co-Chairs Alain Bernard, Ecole Centrale de Nantes, France Oleg Gusikhin, Ford Research & Adv. Engineering, U.S.A. Kurosh Madani, University of PARIS-EST Créteil (UPEC), France

Proceedings Production Helder Coelhas, INSTICC, Portugal Andreia Costa, INSTICC, Portugal Patrícia Duarte, INSTICC, Portugal Bruno Encarnação, INSTICC, Portugal Liliana Medina, INSTICC, Portugal Carla Mota, INSTICC, Portugal Raquel Pedrosa, INSTICC, Portugal Vitor Pedrosa, INSTICC, Portugal Cláudia Pinto, INSTICC, Portugal José Varela, INSTICC, Portugal

CD-ROM Production Pedro Varela, INSTICC, Portugal

Graphics Production and Webdesigner Daniel Pereira, INSTICC, Portugal

Secretariat Marina Carvalho, INSTICC, Portugal

Webmaster Susana Ribeiro, INSTICC, Portugal

PROGRAM COMMITTEE ICINCO 2012


Program Committee

Arvin Agah, The University of Kansas, U.S.A. El-Houssaine Aghezzaf, Ghent University, Faculty of Engineering and Architecture, Belgium Arturo Hernandez Aguirre, Centre for Research in Mathematics, Mexico Eugenio Aguirre, University of Granada, Spain Hyo-Sung Ahn, Gwangju Institute of Science and Technology (GIST), Korea, Republic of Adel Al-Jumaily, University of Technology, Sydney, Australia Frank Allgower, University of Stuttgart, Germany Yacine Amirat, University Paris Est Créteil (UPEC), France Stefan Andrei, Lamar University, U.S.A. Peter Arato, Budapest University of Technology and Economics, Hungary Rui Araujo, University of Coimbra, Portugal Helder Araújo, University of Coimbra, Portugal Carles Batlle Arnau, Universitat Politècnica de Catalunya, Spain Tomas Arredondo, Universidad Tecnica Federico Santa Maria, Chile Ronald Askin, Arizona State University, U.S.A. T. Asokan, Indian Institute of Technology Madras, India Wudhichai Assawinchaichote, King Mongkut’s University of Technology Thonburi, Thailand Ali Bab-Hadiashar, RMIT University, Australia Jacky Baltes, University of Manitoba, Canada Ruth Bars, Budapest University of Technology and Economics, Hungary Laxmidhar Behera, Indian Institute of Technology, Kanpur, India Karsten Berns, University of Kaiserslautern, Germany Arijit Bhattacharya, Dublin City University, Ireland Robert Bicker, Newcastle University, U.K. Mauro Birattari, IRIDIA-CoDE, Université Libre de Bruxelles, Belgium Christian Blum, Universitat Politècnica de Catalunya, Spain Jean-Louis Boimond, ISTIA - LISA, France Magnus Boman, The Royal Institute of Technology, Sweden Alan Bowling, University of Texas at Arlington, U.S.A. Thomas Braunl, The University of Western Australia, Australia Mietek Brdys, University of Birmingham, U.K. Kevin Burn, University of Sunderland, U.K. Amaury Caballero, Florida International University, U.S.A. Javier Fernandez de Canete, University of Malaga, Spain Giuseppe Carbone, LARM - Laboratorio di Robotica e Meccatronica, Italy Alessandro Casavola, University of Calabria, Italy Riccardo Cassinis, University of Brescia, Italy Ratchatin Chanchareon, Chulalongkorn University, Thailand Siam Charoenseang, Institute of Field Robotics, Thailand Antonio Chella, Università di Palermo, Italy Peter C. Y. Chen, National University of Singapore, Singapore Wen-Hua Chen, Loughborough University, U.K. Yuh-Min Chen, National Cheng Kung University, Taiwan Albert Cheng, University of Houston, U.S.A. Graziano Chesi, University of Hong Kong, China Tsung-Che Chiang, National Taiwan Normal University, Taiwan Sung-Bae Cho, Yonsei University, Korea, Republic of Ryszard S. Choras, University of Technology & Life Sciences, Poland Chengbin Chu, Ecole Centrale Paris, France Carlos Coello Coello, CINVESTAV-IPN, Mexico James M. Conrad, University of North Carolina Charlotte, U.S.A. Yechiel Crispin, Embry-riddle Aeronautical University, U.S.A. José Boaventura Cunha, University of Trás-os-montes and Alto Douro, Portugal Prithviraj (Raj) Dasgupta, University of Nebraska, U.S.A. Michael A. Demetriou, Worcester Polytechnic Institute, U.S.A. Mingcong Deng, Tokyo University of Agriculture and Technology, Japan

ICINCO 2012 PROGRAM COMMITTEE


Rüdiger Dillmann, University of Karlsruhe, Germany Dimos Dimarogonas, KTH, Sweden Tony Dodd, The University of Sheffield, U.K. Alexandre Dolgui, Ecole des Mines de Saint-Etienne, France António Dourado, University of Coimbra, Portugal Venky Dubey, Bournemouth University, U.K. Marc Ebner, Ernst-Moritz-Arndt-Universität Greifswald, Germany Petr Ekel, Pontifical Catholic University of Minas Gerais, Brazil Mohammed El-Abd, American University of Kuwait, Kuwait Atilla Elci, Suleyman Demirel University, Turkey Simon G. Fabri, University of Malta, Malta Paolo Falcone, Chalmers University of Technology, Sweden David Fernández-Llorca, University of Alcalá, Spain Paolo Fiorini, University of Verona, Italy Juan J. Flores, University of Michoacan, Mexico Heinz Frank, Heilbronn University Campus Kuenzelsau, Germany Georg Frey, Saarland University, Germany Ponnambalam S. G., Monash University Sunway Campus, Malaysia Dragan Gamberger, Rudjer Boskovic Institute, Croatia Nicholas Gans, University of Texas at Dallas, U.S.A. Maria I. Garcia-Planas, Universitat Politecnica de Catalunya, Spain Ryszard Gessing, Silesian University of Technology, Poland Lazea Gheorghe, Technical University of Cluj-Napoca, Romania Paulo Gil, Universidade Nova de Lisboa, Portugal Maria Gini, University of Minnesota, U.S.A. Alessandro Giua, University of Cagliari, Italy Luis Gomes, Universidade Nova de Lisboa, Portugal Bhaskaran Gopalakrishnan, West Virginia University, U.S.A. Frans C. A Groen, University of Amsterdam, The Netherlands Jason Gu, Dalhousie University, Canada Kevin Guelton, University of Reims Champagne-Ardenne, France José J. Guerrero, Universidad de Zaragoza, Spain Eugenio Guglielmelli, Università Campus Bio-Medico, Italy Lei Guo, Chinese Academy of Sciences (CAS), China Thomas Gustafsson, Luleå University of Technology, Sweden Maki K. Habib, The American University in Cairo, Egypt Wolfgang Halang, Fernuniversitaet, Germany John Hallam, University of Southern Denmark, Denmark Jennifer Harding, Loughborough University, U.K. John Harris, University of Florida, U.S.A. Inman Harvey, University of Sussex, U.K. Dominik Henrich, University of Bayreuth, Germany Suranga Hettiarachchi, Indiana University Southeast, U.S.A. Victor Hinostroza, University of Ciudad Juarez, Mexico Kaoru Hirota, Tokyo Institute of Technology, Japan Wladyslaw Homenda, Warsaw University of Technology, Poland Fumiya Iida, Institute of Robotics and Intelligent Systems, ETH Zurich, Switzerland Mirjana Ivanovic, Faculty of Sciences, University of Novi Sad, Serbia Sarangapani Jagannathan, Missouri University of Science and Technology, U.S.A. Masoud Jamei, Simcyp Limited, U.K. Ray Jarvis, Monash University, Australia Thira Jearsiripongkul, Thammasat University, Thailand Myong K. Jeong, Rutgers University, U.S.A. Ping Jiang, The University of Bradford, U.K. Ivan Kalaykov, Örebro University, Sweden Michail Kalogiannakis, University of Crete, Greece Mansour Karkoub, Texas A&M University, Qatar Dusko Katic, Mihailo Pupin Institute, Serbia Behrokh Khoshnevis, University of Southern California, U.S.A. DaeEun Kim, Yonsei University, Korea, Republic of Jonghwa Kim, University of Augsburg, Germany

PROGRAM COMMITTEE ICINCO 2012


Diego Klabjan, Northwestern University, U.S.A. Ashok K. Kochhar, Aston University, U.K. Lingxue Kong, Deakin University, Australia Waree Kongprawechnon, Thammasat University, Thailand Israel Koren, University of Massachusetts, U.S.A. George L. Kovács, Hungarian Academy of Sciences, Hungary Krzysztof Kozlowski, Poznan University of Technology, Poland Mianowski Krzysztof, Politechnika Warszawska, Poland Masao Kubo, Narional Defense Academy of Japan, Japan Kin Keung Lai, City University of Hong Kong, Hong Kong Kathryn J. De Laurentis, University of South Florida, U.S.A. Soo-Young Lee, KAIST, Korea, Republic of Graham Leedham, University of New England, Australia Kauko Leiviskä, University of Oulu, Finland Kang Li, Queen’s University Belfast, U.K. Tsai-Yen Li, National Chengchi University, Taiwan Yangmin Li, University of Macau, China Gordon Lightbody, University College Cork, Ireland Huei-Yung Lin, National Chung Cheng University, T aiwan Józef Lisowski, Gdynia Maritime University, Poland Changchun Liu, The MathWorks, Inc., U.S.A. Guoping Liu, University of Glamorgan, U.K. Honghai Liu, University of Portsmouth, U.K. Savvas Loizou, Cyprus University of Technology, Cyprus Luís Seabra Lopes, University of Aveiro, Portugal Gonzalo Lopez-Nicolas, Universidad de Zaragoza, Spain Iuan-Yuan Lu, National Sun Yat-sen University, Taiwan Edwin Lughofer, Johannes Kepler University, Austria Martin Lukac, Tohoku University, Japan José Tenreiro Machado, Institute of Engineering of Porto, Portugal Anthony Maciejewski, Colorado State University, U.S.A. Nitaigour Mahalik, California State University - Fresno, U.S.A. Bruno Maione, Politecnico Di Bari, Italy Frederic Maire, Queensland University of Technology, Australia Om Malik, University of Calgary, Canada Jacek Mandziuk, Warsaw University of Technology, Poland Philippe Martinet, Ecole Centrale de Nantes, France Sonia Martínez, Mechanical and Aerospace Engineering Department, UC San Diego, U.S.A. Aníbal Matos, Faculdade de Engenharia da Universidade do Porto (FEUP), Portugal Ross McAree, University of Queensland, Australia Ken Mcgarry, University of Sunderland, U.K. Carlo Menon, Simon Fraser University, Canada Patrick Millot, University of Valenciennes, France Sara Moein, Multimedia University, Malaysia António Paulo Moreira, INESC Porto / FEUP, Portugal Vladimir Mostyn, VSB - Technical University of Ostrava, Czech Republic Zoltan Nagy, Loughborough University, U.K. Saeid Nahavandi, Deakin University, Australia Andreas Nearchou, University of Patras, Greece Luciana Nedel, Universidade Federal do Rio Grande do Sul (UFRGS), Brazil Sergiu Nedevschi, Technical University of Cluj-Napoca, Romania Maria Neves, Instituto Superior de Engenharia do Porto, Portugal Juan A. Nolazco-Flores, ITESM, Campus Monterrey, Mexico Andrzej Obuchowicz, University of Zielona Góra, Poland José Valente de Oliveira, Universidade do Algarve, Portugal Christos Panayiotou, University of Cyprus, Cyprus Igor Paromtchik, INRIA, France Bozenna Pasik-Duncan, University of Kansas, U.S.A. Pierre Payeur, University of Ottawa, Canada Fernando Lobo Pereira, University of Porto, Portugal Marco Antonio Arteaga Perez, Universidad Nacional Autonoma de Mexico, Mexico

ICINCO 2012 PROGRAM COMMITTEE


Jeff Pieper, University of Calgary, Canada Selwyn Piramuthu, University of Florida, U.S.A. Angel P. Del Pobil, Universitat Jaume I, Spain Marie-Noëlle Pons, CNRS, France Raul Marin Prades, Jaume I University, Spain Robert Richardson, University of Leeds, U.K. Julio Elias Normey Rico, Universidade Federal de Santa Catarina, Brazil Juha Röning, University of Oulu, Finland Mehdi Roopaei, University of Texas at San Antonio, U.S.A. Danilo De Rossi, University of Pisa, Italy Carlos Sagüés, University of Zaragoza, Spain Mehmet Sahinkaya, University of Bath, U.K. Antonio Sala, Universidad Politecnica de Valencia, Spain Abdel-Badeeh Mohamed Salem, Ain Shams University, Egypt Marcello Sanguineti, University of Genoa, Italy Elena De Santis, University of L’Aquila, Italy Srikanth Saripalli, Arizona State University, U.S.A. Medha Sarkar, Middle Tennessee State University, U.S.A. Nilanjan Sarkar, Vanderbilt University, U.S.A. Jurek Sasiadek, Carleton University, Canada Hideyuki Sawada, Kagawa University, Japan Daniel Sbarbaro, Universidad de Concepcion, Chile Klaus Schilling, University Würzburg, Germany Carla Seatzu, University of Cagliari, Italy Ali Selamat, Universiti Teknologi Malaysia, Malaysia João Sequeira, Instituto Superior Técnico / Institute for Systems and Robotics, Portugal Michael Short, Teesside University, U.K. Dan Simon, Cleveland State University, U.S.A. Olivier Simonin, INRIA - LORIA, France Joaquin Sitte, Queensland University of Technology, New Zealand Adam Slowik, Koszalin University of Technology, Poland Andrzej Sluzek, Khalifa University of Science Technology and Research, U.A.E. Qing Song, Nanyang Technological University, Singapore Stefano Squartini, Università Politecnica delle Marche, Italy Burkhard Stadlmann, Upper Austria University of Applied Sciences, Wels, Austria Sergiu Stan, Technical University of Cluj-Napoca, Romania Tarasiewicz Stanislaw, Université Laval, Canada A. Frank Van Der Stappen, Utrecht University, The Netherlands J. D. Stigter, Wageningen University, The Netherlands Olaf Stursberg, University of Kassel, Germany Kazuya Takeda, Nagoya University, Japan Tianhao Tang, Shanghai Maritime University, China József K. Tar, Óbuda University, Hungary Abdelhamid Tayebi, Lakehead University, Canada Daniel Thalmann, Nanyang Technological University, Singapore Haydn A. Thompson, University of Sheffield, U.K. Gui Yun Tian, Newcastle University, U.K. N. G. Tsagarakis, Istituto Italiano di Tecnologia, Italy Avgoustos Tsinakos, University of Kavala Institute of Technology / Athabasca University of Canada, Greece Antonios Tsourdos, Cranfield University (Cranfield Defence and Security), U.K. Ali Emre Turgut, Universite Libre de Bruxelle, Belgium Angel Valera, Universidad Politécnica de Valencia, Spain Bram Vanderborght, Vrije Universiteit Brussel, Belgium Eloisa Vargiu, Barcelona Digital, Spain Ramiro Velazquez, Universidad Panamericana, Mexico Damir Vrancic, Jožef Stefan Institute, Slovenia Bernardo Wagner, Leibniz Universität Hannover, Germany James Whidborne, Cranfield University, U.K. Sangchul Won, Pohang University of Science and Technology, Korea, Republic of Peter Xu, The University of Auckland, New Zealand

PROGRAM COMMITTEE AND AUXILIARY REVIEWERS ICINCO 2012


Xun W. Xu, University of Auckland, New Zealand Marek Zaremba, Université du Québec (UQO), Canada Janan Zaytoon, University of Reims Champagne Ardennes, France Du Zhang, California State University, U.S.A. Primo Zingaretti, Università Politecnica delle Marche, Italy Loredana Zollo, Università Campus Bio-Medico, Italy Argyrios Zolotas, Loughborough University, U.K.

Auxiliary Reviewers

Francesca Cordella, University of Naples Federico II, Italy Ayan Dutta, University of Nebraska at Omaha, U.S.A. Daniel Gutierrez Gomez, Universidad de Zaragoza, Spain K. R. Guruprasad, University of Nebraska, U.S.A. Chin-Tien Huang, NCKU, Taiwan Janyl Jumadinova, University of Nebraska at Omaha, U.S.A. Aravindhan Krishnan, ASU, U.S.A. Paolo Lino, Politecnico Di Bari, Italy Angelica Muñoz-Meléndez, INAOE, Mexico Gulherme Raffo, Federal University of Santa Catarina, Brazil Daniele Rosa, Università degli Studi di Cagliari, Italy Jinglin Shen, University of Texas at Dallas, U.S.A. Merhan Zareh, University of Cagliari, Italy

ICINCO 2012 PANEL


Panel

Saturday, 28 15:15 – 16:30 Room: Plenary

Title: (To be defined)

Panel Chair: Kurosh Madani, University of PARIS-EST Créteil (UPEC), France

KEYNOTE LECTURES ICINCO 2012


Keynote Lectures

Sunday 29 09:45 – 10:45 Room: Plenary

Resilience of Dynamical Transportation Networks

Munther A. Dahleh MIT

U.S.A.

In this talk, we present recent results on the stability and robustness properties of transportation networks for various agents' route-choice behavior. We perform the analysis within a dynamical system framework over a directed acyclic graph between a single origin-destination pair. The dynamical system is composed of ordinary differential equations (ODEs), one for every link of the graph. Every ODE is a mass balance equation for the corresponding link, where the inflow term is a function of the agents’ route-choice behaviour and the arrival rate at the base node of that link, and the outflow term is a function of the congestion properties of the link. We propose a novel decision framework, where the drivers combine their historical knowledge about the global congestion levels with real-time local information to make route choice decisions at every node. We show that, if the rate of update of global information is sufficiently slow and if the drivers make route choice decisions cooperatively, then the Wardrop equilibrium is globally asymptotically stable. We then study the resilience of the flow transferring capability of the whole network under disturbances that reduce the flow carrying capacity of the links. In particular, we characterize various margins of resilience of the network with respect to the topology, 'pre-disturbance' equilibrium, and agents' local route-choice behavior. We show that the cooperative route choice behavior is maximally resilient in this setting. We also setup a simple convex optimization problem to find the most resilient 'pre-disturbance' equilibrium for the network and determine link-wise tolls that yield such an equilibrium. Finally, we extend our analysis to link-wise outflow functions that accommodate the possibility of cascaded failures and study the effect of such phenomena on the margins of resilience of the network.

Munther A. Dahleh received the B.S. degree from Texas A \& M university, College Station, Texas in 1983, and his Ph.D. degree from Rice University, Houston, TX, in 1987, all in Electrical Engineering. Since then, he has been with the Department of Electrical Engineering and Computer Science, MIT, Cambridge, MA, where he is now the Associate Department Head. Previously, he was the acting director of the Laboratory for Information and Decision Systems. He has been a visiting Professor at the Department of Electrical Engineering, California Institute of Technology, Pasadena, CA, for the Spring of 1993. He has held consulting positions with several companies in the US and abroad. Dr. Dahleh is interested in problems at the interface of robust control, filtering, information theory, and computation which include control problems with communication constraints and distributed agents with local decision capabilities. In addition to methodology development, he has been interested in the application of distributed control in the future electric grid and the future transportation system with particular emphasis in the management of systemic risk. He is also interested in various problems in network science including distributed computation over noisy network as well as information propagation over complex engineering and social networks. He is the co-author (with Ignacio Diaz-Bobillo) of the book "Control of Uncertain Systems: A Linear Programming Approach", published by Prentice-Hall, and the co-author (with Nicola Elia) of the book "Computational Methods for Controller Design" published by Springer.

ICINCO 2012 KEYNOTE LECTURES


Sunday 29 14:30 – 15:30 Room: Plenary

Space Robotics - Guidance, Navigation and Control Challenges

Jurek Sasiadek Carleton University

Canada

This paper presents methods and technologies usedin guidance, navigation and control of space robotics. It addresses some basic issues related to control of space manipulators and autonomous flying and ground travelling vehicles. Space manipulators such as, Remote Manipulator System (RMS), Space Station Remote manipulator System (SSRMS) and Special Purpose Dextrous Manipulator (SPDMD) were used successfully for variety of space tasks. Robotic manipulators in Space pose several navigation and control challenges. Robots structures and joints are flexible and highly nonlinear and their designs include several degrees of freedom (DOF). Each degree of freedom has to be controlled with high accuracy to achieve satisfactory positioning.

Planetary rovers are used in space to explore the planets and moons. There are many types of robotics vehicles. Their main objective is to explore the planets, take samples of ground and rocks at desired location. Often robotics rovers are directed to take images of the surroundings, mapping, as well as, performing other scientific measurements. Accurate localization of autonomous robots in featureless environment is very difficult and often requires advanced methods of guidance and navigation. It isworth noting that existing sensor as gyroscopes, accelerometers and other inertial systems may not work properly when robot is travelling with low speed. Moreover, odometric sensors such as, encoders and tachometers may also not work properly in the hostile environment of outer planets. In such environment, sensor fusion becomes one of the most important technologies used for guidance, localization, and navigation of planetary robotic vehicles. Integrating several measurements and several types of sensors to retrieve accurate data is one of the most successful technologies used in autonomous robotics.

Free – flying and orbiting robots also will play an important role in Space exploration and servicing. De‐orbiting satellites and servicing space floating vehicles on orbit is an important part of robotics activities in Space. Although, there are not too many successful applications of such types of robots, it is expected that in future they will play much more important role. General methods of navigation are shown and discussed and their possible application in Space are presented and discussed.

Jurek Sasiadek received his Master’s and Ph.D. degree from University of Technology in Wroclaw. He has held academic appointments with several academic institutions in Canada as University of Alberta in Edmonton, Concordia University and McGill University in Montreal, as well as Carleton University in Ottawa. For several years, he was a Technical Director at the Alberta Research Council, Department of Advanced Technologies in Calgary and Canadian Space Agency, Robotics and Automation Group in Ottawa. He has worked as a Consultant for many industrial companies like General Dynamics, Lockheed-Martin, Nortel and Bombardier. Since 1982, he has been at Carleton University, Ottawa, Ontario, Canada where he is a Professor of Aerospace Engineering in the Department of Mechanical and Aerospace Engineering. His research interests includes: robotics, space robotics, guidance, navigation and control of mobile and flying robots, sensor and data fusion, adaptive control, aircraft and spacecraft control, flexible structure control, nonlinear control. Professor J. Sasiadek is a Director of American Automatic Control Council (AACC) and member of AIAA Guidance, Navigation and Control Technical Committee (since 1989). He is also an Associate Editor and International Advisor of AIAA Journal of Guidance, Dynamics and Control. He was a Program Chair (1992) in Phoenix, Arizona and General Chair (2002) in Montreal, Quebec of AIAA Guidance, Navigation and Control Conference. He was also many times a member of ACC (American Control Conference) Program Committee. Professor Jurek Sasiadek is currently serving as a member of IFAC Council and he is a Past Chair of IFAC Robotics Technical Committee (2000-2006). Also, he is an Associate Editor of the IFAC CEP and he is a past Associate Editor of IFAC Automatica (1994-2004). He is also involved in IFAC Intelligent Autonomous Vehicles TC and Aerospace TC. In 1998 he organized as part of IFAC Canada and IFAC Aerospace TC activities a successful IFAC Workshop on Space Robotics (SPRO’98). This event took place at Canadian Space Agency in Montreal. He chaired an International Program Committee for IFAC SYROCO’03 in Wroclaw. At the national level, he is a longtime member of IFAC Canada Board of Directors. Currently, he is a Vice Chair of the IFAC Canada. At the local level, Professor J. Sasiadek is a Chair of IEEE Joint Robotics and Control Systems Chapter in Ottawa.

KEYNOTE LECTURES ICINCO 2012


Monday 30 09:45 – 10:45 Room: Plenary

Recent Advances in Physical Human-Robot Interaction

Alessandro De Luca Università di Roma “La Sapienza”

Italy

In this lecture, some control issues related to physical Human-Robot Interaction (pHRI) will be presented and discussed. In order to achieve the goal of a safer, dependable, and high-performance collaboration of robots and humans in industrial and professional service tasks, an integrated approach is needed, where mechatronic aspects, sensory information, online planning, and reactive control are combined in a single framework. At the level of mechanical design, lightweight robots with compliant joints are being used for safety purposes in pHRI, possibly driven by variable stiffness actuation (VSA). Compliance allows accommodating in a better way the energy transfer in unforeseen impacts. For recovering performance in terms of dynamic accuracy, feedback linearization control is a viable choice. With this control layer, robots having joints with constant stiffness can execute smooth desired trajectories as if they were rigid. For VSA-based robots, the same nonlinear design is able to achieve simultaneous and decoupled control of both the link motion and the desired evolution of joint compliance. It is then possible to implement a strategy in which mechanical compliance is suitably increased when the robot is moving at fast speed, so as to reduce the risk of potential injuries due to accidental collisions. Robots in physical HRI tasks have to deal with highly dynamic, time-varying, and uncertain environments. Two additional control layers are introduced for this purpose. The first is driven by exteroceptive sensing (e.g., cameras or depth sensors) and addresses the on-line collision avoidance of humans and static/moving obstacles. We will present in particular an efficient method that relies on distance computations in the depth space and whole-body artificial potential schemes for robot reaction. The second layer uses only proprioceptive sensing (as obtained by robot encoders, and possibly joint torque sensors) to detect and handle, under strict real-time constraints, unintentional physical collisions that may be unavoidable. The collision detection and the associated reflex reaction of the robot are determined by a vector signal of residuals (as in fault detection), which monitors the generalized momentum of the robot. The final layer shifts the control objective from a safe coexistence of human and robot to their actual physical collaboration, i.e., the robot should act as a co-worker performing useful tasks jointly with the human. In these situations, control of the exchanged forces and/or of human-robot impedance in the contact area is needed, typically resorting to combined information from external and internal sensors. In addition, in order to activate a collaborative phase, the robot needs to recognize when the human intention involves an explicit physical interaction. As a result, the above range of pHRI situations asks for a portfolio of consistent robot control strategies that embed the outlined hierarchy of safety, co-existence, and active collaboration. Some illustrative examples will be given, including recent results on recognition of human motion intention, kinestethic learning of robot motion, and reactive action generation patterns from the on-going European research project SAPHARI.

Alessandro De Luca is Professor of Robotics in the Department of Computer, Control, and Management Engineering at the University of Roma "La Sapienza". His research interests include control of human-robot interaction, dynamic modeling and control of robots with flexible components, visual servoing, motion planning and nonlinear control of robot manipulators and wheeled mobile robots, fault detection and isolation in robotics. He has published over 170 journal and conference papers and book chapters, receiving two best paper awards at the ICRA 1998 and IROS 2006 conferences. He was a recipient of the Helmoltz-Humboldt Award in 2005 and is a Fellow of IEEE (class of 2007). He has been the first Editor-in-Chief of the IEEE Transactions on Robotics, General Chair of IEEE ICRA 2007, and is the current Vice-President for Publications of the IEEE Robotics and Automation Society. He is a member of the Search Committee for Physical Sciences of the Körber European Science Award and a Panel Chair of the European Research Council for the Advanced Grants. He is the coordinator of the FP7-ICT European project SAPHARI. For more details: http://www.dis.uniroma1.it/~deluca.

ICINCO 2012 KEYNOTE LECTURES


Tuesday 31 16:00 – 17:00 Room: Plenary

Optimization in Design of Automated Machining Systems

Alexandre Dolgui Ecole des Mines de Saint-Etienne

France

The aim is to develop advances optimization techniques from the perspective of combinatorial design ofautomated manufacturing systems. The suggested methodology is based upon advanced line balancingmethods and process planning and equipment selection algorithms. The main results are based on exact mathematical programming methods and their intelligent coupling with heuristics and metaheuristics. A decision-aid system based on these methods will be presented. This system is employed for the design of mass production dedicated machining lines as well as for reconfigurable manufacturing systems.

Alexandre Dolgui is the Deputy Director for Research of the Henri Fayol Institute at the Ecole des Mines de Saint-Etienne (France). He has coauthored 5 books, edited 11 additional books or conference proceedings, and published about 135 papers in refereed journals, 24 book chapters and over 250 papers in conference proceedings. Pr. Dolgui is the Editor-in-Chief of the International Journal of Production Research, an Area Editor of the Computers & Industrial Engineering, Associate Editor of Omega–The International Journal of Management Science and past Associate Editor of IEEE Transactions in Industrial Informatics, Board Member of numerous other journals, Chair of the IFAC Technical Committees on Manufacturing Modelling for Management and Control.

ANNIIP SPECIAL SESSION SCOPE AND TOPICS ICINCO 2012


Special Sessions Sunday 29 16:45 – 18:15 Room: Barcelona

Special Session on Artificial Neural Networks and Intelligent Information Processing – ANNIIP 2012

Chair Kurosh Madani The University of PARIS-EST Créteil (UPEC), France

SCOPE AND TOPICS Theoretical, applicative and technological challenges, emanating from nowadays' industrial, socioeconomic or environment needs, open every day new dilemmas to solved and new challenges to defeat. Bio-inspired Artificial Intelligence and related topic have shown its astounding potential in overcoming the above-mentioned needs. It is a fact and at the same time a great pleasure to notice that the ever-increasing interest of both confirmed and young researchers on this relatively juvenile science, upholds a reach multidisciplinary synergy between a large number of scientific communities making conceivable a forthcoming emergence of viable solutions to these real-world complex challenges. ANNIIP takes part in appealing intellectual dynamics created around bio-inspired Artificial Intelligence by offering a privileged space to refit and exchange the knowledge about state of the art and further theoretical advances, new experimental discoveries and novel technological improvements in this promising area. The goal is to bring together different representative actors (from academia, industry, government agencies, etc...) to exchange ideas, to debate divergences and to construct convergences around these propitious concepts.

Topics of Interest Topics of interest include, but are not limited to:

- Bio-inspired Artificial Neural Networks - Hybrid Information Processing - Hierarchical Artificial Neural Network Models and Systems - Intelligent Sensors and Smart Instrumentation - Self-organizing Systems - Self-diagnosable Machines - Self-optimizing Systems - Artificial Neural Networks based Cooperative Systems - Social Behaviour based Systems - Multi-agent and Distributed Intelligent Systems - Neuro-Fuzzy and Fuzzy Logic based Systems - Complex Intelligent Artificial Systems - Artificial Neural Networks' Software and Hardware Issues - Artificial Intelligent Systems' Software and Hardware Issues - Modular Artificial Neural Network based Systems - Modular Implementation of Artificial Neural Networks - Stability and Instability in Artificial Neural Networks - Cooperative Robots and Applications - Humanoid Robots - Application of Artificial Neural Networks and Intelligent Systems:

- Artificial Neural Networks based Pattern Recognition - Artificial Neural Networks based Signal Processing - Artificial Neural Networks based Image Processing - Artificial Neural Networks based Data Fusion

ICINCO 2012 ANNIIP SPECIAL SESSION TOPICS AND PROGRAM COMMITTEE


- Artificial Neural Networks based Data Mining - Artificial Neural Networks based Decision - Artificial Neural Network based Control - Artificial Neural Network based System Identification - Artificial Neural Network in Robustness and Safety - Artificial Neural Network in Management and Financial Applications - Artificial Neural Network based Solutions for Industrial Environmentt

ANNIIP Program Committee Veronique Amarger, University PARIS-EST Creteil (UPEC), France Amine Chohra, Paris-East University (UPEC), France Khalifa Djemal, University of Evry Val d’Essonne, France Vladimir Golovko, Brest State Technical University, Belarus Kurosh Madani, University of Paris-EST Créteil (UPEC), France Jean-Jacques Mariage, PARIS 8 University, France Dominik Maximilián Ramík, UPEC / LISSI, France Christophe Sabourin, IUT Sénart, University Paris-Est Creteil (UPEC), France Weiwei Yu, Northwestern Polytechnical University, China

IVC&ITS SPECIAL SESSION SCOPE, TOPICS AND PROGRAM COMMITTEE ICINCO 2012


Monday 30 15:15 – 16:30 Room: Barcelona

Special Session on Intelligent Vehicle Controls & Intelligent Transportation Systems – IVC&ITS 2012

Chair Oleg Gusikhin Ford Research & Adv. Engineering, U.S.A.

SCOPE AND TOPICS The subject of Intelligent Vehicle Controls & Intelligent Transportation Systems covers a broad interdisciplinary area of the research and development toward next generation mobility solutions. The topics of interest range from computational intelligence methods in vehicle safety applications and autonomous vehicles technologies to new business models based on advancements of transportation information support infrastructure. The goal of the special session is to bring together representatives from academia, industry and government agencies to exchange ideas on state of the art intelligent vehicle controls and intelligent transportation systems.


- Vehicle Communication Technologies - Advanced Driver Assistance Systems - Hierarchical Artificial Neural Network Models and Systems - Active Safety Systems - Autonomous Vehicles - Navigation and Guidance - Eco-routing and Eco-driving - Vision-based Applications - Speech Interface - Sensor Fusion - Vehicle Application of Affective Computing Examples of Commercial Implementations of Vehicle

Intelligence Case Studies from Autonomous Vehicle Competitions

IVC&ITS Program Committee Vishi Gupta, Ford Motor Company, U.S.A. Oleg Gusikhin, Ford Research & Adv. Engineering, U.S.A. Ilya Kolmanovsky, University of Michigan, U.S.A. Anatoli Koulinitch, Meritor, Inc., U.S.A. Perry MacNeille, Ford Motor Company, U.S.A. Danil Prokhorov, Toyota Tech Center, U.S.A.

ICINCO 2012 OMDM SPECIAL SESSION SCOPE, TOPICS AND PROGRAM COMMITTEE


Tuesday 31 10:00 – 13:00 Room: Barcelona

Special Session on Operations Management and Decision Making in Today’s Competitive Environment - OMDM 2012

Chair Amy Lee Chung Hua University, Taiwan

SCOPE AND TOPICS With increasingly fierce global competition, firms today need to confront a variety of issues in order to survive and to acquire reasonable profit. Not only the service industry is service-oriented, the manufacturing industry is undergoing a major transition to services, both in sales and in business structure. The move toward more extensive, customer-oriented products and services has become common practices among the firms. Thus, firms often have to provide customers with integrated product service packages and integrated solutions rather than traditional products and services. For instance, many high-technology firms are evolving and enhancing their service offerings to meet market demands through providing services, such as original design manufacturing (ODM) and original equipment manufacturing (OEM), and under various requirements, such as emergency orders and quality needs. Many leading firms are also striving toward the successful implementation of integrated supply chain management and service strategies. Due to the complexity of the business environment, well-constructed models are necessary to facilitate relevant operations management and decision making. The special session will consists of eight papers, and each paper has a topic that is related to the fields, such as operations management, business decision making and supply chain management, in today’s competitive business environment.


- Operations Management - Decision Making and Performance Evaluation - Production Planning, Scheduling and Control - Supply Chain Management - Systems Modelling and Simulation - Quality Control and Management - Energy Efficiency and Green Manufacturing

OMDM Program Committee Ying Chung Chang, Ching Yun University, Taiwan Cheng-Che Chen, Far East University, Taiwan Hsing Hung Chen, Macau University of Science and Technology, Macau Mou-Yuan Liao, Yuanpei University, Taiwan Yu-Hsin Lin, Minghsin University of Science and Technology, Taiwan

AWARDS AND SELECTED PAPERS BOOK ICINCO 2012


Awards

Best Paper Awards A "Best Paper Award" and a "Best Student Paper Award" will be conferred to the author(s) of a full paper presented at the conference, selected by the Program/Conference Chairs based on the best combined marks of paper reviewing, assessed by the Program Committee, and paper presentation quality, assessed by session chairs at the conference venue. The "Best Student Paper Award" will be given to a paper in which the first author is a registered MSc or PhD student. The awards will be announced and bestowed at the conference closing session.

Selected Papers Book A number of selected papers presented at ICINCO 2012 will be published by Springer-Verlag in a LNEE Series book. This selection will be done by the Conference Chair and Program Co-chairs, among the papers actually presented at the conference, based on a rigorous review by the ICINCO 2012 Program Committee members.

ICINCO 2012 SOCIAL EVENT AND BANQUET


Social Event and Banquet

Venue: Bus Tour in Rome followed by a Dinner at the “Il Borgo di Tragliata” Date: Monday 30, 18:15 - 23:30

Located at the entrance to Rome, “IL borgo di Tragliata,” rises above an impressive tufa buttress. Archeological sources provide evidence that this area has been inhabited since ancient times. The discovery of the famous, “Oinochoe of Tagliatella” vase confirms the existence of human settlements since the Etruscan era within an area subject to control by either Ceri or Veio. The place name “Tragliata” takes note of the place names, Talianum Tagliata or Terlata, during medieval times and appears to be derived from “Tagliata, (meaning, “cut”), which is the word given to the paths dug into the tufa by the Etruscans.

The presence of several tombs dug into the tufa along the East slope of the hill on which the village sits, along with several clay artifacts found in the area, are evidence which suggest the presence of a small agricultural settlement. In addition, other documentation reports the findings of the remains of a Roman villa on Tragliata property. It is also known that the two marble memorial stones found in this area have inscriptions dating back to the third century AD.

Midway through the eighth century, this area of the Roman countryside saw a period of repopulation thanks to the intelligence and will of Pope San Zaccaria (741-752) and Pope Adriano I (772-793). Encouraged by political and religious motives, these two Pontiffs presented an energetic revival and control of the territory.

SOCIAL EVENT AND BANQUET ICINCO 2012


During ninth and tenth centuries the historical scene began to change, the Roman countryside, with less support for the Papacy by the Carolugian empire, was made subject to continual and bloody raids by the Saracen pirates. The system of the “Domuscultae” entered into definite crisis, superceded by a strong defense system of towers and small castles; several coastal light towers were constructed to be used as bright defense signals to alert the inland region upon the pirates’ approach.

The construction of Tragliata’s small castle and tower date back to the ninth century, according to sources at the nearby Boccea castle.

The estate still belongs to the Vatican Basilica, even if time after time it was more or less controlled directly by others. In 1201, for example, it was ruled by a certain Jocobus de Traliata who occupied it, possibly as a lord. Several years later Tragliata, together with nearby “castium” Loterni, became subject to the interests of the turbulent Normanni family.

In 1885, the Chapter granted the Tragliata estate to Mr. Nicola Santovetti as the perpetual leaseholder. Consequently, Santovetti sold the lease to Mr. Domenico Lanza in 1917, (the great grand father of the present proprietor, Andrea de Gallo di Roccagiovane) who then took over as a tenant to finally gain possession of the estate in the following years.

ICINCO 2012 GENERAL INFORMATION


General Information

Welcome Desk/On-site Registration Friday 27 – Open from 15:00 to 17:30 Saturday 28 – Open from 14:30 to 18:15 Sunday 29 – Open from 9:00 to 18:45 Monday 30 – Open from 9:00 to 17:30 Tuesday 31 – Open from 9:45 to 17:15 Opening Session Saturday 28, at 15:00 in the Plenary room. Welcome Drink Saturday 28, at 18:15 in the Picasso Foyer. Closing Session Tuesday 31, at 17:00 in the Plenary room. Farewell Drink Tuesday 31, at 17:15 in the Picasso Foyer. Meals Coffee-breaks will be served in the Picasso Foyer next to the conference rooms to all registered participants. Lunches will be served in the Murillo room from 13:00 to 14:30 to all registered participants. Communications Wireless access will be provided free of charge to all registered participants, during the conference business hours. Secretariat Contacts ICINCO Secretariat Address: Av. D. Manuel I, 27A 2º Esq. 2910-595 Setúbal, Portugal Tel.: +351 265 520 185 Fax: +44 203 014 5435 e-mail: [email protected] website: http://www.icinco.org

ROOMS LAYOUT ICINCO 2012


Rooms Layout

ICINCO 2012 PROGRAM LAYOUT


Program Layout

09:00 09:00

09:45 09:45 09:45

10:00

10:30

10:45 10:45 10:45

11:00 11:00 11:00

11:30

13:00 13:00 13:00

14:30 14:30 14:30 14:30

15:00 15:00

15:15 15:15

15:30

16:00

16:30 16:30 16:30

16:45 16:45 16:45

17:00

17:15

17:30 17:30

17:45

18:15 18:15 18:15

18:45 18:45

23:30

Keynote LectureAlexandre Dolgui

Posters Session 3

Session 7

C.-Break

Special Session OMDM

Special Session OMDM

WD/ Registration

Friday 27 Saturday 28

Welcome Desk / Registration

Keynote LectureMunther A. Dahleh

Coffee-Break

LunchLunchLunch

Monday 30 Tuesday 31

Buses back to hotel


Keynote LectureAlessandro De Luca

Coffee-Break

Session 4

Keynote LectureJurek Sasiadek

C.-Break

Social Event and Banquet

Posters Session 2

Session 5


Coffee-Break

Closing Session

Farewell Cocktail

Sunday 29

Session 2

Session 6

Panel

Opening Session

Session 3

Special Session IVC&ITS


Special Session ANNIIP

Buses to BanquetWelcome Cocktail

Session 1

Coffee-Break

Posters Session 1

Final Programand Book of Abstracts

Contents

Saturday Sessions 37

Opening Session (15:00 - 15:15) 39Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Panel: ICINCO (15:15 - 16:30) 39Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Parallel Session 1 (16:45 - 18:15) 39Room Sevilla - Intelligent Control Systems and Optimization . . . . . . . . . . . . . . . . . . . . . . . 39

61: Emotion Recognition of Violin Music based on Strings Music Theory for Mascot Robot System 39133: A Proposal of Multiobjective Fuzzy Regulator Design for State Space Nonlinear Systems . 39306: Robust Arbitrary Reference Command Tracking with Application to Hydraulic Actuators . . . 39

Parallel Session 1 (16:45 - 18:15) 40Room Valencia - Robotics and Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

265: LGMD based Neural Network for Automatic Collision Detection . . . . . . . . . . . . . . . . . 4011: Particle Filtering for Position based 6DOF Visual Servoing in Industrial Environments . . . . . 4048: How to Use an Adaptive High-gain Observer in Diagnosis Problems . . . . . . . . . . . . . . . 4057: Semi-static Object Detection using Polygonal Maps for Safe Navigation of Industrial Robots . 40

Parallel Session 1a (16:45 - 18:15) 41Room Madrid - Robotics and Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

169: Two-view Epipole-based Guidance Control for Autonomous Unmanned Aerial Vehicles . . . 41319: Discrete Asymptotic Reachability via Expansions in Non-integer Bases . . . . . . . . . . . . 41337: Force Control of a Duct Cleaning Robot Brush using a Compliance Device . . . . . . . . . . 41

Parallel Session 1 (16:45 - 18:15) 42Room Dali - Signal Processing, Sensors, Systems Modelling and Control . . . . . . . . . . . . . . 42

170: Identification of Polytopic Models for a Linear Parameter-varying System Performed on aVehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

215: State Estimation and Send on Delta Strategy Codesign for Networked Control Systems . . 42220: Performance Evaluation of Discrete Event Systems Thanks to New Representations for

(max,+) Automata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42233: Using Forward-backward Contractors to Identify Parasitic Parameters of Electrical Circuits

Working in High Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Parallel Session 1 (16:45 - 18:15) 43Room Barcelona - Industrial Engineering, Production and Management . . . . . . . . . . . . . . . . 43

62: Assessing the Efficacy of Improvements in User Satisfaction for Mobile Applications - UserFeedback from the Review Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

209: Sloshing Suppression Control by using Physical Boundary Element Model and PredictiveControl in Liquid Container Transfer System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

211: Process-oriented Approach to Verification in Engineering . . . . . . . . . . . . . . . . . . . . . 43

Sunday Sessions 45

Keynote Speaker: Munther A. Dahleh (09:45 - 10:45) 47Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Resilience of Dynamical Transportation Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

29

Con

tent

s


23: Risk Prediction of a Behavior-based Adhesion Control Network for Online Safety Analysis ofWall-climbing Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

76: A Particle Swarm Optimization Algorithm for the Grasp Planning Problem . . . . . . . . . . . . 47149: Hardware on-Board Implementation of a Possible Solution to Introduce UAVs into Non

Segregated Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48228: Optimal Distributed Controller Synthesis for Chain Structures - Applications to Vehicle

Formations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48307: A Stochastic Queueing Model for Multi-robot Task Allocation . . . . . . . . . . . . . . . . . . . 48344: Attention Capabilities for AI Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Parallel Session 2a (11:00 - 13:00) 49Room Barcelona - Intelligent Control Systems and Optimization . . . . . . . . . . . . . . . . . . . . . 49

97: Automated Design of Reconfigurable Manufacturing Systems - A Framework based on aGenetic Algorithm and Discrete System Simulation . . . . . . . . . . . . . . . . . . . . . . . 49

107: Evolutionary Symbiotic Feature Selection for Email Spam Detection . . . . . . . . . . . . . . 49110: Predictive Data Reduction in Wireless Sensor Networks using Selective Filtering . . . . . . . 50153: A Hybrid Metaheuristic Approach to Solve the Vehicle Routing Problem with Time Windows 50218: A Hybrid Algorithm using Metaheuristics Applied to H.264/AVC Video Encoder . . . . . . . . 50

Parallel Session 2 (11:00 - 13:00) 51Room Madrid - Robotics and Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

164: View-based SLAM using Omnidirectional Images . . . . . . . . . . . . . . . . . . . . . . . . . . 51216: Link Quality Estimator for a Mobile Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51235: Global Visual Features based on Random Process - Application to Visual Servoing . . . . . 51239: A Bug-based Path Planner Guided with Homotopy Classes . . . . . . . . . . . . . . . . . . . . 52

Parallel Session 2 (11:00 - 13:00) 52Room Dali - Signal Processing, Sensors, Systems Modelling and Control . . . . . . . . . . . . . . . 52

18: Efficient Distributed Fusion Filtering Algorithms for Multiple Time Delayed Systems . . . . . . 5229: On Specification Informatics in Discrete-event Systems - State-transparency for Clarity of

Finite Automata as Control Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52282: Self-Scheduled H∞ Control of a Wind Turbine - A Real Time Implementation . . . . . . . . . 53289: Cost Functions for Scheduling Tasks in Cyber-physical Systems . . . . . . . . . . . . . . . . . 53

Keynote Speaker: Jurek Sasiadek (14:30 - 15:30) 54Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Space Robotics - Guidance, Navigation and Control Challenges . . . . . . . . . . . . . . . . . . . . 54

Poster Session 1 (15:30 - 16:30) 54Foyer - ICINCO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

19: Gesture and Body Movement Recognition in the Military Decision Support System . . . . . . 5460: Use of Flow Control on Car Dumpers - A Case of Success at Vale . . . . . . . . . . . . . . . . 5471: Predictive Control of Unmanned Formations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5587: Business Process Development for Industrial Cluster . . . . . . . . . . . . . . . . . . . . . . . . 55118: Quantitative Estimates of Stability in Controlled GI | D | 1 | ∞ Queueing Systems and in

Control of Water Release in a Simple Dam Model . . . . . . . . . . . . . . . . . . . . . . . . 55127: From a Multi-robot Global Plan to Single-robot Actions . . . . . . . . . . . . . . . . . . . . . . . 55165: Monte Carlo Localization using the Global Appearance of Omnidirectional Images -

Algorithm Optimization to Large Indoor Environments . . . . . . . . . . . . . . . . . . . . . . 56178: Compensation of Unknown Input Dead Zone using Equivalent-Input-Disturbance Approach 56251: Feature Selection Combined with Neural Network for Diesel Engine Diagnosis . . . . . . . . 56252: Distributed Control of Dangerous Goods Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . 57266: Development of Mobile Research Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57267: Modified Hybrid Evolutionary Strategies Method for Termination Control Problem with Relay

Actuator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57290: An Application of Goal Programming Technique for Reconfiguration of Transfer Lines . . . . 57305: On the Temperature Control for a Test Case Short Pipe Network Central Heating System . 58321: Modelling and Simulation of Human-like Movements for Humanoid Robots . . . . . . . . . . 58341: A Control Strategy for Reducing Fuel Consumption in a Hybrid Electric Vehicle . . . . . . . . 58345: Observability of Transportation Systems - A Methodology for Reliability Analysis in Logistics

and Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58


Contents

349: Towards Manufacturing Execution Systems for the Food and Beverage Packaging Industry 59Foyer - ANNIIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

8: Comparison the Performance of Hybrid HMM/MLP and RBF/LVQ ANN Models - Applicationfor Speech and Medical Pattern Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Special Session on Artificial Neural Networks and Intelligent Information Processing - ANNIIP(16:45 - 18:15) 59Room Barcelona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

9: GPU-based Parallel Implementation of a Growing Self-organizing Network . . . . . . . . . . . . 5912: Parallel Batch Pattern Training of Recirculation Neural Network . . . . . . . . . . . . . . . . . . 603: Forecasting Financial Success of Hollywood Movies - A Comparative Analysis of Machine

Learning Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60


64: Fuzzy Control of a Hybrid Renewable Power System based on Real-time Matlab-PLCCommunication through OPC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

80: Data-based Tuning of PI Controllers for Vertical Three-Tank Systems . . . . . . . . . . . . . . . 61123: Differential Evolution in Parameter Identification - Fuel Cell as an Example . . . . . . . . . . 61248: Spacecrafts’ Control Systems Effective Variants Choice with Self-configuring Genetic

Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Parallel Session 3 (16:45 - 18:45) 62Room Valencia - Robotics and Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

176: Autonomously Traversing Obstacles - Metrics for Path Planning of Reconfigurable Robotson Rough Terrain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

206: Design and Control of a Novel Unmanned Ground Vehicle . . . . . . . . . . . . . . . . . . . . 62227: Trajectory Tracking Control of Nonholonomic Wheeled Mobile Robots - Combined Direct

and Indirect Adaptive Control using Multiple Models Approach . . . . . . . . . . . . . . . . . 62285: From Robot Commands to Real-time Robot Control - Transforming High-level Robot

Commands into Real-time Dataflow Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63


67: Modelling of a Grasping and Manipulation Controller . . . . . . . . . . . . . . . . . . . . . . . . . 63120: Dynamic Model of a 7-DOF Whole Arm Manipulator and Validation from Experimental Data 63213: False Positive Outliers Rejection for Improving Image Registration Accuracy - Application

to Road Traffic Aerial Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63347: Towards the Integrated Simulation and Programming of Palletizing Lines . . . . . . . . . . . . 64351: Task-based Method for Designing Underactuated Elastic Mechanisms . . . . . . . . . . . . . 64354: Supporting Mobile Robot’s Tasks through Qualitative Spatial Reasoning . . . . . . . . . . . . 64


25: Flow Optimization for Iron Ore Reclaiming Process . . . . . . . . . . . . . . . . . . . . . . . . . 6453: Using Linear Systems Theory to Study Nonlinear Dynamics of Relay Cells . . . . . . . . . . . 6556: Nonparametric Identification of Nonlinearity in Wiener-Hammerstein Systems . . . . . . . . . 65147: From PID to Extended Learning Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65187: State Dependent Parameter Modelling of a DC-DC Boost Converter in Discontinuous

Conduction Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66260: Applying Hyperbolic Wavelets in Frequency Domain Identification . . . . . . . . . . . . . . . . 66

Monday Sessions 67

Keynote Speaker: Alessandro De Luca (09:45 - 10:45) 69Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Recent Advances in Physical Human-Robot Interaction . . . . . . . . . . . . . . . . . . . . . . . . . 69


77: Visual Servoing Path-planning with Spheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69138: Rotor Speed Sensor Fault Detection in Induction Motors . . . . . . . . . . . . . . . . . . . . . 69225: Visual Anomaly Detection in Production Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . 70


Con

tent

s

230: Global Optimal Solution to SLAM Problem with Unknown Initial Estimates . . . . . . . . . . . 70


104: Design and Analysis of an Automated Heavy Vehicle Platoon . . . . . . . . . . . . . . . . . . 70131: Trajectory Tracking Control by LMI-based Approach for Car-like Robots . . . . . . . . . . . . 70200: Development of Parallel Two-wheel Vehicle with Lower Gravity Center of Vehicle Body . . . 71277: Vision based Real-time Modeling of Dynamic Unstructured Environments in Driving

Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Parallel Session 4a (11:00 - 13:00) 71Room Barcelona - Robotics and Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

40: Hierarchical Planning of Modular Behaviour Networks for Office Delivery Robot . . . . . . . . 71109: An Integrated Approach for Efficient Mobile Robot Trajectory Tracking and Obstacle

Avoidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72224: Construction and Modeling of a Variable Collective Pitch Coaxial UAV . . . . . . . . . . . . . 72226: Mobile Robots Pose Tracking - A Set-membership Approach using a Visibility Information . 72353: A Joint Segmentation and Classification of Object Shapes with Feedback for 3D Point Clouds 72


72: Contact-free Magnetic Clutch Applied for Flywheel Cell System . . . . . . . . . . . . . . . . . . 73103: Modeling and Simulation of a Wastewater Pumping Plant . . . . . . . . . . . . . . . . . . . . . 73195: Support Vector Machines for Identification of HCCI Combustion Dynamics . . . . . . . . . . 73240: Handling Delayed Fusion in Vision-Augmented Inertial Navigation . . . . . . . . . . . . . . . . 74


10: Multivariable Discrete Time Repetitive Control System . . . . . . . . . . . . . . . . . . . . . . . 7420: Reconstruction-based Set-valued Observer - A New Perspective for Fault Detection within

Uncertain Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7478: Adaptive Gravitational Search Algorithm for PI-fuzzy Controller Tuning . . . . . . . . . . . . . . 75231: Broken Bar Fault Detection based on Set Membership Identification for Three Phase

Induction Motors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75297: Design and Experimentation of a Neural Network Controller for a Spherical Parallel Robot . 75334: Research on Integrated Fault Diagnosis of Steam Turbine based on CPN Neural Network

and D-S Evidence Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76


38: On the Multiple-view Triangulation Problem with Perspective and Non-perspective Cameras- A Virtual Reprojection-based Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

49: for Cell Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76236: Compensation of Tool Deflection in Robotic-based Milling . . . . . . . . . . . . . . . . . . . . . 77136: Evaluation of a Joint Hysteresis Model in a Robot Actuated by Pneumatic Muscles . . . . . 77


152: Semantic Place Recognition based on Deep Belief Networks and Tiny Images . . . . . . . . 77212: Integral Sliding Mode and Second Order Sliding Mode Attitude and Altitude Tracking of a

Quadrotor System - Theory and Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78293: Low-speed Modeling and Simulation of Torpedo-shaped AUVs . . . . . . . . . . . . . . . . . . 78301: Cooperative Autonomous Driving for Vehicular Networks . . . . . . . . . . . . . . . . . . . . . 78304: Flocking for Networks of Mobile Robots with Nonlinear Dynamics . . . . . . . . . . . . . . . . 79


89: Load-following Control of APR+ Nuclear Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . 7990: A Fault-Tolerant Controller for an SP-100 Space Nuclear Reactor . . . . . . . . . . . . . . . . . 79145: Smart Walker Control through the Inference of the User’s Command Intentions . . . . . . . . 80223: Nonlinear Deterministic Methods for Computer Aided Diagnosis in Case of Kidney Diseases 80302: Block Triangular Decoupling of General Neutral Multi Delay Systems . . . . . . . . . . . . . . 80


Contents

Special Session on Intelligent Vehicle Controls & Intelligent Transportation Systems - IVC&ITS(15:15 - 16:30) 81Room Barcelona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

4: A Warping Window Approach to Real-time Vision-based Pedestrian Detection in a Truck’sBlind Spot Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

7: Road Safety at Intersections Controlled by Traffic Lights - IVC and Risk Indexes . . . . . . . . . 81

Poster Session 2 (16:30 - 17:30) 81Foyer - ICINCO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

26: Nonlinear Analysis of Costas Loop Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8139: Positive Realization of Continuous Linear Systems with Order Bound . . . . . . . . . . . . . . 8150: Fluidity Measuring Device for the Concrete using Laser Diode Controller via WSN . . . . . . . 82124: Modeling and Simulation of Humanoid Robot Spine Vertebra . . . . . . . . . . . . . . . . . . . 82130: On the Problem of Task Planning in Multi-robot Systems . . . . . . . . . . . . . . . . . . . . . 82162: Disassembly Planning using Visual Servoing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82172: Observer-based Adaptive Sliding Mode Control for a Pneumatic Servo System . . . . . . . . 83175: Design of a Rectangular-type Finger Rehabilitation Robot . . . . . . . . . . . . . . . . . . . . . 83181: Facial Expression Recognition based on Facial Feature and Multi Library Wavelet Neural

Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83194: Simultaneous Control of Translational and Rotational Motion for Autonomous

Omnidirectional Mobile Robot - 2nd Report: Robot Model Considering Moving Parts andEvaluation of Movable Area by Heights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

232: EKF based Data Fusion using Interval Analysis via Covariance Intersection, ML and aClass of OGK Covariance Estimators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

247: Improved Leader Follower Formation Control of Autonomous Underwater Vehicles usingState Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

263: Optimized Chipping Processes with a New Mechatronic Tool System - Application of StrainGauge Sensors and Piezoelectric Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

270: Modified Evolutionary Strategies Algorithm in Linear Dynamic System Identification . . . . . 85295: Autonomous Navigation of an Outdoor Mobile Robot in a Cluttered Environment . . . . . . . 85328: Development of a Wheelchair with a Lifting Function . . . . . . . . . . . . . . . . . . . . . . . . 85

Foyer - IVC&ITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 866: Stereo Vision based On-road Vehicle Detection under Illumination Changing Conditions using

Self Quotient Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

Tuesday Sessions 87

Special Session on Operations Management and Decision Making in Today’s CompetitiveEnvironment - OMDM (10:00 - 10:45) 89Room Barcelona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4: An Evaluation Model for Green and Low-Carbon Suppliers . . . . . . . . . . . . . . . . . . . . . . 891: Developing A New Variables Sampling Scheme for Product Acceptance Determination . . . . 89

Poster Session 3 (10:30 - 11:30) 89Foyer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

36: Modeling and Simulation of a Temperature Robust Control in Grain Drying Systems forThermal Damage Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

59: Control and Model Parameters Identification of Inertia Wheel Pendulum . . . . . . . . . . . . . 9082: Solving the Indoor SLAM Problem for a Low-Cost Robot using Sensor Data Fusion and

Autonomous Feature-based Exploration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90115: A Study of the Impact of Computational Delays in Missile Interception Systems . . . . . . . . 90116: A Rescheduling Framework for Airline Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . 91137: Operating System Core as Template for Embedded System Software Development . . . . . 91150: Improvement of Extraction Method for Inter-turn Fault Detection in IPMSM under Transient

Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91155: Mechanical Modeling of an Actuated Platform for Precision Pointing Control - Via Finite-

element Analysis and Normal Mode Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 91161: Timed Trajectory Generation for a Vision-based Autonomous Mobile Robot in Cluttered

Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92196: Total Intelligent Life-Saving System against Earthquake and “Tsunami” - Disaster Medical

Engineering from Protection to Care in Catastrophe . . . . . . . . . . . . . . . . . . . . . . . 92202: A Novel Approach to Measure under Water Vehicle Disturbance Force for Station Keeping

Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92


Con

tent

s

204: Flocking Approach to Spatial Configuration Control in Underwater Swarms . . . . . . . . . . 93234: Hardware in the Loop Module to Calculate Production Indicators . . . . . . . . . . . . . . . . 93238: Programming of a Mobile Robotic Manipulator through Demonstration . . . . . . . . . . . . . 93258: Improving Stereo Vision Odometry for Mobile Robots in Outdoor Dynamic Environments . . 94269: Robust Control of Excavation Mobile Robot with Dynamic Triangulation Vision . . . . . . . . 94

Special Session on Operations Management and Decision Making in Today’s CompetitiveEnvironment - OMDM (11:00 - 13:00) 94Room Barcelona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

2: Supplier Selection based on Process Yield for LED Manufacturing Processes . . . . . . . . . . 945: An Integrated Model to Evaluate the Performance of Solar PV Firms . . . . . . . . . . . . . . . . 956: An Integrated Replemishment Model under Dynamic Demand Conditions . . . . . . . . . . . . 959: A Study on a Decision Support Model for Strategic Alliance in Express Courier Service . . . . 963: Production Planning and Control Model of Technology Migration for DRAM Industry . . . . . . 96


96: Development of Robust Learning Control and Application to Motion Control . . . . . . . . . . . 96250: Statistical Models for Emotion Recognition using Facial Expression Analisys . . . . . . . . . 97279: New Robust Controller Synthesis Optimization Methodology under Six Sigma Constraint . . 97348: Flexibility in Organic Systems - Remarks on Mechanisms for Adapting System Goals at

Runtime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

Parallel Session 6a (11:30 - 13:00) 98Room Velazquez - Intelligent Control Systems and Optimization . . . . . . . . . . . . . . . . . . . . . 98

180: Distributed Control System for Crystal Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98189: Multi-Agent System for Adaptation of Distributed Control System . . . . . . . . . . . . . . . . 98333: An Optimization Method for Training Generalized Hidden Markov Model based on

Generalized Jensen Inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98355: Constrained Predictive Control of MIMO System - Application to a Two Link Manipulator . . 99


33: An Autonomous Biped - Concept and Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99222: Kinematic Analysis of Lower Mobility Cooperative Arms by Screw Theory . . . . . . . . . . . 99284: Virtual Mechanism Approach for Dual-arm Manipulation . . . . . . . . . . . . . . . . . . . . . . 99287: Two-arm Robot Teleoperation using a Multi-touch Tangible User Interface . . . . . . . . . . . 100


207: Motion Detection and Velocity Estimation for Obstacle Avoidance using 3D Point Clouds . . 100243: The Modular Behavioral Environment for Humanoids and other Robots (MoBeE) . . . . . . . 100330: Robot-Dog – Human Interaction in Urban Search and Rescue Scenarios - Improving the

Efficiency of Rescue Teams in Hazardous Environments . . . . . . . . . . . . . . . . . . . . 101


184: Adaptive Data Update Management in Sensor Networks . . . . . . . . . . . . . . . . . . . . . 101214: Hyper-elastic Pressure Sensors - Temperature Dependence of Piezoresistivity of

Polyisoprene – Nanostructured Carbon Composite . . . . . . . . . . . . . . . . . . . . . . . . 101346: On the Design of Change-driven Data-flow Algorithms and Architectures for High-speed

Motion Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102


148: Optimal Control Strategy of NG Piston Engine as a DG Unit Obtained by an Utilization ofArtificial Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

166: Neuro-fuzzy Sliding Mode Control for a Two Link Flexible Robot . . . . . . . . . . . . . . . . . 102246: Multi-device Power-saving - An Investigation in Energy Consumption Optimisation . . . . . . 103325: Implementation of Cognitive Chips in Machining Error Attenuation . . . . . . . . . . . . . . . . 103


Contents

Parallel Session 7a (14:30 - 16:00) 103Room Velazquez - Intelligent Control Systems and Optimization . . . . . . . . . . . . . . . . . . . . . 103

75: Nonlinear Model Validation of an Internal Combustion Engine with Few Sensors . . . . . . . . 10386: Modeling and Active Vibration Control of a Smart Structure . . . . . . . . . . . . . . . . . . . . 103167: Performance Shaping through Cost Cumulants and Neural Networks-based Series

Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104


210: Robot Phase Entrainment on Quadruped CPG Controller . . . . . . . . . . . . . . . . . . . . . 104242: A Mobile Service Robot for Industrial Applications . . . . . . . . . . . . . . . . . . . . . . . . . 104300: Modeling and Visualizing Individual and Global Trends of a Multi-agent System . . . . . . . 105


44: A Generative Traversability Model for Monocular Robot Self-guidance . . . . . . . . . . . . . . 10593: 3D Dense Reconstruction from Two-Frame Stereo Correspondences . . . . . . . . . . . . . . 105121: Object Recognition based on a Simplified PCNN . . . . . . . . . . . . . . . . . . . . . . . . . . 106275: Integrating Kinect Depth Data with a Stochastic Object Classification Framework for

Forestry Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106


191: Wave Vibration Analysis of Classical Multi-story Planar Frames . . . . . . . . . . . . . . . . . 106257: On the Bus Priority Dilemma: Modelling and Analysis of Congested Traffic Network using

Coloured Petri Nets and (Min,+) Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106280: GFIS: Genetic Fuzzy Inference System for Speech Recognition . . . . . . . . . . . . . . . . . 107

Parallel Session 7 (14:30 - 16:00) 107Room Barcelona - Industrial Engineering, Production and Management . . . . . . . . . . . . . . . . 107

17: Control of Variable Parametric Actuator in Building Temperature Regulation . . . . . . . . . . . 107298: Productivity Improvement through Layout Redesign - A Lean Approach Case Study . . . . . 107358: The Optimal Gasoline Blending into Romanian Refinery - Case Study . . . . . . . . . . . . . 108

Keynote Speaker: Alexandre Dolgui (16:00 - 17:00) 108Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Optimization in Design of Automated Machining Systems . . . . . . . . . . . . . . . . . . . . . . . . 108

Closing Session (17:00 - 17:15) 108Room Plenary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108


Con

tent

s

Contents

Saturday Sessions

FINAL PROGRAM AND BOOK OF ABSTRACTS: SATURDAY SESSIONS ICINCO 2012

Opening Session15:00 - 15:15 Room Plenary

Panel ICINCO15:15 - 16:30 Room Plenary

Paper 6116:45 - 18:15 Room SevillaParallel Session 1 - Intelligent Control Systemsand Optimization

Emotion Recognition of Violin Music based onStrings Music Theory for Mascot Robot System

Z.-T. Liu1,2, Z. Mu1, L.-F. Chen1,2, P. Q. Le1, C.Fatichah1, Y.-K. Tang1, M. L. Tangel1, F. Yan1, K.Ohnishi1, M. Yamaguchi1, Y. Adachi1, J.-J. Lu1,

T.-Y. Li1, Y. Yamazaki3, F.-Y. Dong1 and K. Hirota11 Tokyo Institute of Technology, Yokohama, Japan

2 Central South University, Changsha, China3 Kanto Gakuin University, Yokohama, Japan

Keywords: Emotion Recognition, Violin, Music,Support Vector Regression, Fuzzy Logic.

Abstract: Emotion recognition of violin music isproposed based on strings music theory, wherethe emotional state of violin music is expressed byAffinity-Pleasure-Arousal emotion space. Besidesthe music features from audio processing, threefeatures (i.e., left-hand feature, right-hand feature,and dynamics) with regard to both compositionand performance of violin music, are extracted toimprove the emotion recognition of violin music. Todemonstrate the validity of this proposal, a datasetcomposing of 120 pieces of author-performedviolin music with six primary emotion categories isestablished, by which the experimental results ofemotion recognition using Support Vector Regressionreport overall recognition accuracy of 86.67%. Theproposal could be an integral part for analyzing thecommunication atmosphere with background music,or be used by a music recommendation system forvarious occasions.


A Proposal of Multiobjective Fuzzy RegulatorDesign for State Space Nonlinear Systems

Rafael J. M. Santos and Ginalber L. O. SerraFederal Institute of Education, Science and Technology,

São Luis, Brazil

Keywords: Fuzzy Regulator, Optimal Control, LinearQuadratic Regulator, Pole Placement, MultiobjectiveFeasible Region.

Abstract: This paper proposes a Takagi-Sugeno(TS) fuzzy regulator design methodology fornonlinear dynamic systems. The Linear QuadraticRegulator (LQR) and Pole Placement (PP)techniques are combined in a TS fuzzy structurein order to guarantee an optimal controller withsatisfactory transient response based on polesallocated properly. The definition and analysis ofthe multiobjective feasible region, considering theinfluence of the desired poles on the weightingmatrices Q and R in the quadratic cost function, arepresented. Lyapunov based stability analysis andsimulations results on fuzzy regulator design for arobotic manipulator illustrates the efficience of theproposed methodology.


Robust Arbitrary Reference Command Trackingwith Application to Hydraulic Actuators

M. G. Skarpetis, F. N. Koumboulis and A. S. NtellisHalkis Institute of Technology, Halkis, Greece

Keywords: Robust Control, Arbitrary ReferenceCommand Tracking, Hydraulic Actuator.

Abstract: In this paper a robust tracking controlleris proposed in order to track arbitrary referencesignals in the presence of same type disturbancesignals. The robust tracking controller is based onthe well known Internal Model Principle appropriatelymodified with a Hurwitz invariability technique. Thecontroller parameters are computed using a finitestep algorithm. Solvability conditions are derived.The proposed controller is successfully appliedto a hydraulic actuator uncertain model includinguncertain parameters arising from changes of theoperating conditions and other physical reasons.Simulation results for all the expected range of theactuator model uncertainties are presented indicating


Satu

rday

, 28

ICINCO 2012 FINAL PROGRAM AND BOOK OF ABSTRACTS: SATURDAY SESSIONS

the satisfactory performance of the robust controllerin the presence of external disturbances.

Paper 26516:45 - 18:15 Room ValenciaParallel Session 1 - Robotics and Automation

LGMD based Neural Network for AutomaticCollision Detection

Ana Silva, Jorge Silva and Cristina SantosUniversity of Minho, Guimarães, Portugal

Keywords: Bio-inspired Model, Lobula Giant Move-ment Detector Neuron, Artificial Neural Networks,Collision Avoidance.

Abstract: Real-time collision detection in dynamicscenarios is a hard task if the algorithms usedare based on conventional techniques of computervision, since these are computationally complex and,consequently, time-consuming. On the other hand,bio-inspired visual sensors are suitable candidatesfor mobile robot navigation in unknown environments,due to their computational simplicity. The LobulaGiant Movement Detector (LGMD) neuron, locatedin the locust optic lobe, responds selectively toapproaching objects. This neuron has been usedto develop bio-inspired neural networks for collisionavoidance. In this work, we propose a new LGMDmodel based on two previous models, in order toimprove over them by incorporating other algorithms.To assess the real-time properties of the proposedmodel, it was applied to a real robot. Results shownthat the LGMD neuron model can robustly supportcollision avoidance in complex visual scenarios.


Particle Filtering for Position based 6DOF VisualServoing in Industrial Environments

Aitor Ibarguren, José María Martínez-Otzeta andIñaki Maurtua

Fundación Tekniker, Eibar, Spain

Keywords: Robotics, Visual Servoing, Particle Filter.

Abstract: Visual servoing allows the introduction ofrobotic manipulation in dynamic and uncontrolledenvironments. This paper presents a position-basedvisual servoing algorithm using particle filtering.The objective is the grasping of objects using the6 degrees of freedom of the robot manipulator(position and orientation) in non-automated industrialenvironments using monocular vision. A particle filterhas been added to the position-based visual servoing

algorithm to deal with the different noise sourcesof those industrial environments (metallic nature ofthe objects, dirt or illumination problems. . . ). Thisaddition allows dealing with those uncertainties andbeing able to recover from errors in the graspingprocess. Experiments performed in the real industrialscenario of ROBOFOOT project showed accurategrasping and high level of stability in the visualservoing process.


How to Use an Adaptive High-gain Observer inDiagnosis Problems

Frédéric Lafont1,2, Jean-François Balmat1, NathaliePessel1,2 and Jean-Paul Gauthier1,2

1 Université du Sud-Toulon-Var, LSIS, La Garde, France2 Institut Universitaire de Technologie de Toulon, La

Garde, France

Keywords: Observer, Diagnosis, Sensor.

Abstract: This paper explains how to use anadaptive High-Gain observer in sensor diagnosisproblems. This type of observer allows to switchbetween a classical Extended Kalman Filter andHigh-Gain observer according to an innovationfunction. Combined with a standard technique ofresidual generation, this approach is very efficientto determine fault occurence in the non-lineardynamical systems. We present the obtained resultson a wastewater treatment system.


Semi-static Object Detection using PolygonalMaps for Safe Navigation of Industrial Robots

Dario Lodi Rizzini, Gionata Boccalini and StefanoCaselli

University of Parma, Parma, Italy

Keywords: Mapping, Range Sensing.

Abstract: The collision and safety control ofindustrial UGVs equipped with laser range finders isoften based on conservative area-oriented policiesthat lack in flexibility and does not deal well with nonephemeral environment changes due to semi-staticobjects (e.g. passive misplaced objects). In thispaper, we propose a method to detect and representsemi-static objects using polygonal local maps inorder to improve robot navigation. Each local mapconsists of polylines representing the boundary of


Saturday, 28


an object detected inside a safety area. Polylinesare extracted from laser scans and associated withthe polylines of a reference map using a similaritymeasure criterion. Finally, the map is updated bymerging the new polylines. The proposed polygonalrepresentation allows the recognition of new semi-static obstacles in the environment and supportsmore flexible policies for safe navigation. An EKFlocalizer using artificial landmarks and a fixedpath navigation system have been implemented toreplicate the navigation system of industrial UGVs.The precision of environment reconstruction hasbeen assessed with experiments in simulated andreal environments.

Paper 16916:45 - 18:15 Room MadridParallel Session 1a - Robotics and Automation

Two-view Epipole-based Guidance Control forAutonomous Unmanned Aerial Vehicles

W. Achicanoy1, C. Sagüés2, G. López-Nicolás2 andC. Rodríguez1

1 Universidad de los Andes, Bogotá, Colombia2 Universidad de Zaragoza, Zaragoza, Spain

Keywords: UAV Guidance, Epipolar Geometry,Nonlinear Engagement, Input-output Linearization,State Feedback.

Abstract: A visual control based on epipolargeometry is proposed to guide an autonomousunmanned aerial vehicle (UAV) to a target position.The interest of this contribution resides in a newcontroller that allows purely vision-based guidancereducing the dependence on the accuracy ofthe system’s state estimation using sensors asInertialMeasurement Units (IMU) and GlobalPositioning Systems (GPS). A current view anda target view are defined by a camera on-boardthe vehicle and a camera located at the target’sposition, respectively. The epipolar coordinates fromthese views are used to design a nonlinear controlbased on input-output linearization of the nonlinearengagement rule that relates the cameras’ positionsin time. An integrator is included to force the outputs(epipolar coordinates) to follow an equilibrium pointand a state feedback control law is proposed tostabilize the outcome of the linearized input-outputmapping. Simulation experiments for guidance of ansmall autonomous UAV with a classical three-loopautopilot are presented.

Paper 31916:45 - 18:15 Room MadridParallel Session 1 - Robotics and Automation

Discrete Asymptotic Reachability viaExpansions in Non-integer Bases

Anna Chiara Lai and Paola LoretiSapienza Università di Roma, Roma, Italy

Keywords: Discrete Control Theory, Expansion inNon-integer Bases, Robot Hand.

Abstract: Aim of this paper is to show the connectionbetween the theory of expansions in non-integerbases and discrete control systems. This idea issupported by an example of application, in theframework of robotics. We show how a model ofmulti-phalanx self-similar robot hand may be studiedby means of results and techniques coming fromnon-standard numeration systems and related tools,like Iterated Function Systems (IFS) and, moregenerally, fractal geometry.


Force Control of a Duct Cleaning Robot Brushusing a Compliance Device

Wootae Jeong1, Seung-Woo Jeon2, DuckshinPark1 and Soon-Bark Kwon1

1 Korea Railroad Research Institute, Uiwang, Korea,Republic of

2 University of Science and Technology, Daejeon, Korea,Republic of

Keywords: Compliance Device, Service Robot, AirDuct Cleaning, Air Quality, Force Control.

Abstract: Conserving clean air and removingcontaminants and particular matters accumulatedin the ventilation system of the subway stations arekey issue for high air quality and green environment.Accumulated various pollutants at inner duct surfacecan cause secondary air contamination and injuresubway passengers’ respiratory system and health.In fact, periodic duct cleaning works can improveindoor air quality, but cleaning entire ventilationsystem takes high cost and manpower. This studyproposes a newly developed duct cleaning robot toprovide autonomous air duct cleaning. In addition,effective cleaning method with an automated robotdevice is developed. In particular, the new ductcleaning robot has functionality that cleans four sidesof inner duct surface simultaneously with a constantpressure by using a force compliance brush. Controlmethod with the compliant device has also beenanalysed. The proposed design of autonomous duct


Satu

rday

, 28


cleaning robot is expected to save the operatingcost of subway ventilation system and sustain cleanindoor air quality by providing easier and fastercleaning tools.

Paper 17016:45 - 18:15 Room DaliParallel Session 1 - Signal Processing, Sensors,Systems Modelling and Control

Identification of Polytopic Models for a LinearParameter-varying System Performed on a

Vehicle

Raluca Liacu, Dominique Beauvois and EmmanuelGodoy

Supelec Sciences of Systems (E3S), Paris, France

Keywords: LPV Model Identification, PolytopicModel, Automotive Identification.

Abstract: This paper deals with the parameteridentification of continuous time polytopic modelsfor a linear parameter-varying system (LPV). Acontinuous-time nonlinear identification approach ispresented, a mix between a local approach and aglobal one is introduced in order to identify a LPVmodel for the lateral comportement of a vehicle. Theproposed approach is based on the prediction errormethod for LTI systems, which is modified to take intoaccount polytopic models and regularization terms.Using experimental data, different parameter-varyingstructures, explaining the lateral behavior of thevehicle, were identified by the proposed methodconsidering the velocity as the scheduling parameter.


State Estimation and Send on Delta StrategyCodesign for Networked Control Systems

Ignacio Peñarrocha, Daniel Dolz, Julio A. Romeroand Roberto Sanchis

University Jaume I, Castellón, Spain

Keywords: State Estimation, Networked Control,Send on Delta, Event based Estimation and Control.

Abstract: In this work, a new strategy to minimize theuse of the network in state estimation over networksis addressed, leading to a co-design procedureof both the observer and the policy for messagesending. The sensor nodes implements a send-on-delta approach, sending new data only whenthere is a considerable deviation from the last sentmeasurement. The estimator node implements a

gain scheduling approach that takes into account theavailability of new received data. The performanceof the observer is analyzed through H∞ norm inboth deterministic and stochastic data transfer rate.This norm is used to design both the observer gainsand the output variations that induce the sensorsto send new outputs to the estimator node, whileguaranteeing a given level of performance on thestate estimation error. The design approach isbased on an optimization procedure with linear andbilinear matrix inequalities constraints that is solvediteratively.


Performance Evaluation of Discrete EventSystems Thanks to New Representations for

(max,+) Automata

Rabah Boukra, Sébastien Lahaye and Jean-LouisBoimond

Université d’Angers, Angers, France

Keywords: Discrete Event System, PerformanceEvaluation, (max,+) automaton.

Abstract: In this contribution, we study theperformances of discrete event systems modeledby (max,+) automata. More precisely, newrepresentations for (max,+) automata are firstproposed. From these, several performanceindicators can be derived, in particular the maximumtime execution and a minorant of the minimumexecution time for a sequence of length n. Finallythese results are discussed in comparison withseveral studies of the literature also dealing withperformance evaluation of (max,+) automata.


Using Forward-backward Contractors to IdentifyParasitic Parameters of Electrical Circuits

Working in High Frequency

Nacim Meslem, Cécile Labarre and StéphaneLecoeuche

Univ Lille Nord de France, Douai, France

Keywords: Electrical Circuit, Electrical ParasiticParameters, Input Impedance, Interval Analysis,Forward-backward Contractor, Set-membershipEstimation.


Saturday, 28


Abstract: Parasitic parameters in electrical networksare usually sources of intolerant electromagneticinterference in their near environment. In order tounderstand better the undesirable phenomenon,the values of these unknown parameters mustbe estimated with a good accuracy. This workshows how interval analysis can help designingset-membership algorithm that is able to solve withnumerical guarantee the kind of issue. A simpleexample, namely second order filter, is studiedand our method shows promising performances fordealing with complex circuits.

Paper 6216:45 - 18:15 Room BarcelonaParallel Session 1 - Industrial Engineering,Production and Management

Assessing the Efficacy of Improvements in UserSatisfaction for Mobile ApplicationsUser Feedback from the Review Data

Bomi Song, Woori Han and Yongtae ParkSeoul National University, Seoul, Korea, Republic of

Keywords: Improvement of Mobile Application, UserSatisfaction, User Feedback, Review Data.

Abstract: Monitoring user feedback is of centralimportance in the iterative and incremental appraochto improvement of mobile applications. Knowledgefrom user review on mobile applications can be fruitfulsource of user feedback on firms’ effort to improvetheir mobile applications. This paper proposes anapproach to assessing the efficacy of improvementsin user satisfaction for mobile applications using userreview data. Specifically, overall satisfaction scoreand frequencies of updated features in review dataare compared before and after updates of mobileapplications. We believe our method can facilitateutilizing user reviews to track user feedback andobtain useful knowledge in planning and managingupdates of mobile applications, and serve as astarting point of more general model.


Sloshing Suppression Control by using PhysicalBoundary Element Model and Predictive Control

in Liquid Container Transfer System

Hisashi Okatsuka, Ryota Shibuya, KazuhikoTerashima

Toyohashi University of Technology, Aichi, Japan

Yoshiyuki NodaUniversity of Yamanashi, Yamanashi, Japan

Yoshiki MatsuoTokyo University of Technology, Tokyo, Japan

Keywords: Sloshing Control, Boundary ElementMethod, Generalized Predictive Control.

Abstract: This paper presents sloshing suppressioncontrol of liquid transferred at a desirable-speed. Inorder to suppress sloshing, a mathematical modelconsisting of continuous equations and the pressureequation is used and the sloshing phenomena areanalyzed by using the Boundary Element Method(BEM). Further, the BEM model is transformed intothe state-variable model. The proposed model canestimate not only first-order mode sloshing but alsohigher-order mode sloshing and predict the futurebehavior of liquid level more precisely. Moreover,sloshing is suppressed by using Model PredictiveControl (MPC).We were able to transfer the containerwhile minimizing sloshing and maintaining a desirablespeed.


Process-oriented Approach to Verification inEngineering

Ekaterina Auer, Roger Cuypers and Wolfram LutherUniversity of Duisburg-Essen, Duisburg, Germany

Keywords: Numerical Verification Assessment,Validation, Uncertainty, Result Verification.

Abstract: Common verification and validationmethodologies suggest using verificationbenchmarks in the cases where no theoreticalproof of the model and algorithm accuracy is possibleand interpreting the outcome of the process ofinterest on these. In this paper, we propose to shiftthe focus on the process itself. Helped by appropriatequestionnaires, users are urged to subdivide their


Satu

rday

, 28


processes into smaller tasks and to classify theinput/output data and algorithms, which allows us toassign a certain initial verification degree to the (sub-)process from a recently proposed four-tier hierarchy.After that, the initial degree can be improved aswell as uncertainty quantified by with the help of aselection of specialized interoperable data types andtools. Besides, we address issues of software qualityand user support by describing a comparison systemfor verified initial value problem solvers.


Saturday, 28

Sunday Sessions

FINAL PROGRAM AND BOOK OF ABSTRACTS: SUNDAY SESSIONS ICINCO 2012

09:45 - 10:45 Room PlenaryResilience of Dynamical Transportation NetworksKeynote Speaker: Munther A. Dahleh

Resilience of Dynamical TransportationNetworks

Munther A. DahlehMIT, Cambridge, U.S.A.

Abstract: In this talk, we present recent resultson the stability and robustness properties oftransportation networks for various agents’ route-choice behavior. We perform the analysis withina dynamical system framework over a directedacyclic graph between a single origin-destinationpair. The dynamical system is composed of ordinarydifferential equations (ODEs), one for every link ofthe graph. Every ODE is a mass balance equationfor the corresponding link, where the inflow term isa function of the agents’ route-choice behavior andthe arrival rate at the base node of that link, and theoutflow term is a function of the congestion propertiesof the link. We propose a novel decision framework,where the drivers combine their historical knowledgeabout the global congestion levels with real-time localinformation to make route choice decisions at everynode. We show that, if the rate of update of globalinformation is sufficiently slow and if the driversmake route choice decisions cooperatively, then theWardrop equilibrium is globally asymptotically stable.We then study the resilience of the flow transferringcapability of the whole network under disturbancesthat reduce the flow carrying capacity of the links.In particular, we characterize various margins ofresilience of the network with respect to the topology,’pre-disturbance’ equilibrium, and agents’ local route-choice behavior. We show that the cooperative routechoice behavior is maximally resilient in this setting.We also setup a simple convex optimization problemto find the most resilient ’pre-disturbance’ equilibriumfor the network and determine link-wise tolls that yieldsuch an equilibrium. Finally, we extend our analysisto link-wise outflow functions that accommodate thepossibility of cascaded failures and study the effectof such phenomena on the margins of resilience ofthe network.**This work is done in collaboration with Giacomo Como,Ketan Savla, Daron Acemoglu and Emilio Frazzoli.


Risk Prediction of a Behavior-based AdhesionControl Network for Online Safety Analysis of

Wall-climbing Robots

Daniel Schmidt and Karsten BernsUniversity of Kaiserslautern, Kaiserslautern, Germany

Keywords: Genetic Algorithm, Behavior-based,Risk Prediction, Climbing Robot.

Abstract: Risk analysis in combination with terrainclassification is a common approach in mobilerobotics to adapt robot control to surface conditions.But for climbing robots it is hard to specify, howthe robotic system and especially the adhesion isaffected by different surfaces and environmentalfeatures. This paper will introduce the climbingrobot CROMSCI using negative pressure adhesionvia multiple chambers, adaptive inflatable sealingsand an omnidirectional drive system. It presents theused behavior-based control network which allowsthe balancing of adhesion force, but fails in extremesituations. Therefore, a risk prediction has beendeveloped which evaluates behavioral meta-data andallows an estimation of current hazards caused bythe environment. This prediction is used to performevasive actions to prevent the robot from fallingdown.


A Particle Swarm Optimization Algorithm for theGrasp Planning Problem

Chiraz Walha, Hala Bezine and Adel M. AlimiUniversity of Sfax, Sfax, Tunisia

Keywords: Grasp Planning, Robotic Hand, ParticleSwarm Optimization (PSO).

Abstract: Computing a set of contact points betweena robotic hand and an object in order to fulfill somecriteria is the main problem of the grasp planning.An automatic grasp planning can produce a set ofjoint angles defining a configuration of the robotichand. The huge number of solutions that satisfy agood grasp is the main difficulty of such a planner. Inthis paper, we represent the grasp planning problemas an optimization problem and we propose a newalgorithm based on a Particle Swarm Optimization(PSO) technique. To generate the positions of thefingertips, the kinematic of the hand is modeled.


Sund

ay, 2

9

ICINCO 2012 FINAL PROGRAM AND BOOK OF ABSTRACTS: SUNDAY SESSIONS

Therefore, a simple PSO algorithm is described tooptimize the workspace of the operating hand basedon a quality of measure of the grasp. The simulationresults support the effectiveness of our approach.


Hardware on-Board Implementation of aPossible Solution to Introduce UAVs into Non

Segregated Areas

Sergio Taraglio, Vincenzo NanniENEA, Rome, Italy

Damiano TaurinoDeep Blue, Rome, Italy

University of Florence, Florence, Italy

Keywords: Game Theory, Software Agents for Intel-ligent Control Systems, Distributed Control Systems,Separation Assurance, Unmanned Aerial Vehicles,Engineering Applications.

Abstract: The implementation of a distributedcontrol algorithm for the safe flight of UnmannedAerial Vehicles (UAVs) in traffic conditions in anembedded processor unit is presented. Details onthe implementation are given. The control algorithmis designed in order to avoid separation infringementswith aircraft in the neighbourhood. The algorithm isgrounded on Satisficing Game Theory and suppliesmanoeuvres aimed at providing separation distanceamong aircraft. Some results of simulated flight trialsof UAVs in segregated and non segregated areasare reviewed. Some experiments on the capability ofthe embedded algorithm to pilot a simulated UAV indifferent traffic condition are presented.


Optimal Distributed Controller Synthesis forChain Structures

Applications to Vehicle Formations

Omid Khorsand, Assad Alam and Ather GattamiKTH-Royal Institue of Technology, Stockholm, Sweden

Keywords: Optimal Control, Distributed Control,Linear Quadratic Gaussian Control (LQG), IntelligentTransportation Systems (ITS), Heavy Duty Vehicle(HDV), Platooning.

Abstract: We consider optimal distributed controller

synthesis for an interconnected system subjectto communication constraints, in linear quadraticsettings. Motivated by the problem of finite heavyduty vehicle platooning, we study systems composedof interconnected subsystems over a chain graph.By decomposing the system into orthogonal modes,the cost function can be separated into individualcomponents. Thereby, derivation of the optimalcontrollers in state-space follows immediately.The optimal controllers are evaluated under thepractical setting of heavy duty vehicle platooningwith communication constraints. It is shown that theperformance can be significantly improved by addinga few communication links. The results show that theproposed optimal distributed controller outperforms asuboptimal controller in terms of control input energy.


A Stochastic Queueing Model for Multi-robotTask Allocation

Angelica Muñoz-Meléndez1,2, Pritviraj Dasgupta2

and William Lenagh21 National Institute of Astrophysics, Optics, and

Electronics, Sta Ma Tonantzintla, Mexico2 University of Nebraska at Omaha, Omaha, U.S.A.

Keywords: Robot Team, Task Allocation, Demand,Stochastic Queue.

Abstract: A central problem in multi-robot systemsis to solve the multi-robot task allocation problem.In this paper, a decentralized stochastic modelbased on stochastic queueing processes isapplied for an application of collective detectionof underground landmines where the robots are nottold the distribution or number of landmines to beencountered in the environment. Repeat demands ofinspection in the environment to ensure the accuracyof robot findings are necessary in this application.The proposed model is based on the estimationof a stochastic queue of pending demands thatrepresents the alternatives of action for a robot andis used to negotiate possible conflicts with otherrobots. We compare and contrast this methodwith a decentralized greedy approach based on thedistance towards the sites where inspection demandsare required. Experimental results obtained usingsimulated robots in the Webots© environment arepresented. The performance of robots is measuredin terms of two metrics, completion time and distancetraveled for processing a demand. Robots applyingthe stochastic queueing model obtained competitiveresults.


Sunday, 29



Attention Capabilities for AI Systems

Helgi Páll Helgason1 and Kristinn R. Thórisson1,21 Reykjavik University, Reykjavik, Iceland

2 Icelandic Institute for Intelligent Machines, Reykjavik,Iceland

Keywords: Artificial Intelligence, Attention, Re-source Management.

Abstract: Much of present AI research is basedon the assumption of computational systems withinfinite resources, an assumption that is eitherexplicitly stated or implicit in the work as researchersignore the fact that most real-world tasks must befinished within certain time limits, and it is the role ofintelligence to effectively deal with such limitations.Expecting AI systems to give equal treatment toevery piece of data they encounter is not appropriatein most real-world cases; available resources arelikely to be insufficient for keeping up with availabledata in even moderately complex environments.Even if sufficient resources are available, they mightpossibly be put to better use than blindly applyingthem to every possible piece of data. Findinginspiration for more intelligent resource managementschemes is not hard, we need to look no further thanourselves. This paper explores what human attentionhas to offer in terms of ideas and concepts forimplementing intelligent resource management andhow the resulting principles can be extended to levelsbeyond human attention. We also discuss someideas for the principles behind attention mechanismsfor artificial (general) intelligences.

Paper 9711:00 - 13:00 Room BarcelonaParallel Session 2a - Intelligent Control Systemsand Optimization

Automated Design of ReconfigurableManufacturing Systems

A Framework based on a Genetic Algorithm andDiscrete System Simulation

J. Padayachee and G. BrightUniversity of KwaZulu-Natal, Durban, South Africa

Keywords: Reconfigurable Manufacturing Systems,Discrete System Simulation, Genetic Algorithms,Manufacturing System Design.

Abstract: The concept of ReconfigurableManufacturing Systems (RMSs) was formulated

due to the global necessity for production systemsthat are able to economically evolve according tochanges in markets and products. Technologiesand design methods are under development toenable RMSs to exhibit transformable systemlayouts, reconfigurable processes, cells andmachines. Existing manufacturing design systemsdo not encapsulate concepts of reconfigurabilityin design mechanisms to obtain optimal RMSconfigurations. This paper presents a frameworkfor a resource allocation and shop floor designsystem within the context of RMSs. The frameworkfocuses on the automated generation of shopfloor configurations for systems with high productvariety and shared resources. The DEVS, (DiscreteEvent System Specification), formalism is usedto model reconfigurable equipment and simulatemanufacturing processes. The “design engine” in theproposed framework, implements a genetic algorithmfor the assembly, evaluation and optimisationof candidate shop floor configurations and theircorresponding DEVS models.


Evolutionary Symbiotic Feature Selection forEmail Spam Detection

Paulo Cortez1, Rui Vaz1, Miguel Rocha2, MiguelRio3 and Pedro Sousa2

1 Universidade do Minho, Guimarães, Portugal2 Universidade do Minho, Braga, Portugal3 University College London, London, U.K.

Keywords: Collaborative Filtering, Content-basedFiltering, Evolutionary Algorithms, Feature Selection,Naive Bayes, Spam Email, Symbiotic Filtering, TextClassification.

Abstract: This work presents a symbiotic filteringapproach enabling the exchange of relevant wordfeatures among different users in order to improvelocal anti-spam filters. The local spam filtering isbased on a Content-Based Filtering strategy, whereword frequencies are fed into a Naive Bayes learner.Several Evolutionary Algorithms are explored forfeature selection, including the proposed symbioticexchange of the most relevant features amongdifferent users. The experiments were conductedusing a novel corpus based on the well knownEnron datasets mixed with recent spam. Theobtained results show that the symbiotic approach iscompetitive.


Sund

ay, 2

9



Predictive Data Reduction in Wireless SensorNetworks using Selective Filtering

David James McCorrie, Elena Gaura, KeithBurnham and Nigel Poole

Coventry University, Coventry, U.K.

Keywords: Signal Reconstruction, OptimizationProblems in Signal Processing, Change DetectionProblems Instrumentation Networks and Software.

Abstract: In a wireless sensor network,transmissions consume a large portion of anodeâC™s energy budget. Data reduction isgenerally acknowledged as an effective means toreduce the number of network transmissions, therebyincreasing the overall network lifetime. This paperbuilds on the Spanish Inquisition Protocol, to furtherreduce transmissions in a single-hop wireless sensorsystem aimed at a gas turbine engine exhaustgas temperature (EGT) monitoring application. Anew method for selective filtering of sensed databased on state identification has been devised foraccurate state predictions. Low transmission ratesare achieved even when significant temperature stepchanges occur. A simulator was implemented togenerate flight temperature profiles similar to thoseencountered in real-life, which enabled tuning andevaluation of the algorithm. The results, summarizedover 280 simulated flights of variable duration (fromapproximately 58 minutes to 14 hours) show anaverage reduction in the number of transmissionsby 95%, 99.8% and 91% in the take-off, cruiseand landing phases respectively, compared totransmissions encountered by a sense-and-sendsystem sampling at the same rate. The algorithmgenerates an average error of 0.11 ± 0.04 °C overa 927 °C range.


A Hybrid Metaheuristic Approach to Solve theVehicle Routing Problem with Time Windows

Arthur T. GómezUniversidade do Vale do Rio dos Sinos, Porto Alegre,

Brazil

Cristiano Galafassi, Leonardo ChiwiacowskyUniversity of the Sinos Valley„ São Leopoldo, Brazil

Keywords: Metaheuristics, Tabu Search, GeneticAlgorithms, Vehicle Routing Problem.

Abstract: This paper addresses the CapacitatedVehicle Routing Problem with Time Windows, withconstraints related to the vehicle capacity andtime windows for customer service. To solve thisproblem two different metaheuristics are used:Tabu Search and Genetic Algorithms. Based onthese techniques a hybrid algorithm is developed.The main goal is the development of a HybridAlgorithm focused on the Vehicle Routing Problemwhich uses the intensification power of the TabuSearch and the diversification power of the GeneticAlgorithms, in order to obtain good quality solutionswithout compromising the computational time. In theexperiments are combined policies of diversificationand intensification in Tabu Search and GeneticAlgorithm to verify the efficiency and robustness ofthe proposed hybrid algorithm. Finally, the results arecompared with the best heuristic and exact methodsresults found in the literature. The Hybrid Algorithmhere proposed shows efficiency and robustness, withseveral optimal solutions achieved.


A Hybrid Algorithm using MetaheuristicsApplied to H.264/AVC Video Encoder

Arthur Tórgo Gómez and Iris Correa C. LinckUniversidade do Vale do Rio dos Sinos, São Leopoldo,

Brazil

Keywords: Metaheuristics, Tabu Search, GeneticAlgorithm, H.264 Video Encoder, CODEC.

Abstract: This paper focuses on the study andthe analysis of the dynamic relationship among sixparameters of the H.264/AVC video encoder, thatare: frame rate, bit rate, quantization parameterfor I slice, B slice, and P slice, and the number of


Sunday, 29


B slices in the GOP (Group of Pictures). For thisstudy, it was developed and implemented a hybridalgorithm called Simulator of Metaheuristics appliedto a CODEC (SMC). The SMC algorithm consists oftwo metaheuristics that are Tabu Search and GeneticAlgorithm. It tries to find the best configuration ofthe studied parameters in order to obtain a goodquality and compression of the encoded video inthe H.264/AVC standard. The SMC algorithm usesa maximization objective function as an objectiveevaluation method to reach the proposed goals.


View-based SLAM using Omnidirectional Images

D. Valiente, A. Gil, L. Fernández and O. ReinosoMiguel Hernández University, Elche, Spain

Keywords: Visual SLAM, Omnidirectional Images.

Abstract: In this paper we focus on the problem ofSimultaneous Localization and Mapping (SLAM)using visual information obtained from theenvironment. In particular, we propose the useof a single omnidirectional camera to carry out thistask. Many approaches to visual SLAM concentrateon the estimation of the position of a set of 3D points,commonly denoted as visual landmarks which areextracted from images acquired at the environment.Thus the complexity of the map computation grows asthe number of visual landmarks in the map increases.In this paper we propose a different representation ofthe environment that presents a series of advantagescompared to the before mentioned approaches,such as a simplified computation of the map and amore compact representation of the environment.Concretely, the map is represented by a set of viewscaptured from particular places in the environment.Each view is composed by its position and orientationin the map and a set of 2D interest points representedin the image reference frame. Thus, in each viewthe relative orientation of a set of visual landmarksis stored. During the map building stage, the robotcaptures an image and finds corresponding pointsbetween the current view and the views stored inthe map. Assuming that a set of correspondingpoints is found, the transformation between bothviews can be computed, thus allowing us to buildthe map and estimate the pose of the robot. Inthe suggested framework, the problem of findingcorrespondences between views is troublesome.Consequently, with the aim of performing a morereliable approach, we propose a new method tofind correspondences between two omnidirectionalimages when the relative error between them ismodeled by a gaussian distribution which correlates

the current error on the map. In order to validate theideas presented here, we have carried out a seriesof experiment in a real environment using real data.Experiment results are presented to demonstrate thevalidity of the proposed solution.


Link Quality Estimator for a Mobile Robot

Narek Pezeshkian, Joseph D. Neff and AbrahamHart

SPAWAR Systems Center, Pacific, San Diego, U.S.A.

Keywords: Link Quality, Video Quality, Estimator,Metric, Mobile Robot.

Abstract: Maintaining link connectivity between amobile robot and its control station in a non-line-of-sight environment is challenging. One solution is touse intermediate relay radios that the robot can carryand deploy when and where needed to maintain thelink. However, the precise placement locations forthe relays are not known ahead of time. Therefore,the deployment decision must be formulated onlineand the relays deployed before the link with thecontrol station breaks. A link-quality estimator isdeveloped based on video throughput and receivedsignal strength indicator data. The estimator takesinto account human perception of video quality thatis obtained via subjective testing by an operator. Thedata is used to train the link-quality estimator, whichissues an alert that can be used as a trigger for anautomatic relay deployment mechanism or to advisethe operator to manually deploy relays before the linkbetween the robot and control station fails.


Global Visual Features based on RandomProcess

Application to Visual Servoing

Laroussi Hammouda1, Khaled Kaaniche1,2, HassenMekki1,2 and Mohamed Chtourou1

1 University of Sfax, Sfax, Tunisia2 University of Sousse, Sousse, Tunisia

Keywords: Visual Servoing, Global Visual Features,Mobile Robot.

Abstract: This paper presents new global visualfeatures: random distribution of limited set of pixelsluminance. Our approach aims to improve the real-time performance of visual servoing applications.


Sund

ay, 2

9


In fact, using these new features, we reduce thecomputation time of the visual servoing scheme.Our method is based on a random process whichensures efficient and fast convergence of the robot.The use of our new features removes the matchingand tracking process. Experimental results arepresented to validate our approach.


A Bug-based Path Planner Guided withHomotopy Classes

Emili Hernández, Marc Carreras and Pere RidaoUniversity of Girona, Girona, Spain

Keywords: Path Planning, Homotopy Classes,Robotics.

Abstract: This paper proposes a bug-basedpath planning algorithm guided topologically withhomotopy classes. Homotopy classes provide atopological description of how paths avoid obstaclesin the workspace. They are generated with amethod we developed, which builds a topologicalenvironment based on the workspace that allowsto compute homotopy classes systematically. Thehomotopy classes are sorted according to a heuristicestimation of their lower bound. Only those withthe smaller lower bound are used to guide the pathplanner we propose, called Homotopic Bug (HBug),which efficiently computes paths in the workspacethat accomplish homotopy classes. Results showthe feasibility of our method. A comparison withwell-known path planners has also been included.


Efficient Distributed Fusion Filtering Algorithmsfor Multiple Time Delayed Systems

Il Young Song and Moongu JeonGwangju Institute of Science and Technology, Gwangju,

Korea, Republic of

Keywords: Distribute Fusion, Multi Sensor, KalmanFilter, Time-delayed System, Receding Horizon.

Abstract: In this paper, we provide twocomputational effective multi sensor fusion filteringalgorithms for discrete-time linear uncertain systemswith state and observation time delays. The firstalgorithm is shaped by algebraic forms for multirate sensor systems, and then we propose a matrix

form of filtering equations using block matrices.The second algorithm is based on exact cross-covariance equations. These equations are usefulto compute matrix weights for fusion estimation in amultidimensional-multisensor environment. Also, ourproposed filtering algorithm is based on the recedinghorizon strategy in order to achieve high estimationaccuracy and stability under parametric uncertainties.We demonstrate the low computational complexitiesof the proposed fusion filtering algorithm and how theproposed algorithm robust against dynamic modeluncertainties comparing with Kalman filter with timedelays.


On Specification Informatics in Discrete-eventSystems

State-transparency for Clarity of Finite Automataas Control Specifications

Amrith DhananjayanSchool of Computer Engineering, Singapore, Singapore

Kiam Tian SeowNanyang Technological University, Singapore, Singapore

Keywords: Discrete-event Systems, SupervisoryControl, Finite Automata, Specification Informatics,State Transparency, Human Perception.

Abstract: In control of discrete-event systems(DES’s), the formalization of control requirementsfrom natural language statements is essentially ahuman endeavor. Without automated support tools,human designers often face the uncertainty of notknowing if the control requirements formalized inthe rudimentary DES formalism of automata areas intended, motivating the automata-theoreticstudy of specification informatics in the field ofDES’s. A specification automaton that renders itslinguistic description more transparent should helpdesigners ascertain the prescribed requirement.Such transparency may be formalized in eitherthe state space or the event space of the DES.In this paper, treating the former as fundamental,a state-transparent specification automaton isconceptualized with respect to a full specificationautomaton (‘full’ in the sense of having all the apriori transitional constraints of the DES embeddedin it). It contains only specification relevant statescalled specification epochs. Each epoch denotesa “well-defined" disjoint subset of states of thefull specification automaton in the same phaseof execution, meaningfully aggregated such thatthe resultant specification automaton retains


Sunday, 29


the original restrictiveness on the DES. Theproblem of maximizing the state-transparency ofspecification automata for DES’s is then theoreticallyformulated. Subject to human perceptive or cognitionlimits, we believe that such a maximally state-transparent specification automaton could be morecomprehensible, showing clearly the compliantexecution of the system through a minimum numberof specification epochs, and should aid designersin clarifying if the requirement prescribed is the oneintended.


Self-Scheduled H∞ Control of a Wind TurbineA Real Time Implementation

Florin Sebastian Tudor, Dumitru Popescu and DanStefanoiu

”Politehnica” University of Bucharest, Bucharest, Romania

Keywords: Wind Power, Renewable Energy, Ro-bust, Control, Gain-scheduling, LPV, Modeling, RealTime Implementation.

Abstract: This paper is concerned with the designof robust gain-scheduled controllers with guaranteedH∞ performance for a horizontal axis wind turbine(HAWT) with variable-speed and fixed-pitch. Thecontrol problem in terms of Linear Parameter-Varying (LPV) plants is stated and the theoreticalbackground of the design method is given. Due tosome interesting properties outlined in this paper,the synthesis problem is reduced to solving off-linea finite-dimensional set of Linear Matrix Inequalities(LMIs), making the controller suited for real-timeapplications. The computed LPV controller focuseson multiple objectives such as mechanical fatiguereduction, speed regulation and mode stabilizationwith simultaneously maximizing energy capture. Theperformances obtained through this control methodare discussed and presented by means of a set ofsimulations. A real-time control algorithm for thelarge-scale wind turbines is also proposed.


Cost Functions for Scheduling Tasks inCyber-physical Systems

Abhinna Jain, C. M. Krishna, Israel Koren andZahava Koren

University of Massachusetts, Amherst, U.S.A.

Keywords: Cyber-physical Systems, Task Schedul-ing, Cost Functions, Controlled Plant Dynamics.

Abstract: In Cyber Physical Systems (CPS),computational delays can cause the controlledplant to exhibit degraded control. The traditionalapproach to scheduling in such systems has beento define controller task deadlines, based on thedynamics of the controlled plant. Controller tasks arethen scheduled to meet these deadlines; meetingthe deadline is considered the sole criterion forscheduling success.This traditional approach has the advantage ofsimplicity, but overlooks the fact that the quality ofcontrol depends on the actual task response times.Two different schedules, each satisfying the taskdeadlines, can provide very different levels of controlquality, if their task response times are different.In this paper, we consider using cost functionsof task response time to capture the impact ofcomputational delay on the quality of control. Sincethe controller workload typically consists of multipletasks, these cost functions are multivariate innature. Furthermore, since these tasks are generallycoupled, the response time of one control task canaffect the sensitivity of the controlled plant to theresponse times of other tasks.In this paper, we first demonstrate how a multivariatecost function can be formulated to quantify theeffect of computational delays in vehicles. Wethen develop cost-sensitive real-time control taskscheduling algorithms. We use as an applicationexample an automobile: the controller workloadconsists of steering and torque control. Our resultsindicate that cost-function-based scheduling providessuperior control to the traditional deadline-only-basedapproach.


Sund

ay, 2

9


14:30 - 15:30 Room PlenarySpace Robotics - Guidance, Navigation andControl ChallengesKeynote Speaker: Jurek Sasiadek

Space RoboticsGuidance, Navigation and Control Challenges

Jurek SasiadekCarleton University, Ottawa, Canada

Abstract: This keynote presents methods andtechnologies used in guidance, navigation andcontrol of space robotics. It addresses some basicissues related to control of space manipulators andautonomous flying and ground travelling vehicles.Space manipulators such as, Remote ManipulatorSystem (RMS), Space Station Remote manipulatorSystem (SSRMS) and Special Purpose DextrousManipulator (SPDMD) were used successfullyfor variety of space tasks. Robotic manipulatorsin Space pose several navigation and controlchallenges. Robots structures and joints are flexibleand highly nonlinear and their designs include severaldegrees of freedom (DOF). Each degree of freedomhas to be controlled with high accuracy to achievesatisfactory positioning. Planetary rovers are used inspace to explore the planets and moons. There aremany types of robotics vehicles. Their main objectiveis to explore the planets, take samples of groundand rocks at desired location. Often robotics roversare directed to take images of the surroundings,mapping, as well as, performing other scientificmeasurements. Accurate localization of autonomousrobots in featureless environment is very difficultand often requires advanced methods of guidanceand navigation. It is worth noting that existingsensor as gyroscopes, accelerometers and otherinertial systems may not work properly when robotis travelling with low speed. Moreover, odometricsensors such as, encoders and tachometers mayalso not work properly in the hostile environment ofouter planets. In such environment, sensor fusionbecomes one of the most important technologiesused for guidance, localization, and navigationof planetary robotic vehicles. Integrating severalmeasurements and several types of sensors toretrieve accurate data is one of the most successfultechnologies used in autonomous robotics. Free –flying and orbiting robots also will play an importantrole in Space exploration and servicing. De-orbitingsatellites and servicing space floating vehicles onorbit is an important part of robotics activities inSpace. Although, there are not too many successfulapplications of such types of robots, it is expectedthat in future they will play much more importantrole. General methods of navigation are shown anddiscussed and their possible application in Space arepresented and discussed.

Paper 1915:30 - 16:30 Foyer - ICINCOPoster Session 1

Gesture and Body Movement Recognition in theMilitary Decision Support System

Jan Hodicky and Petr FrantisUniversity of Defense, Brno, Czech Republic

Keywords: Military Decision Support System, Com-mon Operational Picture, Visualization, Kinect.

Abstract: The paper deals with the result ofthe research activity in the field of militarydecision support system. It brings a new wayof communication between system and commander.Kinect - the low cost gesture and body movementrecognition device was employed to control 3Dvisualization of real-time battlefield situation.Experiments confirmed the correctness of Kinectusing to support all phases of decision makingprocess. The quality of operation planning andcontrol was increased.

Paper 6015:30 - 16:30 Foyer ICINCOPoster Session 1

Use of Flow Control on Car DumpersA Case of Success at Vale

Bruno Eduardo Lopes and Luis Freitas CoutinhoVale, Itaqui, Brazil

Keywords: Car Dumper, Flow Control.

Abstract: The flow control of the car dumper (VV,in Portuguese) used to be done in open loop, i.e.,operators used to manually adjust the speed ofthe feeders (AL, in Portuguese) to make the actualflow equal to the desired flow. The problem wasthat these manual adjustments were not alwaysperformed properly and the result was a loss inproductivity and the occurrence of overflow on thefeeders. This work demonstrates the techniquesof control, modelling and identification used for theimplementation of flow control in closed loop inthe car dumpers at the Ponta da Madeira MaritimeTerminal, Vale.


Sunday, 29



Predictive Control of Unmanned Formations

Martin Saska and Libor PreucilCzech Technical University in Prague, Prague, Czech

Republic

Keywords: Mobile Robots, Formation Control,Receding Horizon Control, Obstacle Avoidance.

Abstract: A receding horizon control basedapproach for guiding of autonomous formationsof nonholonomic robots in a leader-followerconstellation is proposed in this paper. Thepresented method ensures dynamic obstacleavoidance, formation coordination as well as failuretolerance. The robustness of the algorithm is verifiedby numerical multi-robot experiments. Besides,effects of system’s parameters on the algorithmperformance are investigated.


Business Process Development for IndustrialCluster

Taivo KangilaskiTallinn University of Technology, Tallinn, Estonia

Keywords: Industrial Cluster, Partner Network,Enterprise Architecture Management, Business Pro-cess Management, Quality Management System.

Abstract: Small production companies frequentlyhave only limited amount of customers, especiallyin country sides. To have more orders and to bemore competitive, they need to cooperate andform Industrial Clusters. To manage IndustrialCluster, there should be firm Quality ManagementSystem. The current article proposes approachabout handling that complicity, which is related to theIndustrial Cluster Management.


Quantitative Estimates of Stability in ControlledGI | D | 1 | ∞ Queueing Systems and in Control

of Water Release in a Simple Dam Model

Evgueni Gordienko and Juan Ruiz de ChávezAutonomous Metropolitan University - Iztapalapa, Mexico

City, Mexico

Keywords: Queueing, Inventory and Water ReleaseModels, Markov Control Process, Stability Index,Prokhorov Metric.

Abstract: We consider two applied discrete-timeMarkov control models: waiting times process inGI | D | 1 | ∞ queues with a controlled service rateand water release control in a simple dam modelwith independent water inflows. The stochasticdynamics of both models is determinated by asequence of independent and identically distributedrandom variables with a distribution function F. Inthe situation when an available approximation Fis used in place of the unknown F, we estimatethe deterioration of performance of control policiesoptimal with respect to the total discounted cost andthe average cost per unit of time. For this purposewe introduce a stability index and find uppers boundsfor this index expressed in terms of the Prokhorovdistance between the distributions functions F and F.When F ≡ Fm is the empirical distribution functionobtained from a sample of size m in average the

stability index is less than a constant times m−13 .


From a Multi-robot Global Plan to Single-robotActions

Bernd BrüeggemannFraunhofer FKIE, Wachtberg, Germany

Elmar Langetepe, Andreas LenerzUniversity of Bonn, Bonn, Germany

Dirk SchulzFraunhofer Institute for Communication, InformationProcessing and Ergonomics, Wachtberg, Germany

Keywords: Multi-robot Systems, Planning, PlanExecution, Coordination, Algorithmic Geometry.

Abstract: Planning for and coordinating robots ina multi-robot system (MRS) is crucial for optimizingthe performance of the whole MRS. Thus a planfor the movement of the MRS must exist. If some


Sund

ay, 2

9


centralized entity calculates the plan, it may result inone plan for the whole group. Such a global plan,which shows how the MRS can reach a goal state,has to be transformed into action guidelines for eachrobot. This task becomes harder if such a globalplan includes dependencies caused by necessarycooperation of the robots. In this paper we presentan approach to transform a global multi-robot planinto single-robot actions. We also provide a methodto determine how many robots are needed to fulfilthe global plan while obeying some constraints. Herewe use plans generated by a coordinated navigationplanner with spatial constraints, but the method couldbe expanded to a more general class of plans builtfrom a centralized entity.


Monte Carlo Localization using the GlobalAppearance of Omnidirectional ImagesAlgorithm Optimization to Large Indoor

Environments

Lorenzo Fernández, Luis Payá, David Valiente,Arturo Gil and Oscar Reinoso

Miguel Hernández University, Elche, Spain

Keywords: Autonomous Robot, Global Appearance,Panoramic Images, Fourier Signature, Monte-CarloLocalization.

Abstract: In this paper we deal with the problemof robot localization using the visual informationprovided by a single omnidirectional cameramounted on the robot, using techniques basedon the global appearance of panoramic images. Ourmain objective consists in showing the feasibility ofthe appearance-based approaches in a localizationtask in a relatively large and real environment. First,we study the approaches that permit us to describeglobally the visual information so that it representswith accuracy locations in the environment. Then, wepresent the probabilistic approach we have used tocompute the most probable pose of the robot when itperforms a trajectory within the map. At the end, wedescribe the kind of environments and maps we haveused to test our localization algorithms and the finalresults. The experimental results we show have beenobtained using real indoor omnidirectional images,captured in an office building under real conditions.


Compensation of Unknown Input Dead Zoneusing Equivalent-Input-Disturbance Approach

Liyu Ouyang1,2, Jinhua She3, Min Wu1,2 andHiroshi Hashimoto4

1 Central South University, Changsha, China2 Hunan Engineering Laboratory for Advanced Control

and Intelligent Automation, Changsha, China3 Tokyo University of Technology, Tokyo, Japan

4 Advanced Inst. of Industrial Tech., Tokyo, Japan

Keywords: Compensation of Nonlinearity, DeadZone, Distortion Factor, Equivalent Input Disturbance(EID), Nonlinearity.

Abstract: This paper considers the problem of thecompensation of an unknown dead zone in theinput of a plant. A new compensation method ispresented based on the estimation of an equivalentinput disturbance (EID). Unlike other methods, thismethod does not require the exact information ofa dead zone. First, we consider the dead zone asan input-dependent disturbance and employ an EIDestimator to estimate it. Then, we incorporate theestimate in the control input and compensate theeffect of the dead zone almost completely. Simulationresults demonstrate the validity of the method.


Feature Selection Combined with NeuralNetwork for Diesel Engine Diagnosis

M. Benkaci1 and G. Hoblos1,21 Institut de Recherche en Systèmes Electroniques

Embarqués, Saint-Etienne du Rouvray, France2 Ecole Supérieure d’Ingénieurs en Génie Électrique,

Saint-Etienne du Rouvray, France

Keywords: Leaks Detection, Automotive Diagnosis,Feature Selection, Neural Data Classification, DieselAir Path.

Abstract: The Feature selection is an essential stepfor data classification used in fault detection anddiagnosis process. In this work, a new approachis proposed which combines a feature selectionalgorithm and neural network tool for leaks detectiontask in diesel engine air path. The Chi2 is used asfeature selection algorithm and the neural networkbased on Levenberg-Marquardt is used in systembehaviour modelling. The obtained neural network isused for leaks detection. The model is learned and


Sunday, 29


validated using data generated by xMOD. This toolis used again for test. The effectiveness of proposedapproach is illustrated in simulation when the systemoperates on a low speed/load and the consideredleak affecting the air path is very small.


Distributed Control of Dangerous Goods Flows

Claudio Roncoli, Chiara Bersani and Roberto SacileUniversity of Genova, Genova, Italy

Keywords: Dangerous Good Transport, Opti-misation Problem, Decentralised Control, LinearQuadratic Regulator.

Abstract: A risk-based approach to managingdangerous goods (DG) transport flows by roadis proposed, solving a real-time flow assignmentproblem. The model assumes the plannedscheduling of the fleets and the medium plannedspeed for vehicles known a priori. The objective isto plan in the vehicle tour schedules in base on DGand general vehicle flows data on the infrastructuresacquired in real time. The model minimises both thetotal risk on the road network and the gap betweenthe real delivery times with respect to the plannedones. The first objective is a social intent of aNational Authority and the second one representsthe main important cost minimisation for DG carriers.The proposed model is formulated according toan original distributed control approach, based onthe decomposition of the original centralised linearquadratic problem.


Development of Mobile Research Robot

A. Baums, A. Gordjusins and G. KanonirsInstitute of Electronic and Computer Science, Riga, Latvia

Keywords: Research Robot, Real Time System,Timelines, TUF, Physical Model.

Abstract: For new autonomous mobile robot design,the real time problem analysis at different periodsof robot activity phases is made. The robot sensorand actuator cluster structure is used. At first therobot is determined as a hard real time system whenall phases defined and executed sequentially are inhard deadlines. At second for the robot activity withhard and soft deadline execution phases is proposedusing of the time/utility function (TUF). For time and

energy consumption estimation, the flexible robotphysical model is developed and used. Wirelesstechnology is proposed for new autonomous mobilerobot design.


Modified Hybrid Evolutionary Strategies Methodfor Termination Control Problem with Relay

Actuator

Ivan Ryzhikov and Eugene SemenkinSiberian State Aerospace University, Krasnoyarsk,

Russian Federation

Keywords: Termination Control Problem, Evolution-ary Strategies, Relay Control, Two-point BoundaryProblem, Dynamic Systems.

Abstract: The termination control problem, i.e., theproblem with finite time and relay control function,is considered. The proposed approach fits differentcontrol problems, e.g., problems with the fixedor free time, problems with the set-up actuatorcharacteristics and problems where the actuator canbe tuned. The considered system is the nonlineardynamic one and the number of relay switch pointsassumed to be tuned indirectly. To find out thesolution of the given problem, the modified evolutionstrategies method is suggested. The proposedapproach is useful also for the non-analyticalsystem models and systems that can be evaluatednumerically.


An Application of Goal Programming Techniquefor Reconfiguration of Transfer Lines

Fatme Makssoud, Olga Battaïa and AlexandreDolgui

Ecole des Mines de Saint Etienne, Saint Etienne Cedex 2,France

Keywords: Transfer Lines, Reconfiguration, GoalProgramming, Multi-objective Optimization.

Abstract: In this paper, the reconfiguration problemof transfer machining lines is addressed. Thisproblem appears when an existing line hasto be adapted for the production of a new ormodified product. The objective is to minimizethe reconfiguration line cost. The compatibilityconstraints between old and new operations haveto be taken into account. Therefore, a compromisebetween introducing new equipment and reusing


Sund

ay, 2

9


old one is to be found. A goal programming modelfor this optimization problem is developed. Thismathematical model minimizes the reconfigurationcost of transfer line as the primary objective andmaximizes the reusability of old equipment as thesecond objective.


On the Temperature Control for a Test CaseShort Pipe Network Central Heating System

Nikolaos D. Kouvakas and Fotis N. KoumboulisHalkis Institute of Technology, Psahna Evoias, Greece

Keywords: Central Heating Systems, TemperatureControl, Autonomous Heating, Dynamic Controllers.

Abstract: In the present paper the mathematicalrepresentation of a test case central heating systemwith a short pipng network, three radiators and oneboiler heating two apartments is developed in theform of a nonlinear model. A linear dynamic controllerachieving independent apartment temperaturecontrol and being unaffected from the externaltemperature is proposed. The controller is developedon the basis of a linear approximant. The closed loopperformance is tested through simulations on theoriginal nonlinear model.


Modelling and Simulation of Human-likeMovements for Humanoid Robots

Parvin Abedi1 and Ali Leylavi Shoushtari1,21 Shoushtar Branch, Islamic Azad University, Shoushtar,

Iran, Islamic Republic of2 South Tehran Branch, Islamic Azad University, Tehran,

Iran, Islamic Republic of

Keywords: Human Body Dynamics, HumanoidRobots, Optimization-based Simulation.

Abstract: The humanoid robots are bio-inspiredmodels of human body. The mechanical structureof humanoid robots consists of several joints andsegments. Numerous degrees of freedom arecaused the redundancy problem. There is anunanswered question concerning with strategieswhich central nervous system implements to predictthe human posture and gesture during differentmovements. A 7 degree of freedom model is usedfor modelling humanoid robot and an optimization-based method is planned to simulation of human

motion. The joints angles and torques are subjectedas optimization variables. The joints range ofmotion and limits of actuator torques are used asoptimization constraints. The weight lifting is themotion which is subjected to simulation. Finallythe results presented for two velocity lifting. Theresult shows the body posture varies naturally andthe weight maintain at the end position at final timecorrectly.


A Control Strategy for Reducing FuelConsumption in a Hybrid Electric Vehicle

Babici Leandru Corneliu Cezar and Alexandru Onea“Gh. Asachi” Technical University, Iasi, Romania

Keywords: Control, Command, Hybrid, PowerVehicle.

Abstract: Hybrid electric vehicles are one ofthe most suitable alternative for conventionalautomobiles. This paper describes a control strategyfor a hybrid electric vehicle, in order to reduce thefuel consumption, and to maintain a reasonablestate of charge (SOC), at the end of the drive cycle.The main goal is to split the requested power fromthe driver between the internal combustion engine,and the electric motor, such way to decrease thefuel consumption, and to maintain the dynamicperformances. The algorithm was tested usingMatlab Simulink and ADVISOR interface. The resultsinclude statistical comparisons of the standard drivecycles using default model and the modified controlstrategy.


Observability of Transportation SystemsA Methodology for Reliability Analysis in

Logistics and Manufacturing

Jan PinkowskiOFFIS Institute for Information Technology, Oldenburg,

Germany

Axel HahnCarl von Ossietzky University, Oldenburg, Germany

Keywords: Reliability Analysis, Material Flow,Observability, Information Flow, Logistics.

Abstract: Real world events are observed bysensors since decades, for instance in the logisticswhere packages are identified and tracked. This


Sunday, 29


information result in an information flow. Thisinformation flow is used to control the physicalmaterial flow. Hence, the information flow is a digitalrepresentation of the physical material flow. However,to guarantee that the digital representation is inalignment to the physical world is a challenging task.Especially for scenarios with manual operations, therepresentation is vulnerable for errors. This paperproposes a generic approach to assure consistencybetween digital and physical world. The paperpresents a methodology to model the monitoring ofphysical entities and to analyse the model to evaluatethe risk of unreliable digital representation.


Towards Manufacturing Execution Systems forthe Food and Beverage Packaging Industry

Stefano Caselli, Michele PatteraUniversity of Parma, Parma, Italy

Massimo RicciOCME s.r.l., Parma, Italy

Keywords: Manufacturing Execution Systems,Packaging, SCADA.

Abstract: In this paper, we describe the issuesfaced in bringing advanced supervisory and controlconcepts in the fragmented food and beveragepackaging industry. Although packaging equipmentmanufacturers must cope with tight cost constraints,heterogenous machines, and customers’ requests forspecialized features, we take the position that timeis ready for bringing more advanced features suchas automated production flow and full productionsupervision even in this domain of consumer orientedand highly personalized products. ManufacturingExecution Systems, which have been introducedwith success in other manufacturing areas, canbe applied in food and beverage packaging takingadvantage from recently proposed standards. Thiseffort requires to supplement the software base ofexisting machines with an interface layer ensuringinteroperability according to the existing standards.By consolidating these standards and adopting gooddesign practices, flexible integration, supervision andcontrol of packaging lines can be obtained.

Paper 815:30 - 16:30 FoyerPoster Session 1

Comparison the Performance of HybridHMM/MLP and RBF/LVQ ANN Models

Application for Speech and Medical PatternClassification

Lilia LazliBadji Mokhar University, Annaba, Algeria

Mounir BoukadoumUnivesité du Québec A Montréal (UQAM), Montreal,

Canada

Abdennasser Chebira, Kurosh MadaniPARIS XII University, Lieusaint, France

Keywords: Speech Recognition, Medical Diagnosis,Hybrid RBF/LVQ Model, Hybrid HMM/MLP Model.

Abstract: In the last several years, the hybridmodels have become increasingly popular. We useinvolves multi-network RBF/LVQ structure and hybridHMM/MLP model for speech recognition and medicaldiagnosis.

Paper 916:45 - 18:15 Room BarcelonaSpecial Session on Artificial Neural Networks andIntelligent Information Processing - ANNIIP

GPU-based Parallel Implementation of a GrowingSelf-organizing Network

Giacomo Parigi1, Angelo Stramieri1, Danilo Pau2

and Marco Piastra11 University of Pavia, Pavia, Italy

2 STMicroelectronics, Agrate Brianza, Italy

Keywords: Growing Self-organizing Networks,Graphics Processing Unit, Parallelism, SurfaceReconstruction, Topology Preservation.

Abstract: Self-organizing systems are characterizedby an inherently local behavior, as their configurationis almost exclusively determined by the union ofthe states of each of the units composing thesystem. Moreover, all state changes are mutuallyindependent and governed by the same laws. Inthis work we study the parallel implementation ofa specific subset of this broader family, namelythat of growing self-organizing networks, in relationto parallel computing hardware devices basedon Graphic Processing Units (GPUs), which areincreasingly gaining popularity due to their favourablecost/performance ratio. In order to do so, we firstdefine a new version of the standard, sequentialalgorithm, where the intrinsic parallelism of the


Sund

ay, 2

9


execution is made more explicit and then weperform comparative experiments with the standardalgorithm, together with an optimized variant of thelatter, where an hash index is used for speed. Ourexperiments demonstrates that the parallel versionoutperforms both variants of the sequential algorithmbut also reveals a few interesting differences in theoverall behavior of the system, that might be relevantfor further investigations.


Parallel Batch Pattern Training of RecirculationNeural Network

Volodymyr Turchenko1, Vladimir Golovko2 andAnatoly Sachenko1

1 Ternopil National Economic University, Ternopil, Ukraine2 Brest State Technical University, Brest, Belarus

Keywords: Parallel Batch Pattern Training, Recircu-lation Neural Network, Parallelization Efficiency.

Abstract: The development of a parallel batchpattern back propagation training algorithm of arecirculation neural network is presented in thispaper. The model of a recirculation neural networkand usual sequential batch pattern algorithm of itstraining are theoretically described. An algorithmicdescription of the parallel version of the batch patterntraining method is presented. The parallelizationefficiency of the developed parallel algorithm isinvestigated on the example of data compressionand principal component analysis. The results of theexperimental researches show that the developedparallel algorithm provides high parallelizationefficiency on a parallel symmetric multiprocessorcomputer system. It allows applying the developedparallel software for the facilitation of scientificresearch of neural network-based intrusion detectionsystem for computer networks.


Forecasting Financial Success of HollywoodMovies

A Comparative Analysis of Machine LearningMethods

Dursun DelenOklahoma State University, Tulsa, U.S.A.

Ramesh ShardaOklahoma State University, Stillwater, U.S.A.

Keywords: Prediction, Box-office Receipts, Hol-lywood, Machine Learning, Neural Networks,Sensitivity Analysis.

Abstract: Forecasting financial success of aparticular movie has intrigued many scholars andindustry leaders as a worthy but challenging problem.In this study, we explore the use of machine learningmethods to forecast the financial performance of amovie at the box-office before its theatrical release.In our models, we convert the forecasting probleminto a multinomial classification problem—rather thanforecasting the point estimate of box-office receipts;we classify a movie based on its box-office receiptsin one of nine categories, ranging from a “flop” to a“blockbuster.” Herein, we present our comparativeprediction results along with variable importancemeasures (using sensitivity analysis on trainedprediction models).


Fuzzy Control of a Hybrid Renewable PowerSystem based on Real-time Matlab-PLC

Communication through OPC

Isaías González Pérez, A. José Calderón Godoyand Manuel Calderón Godoy

University of Extremadura, Badajoz, Spain

Keywords: Hybrid Power System, RenewableEnergy, Electrolyzer, Fuzzy Control, OPC, PLC.

Abstract: This paper presents the design of afuzzy logic controller to operate an electrolyzerof an experimental test-bed of hybrid wind-solarsystem with hydrogen storage. This controller runsin Simulink and is linked through Open ProcessControl interface with the industrial programmablelogic controller responsible of global management ofthe installation. Real-time data exchange and control


Sunday, 29


of the process variables have been successfullyachieved and obtained results under real conditionsare presented.


Data-based Tuning of PI Controllers for VerticalThree-Tank Systems

Mircea-Bogdan Radac1, Bogdan-AlexandruBigher1, Radu-Emil Precup1, Emil M. Petriu2,

Claudia-Adina Dragos1, Stefan Preitl1 andAlexandra-Iulia Stînean1

1 “Politehnica” University of Timisoara, Timisoara,Romania

2 University of Ottawa, Ottawa, Canada

Keywords: Iterative Feedback Tuning, Level Control,PI Controllers, Vertical Three-tank Systems.

Abstract: This paper suggests the applicationof Iterative Feedback Tuning (IFT) as a data-basedcontrol technique to parameter tuning of PI controllersdedicated to vertical three-tank systems. The levelcontrol in the first two tanks is carried out using amultivariable control system structure which consistsof two control loops, one for each level. The twoPI controllers in these control loops are first tunedin terms of the Modulus Optimum method. NewIFT algorithms are proposed in order to ensure theperformance improvement of level control systemsby means of six steps assisted by experiments. Theexperimental results show the strong performanceimprovement obtained after few iterations of IFTalgorithms in terms of a model reference trackingoptimization problem.


Differential Evolution in Parameter IdentificationFuel Cell as an Example

Aki Sorsa, Anssi Koskenniemi and Kauko LeiviskäUniversity of Oulu, Oulu, Finland

Keywords: Differential Evolution, Identification,Nonlinear Model, Fuel Cell.

Abstract: Evolutionary algorithms are optimizationmethods and their basic idea lies in biologicalevolution. They suit well for large and complexoptimization problems. In this study, differentialevolution is applied for identifying the parameters of

the nonlinear fuel cell model. Different versions of thealgorithm are used to compare the genetic operatorsthey use. One problem with the studied algorithms isalso in defining the internal parameters that regulatethe development of the population. In this paper,entropy is used for defining the population size andother parameters are tuned using recommendationsfrom the literature and by trial-and-error. The resultsshow that DE/rand-to-best/1/bin is the most suitablealgorithm for the studied problem. Selection of thecrossover operator has no considerable effect onthe results. The results also show that the studiedidentification problem has a lot of local minimathat are very close to each other that makes theoptimization problem even more challenging.


Spacecrafts’ Control Systems Effective VariantsChoice with Self-configuring Genetic Algorithm

Eugene Semenkin and Maria SemenkinaSiberian State Aerospace University, Krasnoyarsk,

Russian Federation

Keywords: Spacecraft Control System, ControlContours Models, Markov Chains, Effective VariantChoice, Optimization, Self-configuring Genetic Algo-rithm.

Abstract: The work of the spacecraft controlsystem is modeled with Markov chains. Small andlarge models for the technological and command-programming control contours are developed.The way of the calculation of the control contoureffectiveness indicators is described. Specialself-configuring genetic algorithm that requiresno settings determination and parameter tuningis proposed for choosing effective variants ofspacecraft control system. The high performanceof the suggested algorithm is demonstrated throughexperiments with test problems and then is validatedby the solving hard optimization problems.


Sund

ay, 2

9



Autonomously Traversing ObstaclesMetrics for Path Planning of Reconfigurable

Robots on Rough Terrain

Michael Brunner, Bernd Brüeggemann and DirkSchulz

Fraunhofer Institute for Communication, InformationProcessing and Ergonomics FKIE, Wachtberg, Germany

Keywords: Metric, Traversability, Obstacle, RoughTerrain, Reconfigurable Chassis, Motion Planning,Mobile Robot, Autonomy.

Abstract: The fixed chassis design of commonlyemployed mobile robots restricts their application tofairly flat environments, as the wheel diameters orthe track heights impose hard limits on their mobility.Unstructured outdoor and urban environments alikecomprehend many different invincible obstacles formost of those systems, like stairs, boulders or rubble.However, there are mobile robots with reconfigurablechassis providing a higher degree of mobility andenabling them to overcome such obstacles. Yet,current planning algorithms rarely exploit thoseenhanced capabilities, limiting these systems to thesame environments as the fixed chassis robots.This paper focuses on the metrics used by our motionplanner. The employment of a two-stage planningapproach allows us to use different cost functionsfor the initial path search and the detailed motionplanning step. The purpose of the initial search isto quickly find a fast environment-driven path to thegoal. Hence, it uses fast computable heuristics toassess the drivability, i.e. a risk quantification andthe utmost operation limits of the robot model. Thedetailed planning step determines the desired robotconfigurations. For this purpose, we consider theactuator controls, the system’s stability, an estimateof the traction, and the driving speed in addition tothe quantities used in the first stage.We present experiments to illustrate the influence ofthe safety weights and real world experiments whichprove the validity and feasibility of the metrics usedby our motion planning algorithm.


Design and Control of a Novel UnmannedGround Vehicle

M. Osinuga and A. LanzonThe University of Manchester, Manchester, U.K.

Keywords: Unmanned Ground Vehicle, UnmannedAerial Vehicle, Modelling, Inverted Pendulum, Ro-bust Control.

Abstract: This paper presents a novel design of afull six-degree-of-freedom, tri-rotor-actuated single-wheeled vehicle. When stationary on the ground, themodel of the vehicle mimics an inverted pendulumsystem, a typical representative of highly nonlinearsystems with non-minimum phase and unstablecharacteristics. The full nonlinear model of thevehicle is derived using Newton-Euler approach, butthe objective herein is to balance the single-wheeledvehicle in its unstable upright position as this is aprecursor to rolling motion. Robust feedback andJacobi linearization techniques are invoked on thenonlinear dynamics of the vehicle in the relevant axisfor subsequent controller synthesis. The synthesizedcontrollers are verified and compared by meansof numerical simulations, where it is shown thatthe design objective of stabilizing the vehicle in itsunstable equilibrium (upright) position is robustlyachieved.


Trajectory Tracking Control of NonholonomicWheeled Mobile Robots

Combined Direct and Indirect Adaptive Controlusing Multiple Models Approach

Altan OnatAnadolu University, Eskisehir, Turkey

Metin OzkanEskisehir Osmangazi University, Eskisehir, Turkey

Keywords: Combined Direct and Indirect AdaptiveControl, Trajectory Tracking Control, NonholonomicWheeled Mobile Robots, Multiple Models Approach.

Abstract: This paper presents a novel methodologyfor the trajectory tracking control of nonholonomicwheeled mobile robots using multiple identificationmodels. The overall control system includes twostages. In the first stage, a kinematic controllerdeveloped by using kinematic model provides therequired linear and angular velocities of the robot for


Sunday, 29


tracking a reference trajectory. In the second stage,the required velocities are taken as the inputs toan adaptive dynamic controller which uses multipleadaptive models for the parameter identification. Theproposed adaptive dynamic controller is developedusing a combined direct and indirect adaptive controlapproach where both prediction and tracking errorsare used for identification. Simulation results showthe effectiveness of the proposed combined directand indirect control scheme and multiple modelsapproach.


From Robot Commands to Real-time RobotControl

Transforming High-level Robot Commands intoReal-time Dataflow Graphs

Andreas Schierl, Andreas Angerer, Alwin Hoffmann,Michael Vistein and Wolfgang Reif

University of Augsburg, Augsburg, Germany

Keywords: Industrial Robotics, Robot Programming,Real-time Robot Control.

Abstract: Task descriptions in robotics alwaysprovide a level of abstraction in order to simplifythe use of robots. Nevertheless, aspects such asexecution time determinism and closed-loop controlare still essential for industrial-strength roboticssystems. For this reason, we propose an approachto combine high-level task description with real-timerobot control. At application runtime, coordinated andsensor-guided robot actions are composed using anobject-oriented application programming interface.The resulting high-level command descriptions arethen automatically transformed into dataflow graphsand executed with real-time guarantees on robothardware. The approach is illustrated with severalexamples.


Modelling of a Grasping and ManipulationController

Pavel Dzitac and Abdul Md MazidCentral Queensland University, Rockhampton, Australia

Keywords: Grasping, Slippage, Grasp Control,Tactile Sensor.

Abstract: This paper presents the development ofa robot grasping and manipulation control system,

including the modelling approach for the controlfunctions, and the criteria used for functional moduledesign in order to achieve the required functionalityand allow its integration into the overall control model.This work is an example of a practical implementationof a robotic grasping and manipulation controller andmay be relevant to researchers looking for anexample of a practical controller design “fromscratch”.


Dynamic Model of a 7-DOF Whole ArmManipulator and Validation from Experimental

Data

Zaira Pineda Rico, Andrea Lecchini-Visintini andRodrigo Quian Quiroga

University of Leicester, Leicester, U.K.

Keywords: Whole Arm Manipulator Model, FrictionModel, Friction Identification.

Abstract: The present paper describes the design ofthe dynamic model of a 7 degrees of freedom wholearm manipulator implemented in SimMechanics. Thefriction phenomena of the manipulator is identified,represented through a fitted model and includedin the system model with the aim of increment theaccuracy of the model with respect to the real system.The characteristics of the model make it suitableto test and design control strategies for motion andfriction compensation in MATLAB/Simulink.


False Positive Outliers Rejection for ImprovingImage Registration Accuracy

Application to Road Traffic Aerial Sequences

Ines Hadj Mtir1, Khaled Kaâniche1, PascalVasseur2 and Mohamed Chtourou1

1 University of Sfax, Sfax, Tunisia2 University of Rouen, Rouen, France

Keywords: Outliers Rejection, Features Matching,Image Registration, Motion Compensation, VehiclesDetection, Aerial Sequences.

Abstract: The objective of our system is to detectvehicles from aerial sequences. Theses sequencesare taken from a camera mounted on UAV which fliesover roads and highways. Our approach is to firstlycompensate the motion introduced by the dynamicbehaviour of the camera. This leads us to a problem


Sund

ay, 2

9


of image registration. The moving regions (vehicles)are after that extracted using residual motion. Theaim of this paper is to present a combined method forfeatures matching and outliers rejection to increasethe accuracy of the registration phase. We use first,the SIFT descriptors and then outliers are rejectedusing geometric constraints. This leads to a betterregistration and a minimum of false alarms in thedetection phase.


Towards the Integrated Simulation andProgramming of Palletizing Lines

Antonello Calò, Davide BurattiOCME S.r.l., Parma, Italy

Dario Lodi Rizzini, Stefano CaselliUniversity of Parma, Parma, Italy

Keywords: Industrial Machine, Simulation, VisualProgramming.

Abstract: In this paper, we discuss the advantagesand the issues of simulation and visual programmingof palletizing machines and, in general, palletizinglines, and we illustrate an integrated software toolsuite that meets such requirements. The increasingcomplexity of lines and the variability of productformats require a common machine model for allthe tools, together with the independence fromvisualization to allow software reuse and extendibility.Furthermore, the model should also be able toinclude palletizing line components external to themachine that are critical for performance and whosemodel is only partially known.


Task-based Method for Designing UnderactuatedElastic Mechanisms

Shoichiro Kamada, Youngwoo Kim and GoroObinata

Nagoya University, Nagoya, Japan

Keywords: Underactuated Mechanism, Task-basedDesign, Principal Component Analysis, ElasticElement.

Abstract: In this paper, we introduce a task-basedmethod for designing underactuated mechanismswhich actuators are linked with the joints via elasticelements. We consider multi-joint mechanisms thatcontain fewer independent actuators than the joints.

The elastic elements work as convertors from thedisplacement of the actuators to the joint torquesof the mechanisms. In our method, we analyzethe joint motions of the mechanisms during thecompletion of each task and the level of participationof each joint for few specific tasks. The results ofthis study can be used for the synthesis of dedicatedunderactuated mechanisms that can operate in a lowtask coordinate space and for the systematic designof underactuated mechanisms.


Supporting Mobile Robot’s Tasks throughQualitative Spatial Reasoning

Pascal Rost, Lothar Hotz and Stephanie von RiegenUniversität Hamburg, Hamburg, Germany

Keywords: Qualitative Spatial Reasoning, Onto-logical Reasoning, Cognitive Robotics, Knowledge-based Systems Applications.

Abstract: In this paper, we present an applicationof qualitative spatial reasoning technologies forsupporting mobile robot tasks. While focusingon detection of interaction ability, we provide acombination of the spatial reasoning calculi RCC-8and CDC as well as their integration with OWL-basedontologies. An architecture that uses Prolog andcomplex-event processing implements our approach.We illustrate the results with a mobile robot scenarioin a restaurant.


Flow Optimization for Iron Ore ReclaimingProcess

Bruno Eduardo Lopes, José Pinheiro de Moura,Denis Anderson Ribeiro, Fernando Henrique Costa

e Borges and Marco Antônio de SouzaVALE, São Luis, Brazil

Keywords: Iron Ore Reclaimers, PID Control,Process Flow, Predictive Control.

Abstract: The purpose of the this paper is todemonstrate the optimization of the flow for theiron ore reclaiming process by reclaimers over railsusing implementation of PID control algorithms,identification techniques, Predictive Control anda new effort-based learning method herein calledreinforcement by difference learning method andproportional reinforcement learning method. The


Sunday, 29


outcome was an increase of productivity, withreduction of the flow variability and on the amount ofoverflow occurrences.


Using Linear Systems Theory to Study NonlinearDynamics of Relay Cells

Rahul Agarwal and Sridevi V. SarmaJohns Hopkins University, Baltimore, U.S.A.

Keywords: lay Neurons, Thalamus, Reliability,Hodgkin Huxley Type Models.

Abstract: Relay cells are prevalent throughoutsensory systems and receive two types of inputs:driving and modulating. The driving input containsreceptive field properties that must be relayed whilethe modulating input alters the reliability of thisrelay. In this paper, we analyze a biophysical basednonlinear model of a relay cell and use systemstheoretic tools to construct analytic bounds on howwell the cell transmits a driving input as a functionof the neuron’s electrophysiological properties, themodulating input, and the driving signal parameters.Our analysis applies to both 2nd & 3rd order modelas long as the neuron does not spike without a drivinginput pulse and exhibits a refractory period. Ourbounds suggest, for instance, that if the frequencyof the modulating input increases and the DC offsetdecreases, then reliability increases. Our analysisalso shows how the biophysical properties of theneuron (e.g. ion channel dynamics) define theoscillatory patterns needed in the modulating inputfor appropriately timed relay of sensory information.


Nonparametric Identification of Nonlinearity inWiener-Hammerstein Systems

Grzegorz MzykWroclaw University of Technology, Wroclaw, Poland

Keywords: System Identification, Wiener-Hammerstein System, Nonparametric Methods,Kernel Estimate, Convergence Analysis.

Abstract: In the paper we recover the staticcharacteristic of Wiener-Hammerstein (sandwich)system from input-output data. The system isexcited and disturbed by random processes with

arbitrary distribution. Two kernel-based estimatesare proposed and compared. It is shown that theycan successfully recover the system characteristicunder small amount of a priori information about thestatic characteristic and the surrounding dynamicblocks. The identified nonlinear function is notparametrized and is not assumed to be invertible,which is common restriction in the literature. Theorders of linear dynamic blocks are also unknown.The convergence of the estimates take place for thepoints in which the input probability density functionin positive. The effectiveness of the algorithms isillustrated in simulation example.


From PID to Extended Learning Control

Cristiano Maria Verrelli”Tor Vergata” University, Rome, Italy

Keywords: Learning Control, Uncertain NonlinearSystems, Minimum Phase, Output Feedback Form,PID Control, Permanent Magnet Step Motors.

Abstract: It has been recently shown in Marino,Tomei, and Verrelli (2011) that the output errorfeedback regulation problem with (unknown) periodicreference and/or disturbance signals of knowncommon period can be effectively solved for theclass of single-input, single-output, minimumphase, nonlinear, time-invariant systems in outputfeedback form (of known relative degree one ortwo) which are affected by unknown parametersand unknown output-dependent nonlinearities. Theresulting nonlinear control, which relies on advancedlearning control techniques, can be interpretedas a generalization of the classical PID controlwhich solves the problem when both reference anddisturbance signals are constant. In this paper, wepresent sophisticated analytical arguments whichprove that the learning control designed in Marino,Tomei, and Verrelli (2011) can be endowed with aperiod identifier when the output reference signal isperiodic of uncertain period but available at each timeinstant. The generalized resulting control preservesthe achievement of the closed loop propertiesobtained in Marino, Tomei, and Verrelli (2011)while maintaining an overall simple structure. Theapplication of the presented control techniques tothe position synchronization problem for current-fedpermanent magnet step motors with non-sinusoidalflux distribution and uncertain position-dependentload torque allows us to provide a solution to a yetunsolved problem.


Sund

ay, 2

9



State Dependent Parameter Modelling of aDC-DC Boost Converter in Discontinuous

Conduction Mode

U. Hitzemann1 and K. J. Burnham1,2

1 Coventry University, Coventry, U.K.2 Wroclaw University of Technology, Wroclaw, Poland

Keywords: DC-DC Boost Converter, NonlinearCircuits, State-dependent Parameter Modelling,System Identification, System Modelling.

Abstract: This paper is concerned with themodellingof a DC-DC boost converter, operating indiscontinuous conduction mode (DCM). Theapproach chosen is to model the converter using astate-dependent parameter (SDP) model approachwhich is expected to be able to deal with thenonlinearities of the system, as well as a varyingload. The modelling procedure presented, makesuse of input-output data only and no physical insightinto the system is required. Results are verified vialaboratory experiments.


Applying Hyperbolic Wavelets in FrequencyDomain Identification

Alexandros Soumelidis, József BokorComputer and Automation Research Institute, Budapest,

Hungary

Ferenc SchippEötvös Loránd University, Budapest, Hungary

Keywords: System Identification, Discrete–timeSystems, Frequency–domain Representations,Wavelets, Hyperbolic Geometry.

Abstract: The paper elaborates a hyperbolicwavelet construction for representing signals inthe Hardy space H2 on the unit disc. An efficientcomputing scheme based on the matrix form ofthe representation is worked out. The waveletcoefficients can be computed on the basis of discretetime–domain measurements. This wavelet is usedto reconstruct poles of functions in H2 as the basisof nonparametric frequency–domain identification ofdiscrete–time signals and systems.


Sunday, 29

Monday Sessions

FINAL PROGRAM AND BOOK OF ABSTRACTS: MONDAY SESSIONS ICINCO 2012

09:45 - 10:45 Room PlenaryRecent Advances in Physical Human-RobotInteractionKeynote Speaker: Alessandro De Luca

Recent Advances in Physical Human-RobotInteraction

Alessandro De LucaUniversità di Roma "La Sapienza", Roma, Italy

Abstract: In this talk I will present some controlaspects related to physical Human-Robot Interaction(pHRI). In order to achieve the goal of a safer,dependable, and high-performance collaboration ofrobots and humans in industrial and professionalservice tasks, an integrated approach is needed,where mechatronic design, sensory information,on-line task planning, and reactive control issues arecombined in a single framework. At lowest level, I willillustrate the control design for lightweight robots withcompliant joints, possibly driven by variable stiffnessactuation (VSA). In the latter case, simultaneousand decoupled control of both robot motion andcompliance is possible. This allows implementinga safe strategy by increasing compliance whenthe robot is moving fast, so as to reduce the riskof potential injuries in undesired collisions. Fordynamically varying and uncertain environments,the typical case in HRI tasks, on-line collisionavoidance methods driven by exteroceptive sensorsand rapid detection of unavoidable physical contactsusing only encoder sensing will be considered.These situations ask for a portfolio of consistentrobot reaction strategies in a hierarchy of safety,co-existence, and active collaboration in the human-robot interaction. Some illustrative examples will begiven. The presentation will mainly highlight the latestresults obtained in the on-going European projectSAPHARI, including also recognition of humanmotion intentions, human-driven kinestethic learningof robot motion, and reactive action generationpatterns.


Visual Servoing Path-planning with Spheres

Tiantian Shen and Graziano ChesiUniversity of Hong Kong, Pokfulam, Hong Kong

Keywords: Sphere, Path-planning, Visual Servoing.

Abstract: This paper proposes a path-planningapproach for visual servoing in the case where theobserved object features are points and spheres. Two

main situations are considered. In the first situation,it is supposed that at least two points and at least asphere are observed. In the second situation, it issupposed that at least three spheres are observed.The problem consists of planning a trajectory in orderto ensure the convergence of the robot end-pointto the desired location while satisfying visibilityand workspace constraints, in particular includingocclusion and collision avoidance. A solution basedon polynomial parametrizations is proposed in orderto determine a feasible path in the 3D space, andsuch a path is then followed by tracking its imageprojection through image-based visual servoing.Some simulation results illustrate the proposedapproach.


Rotor Speed Sensor Fault Detection in InductionMotors

Riccardo Marino, Stefano Scalzi, Patrizio Tomei andCristiano Maria Verrelli

University of Rome “Tor Vergata”, Rome, Italy

Keywords: Fault Detection, Speed Sensor Faults,Induction Motors.

Abstract: The problem of detecting a speed sensorfault in induction motor applications with loadtorque and rotor/stator resistances uncertaintiesis addressed. It is shown that in typical operatingconditions involving constant rotor speed and fluxmodulus and non-zero load torque, a constantnon-zero (sufficiently large) difference between themeasured speed and the actual speed may beon-line identified by an adaptive flux observer whichincorporates a convergent rotor resistance identifierand relies on the measured rotor speed and statorcurrents/voltages.


Mon

day,

30

ICINCO 2012 FINAL PROGRAM AND BOOK OF ABSTRACTS: MONDAY SESSIONS


Visual Anomaly Detection in Production Plants

Alexander Maier1, Tim Tack1 and OliverNiggemann1,2

1 OWL University of Applied Sciences, Lemgo, Germany2 Fraunhofer IOSB-INA, Lemgo, Germany

Keywords: Anomaly Detection, Production Plant,Automation System, Visualization Technique, VisualAnalytics.

Abstract: This paper presents a novel method forvisual anomaly detection in production plants. Sincethe complexity of the plants and the number ofsignals that have to be monitored by the operatorgrows, there is a need of tools to overcome theinformation overflow. The human is highly able torecognize irregularities in figures. More than 80% ofthe perceived information is captured visually. Theapproach proposed in this paper exploits this factand subjects data to make the operator able to findanomalies in the displayed figures. In three stepsthe operator is lead from the visualization of thenormal behavior over the anomaly detection and thelocalization of the faulty module to the anomaloussignal.


Global Optimal Solution to SLAM Problem withUnknown Initial Estimates

Usman Qayyum and Jonghyuk KimAustralian National University, Canberra, Australia

Keywords: Greedy Random Adaptive SearchProcedure, Gauss-Newton Optimization, OptimalSolution, Map-joining.

Abstract: The paper presents a practicalapproach for finding the globally optimal solutionto SLAM. Traditional methods are based uponlocal optimization based strategies and are highlysusceptible to local minima due to non-convex natureof the SLAM problem. We employed the nonlinearglobal optimization based approach to SLAM byexploiting the theoretical limit on the numbers of localminima. Our work is not reliant on good initial guess,whereas existing approaches in SLAM literatureassume good starting point to avoid local minimaproblem. The paper presents experimental results

on different datasets to validate the robustness of theapproach, finding the global basin of attraction withunknown initial guess.


Design and Analysis of an Automated HeavyVehicle Platoon

Gábor Rödönyi, Péter Gaspár, József BokorHungarian Academy of Sciences, Budapest, Hungary

László PalkovicsKnorr-Bremse Brake-systems Gmbh, Budapest, Hungary

Keywords: Vehicle Platoon, Peak-to-peak Gain,Model Set Identification, Unfalsification.

Abstract: From the model set identification throughthe control design and robust performance analysisto the implementation and experimental verification,the whole design process for an automated vehicleplatoon is presented. The goal is to demonstrate thatsafe platooning with acceptable performance can beachieved by utilizing the services already availableon every commercial heavy trucks with automatedgearbox. Using the services, the control designreduces to the selection of four design parameters,the static gain coefficients of the output-feedback-input-feed-forward controller common for everyvehicle. It is experimentally demonstrated, that innormal driving maneuvers, the spacing errors areless than three meters.


Trajectory Tracking Control by LMI-basedApproach for Car-like Robots

Nicoleta Minoiu EnacheRenault SAS, Guyancourt, France

Keywords: Trajectory Tracking, Car-like Vehicle,Passenger Vehicle, LMI Optimization.

Abstract: A lot of research has been doneconcerning motion planning and trajectory trackingcontrol for robots, including car-like robots.Nevertheless, most of the methods require verywell defined trajectories, continuous and severaltimes derivable. Frequently, the system has to bewritten in a specific form, like the chained form,and the path obtained in the original space may notsatisfy additional constraints in the original space orsingularities can occur in the control law. The goalof this work is to investigate a control method to do


Monday, 30


the trajectory tracking control of car-like robots in theoriginal space, which does not require trajectoriesseveral times derivable, but with good robustnessand pursuit properties in order to implement it ona passenger vehicle. The control law for trajectorytracking presented here is derived from a methoddeveloped for unicycles. This is based on a realtime combination of static linear feedbacks that areobtained by an off line LMI (Linear Matrix Inequalities)approach.


Development of Parallel Two-wheel Vehicle withLower Gravity Center of Vehicle Body

Yoshiyuki NodaUniversity of Yamanashi, Kofu, Japan

Yukinori Sago, Kazuhiko TerashimaToyohashi University of Tech., Toyohashi, Japan

Kiyoaki KakiharaKER Co., Ltd., Toyokawa, Japan

Hirotoshi KawamuraSinfonia Technology Co., Ltd., Toyohashi, Japan

Keywords: Parallel Two-wheel Vehicle, LowerGravity Center, Sway Suppression Control of VehicleBody, Active Mass Damper, Backstepping Control.

Abstract: This paper presents an advanced paralleltwo-wheel vehicle which has lower gravity centerof vehicle body. The gravity center is assigned atthe lower position than the wheel axis. Therefore,the vehicle has a structure of the pendulum, andenables the vehicle body with the passenger toalways keep the stable posture, even if the vehicle isin the power-off or control-off condition. And, 2-DOFjoystick which has operation with back-and-forthdirection and rotation is applied to the proposedvehicle. The elderly or handicapped passenger canoperate easily the vehicle by this joystick. Moreover,in order to suppress the sway of the vehicle body asa pitching oscillation while driving the vehicle, thesway suppression control system with an active massdamper system is proposed in this paper. The controlsystem is designed by a backstepping method. Theeffectiveness of the proposed sway suppressioncontrol system with the active mass damper systemis verified by the experiments using the proposedparallel two-wheel vehicle with lower gravity center.


Vision based Real-time Modeling of DynamicUnstructured Environments in Driving Scenarios

Andrei Vatavu and Sergiu NedevschiTechnical University of Cluj-Napoca, Cluj-Napoca,

Romania

Keywords: Environment Representation, Polyline,Object Contour, Iterative Closest Point, DrivingAssistance, Stereo-vision, Object Delimiters, MotionDetection.

Abstract: The detection of moving traffic participantsis an essential intermediate step for higher leveldriving technology tasks. Regardless of the typeof used sensors, dynamic environment modelingbecomes even more difficult when the surroundingworld is unstructured and heterogeneous. In suchcomplex environments the representation systemcan be affected by noisy measurements, occlusions,wrong data association or unpredictable nature ofthe traffic participants. We propose a solution ofrepresenting the dynamic environment in real-timeby using the pairwise alignment of free-form modelsand considering the advantages provided by adense stereovision system. Instead of registeringthe whole 3D point cloud, our method is based onextracting and registering a more compact model ofthe environment taking into consideration the mostvisible object cells from the ego car. The proposedmethod is based on information provided by a DigitalElevation-Map, but can be easily adapted for othertypes of intermediate representations.

Paper 4011:00 - 13:00 Room BarcelonaParallel Session 4a - Robotics and Automation

Hierarchical Planning of Modular BehaviourNetworks for Office Delivery Robot

Jong-Won Yoon and Sung-Bae ChoYonsei University, Seoul, Korea, Republic of

Keywords: Office Delivery Robot, Hybrid RobotControl, Behaviour Networks.

Abstract: This paper proposes a hybrid architecturebased on hierarchical planning of modular behaviournetworks for generating autonomous behaviours ofthe office delivery robot. Behaviour networks suitablefor goal-oriented problems are exploited for thearchitecture, where a monolithic behaviour networkis decomposed into several smaller behaviourmodules. In order to construct and adjust sequencesof the modules the planning method considers


Mon

day,

30


the sub-goals, the priority in each task and theuser feedback. It helps a robot to quickly react indynamic situations as well as achieve global goalsefficiently. The proposed architecture is verified onboth the Webot simulator and Khepera II robot inoffice environment with delivery tasks. Experimentalresults confirms that a robot can achieve goals andgenerate module sequences successfully even inunpredictable situations, and the proposed planningmethod reduces the elapsed time during tasks by17.5%.


An Integrated Approach for Efficient MobileRobot Trajectory Tracking and Obstacle

Avoidance

Aleksandar Cosic, Marko Susic and Dusko KaticUniversity of Belgrade, Belgrade, Serbia

Keywords: Mobile Robots, Trajectory Tracking,Obstacle Avoidance, Fuzzy Logic, Intelligent Control.

Abstract: An approach for nonholonomic two-wheeled mobile robot trajectory tracking andobstacle avoiding is presented in this paper. If thedesired trajectory is provided by high level planner,trajectory tracking problem can be solved in variousways. In this paper, tracking is provided usingproportional-integral (PI) or fuzzy logic controller(FLC). Unfortunately, tracking is never perfect,due to uncertainties and obstacles can changetheir positions in time. In order to overcome thesedifficulties, additional correction controller must beused. Here is proposed fuzzy controller, whichslightly changes control action of the trackingcontroller in order to prevent collision with obstacles.This approach is proved to be efficient even indynamic environments. Simulation results arepresented as illustration of the proposed approach.


Construction and Modeling of a VariableCollective Pitch Coaxial UAV

Jinqiang Cui, Fei Wang, Zhengyin Qian, Ben M.Chen and Tong H. Lee

National University of Singapore, Singapore, Singapore

Keywords: Coaxial Helicopter, Unmanned AerialVehicle, Flapping Dynamics, Model Identification.

Abstract: This paper describes the construction andmodeling of a coaxial unmanned aerial vehicle for

in-forest operation. The bare helicopter platform isupgraded and mounted with an onboard navigationsystem, which includes central processing unitsand sensors such as inertial measurement unit,camera and scanning laser range finder. The modelstructure of the helicopter is formulated, in whichthe model of rotor thrust and roll-pitch dynamics aredescribed in details. The flapping dynamics of therotor and the stabilizer bar are presented and lumpedinto a state-space model. The parameters of thestate-space model are identified in frequency domainusing CIFER. Time domain verification with a newset of flight data exhibits excellent agreement withthe prediction of the identified model.


Mobile Robots Pose TrackingA Set-membership Approach using a Visibility

Information

Rémy Guyonneau, Sébastien Lagrange and LaurentHardouin

Université d’Angers, Angers, France

Keywords: Mobile Robots, Localization, PoseTracking, Visibility, Interval Analysis.

Abstract: This paper proposes a set-membershipmethod based on interval analysis to solve the posetracking problem. The originality of this approach isto consider weak sensors data: the visibility betweentwo robots. By using a team of robots and thisboolean information (two robots see each other ornot), the objective is to compensate the odometryerrors and be able to localize, in a guaranteedway, the robots in an indoor environment. Thisenvironment is supposed to be defined by two sets,an inner and an outer characterizations. Simulatedresults allow to evaluate the efficiency and the limitsof the proposed method.


A Joint Segmentation and Classification ofObject Shapes with Feedback for 3D Point

Clouds

Frauke Wübbold and Bernardo WagnerLeibniz Universität Hannover, Hannover, Germany

Keywords: Object Classification, Segmentation,Feedback, 3D Point Cloud, 3D Shape.

Abstract: Limited knowledge and limited deduction


Monday, 30


abilities are among the main restraints of autonomousrobots for acting truly autonomously. This especiallybecomes obvious in the area of object recognitionand classification, where many methods rely onknowledge teached manually in a prior setup step.Self-generating this knowledge from environmentperception with a set of rules would significantlyincrease the robots autonomy as well as supersedemanual training effort. In this paper, we proposea novel approach to rule-based classificationfor 3D point clouds by means of object shape,which additionally overcomes typical problemsfrom a separate prior segmentation by integratingclassification feedback into the segmentationprocess. Although it is still in its conceptual state, weexplain in detail why we consider this approach to bevery promising.


Contact-free Magnetic Clutch Applied forFlywheel Cell System

Nan-Chyuan Tsai and Hong-Seng AwNational Cheng Kung University, Tainan City, Taiwan

Keywords: Servo Gap-Retained Mechanism, HybridMagnetic Actuator, Feedback-Linearized SlidingMode Control.

Abstract: A TDOF (Two Degrees of Freedom) ServoGap-Retained Mechanism (SGRM) is proposed andverified by experiments. It consists of a flywheeland an Intelligent Posture Tracking System (IPTS).The flywheel is regarded as the tracking objectiveof the IPTS. The IPTS is mainly composed by anintelligent disc and two pairs of Hybrid MagneticActuators (HMAs). The posture of the intelligentdisc is controlled by the magnetic forces inducedby the HMAs to retain a constant gap with respectto the eccentric flywheel. Since the HMA is highlynonlinear, a Feedback-Linearized Sliding ModeControl (FLSMC) is synthesized to account forsystem parameter nonlinearities. The proposedSGRM is part of the flywheel cell system. When theMGU (Motor/Generator Unit) in flywheel cell operatesat idle mode, the shaft of flywheel will be separatedfrom MGU in order to avoid the energy loss of theflywheel by the back EMF induced by the magneticfield of MGU. The shaft of flywheel and MGU stillneed to maintain synchronous power transmissionso that a contact-free clutch has to be equipped.The role of SGRM in a flywheel cell is to ensurethe centerline of the flywheel properly is alignedwith the magnetic clutch. Intensive experimentalsimulations are undertaken to verify the feasibility of

the proposed SGRM and FLSMC.


Modeling and Simulation of a WastewaterPumping Plant

Mohamed AbdelatiIUG, Gaza, Palestinian Territory, Occupied

Felix Felgner, Georg FreySaarland University, Saarbrücken, Germany

Keywords: Wastewater System Modeling, Simula-tion, Automation, Modelica.

Abstract: Modeling wastewater pumping plantsis rarely addressed in the literature. Standardcomponent models as found in fluid simulation toollibraries are too complex, due to their projectedgenerality, to be used for these applications. Lack ofmodels results in a burden on engineers who haveto test their control scenarios on real implementedsystems. This may lead to unexpected delaysand painful costs. In this work, easily manageablecomponent-oriented models are derived and appliedto the modeling and simulation of a real wastewaterpumping system. The model derived in this paperis implemented in Modelica, and it helps betterunderstanding the system dynamics. Thereby, atool is provided for evaluating the performance ofpossible control schemes.


Support Vector Machines for Identification ofHCCI Combustion Dynamics

Vijay Manikandan JanakiramanUniversity of Michigan, Ann Arbor, U.S.A.

Jeff SterniakRobert Bosch LLC, Farmington Hills, U.S.A.

Dennis AssanisStony Brook University, Stony Brook, U.S.A.

Keywords: Support Vector, Identification, Combus-tion, Homogeneous Charge Compression Ignition,HCCI, Neural Networks, Nonlinear Regression,Engine Model, Control Model.

Abstract: Homogeneous charge compressionignition (HCCI) is a promising technology for Internal


Mon

day,

30


Combustion Engines to improve efficiency andreduce nitrogen oxides emissions. Control of HCCIcombustion is often model-based, and it is vital tohave a good model of the engine to make controldecisions. The HCCI engine is characterized bycomplex chemical kinetics whose physical modelingis difficult and laborious. Identification is an effectivealternative to quickly develop control oriented modelsfor such systems. This paper formulates a SupportVector Regression (SVR) methodology for developingidentification models capturing HCCI combustionbehavior. Measurable quantities from the enginesuch as net mean effective pressure (NMEP) andcrank angle at 50% mass fraction burned (CA50) canbe used to characterize and control the HCCI engineand are considered for identification in this study. Theselected input variables include injected fuel mass(FM) and valve events {intake valve opening (IVO),exhaust valve closing (EVC)}. Transient data froma gasoline HCCI engine recorded at stable HCCIconditions is used for training, validating and testingthe SVR models. Comparisons with the experimentalresults show that SVR with Gaussian kernels can bea powerful approach for identification of a complexcombustion system like the HCCI engine.


Handling Delayed Fusion in Vision-AugmentedInertial Navigation

Ehsan Asadi and Carlo L. BottassoPolitecnico di Milano, Milano, Italy

Keywords: Delayed Fusion, Vision-Augmented INS,State Estimation.

Abstract: In this paper we consider the effectsof delay caused by real-time image acquisitionand feature tracking in a previously documentedVision-Augmented Inertial Navigation System. Atfirst, the paper illustrates how delay caused by imageprocessing, if not explicitly taken into account, canlead to appreciable performance degradation of theestimator. Next, three different existing methodsof delayed fusion are considered and compared.Simulations and Monte Carlo analyses are used toassess the estimation error and computational effortof the various methods. Finally, a best performingformulation is identified, that properly handles thefusion of delayed measurements in the estimatorwithout increasing the time burden of the filter.


Multivariable Discrete Time Repetitive ControlSystem

Hammoud SaariEcole Nationale Supérieure Maritime, Bou Ismail, Algeria

Bernard CaronUniversité de Savoie, Annecy le Vieux, France

Keywords: Repetitive Control, Multivariable Sys-tems, Invertible Systems, Non Invertible Systems,Tracking.

Abstract: This paper deals with an iterative learningcontrol law for multivariable systems. The desiredinputs are supposed to be known and periodic. Theprinciple of the control is to make outputs as close aspossible to desired inputs at each new period. Afterthe design of multivariable repetitive controller, wegive the stability condition of the algorithm and somesimulation results.


Reconstruction-based Set-valued ObserverA New Perspective for Fault Detection within

Uncertain Systems

Letellier Clément, Chafouk Houcine and HoblosGhaleb

Institut de Recherche en Systèmes ElectroniquesEmbarqués, Rouen, France

Keywords: Uncertain Systems, Set-valued Ob-server, Sensor Reconstruction, Fault Detection.

Abstract: This paper presents an extension of aparticular type of observer called the Set-valuedObserver; this kind of observer is very well suitedfor uncertain fault detection. But some limitationsrestrict its use. Indeed, all the sensors are neededto observe the state and as a consequence thismethod does not allow fault detection when somesensor information is not available. Other work hasfocused on the well-known Luenberger Observerapplied to uncertain systems; but once again, thisoption is limited. Indeed, it is difficult to converge thealgorithm because of the wrapping effect inducedby recursivity. Here a new approach is proposedcombining the power of the two algorithms. TheLuenberger Observer coupled with the Set-ValuedObserver allows us to reconstruct the states without


Monday, 30


divergence. This combination is a substantialcontribution for fault detection within uncertainsystems.


Adaptive Gravitational Search Algorithm forPI-fuzzy Controller Tuning

Radu-Codrut David1, Radu-Emil Precup1, Emil M.Petriu2, Mircea-Bogdan Radac1, Constantin

Purcaru1, Claudia-Adina Dragos1 and Stefan Preitl31 “Politehnica” University of Timisoara, Timisoara,

Romania2 University of Ottawa, Ottawa, Canada

3 Politehnica University of Timisoara, Timisoara, Romania

Keywords: Adaptive Gravitational Search Algo-rithms, Process Gain Sensitivity, Simulation Results,Takagi-Sugeno PI-fuzzy Controllers.

Abstract: This paper proposes an adaptiveGravitational Search Algorithm (aGSA) focusedon tuning of Takagi-Sugeno PI-fuzzy controllers (T-SPI-FCs). The algorithm adapts two depreciation lawsof the gravitational constant to the iteration index, aparameter in the weighted sum of all forces exertedfrom the other agents to the iteration index, and thereset at each stage of agents’ worst fitnesses andpositions to their best values. Two fuzzy logic blockscarry out the adaptation of the ratios of explorationruns and explanation runs using the ratio betweenthe minimum and maximum Popov sums as an inputvariable. A tuning method for T-S PI-FCs dedicatedto a class of nonlinear servo systems with an integralcomponent and is offered, and T-S PI-FCs withreduced process gain sensitivity are tuned. A casestudy and digital simulation results illustrate theoptimal tuning of a T-S PI-FC for the position controlof a laboratory servo system.


Broken Bar Fault Detection based on SetMembership Identification for Three Phase

Induction Motors

Mohammed Obaid Mustafa, George Nikolakopoulos,Thomas Gustafsson

Luleå University of Technology, Luleå, Sweden

Basil M. SaiedUniversity of Mosul, Mosul, Iraq

Keywords: Three Phase Induction Motor, FaultDetection, Set Membership Identification.

Abstract: This article presents a fault detectionscheme for the case of a broken bar occurrence in athree phase induction motor. The proposed schemerelies on Set Membership Identification (SMI) andnovel proposed boundary violation rules for theidentified motor’s parameters. The model of the threephase induction motor is being transformed into anequivalent two phase model, described in the q − dspace, for both the normal and the faulty case. Bythe utilization of the SMI technique, the simplifiedequivalent model of the induction motor is beingidentified during the steady state operation (non–faultcase), while at the same time safety bounds for theidentified variables are being provided, based on ana priori defined corrupting additive noise. On theevent of a fault, specific fault detection conditionsare being proposed that can capture the specifictype of a broken bar fault. The proposed conditionsdepend on: a) abnormal parameter jumps, andb) rapid changes in the volume of the boundinguncertainty, which is being formulated either byellipsoids or orthotopes. Detailed analysis of theproposed approach as also extended simulationresults are being presented that prove the efficiencyof the proposed scheme.


Design and Experimentation of a Neural NetworkController for a Spherical Parallel Robot

Donatello Tina, Luca Carbonari and MassimoCallegari

Polytechnic University of Marche, Ancona, Italy

Keywords: Neural Networks, Robot Control, ParallelKinematics Machines.


Mon

day,

30


Abstract: The paper deals with a neural networkcontrol for the gravity compensation of a parallelkinematics robot. The network has been designedin a simulation environment then it has beenimplemented in robot’s controller by using an FPGAdevice that is part of an embedded system. Afterthe training phase, several experiments have beenperformed on the prototype manipulator and therelated datasets have been collected and elaborated.In the end, a comparative analysis has shownthe improved performance of the neural networkcontroller with respect to the inverse dynamics one,mainly due to the well-known difficulties in derivingexplicit models of friction and play in the joints.


Research on Integrated Fault Diagnosis ofSteam Turbine based on CPN Neural Network

and D-S Evidence Theory

Peng Daogang, Zhang Hao, Huang Hengzi and XiaFei

Shanghai University of Electric Power, Shanghai, China

Keywords: CPN Neural Network, D-S EvidenceTheory, Information Fusion, Fault Diagnosis, Turbine.

Abstract: Counter-propagation Network, combiningthe type of supervised and unsupervised in thelearning process, can overcome the shortcomingsof BP Neural Network fall into local minimum, lowlearning rate and bad convergence performance.D-S evidence theory does not require a prioriinformation and use the interval estimation methodto describe the uncertain information based onreliability function calculating, and it adapted tothe decision level of information fusion. Using thesteam turbine rotor vibration simulation experimentalplatform as the object, the integrated fault diagnosismethod for turbine generator based on CPN neuralnetwork and D-S evidence theory is established inthis paper. This method takes the turbine vibrationcondition parameters as independent samples whichare collected from different sensors. After featureextraction, processing and parameters normalized,these parameters are input independent CPN neuralnetwork to train, which makes the independence ofeach CPN neural network form the fault symptomto failure mode nonlinear mapping, then use theD-S evidence theory method for each CPN neuralnetwork to diagnosis and data fusion, thus realizingthe turbine running state more accurate diagnosis.


On the Multiple-view Triangulation Problem withPerspective and Non-perspective Cameras

A Virtual Reprojection-based Approach

Graziano ChesiUniversity of Hong Kong, Hong Kong, China

Keywords: Vision System, Multiple-view, Perspec-tive Camera, Non-perspective Camera, Triangula-tion.

Abstract: This paper considers the multiple-viewtriangulation problem in a vision system withperspective and non-perspective cameras. Inparticular, cameras that can be modeled through aspherical projection followed by a perspective one,such as perspective cameras and fisheye cameras,are considered. For this problem, an approachbased on reprojecting the available image pointsonto virtual image planes is proposed, which hasthe advantage of transforming the original probleminto a new one for which the existing methods formultiple-view triangulation with perspective camerascan be used. In particular, algebraic and geometricerrors of such methods are now evaluated on thevirtual image planes, and the solution of the newproblem exactly approaches the sought scene pointas image noise and calibration errors tend to zero.The proposed approach is illustrated by severalnumerical investigations with synthetic and real data.


Development of a Piezo-actuated Robot for CellInjection

D. Chakarov, K. Kostadinov, A. Shulev and T.Tiankov

Bulgarian Academy of Sciences, Sofia, Bulgaria

Keywords: Parallel Micromanipulator, Piezo-actuator, Elastic Joint, Preliminary Tension, CellInjection, Simulations, Experimental Investigation.

Abstract: In the presented work model andexperiments of compliant robots with piezo actuatorsare carried out. The robot is designed to performautomatic injection of cells in the range of 10-30μm. A kinematics model of serial-parallel structuresis presented. Pseudo rigid body approach isused, where the elastic joints are modelled asrevolute joints. Models for tension of parallelstructures with elastic joints are developed in orderto eliminate backlashes, to diminish hysteresis,


Monday, 30


and to improve the performance of the piezo-actuators. Two design approaches are proposed.First approach ensures preliminary tensioning byassembly translation along the axes of the drivingjoints. Second approach ensures preliminarytensioning by assembly deflections of the basic serialchain elastic joints. The design of new 3-degrees offreedom (DOF) piezo actuated micro-manipulatorswith serial-parallel structure including elastic jointscapable of performing cell injection is presented.Numerical experiments are done for tensioning ofthe manipulator. An estimation of the manipulatormechanical parameters for different approaches iscarried out. Manipulator simulations with elasticjoints are performed using FEA based function ofa CAD system. The real manipulator prototypeis experimentally investigated using digital imagecorrelation technique.


Compensation of Tool Deflection inRobotic-based Milling

Alexandr Klimchik1,2, Dmitry Bondarenko2,3, AnatolPashkevich1,2, Sebastien Briot2 and Benôit Furet2,4

1 Ecole des Mines de Nantes, Nantes, France2 Institut de Recherches en Communications et

Cybernétique de Nantes, UMR CNRS 6597, Nantes,France

3 Ecole Centrale de Nantes, Nantes, France4 Université de Nantes, Nantes, France

Keywords: Industrial Robot, Milling, ComplianceError Compensation, Dynamic Machining ForceModel, Non-linear Stiffness Model.

Abstract: The paper presents the complianceerrors compensation technique for industrial robots,which are used in milling manufacturing cells. underexternal loading, which is based on the non-linearstiffness model. In contrast to previous works, ittakes into account the interaction between the millingtool and the workpiece that depends on the end-effector position, process parameters and cuttingconditions (spindle rotation, feed rate, geometry ofthe tool, etc.). Within the developed technique, thecompensation errors caused by external loading isbased on the non-linear stiffness model and reducesto a proper adjusting of a target trajectory that ismodified in the off-line mode. The advantages andpractical significance of the proposed technique areillustrated by an example that deals with milling withKuka robot.


Evaluation of a Joint Hysteresis Model in aRobot Actuated by Pneumatic Muscles

Michael Kastner, Hubert GattringerJohannes Kepler University Linz, Linz, Austria

Ronald NadererFerRobotics Compliant Robot Technology GmbH, Linz,

Austria

Keywords: Compliant Robotics, Pneumatic Mus-cles, Hysteresis Model.

Abstract: Passively compliant drives are interestingalternatives to classical stiff actuators in emergingfields like human–robot cooperation, service andrehabilitation robotics. Pneumatic muscles havebeen found to be interesting low–cost actuators forsuch purposes. To fully realize the (desired) highersensitivity and at the same time maintain a goodcontrol quality, detailed models of the robot’s owncomponents are required. For pneumatic muscles,their hysteresis characteristic is a challengingproperty. In this paper we present a hysteresis modelbased on a Prandtl–Ishlinskii operator approachand evaluate the resulting performance when theinverse model is used for compensation in theposition controller. The evaluation is done on a realmulti–axes robot arm.


Semantic Place Recognition based on DeepBelief Networks and Tiny Images

Ahmad Hasasneh1,2, Emmanuelle Frenoux1,2 andPhilippe Tarroux2,3

1 Paris Sud University, Orsay, France2 LIMSI-CNRS, Orsay, France

3 Ecole Normale Supérieure, Paris, France

Keywords: Semantic Place Recognition, RestrictedBoltzmann Machines, Deep Belief Networks, Bag-of-Words, Softmax Regression.

Abstract: This paper presents a novel approach forrobot semantic place recognition (SPR) based onRestricted Boltzmann Machines (RBMs) and a directuse of tiny images. RBMs are able to code imagesas a superposition of a limited number of featurestaken from a larger alphabet. Repeating this processin a deep architecture leads to an efficient sparserepresentation of the initial data in the feature space.A complex problem of classification in the input space


Mon

day,

30


is thus transformed into an easier one in the featurespace. In this article, we show that SPR can thus beachieved using tiny images instead of conventionalBag-of-Words (BoW) methods. After appropriatecoding, a softmax regression in the feature spacesuffices to compute the probability to be in a givenplace according to the input image.


Integral Sliding Mode and Second Order SlidingMode Attitude and Altitude Tracking of a

Quadrotor SystemTheory and Experiment

Mouloud Bouchoucha, Abdessamed Boudane,Kamel Ali and Sofiane Seghour

Ecole Militaire Polytechnique(EMP), Algiers, Algeria

Keywords: dsPIC, Dynamic Modeling, EmbeddedControl System, IMU, Integral Sliding Mode, Quadro-tor, Second Order Sliding Mode, Super- Twisting ,UAV, μC.

Abstract: In this paper Attitude and Altitude trackingcontrol design of the four rotors helicopter will beconsidered. Two robust control algorithms will bedesigned for the case of stabilization and tacking ofattitude and altitude system’s outputs. The attitudecontroller is realized using an inertial measurementunit (IMU) based on MEMS sensors. The altitudecontrol algorithm uses a sonar sensor output. Thecontrol algorithms designed are implemented onan embedded control system based on a dsPICμC. The obtained experimental results demonstratehigh performance of both controllers and robustnessagainst disturbances.


Low-speed Modeling and Simulation ofTorpedo-shaped AUVs

Bjarni Helgason1, Leifur Leifsson1, IndridiRikhardsson1, Helgi Thorgilsson2 and Slawomir

Koziel11 Reykjavik University, Reykjavik, Iceland2 Teledyne Gavia ehf, Kopavogur, Iceland

Keywords: Autonomous Underwater Vehicle,Low-speed Motion, Vehicle Dynamics, Simulation,Experimental Validation.

Abstract: Autonomous underwater vehicles (AUVs)have become important in many marine engineering

applications, such as environmental monitoring,pipeline inspections, or oceanography. For thesetypes of applications, most of the AUVs availablein both academia and industry are shaped like atorpedo and travel at speeds of 3 knots or higher.There is an growing interest in AUVs that are capableof performing tasks at both low-speed as well ashigh speeds. Currently, many torpedo-shaped AUVsare not capable of controlled low-speed motion.This paper presents a simulation model for thelow-speed motion of torpedo-shaped AUVs. Themodel is capable of simulating the surge, sway,heave, and yaw motions. The hydrodynamic forcesacting on the AUV hull are modelled using striptheory, experimental data, and computational fluiddynamics. The simulation model was implementedusing a commercially available software and validatedusing experimental data obtained from the GaviaAUV. The results show that the simulation modelcaptures the AUV motion at low-speed and agreeswell with the experimental data.


Cooperative Autonomous Driving for VehicularNetworks

Lamia Iftekhar and Reza Olfati-SaberDartmouth College, Hanover, U.S.A.

Keywords: Cyber-physical Systems, AutonomousDriving, Flocking Algorithms, Intelligent Transporta-tion Systems.

Abstract: In this paper, we introduce cooperativeautonomous driving algorithms for vehicular networksin urban environments that take human safety intoaccount and are capable of performing vehicle-to-vehicle (V2V) and vehicle-to-pedestrian (V2P)collision avoidance. We argue that “flocks" aremulti-agent models of vehicular traffic on roads andpropose novel autonomous driving architecturesfor cyber-physical vehicles capable of performingautonomous driving tasks such as lane-driving,lane-changing, braking, passing, and making turns.These autonomous driving algorithms are inspiredby the flocking theory of Olfati-Saber (Olfati-Saber,2006), though, there are notable differences betweenautonomous driving on urban roads and flockingbehavior—flocks have a single desired destinationwhereas most drivers on road do not share the samedestination. We demonstrate that lane-driving fora vehicular network with n > 3 vehicles cannotnecessarily be performed using pairwise vehicularinteractions and might require triangular interactionsamong triplets of vehicles. The self-driving vehiclesin our framework turn out to be nonlinear switching


Monday, 30


systems with discrete states that are related to thedriving modes of the vehicles. Complex drivingmaneuvers can be performed using a sequence ofmode switchings. We present several examples ofdriving tasks that can be effectively performed usingour proposed driving algorithms.


Flocking for Networks of Mobile Robots withNonlinear Dynamics

Reza Olfati-Saber and Lamia IftekharDartmouth College, Hanover, U.S.A.

Keywords: Flocking, Nonholonomic Robots, Multi-robot Networks, Near-Identity Transformation,Nonlinear Systems.

Abstract: In this paper, we address the problemof flocking for networks of nonholonomic mobilerobots with nonlinear dynamics given that a flockingalgorithm for particles is known. Our approach relieson the use of near-identity change of coordinatesthat transform the nonlinear dynamics of the robot toa partially-linear normal form with a double-integratorlinear subsystem. The flocking algorithm is thenapplied to the linear part. The inverse of the near-identity transformation provides the flocking algorithmfor the networked nonholonomic robots. We provethe emergence of flocking behavior for roboticnetworks with nonlinear dynamics according to theformal definition of flocking in Olfati-Saber’s flockingpaper (TAC ’06). Simulation results are providedfor large-scale networks of two-wheeled robots withnonlinear dynamics as models of Khepera-III robotsthat demonstrate the effectiveness of our proposedtransformation and algorithm.


Load-following Control of APR+ NuclearReactors

Jae Hwan Kim, Man Gyun NaChosun University, Gwangju, Korea, Republic of

Keuk Jong Yu, Han Gon KimKHNP-Central Research Institute, Yuseong-gu, Daejeon,

Korea, Republic of

Keywords: Load-following Operation, Model Predic-tive Control, APR+ Reactor, Thermal Power Level,Axial Shape Index (ASI), KISPAC-1D Code.

Abstract: The load-following operation of APR+reactor is needed to control the power effectivelyusing the control rods and to restrain the reactivitycontrol from using the boric acid for flexibility of plantoperation. The xenon has a very high absorptioncross-section and makes the impact on the reactordelayed by the iodine precursor. The powermaneuvering using automatically load-followingoperation has advantage in terms of safety andeconomic operation of the reactor. Therefore, anadvanced control method that meets the conditionssuch as automatic control, flexibility, safety, andconvenience is necessary to load-following operationof APR+ reactor. In this paper, the MPC method isapplied to design APR+ reactor’s automatic load-following controller for the integrated average coolanttemperature and ASI control. The KISPAC-1Dcode, which models the APR+ nuclear power plants,is interfaced to the proposed controller to verifythe tracking performance of the average coolanttemperature and ASI. It is known that the proposedcontroller exhibits very fast tracking responses.


A Fault-Tolerant Controller for an SP-100 SpaceNuclear Reactor

Ju Hyun Kim, Dae Seup Kim and Man Gyun NaChosun University, Gwangju, Korea, Republic of

Keywords: Fault Detection and Diagnostics, Fault-Tolerant Control, Fuzzy Model, Model PredictiveControl, Space Reactor Power Control, SequentialProbability Ratio Test.

Abstract: The control system is a key element ofspace reactor design to meet the space missionrequirements of safety, reliability, survivability,economics, and autonomous action. The objectivesof the proposed model predictive control are tominimize both the difference between the predictedTE power and the desired power, and the variation ofcontrol drum angle that adjusts the control reactivity.A genetic algorithm is used to optimize the modelpredictive controller. The model predictive controlleris integrated with a fault detection and diagnosticsalgorithm so that the controller can work properlyeven under input and output measurement faults.Simulation results of the proposed controller showthat the TE generator power level controlled by theproposed controller could track the target power leveleffectively even under measurement faults, satisfyingall control constraints.


Mon

day,

30



Smart Walker Control through the Inference ofthe User’s Command Intentions

M. Martins1, A. Frizera2, C. Santos1 and R. Ceres31 Universidade do Minho, Gualtar, Braga, Portugal

2 Universidade Federal do Espírito Santo, Vitória, Brazil3 Consejo Superior de Investigaciones Científicas,

Arganda del Rey, Madrid, Spain

Keywords: Smart Walker, Assistive Mobility, FuzzyControl.

Abstract: In this work is presented theNeoASAS walker including its conceptual design,implementation and validation with a new interfaceapproach integrated. This interface is basedon a joystick and it is intended to extract theuser’s movement intentions. Eleven healthy usersperformed preliminary sets of experiments with thewalker, which showed the sensibility of the joystickto extract command intentions from the user. Thesesignals presented a higher frequency componentthat was attenuated by a Benedict-Bordner filter.Then, an approach to the control architecture wasdeveloped, in order to obtain stable and safe userassistance. This control architecture is basedon a fuzzy logic control that allows the control ofthe walkers’ motors. Thus, an assistive device toprovide safety and natural manoeuvrability wasconceived and offers a certain degree of intelligencein assistance and decision-making. The motivationis that this will contribute to improve rehabilitationpurposes by promoting ambulatory daily exercisesand thus extend users’ independent living.


Nonlinear Deterministic Methods for ComputerAided Diagnosis in Case of Kidney Diseases

Andreea Udrea, Mihai Tanase and Dumitru PopescuUniversity Politehnica of Bucharest, Bucharest, Romania

Keywords: Computer Aided Diagnosis, NonlinearDeterministic Methods, CT Images.

Abstract: This paper proposes a set of nonlineardeterministic methods derived from chaos theorythat can serve as computed aided diagnosis toolsfor kidney diseases based on computer topographies(CT). These procedures target the classification of

the analyzed tissue samples in normal, malign andbenign affected and also enhanced visualization ofthe CT images. The classification methods consistin estimating the fractal dimension of the kidneytissue and, respectively, the correlation dimensionof the attractor obtained from the spatial seriesassociated to the kidney image. The enhancedvisualization method associates a fractal map to theanalysed image. The methods are tested on 120 CTspresenting normal and modified tissue. The degreeof trustworthiness of the methods while dealingwith classifications is discussed based on statisticalresults and samples of fractal maps associated tothe images are also presented.


Block Triangular Decoupling of General NeutralMulti Delay Systems

Fotis N. Koumboulis and Nikolaos D. KouvakasHalkis Institute of Technology, Psahna Evoias, Greece

Keywords: General Neutral Multi Delay Systems,Dynamic Controllers, Realizable Controllers.

Abstract: The problem of block triangular decouplingis studied for the case of general neutral multi delaysystems. The system is not restricted to be squareand invertible. The controller is of the generalneutral dynamic type involving a dynamic feedbackand dynamic precompensator. Two different casesof feedback are studied. The first is the case ofmeasurable output feedback and second is the caseof performance output feedback. The controlleris restricted to be realizable. The necessary andsufficient conditions for the problem to be solvableare established. The general class of the realizablecontrollers solving the problem is derived. Theclosed loop transfer function is proven to havearbitrary characteristic polynomial thus facilitatingcommand tracking and stability.


Monday, 30


Paper 415:15 - 16:30 Room BarcelonaSpecial Session on Intelligent Vehicle Controls &Intelligent Transportation Systems - IVC&ITS

A Warping Window Approach to Real-timeVision-based Pedestrian Detection in a Truck’s

Blind Spot Zone

Kristof Van Beeck1, Toon Goedemé1,2 and TinneTuytelaars2

1 Lessius Mechelen, Campus De Nayer,Sint-Katelijne-Waver, Belgium

2 KU Leuven, Heverlee, Belgium

Keywords: Computer Vision, Pedestrian Tracking,Real-time, Active Safety Systems.

Abstract: We present a vision-based pedestriantracking system targeting a very specific application:avoiding accidents in the blind spot zone of trucks.Existing blind spot safety systems do not offer acomplete solution to this problem. Therefore wepropose an active alarm system, which warns thetruck driver if vulnerable road users occur in the blindspot zone. Our system is based solely on a visionsensor, and automatically detects vulnerable roadusers in the blind spot camera images. Due to thenature of this specific problem, this is a challengingtask. Besides the demanding time constraint thereis a need for a high accuracy, and we have tocope with the large distortion that a blind spotcamera introduces. To achieve this we propose awarping window multi-pedestrian tracking algorithm.Our algorithm achieves real-time performancewhile maintaining high accuracy. To evaluate ouralgorithms we recorded several datasets with a realblind spot camera mounted on a real truck, consistingof realistic simulated dangerous blind spot situations.

Paper 715:15 - 16:30 Room BarcelonaSpecial Session on Intelligent Vehicle Controls &Intelligent Transportation Systems - IVC&ITS

Road Safety at Intersections Controlled byTraffic Lights

IVC and Risk Indexes

Bruno Dalla Chiara, Francesco Paolo Deflorio andSerena Cuzzola

Politecnico di Torino, Torino, Italy

Keywords: ADAS, Safety Index, Intersection Safety,Driver Reaction Time.

Abstract: The paper reports the results of safetyanalyses conceived to assess the effects and benefitswhich might be generated by the forthcoming use

of the infrastructure-to-vehicle (I2V) or vehicle-to-infrastructure (V2I) communication systems atroad intersections regulated by traffic lights. Roadcrossings are often considered as critical areas forthe occurrence of accidents, because they increasethe likelihood of the event given the confluenceof traffic streams from and to different directions.The analyses are aimed at calculating a real-timeestimate of some risk indexes of accident, whichmight be provided on-board when approachingroad intersection regulated by traffic lights. Thisinformation can then be used by an ADAS for trafficsignal approaching. Two typologies of use of theinformation on the risk indexes can be identified: ifdata can be detected in real time, the driver couldbe informed on-board of a potentially hazardoussituation using algorithms to predict the trend of thevehicle on the basis of the data detected from themonitoring; another use would be detecting – in casethe vehicle were already within the dilemma zone –the lowest risk manoeuvre and sending a messageon board to inform the driver.


Nonlinear Analysis of Costas Loop Circuit

N. V. Kuznetsov1,2, G. A. Leonov1, M. V.Yuldashev1,2 and R. V. Yuldashev1,2

1 University of Jyväskylä, Jyväskylä, Finland2 Saint-Petersburg State University, Saint-Petersburg,

Russian Federation

Keywords: Costas Loop, Phase-locked Loop, PhaseDetector Characteristic, Nonlinear Analysis.

Abstract: Problems of rigorous mathematicalanalysis of Costas Loop are considered. Theanalytical method for phase detector characteristicscomputation is proposed and new classes of phasedetector characteristics are computed for the firsttime. Effective methods for nonlinear analysis ofCostas Loop are discussed.


Positive Realization of Continuous LinearSystems with Order Bound

Kyungsup Kim and Jaecheol RyouChungnam National University, Daejeon, Korea, Republic

of

Keywords: Positive Realization, Positive LinearSystem, Metzler Matrix, Polyhedra Cone.


Mon

day,

30


Abstract: This paper discusses the realizationproblem of a class of linear-invariant system, inwhich state variables, input and output are restrictedto be nonnegative to reflect physical constraints. Thispaper presents an efficient and general algorithmof positive realization for positive continuous-timelinear systems in the case of transfer function with(multiple) real or complex poles. The solution ofthe corresponding problem for continuous-timepositive is deduced from the discrete-time case by atransformation. We deal with the positive realizationproblem through convex cone analysis. We providea simple general and unified construction methodfor the positive realization of the transfer function,which has multiple poles, upper-bound and a sparserealization matrix. We consider a sufficient conditionof positive realization.


Fluidity Measuring Device for the Concrete usingLaser Diode Controller via WSN

Bo Hee LeeSemyung University, Jecheon, Korea, Republic of

Keywords: Measurement of Concrete Fluidity,Wireless Sensor Network (WSN), Laser Sensors,Driving Mechanism.

Abstract: Presented is a high performance devicefor the measurement of concrete fluidity usingWireless Sensor Network (WSN). This deviceis an improvement over the existing method ofmanual measurement which is subject to significanthuman-induced error. Using this device we canmake measurements automatically and analyze theinformation simultaneously for the concrete fluidity.In this paper we present a novel device utilizinglaser sensors and wireless data acquisition includingdriving mechanism. The effectiveness of the deviceis verified through experiment.


Modeling and Simulation of Humanoid RobotSpine Vertebra

M. Souissi, V. Hugel and P. BlazevicUniversité de Versailles Saint Quentin en Yvelines,

Versailles, France

Keywords: ROMEO Robot, Simulation, VertebralColumn, Mechanical Structure, Kinematics.

Abstract: In this paper, a parallel mechanism isproposed for the design of humanoid vertebra.This mechanism is inspired by a flight simulatorsystem, and has been adapted and optimized toenable pitch and roll motion of a humanoid trunkat reduced energy cost. The system consists of abottom platform and a top platform connected bytwo articulated arms and a vertical central rod. A3D model of the system has been elaborated forsimulation and design.


On the Problem of Task Planning in Multi-robotSystems

Zhi Yan, Nicolas Jouandeau and Arab Ali CherifParis 8 University, Saint-Denis, France

Keywords: Task planning, Multi-robot Systems,Coordination.

Abstract: Multi-robot task planning (MRTP) is oneof the fundamental problems for multi-robot systems.An important question facing this research topicis, which robot should execute which task so asthe expected overall system performance can bemaximized? Many approaches have been proposedfor such a purpose. This paper investigates theexisting works in the field. The approaches havebeen surveyed and some representatives arecompared with detailed results. A brief discussionand further research perspectives are also given atthe end of the paper.


Disassembly Planning using Visual Servoing

S. T. Puente, J. Pomares and F. TorresUniversity of Alicante, Alicante, Spain

Keywords: Disassembly, Visual-servoing, Recy-cling, Robotics.

Abstract: The paper presents a disassemblyalgorithm which computes the cost required for theautomatic disassembly of a component of a product.The product is modelled using a hierarchical modelthat represents the relations among components.The characteristics of the product are expressedusing matrices, which represents the hierarchicalmodel and the cost of removal the unions amongcomponents. The disassembly cost of a componentis computed taking the trajectory required to perform


Monday, 30


the separation of a component into consideration.This trajectory is performed using an image-basedvisual-servoing system. To determine the cost ofthese trajectories the concepts of manipulability andperceptibility are employed.


Observer-based Adaptive Sliding Mode Controlfor a Pneumatic Servo System

Kuo-Ming Chang and Tien-Tsai KungNational Kaohsiung University of Applied Sciences,

Kaohsiung, Taiwan

Keywords: Pneumatic Servo System, AdaptiveSliding Mode Control, Extended State Observer,Dead-Zone.

Abstract: This paper combines an extended stateobserver (ESO) and adaptive sliding mode controlto deal with the problem of nonlinear input dead-zone, unkonw system function, external distrubanceand system states unmeasured in a pneumaticservo system. Firstly, an extended state observeris applied to estimate system state variables anduknown system nonlinear function contained theexternal disturbance, and an adaptive law is usedto estimate unknown dead-zone parameters. Then,an observer-based adaptive sliding mode controlcan be derived for such a uncertain pneumatic servosystem. Furturemore, the proposed control schemein this paper is applied to a pneumatic positioningcontrol experimental equipment and it is shown thatthe positioning accuracy with less than 0.05 μm canbe obtained.


Design of a Rectangular-type FingerRehabilitation Robot

Gab-Soon Kim, Hyeon-Min Kim, Young-Guk Kim,Hee-Suk Shin and Jungwon Yoon

Gyeongsang National University, Jinju, Korea, Republic of

Keywords: Finger Rehabilitation, RehabilitationRobot, Force Sensor, Intelligent Robot.

Abstract: This paper describes the developmentof a finger-rehabilitation robot for rehabilitatingstroke patients’ fingers and other patient’s paralyzedfingers. The developed finger rehabilitation robot iscomposed of a thumb-rehabilitation robot instrumentand four finger rehabilitation robot instruments. Thefinger-rehabilitation robot could exercise fingers of

patient for their rehabilitation. A control characteristictest of the developed rectangular-type finger-rehabilitation robot was carried out, and the resultsconfirmed that the robot could be used for theflexibility rehabilitation exercise for the fingers ofnormal person and patients.


Facial Expression Recognition based on FacialFeature and Multi Library Wavelet Neural

Network

Nawel Oussaifi, Wajdi Bellil and Chokri Ben AmarSfax University, Sfax, Tunisia

Keywords: Facial Expressions Classification,Wavelet Network, Facial Landmarking.

Abstract: In this paper, we propose a waveletneural network-based system for automaticallyclassifying facial expressions. This system isbased on Multi Library Wavelet Neural Network(MLWNN) for emotions classification. Like othermethods, our approach relies on facial deformationfeatures. Eyes, mouth and eyebrows are identifiedas the critical features and their feature points areextracted to recognize the emotion. After featureextraction is performed a Multi Library Wavelet NeuralNetwork approach is used to recognize the emotionscontained within the face. This approach differsfrom existing work in that we define two classesof expressions: active emotions (smile, surpriseand fear) and passive emotions (anger, disgust andsadness). In order to demonstrate the efficiencyof the proposed system for the facial expressionrecognition, its performances are compared withother systems.


Simultaneous Control of Translational andRotational Motion for Autonomous

Omnidirectional Mobile Robot2nd Report: Robot Model Considering Moving

Parts and Evaluation of Movable Area by Heights

Ayanori Yorozu, Takafumi Suzuki, MatsumuraTetsuya and Masaki TakahashiKeio University, Yokohama, Japan

Keywords: Service Robot, Obstacle Avoidance,Omnidirectional Platform, Fuzzy Potential Method.

Abstract: This paper presents a real time


Mon

day,

30


collision avoidance method for an autonomousomnidirectional mobile robot considering shape ofthe robot and movable area by heights based onsimultaneous control of translational and rotationalmotion. Service robots which have been developedin recent years have arms to work and executetasks. In these robots, the size of width is sometimesnot equal to that of length by heights. In order toavoid obstacles considering safety and mobility forthe robots, it is necessary to evaluate shape of therobot and movable area by heights. To evaluatethem, the robot model is defined in heights of eachmoving part. Evaluating of the robot model and themovable area for each height, if the robot is unableto move keeping a safe distance from the obstacles,the robot determines the suitable orientation angleconsidering the minimum length from the center ofthe robot model to that outer shape. In this paper, thenovel control method based on the Fuzzy PotentialMethod is presented. To verify the effectiveness ofthe proposed method, several numerical simulationsare carried out.


EKF based Data Fusion using Interval Analysisvia Covariance Intersection, ML and a Class of

OGK Covariance Estimators

Samuel B. Lazarus1, Antonios Tsourdos1, JoãoSequeira2 and Al Savvaris1

1 Cranfield University, Cranfield, U.K.2 Instituto Superior Técnico, Lisboa, Portugal

Keywords: Multiple Sensor Fusion, Data Fusion,EKF based Navigation, Interval Analysis (IA), RobustNavigation, Covariance Intersection (CI), Maxi-mum Likelihood (ML), Orthogonal Gnanadesikan-Kettenring (OGK).

Abstract: This paper addresses the comparison ofrobust estimation of a covariance matrix in vehiclenavigation task to express the uncertainty whenfusing information from multiple sensors. TheEKF estimates are fused with the Interval Analysisestimates and further the results are fused using theCovariance Intersection (CI), Maximum Likelihood(ML) and a class of Orthogonal Gnanadesikan-Kettenring (OGK) estimators. The simulation resultspresented show that the variation between CI andOGK and the correlation between sensors aresignificant in the presence of outliers.


Improved Leader Follower Formation Control ofAutonomous Underwater Vehicles using State

Estimation

Umesh Neettiyath and Asokan ThondiyathIndian Institute of Technology Madras, Chennai, India

Keywords: Underwater Robots, Leader FollowerControl, Formation Control, AUV, Multirobot Sys-tems.

Abstract: Multi robot coordination and controlfor underwater robots is an area of significantimportance in many underwater missions. A newapproach for leader follower formation control ofmulti AUV systems is explored in this paper. Thecontroller estimates the next desired position of thefollower robot from the past and current positions ofthe leader and follower robots. The control signalsare then issued to the follower robot to align it to theestimated trajectory. This control scheme has thecapability to compensate for initial errors and followthe leader under various operational scenarios.The development of the controller and simulationresults for selected scenarios are presented. Theresults show that the proposed method is simple andcomputationally efficient.


Optimized Chipping Processes with a NewMechatronic Tool System

Application of Strain Gauge Sensors andPiezoelectric Actuators

Franz Haas1, Elke Nuspl2 and Manfred Pauritsch11 FH CAMPUS 02, Graz, Austria

2 TCM International Tool Consulting & ManagementGmbH, Stainz, Austria

Keywords: Chipping Technology, Mechatronic ToolSystem, Condition Monitoring, Piezoelectric Actua-tors.

Abstract: The paper introduces a new measurementdevice which allows collecting the chipping forcein three directions and additionally the processtemperature. The apparatus consists of a speciallydesigned tool holder with integrated strain gauges,the electronic measuring equipment and theevaluation software. During the first period anew device for research and development purposeshas been designed, experiments on innovativematerials are presented including the comparison


Monday, 30


with simulation results. The next step of the projectincorporates the product development based onthe previous experience with targets accuracy,cost effectiveness, reliability and the possibility ofactive damping of tool vibrations. To achieve thelast mentioned demand a piezoelectric actuator isintegrated into the tool shank and works against thevibration forces.


Modified Evolutionary Strategies Algorithm inLinear Dynamic System Identification

Ivan Ryzhikov and Eugene SemenkinSiberian State Aerospace University, Krasnoyarsk,

Russian Federation

Keywords: Linear Dynamic System, Linear Differ-ential Equation, Evolutionary Strategies, ParametersIdentification Problem, Structure Identification.

Abstract: The approach to dynamic systemsmodelling in the form of the linear differentialequation that uses only the system output and thecontrol sample is presented. To develop a lineardynamic model as an ordinary differential equationwe need to know the structure of differential equationand its order, so then it would be possible to identifyparameters. It is common that measurements ofthe system output are distorted with a noise. Incase of the non-uniform sample we would needa special output function approximation approachso the unit step function can be estimated. Thedynamic system identification with an ordinary lineardifferential equation allows solving different controltasks, determining the system state with anothercontrol function.


Autonomous Navigation of an Outdoor MobileRobot in a Cluttered Environment

Amir Monjazeb, Jurek Z. SasiadekCarleton University, Ottawa, Canada

Dan NecsulescuOttawa University, Ottawa, Canada

Keywords: Simulataneous Localization and Map-ping (SLAM) Problem, EKF, Unscented KalmanFilter, Fastslam, Hybridslam, Unscented Hybridslam,Cluttered Environment, Loop Closing, Long Trajec-tory.

Abstract: This paper proposes a modification ofHybridSLAM strategy that is used to navigate anoutdoor autonomous mobile robot in a clutteredenvironment. HybridSLAM is combining extendedKalman filter SLAM EKF-SLAM and FastSLAMto take advantage of strengths and to covershortcomings of both filters. By the use of anunscented version of Kalman filter instead ofEKF-SLAM, the formulation of the HybridSLAMis revised. Same as HybridSLAM, the new revisedalgorithm uses the state distribution capabilities of theunscented Kalman filter to keep the uncertainty of thesystem to be remembered for a long trajectory, andat each time step, FastSLAM is used to produce localmaps. Presented simulations and results evaluatethe performance of the proposed approach usingUnscented Kalman filter in a cluttered environment.


Development of a Wheelchair with a LiftingFunction

Yoshikazu MoriIbaraki University, Hitachi, Japan

Norikatsu SakaiSMC Corp., Chiyoda-ku, Japan

Kaoru KatsumuraTamron Co., Ltd., Saitama, Japan

Keywords: Medical and Welfare Assistance,Mechatronics, Design, Disabled Person, Care Lift,Transferring.

Abstract: A wheelchair with a lifting function isdesigned to assist a caregiver when transferring awheelchair user not only indoors but also outdoors.The target user is typically a severely disabled personwith disabled upper and lower limbs, and thereforeneeds the physical support when using a toilet ortransferring from a bed to a wheelchair and so forth.Both the wheelchair and the lift are driven by theirrespective motors. The user can approach above thetoilet stool or the bed from the rear because the largedriving wheels are located in front of the body andthe seat can be folded. This wheelchair is allowedto travel on public roads because of the mechanismof folding the frame for lifting. This paper presentsthe concept design and the experimental resultsof a full-sized prototype wheelchair with the liftingfunction, which confirms the design effectiveness.


Mon

day,

30


Paper 616:30 - 17:30 Foyer - IVC&ITSPoster Session 2

Stereo Vision based On-road Vehicle Detectionunder Illumination Changing Conditions using

Self Quotient Image

Jonghwan Kim, Chung-Hee Lee and Young-ChulLim

Daegu Gyeongbuk Institute of Science & Technology,Dalseong-Gun, Daegu, Korea, Republic of

Keywords: Stereo Vision, Vehicle Detection, Vehi-cle Recognition, Self Quotient Image, IlluminationEqualization.

Abstract: Today the many of automotive researchgroups study how to reduce vehicle accidents. Forthis reason, they have been developing the advanceddriver assistance system (ADAS). In ADAS, thevarious sensors are used for recognizing the drivingsituations. For example, there are supersonic wavesensors and radar sensors and so on. In particular, incomputer vision research groups, the vision sensors(ex. CCD, IR) are used for this. But it has somedifficult problems because the vehicles are mainlydriven in outdoors. The images captured by outdoorshave various illumination conditions due to weather.It makes difficulty to detecting vehicles in images.In this paper, we introduce the vehicle detectionmethod when the input images of system haveillumination changes. We use the self quotient image(SQI) algorithm for illumination equalization. But SQIalgorithm produces many false positive results. Sowe eliminate the false-positive results using stereovision technique. In main section, we explain thismethod in detail. And we prove the proposed methodhas superior performance than existing systemsusing experiments.


Monday, 30

Tuesday Sessions

FINAL PROGRAM AND BOOK OF ABSTRACTS: TUESDAY SESSIONS ICINCO 2012

Paper 410:00 - 10:45 Room BarcelonaSpecial Session on Operations Management andDecision Making in Today’s CompetitiveEnvironment - OMDM

An Evaluation Model for Green and Low-CarbonSuppliers

Amy H. I. Lee1,1,1, He-Yau Kang2, Chun Yu Lin1

and Hsin Wei Wu11 Chung Hua University, Hsinchu, Taiwan

2 National Chin-Yi University of Technology, Taichung,Taiwan

Keywords: Suppliers, Low-Carbon, Fuzzy AnalyticNetwork Process, Fuzzy Set.

Abstract: Under a conventional supplier evaluation,cost, on-time delivery and quality are treated asthe most important factors. However, in today’sincreasingly environmental conscious market withgrowing demands of green products, more and morefirms are aiming to manufacture green productsto reduce the damage to the environment and tolimit the use of energy and other resources at anystage of its life, including raw materials, manufacture,use, and disposal. Thus, a firm needs to select theright suppliers that not only can satisfy the basicrequirements, such as cost and quality, but alsocan provide green and low-carbon materials. Thegoal of this research is to construct a green andlow-carbon supplier evaluation model. The criteriato evaluate green and low-carbon suppliers areanalyzed first, and the most important ones areselected. Fuzzy analytic network process (FANP)model is constructed to evaluate various aspects ofsuppliers. By applying the model, the manufacturercan find the most suitable suppliers for cooperation.Goal programming (GP) is applied next to allocatethe most appropriate amount of orders to each of theselected suppliers.


Developing A New Variables Sampling Schemefor Product Acceptance Determination

Chien-Wei WuNational Taiwan University of Science and Technology,

Taipei, Taiwan

James C. ChenNational Tsing Hua University, Hsinchu, Taiwan

Keywords: Decision Making, Acceptance Sampling,Fraction of Defectives, Quality Assurance.

Abstract: Acceptance sampling is a useful toolfor determining whether submitted lots should beaccepted or rejected. With the current increasein outsourcing production processes and the highquality levels required, it is very desirable to havean efficient and economic sampling scheme. Thispaper develops a variables repetitive group sampling(RGS) plan based on the third generation of processcapability index. The plan parameters are determinedby minimizing the average sample number (ASN)for inspection and fulfilling the classical two-point-condition on the operating characteristic (OC)curve. Besides, the efficiency of the proposed plan isinvestigated and compared with the existing variablessingle sampling plan. Tables of the plan parametersare also provided.


Modeling and Simulation of a TemperatureRobust Control in Grain Drying Systems for

Thermal Damage Reduction

Josenalde B. de OliveiraFederal University of Rio Grande do Norte, Macaiba,

Brazil

Marcus V. A. FernandesFederal Institute of Rio Grande do Norte, Natal, Brazil

Leonardo R. L. TeixeiraFederal Institute of Rio Grande do Norte, Currais Novos,

Brazil

Keywords: Grains Drying, Temperature Control,Industrial Controller, PID, Variable Structure Control,Adaptive Control.

Abstract: Informatics plays an imperative role inthe designing and tuning of new control systems


Tues

day,

31

ICINCO 2012 FINAL PROGRAM AND BOOK OF ABSTRACTS: TUESDAY SESSIONS

strategies, since the computational simulation of suchsystems is part of the entire process of applying analgorithm on a real environment. This paper presentsan alternative to the Proportional-Integrative-Derivative (PID) controller for temperature control ingrain drying systems. The PID controller may presentundesirable oscillations in the presence of externaldisturbances associated with agroindustrial facilities,thus demanding a precise and automatic tuningduring the entire process. Robust controllers aresuitable and recommended for the drying final quality,since the grains are offered a thermal damagereduction when submitted to abrupt temperaturevariations, as fragility and even crack duringprocessing. Simulation results on an experimentalmodel of a nonlinear robust controller, named ShuntIndirect Variable Structure Model Reference AdaptiveController (SIVS-MRAC) are shown. Performanceresults before disturbances and parametric variationsare compared with the PID behaviour.


Control and Model Parameters Identification ofInertia Wheel Pendulum

Paweł Drapikowski, Jarosław Goslinski and AdamOwczarkowski

Poznan University of Technology, Poznan, Poland

Keywords: Inertia Wheel Pendulum, Kalman Filter,Extended Kalman Filter, Linear-Quadratic Regulator.

Abstract: This paper presents control method ofan inverted pendulum with an inertial drive (IWPâC“ Inertia Wheel Pendulum). This is a non-linear,underactuated mechanical system and therefore ithas more degrees of freedom than control variables.In application, electric motor has been placed onthe pendulum top and it is a source of torque, whichaccelerates the flywheel. Position of the pendulumdepends on the acceleration of the flywheel. Thispaper shows algorithm for the real object to keepthe inertia wheel pendulum in the vertical position atequilibrium point. In order to achive this aim, authorsdecided to concentrate on modern and advancedtechniques of control and estimation such as LQRregulator, sensor fusion, extended Kalman Filter andmodel parameters identification.


Solving the Indoor SLAM Problem for aLow-Cost Robot using Sensor Data Fusion and

Autonomous Feature-based Exploration

Luciano Buonocore1, Cairo Lúcio NascimentoJúnior2 and Areolino de Almeida Neto1

1 Universidade Federal do Maranhão, São Luis, Brazil2 Instituto Tecnológico de Aeronáutica, São José dos

Campos, Brazil

Keywords: Sensor Data Fusion, Feature-basedAutonomous Exploration, Visual Sensor, Low-CostRobot, SLAM.

Abstract: This article is concerned with the solutionof the SLAM (Simultaneous Localization AndMapping) problem in an indoor environment usinga low-cost mobile robot that autonomously exploresthe environment. The robot was constructed witha distance measurement subsystem composed ofthree types of sensors: a wireless webcam with alaser pointer (a visual sensor), two infrared sensorsand an ultrasonic TOF (time-of-flight) sensor. Firstly,an algorithm that requires a small computational loadis used to fuse the noisy raw data acquired by thesesensors and generate the environment features.These features are then used by a particle filter tosolve the SLAM problem. An autonomous feature-based exploration algorithm was implementedand is also presented. The results obtained inthe experiments carried out in two small indoorenvironments show that the estimated environmentmap generated when the robot uses the autonomousexploration algorithm is very similar to the onegenerated when the robot poses were manuallychosen.


A Study of the Impact of Computational Delaysin Missile Interception Systems

Ye Xu, Israel Koren and C. M. KrishnaUniversity of Massachusetts at Amherst, Amherst, U.S.A.

Keywords: Computer Response Time, Computa-tional Delay, Single-missile Single-target System,Multiple-missile Multiple-target System.

Abstract: Most publications discussing missileinterception systems assume a zero computerresponse time. This paper studies the impact ofcomputer response time on single-missile single-target and multiple- missile multiple-target systems.


Tuesday, 31


Simulation results for the final miss distance as thecomputer response time increases are presented.A simple online cooperative adjustment model formultiple-missile multiple-target system is presentedfor the purpose of studying the computer delay effect.


A Rescheduling Framework for AirlineScheduling

Birthe Gebhardt and Jürgen SauerCarl von Ossietzky Universität Oldenburg, Oldenburg,

Germany

Keywords: Reschedule, Framework, Airline.

Abstract: Airline scheduling is a sophisticatedarea. If a disruption occurs many tasks have tobe taken into account. In order to structure therescheduling process a framework is useful. Theexisting framework for rescheduling will extendwith further information tasks to use various repairmethods. The framework is cuts into two parts. Thefirst part includes the precondition of the domain. Thesecond parts described the tasks of the reschedulingprocess. The use of this framework allows toimplement more than one rescheduling method.


Operating System Core as Template forEmbedded System Software Development

Leonardo Jelenkovic and Domagoj JakobovicUniversity of Zagreb, Zagreb, Croatia

Keywords: Operating System Core, EmbeddedSystem, Software Development.

Abstract: Software development for embeddedsystems is a fast growing industry. Development fora mid-range complexity embedded system is usuallybased on custom built templates and tools, or oncommercially developed solutions with an operatingsystem as a base. This paper presents possibilitiesfor building customized templates that are operatingsystem primitives. Since many embedded systemsrequire only a few subsystems, and only basicoperations from them, such subsystems couldbe built fast and then used as a base for newsystems. Based on our experience while creatingan educational operating system for embeddedsystems, we propose operating system primitivesthat can be created and then used as templates for

creating new embedded systems.


Improvement of Extraction Method for Inter-turnFault Detection in IPMSM under Transient

Conditions

Gyu Tae Choi1, Je Won Lee1, Minho Choi1 andSang Woo Kim1,1

1 POSTECH, Pohang, Korea, Republic of

Keywords: Extracting Non-stationary Sinusoids,Inter-turn Faults, Interior Permanent Magnet Syn-chronous Motor (IPMSM), Motor Current SignatureAnalysis (MCSA).

Abstract: Most fault detection techniques arefocused on induction motors and are based onsteady-state conditions. In this paper, an extractionmethod for inter-turn fault detection in interiorpermanent magnet synchronous motors (IPMSM)is proposed. The study is focused on an IPMSMunder non-stationary conditions. The technique isformulated by modifying existing fixed frequencysinusoid tracking algorithms, which is based onan adaptive algorithm for extracting non-stationarysinusoids. The faults are determined using the motorcurrent signature analysis technique. Simulationsperformed in this study validate that the proposedalgorithm improves the extraction performance.


Mechanical Modeling of an Actuated Platform forPrecision Pointing Control

Via Finite-element Analysis and Normal ModeAnalysis

Eric U. Diaz, Gerardo Zarate, Helen Boussalis andKhosrow Rad

California State University, Los Angeles, U.S.A.

Keywords: Finite-element Analysis, Normal ModeAnalysis, Space Telescope.

Abstract: A large, segmented space telescoperequires high precision and accuracy in its mirrorshape to obtain clear images. The Structures,Propulsion, and Control Engineering (SPACE)telescope testbed at the NASA sponsored UniversityResearch Center of excellence must maintain apointing control accuracy of 2 arc seconds. APeripheral Pointing Architecture (PPA) has beendesigned to demonstrate the Testbed’s pointing


Tues

day,

31


accuracy. A Finite Element Analysis (FEA) modelof the PPA is developed. Normal mode analysis isperformed to establish the PPA’s natural frequencies,mode shapes, and the mass and stiffness matrices.Utilizing the H-infinity controllers developed toachieve figure maintenance, the pointing control ofthe Testbed structure is achieved.


Timed Trajectory Generation for a Vision-basedAutonomous Mobile Robot in Cluttered

Environments

Jorge B. Silva, Cristina P. SantosUniversity of Minho, Guimarães, Portugal

João SequeiraInstituto Superior Técnico, Lisbon, Portugal

Keywords: Timed Trajectories, Nonlinear DynamicalSystem, Extended Kalman Filter, Autonomous Navi-gation.

Abstract: An autonomous mobile robot should findfeasible trajectories in order to avoid collisions withobstacles in its environment. This ability to plancollision-free trajectories requires two major aspects:modulation and generation of trajectories. Thisis especially important if temporal stabilization ofthe generated trajectories is considered. Temporalstabilization means to conform to the plannedmovement time, in spite of diversified environmentalconditions or perturbations. This timing problemapplied for wheeled vehicles has not been addressedin most current robotic systems. Herein, we extendour previous work by integrating an architectureable to generate timed trajectories for a wheeledmobile robot, whose goal is to reach a target locationwithin a specified time, independently of the worldcomplexity.


Total Intelligent Life-Saving System againstEarthquake and “Tsunami”

Disaster Medical Engineering from Protection toCare in Catastrophe

Ichiro Fukumoto1, Taku Sahashi1,2 and AkiraSahashi2

1 Nagaoka University of Techology, Nagaoka city, Japan2 Project I- Ltd., Nagoya city, Japan

Keywords: Disaster Medicine, Rescue Robot,Evacuation Alarm, Earthquake -Tsunami Protection.

Abstract: Japan has suffered from earthquakesand Tsunami 4 times since 1995. The economicalloss reached to 25 thousand billion Yen and thenumber of the victims was more than 26,000. Inorder to prepare the future catastrophe we havedecided to recruit multidisciplinary regions of scienceand engineering and we have founded a researchgroup “The society of disaster medical engineering”for life saving techniques under natural disastersin 2004. The research subjects are wide from thephase-0 (Protection) to the phase-3 (medical care)that include from emergency alarms to rescue robotas well as the active Tsunami protection systems.The phase-0 includes earthquake/Tsunami proof,safety box and disaster communication system. Asthe phase-1 we have composed universal alarmingsystem as well as handy Braille communicator.Rescue Robot and other life saving systems areimportant instruments for the phase-2. Finally thephase-3 aims automatic support for the medical carein refuge shelters.


A Novel Approach to Measure under WaterVehicle Disturbance Force for Station Keeping

Control

J. Manecius Selvakumar and T. AsokanIndian Institute of Technology, Chennai, India

Keywords: Under Water Vehicles, Sensor Beam,Strain Gauge, Feedback Mechanism, DisturbanceForce, Station Keeping.

Abstract: Maintenance of target position and/ororientation is essential for underwater vehicles(UWV) to successfully complete a mission. However,in the case of work class vehicles, station keepingbecomes an important issue due to the presence ofdisturbance forces and requires effective feedback


Tuesday, 31


mechanism to maintain the pose. Conventionally,the changes in position due to disturbance forceis monitored and fed back to the station keepingcontroller to make necessary corrections. Thisintroduces unnecessary delay in response andcontinuous variations in vehicle position. In thispaper, an attempt has been made to develop adisturbance force measurement setup using straingauges which will directly measure the disturbanceforces which can be used for predicting the vehiclepose disturbance and make necessary correctionseven before the vehicle starts responding to thedisturbance forces. The methodology adopted forforce measurement is presented and experimentalanalysis has shown promising results. This approachcan be used as an alternative feedback mechanismfor station keeping control of underwater Vehicles.


Flocking Approach to Spatial ConfigurationControl in Underwater Swarms

Stefano Chiesa1,2, Sergio Taraglio2, StefanoPagnottelli3 and Paolo Valigi3

1 University of Roma Tre, Rome, Italy2 ENEA, Rome, Italy

3 University of Perugia, Perugia, Italy

Keywords: Distributed Control Systems, MobileRobots and Intelligent Autonomous Systems, Au-tonomous Agents.

Abstract: A modification of the flocking algorithmapproach for a swarm of underwater vehiclesis introduced. The proposed approach relaxesthe symmetry of the inter vehicle interaction. Itis thus possible to change the swarm spatialconfiguration assuming different formations withvarying parameters. The swarm geometry ischanged with a very limited effort, exploiting thecapability of the flocking approach to make emergea large scale arrangement. Examples of proposedvariations are provided. The vehicles are dynamicallymodelled and the relative non holonomic proportionalderivative controller is described. Experimental dataare gathered from many vehicle physical simulationsand graphically presented.


Hardware in the Loop Module to CalculateProduction Indicators

Mildred J. Puerto1, Josu Larrañaga1, Ulrich Doll2

and Damien Sallé11 TECNALIA, San Sebastián, Spain2 HOMAG AG, Schopfloch, Germany

Keywords: HIL, KPI, Remote Module, Manufactur-ing, Matlab, Hybrid Simulation.

Abstract: Currently Hardware in the Loop (HIL) isa powerful tool in manufacturing planning. A HILmodule for manufacturing must include the dynamicsof a critical machine, the logical control signals andthe production sequence information. By this wayHardware in the Loop can provide a complete setof possible cycletimes, due the module capability tochange delays times, sizes of the workpieces andproblems in the parts or in the logic states. All thisinformation could be re-used for manufacturers toimprove factory designs or by other managementmodules to improve production indicators underthe Virtual Factory Framework European projectframework. Due to the level of detailed dynamicsrequired in Hardware in the Loop simulations, it issuggested to integrate information from the multibodydynamic simulations programmed at the design level.


Programming of a Mobile Robotic Manipulatorthrough Demonstration

Lorenzo Peppoloni and Alessandro Di FavaScuola Superiore S. Anna, Pisa, Italy

Keywords: Programming by Demonstration, Manip-ulation, Autonomous Navigation, Semantic Learning.

Abstract: This paper presents an integratedrobotic system capable of learning and executingmanipulation tasks from a single human user’sdemonstration. The system capabilities are threefold.The system learns tasks from perceptual stimuli,models and stores the information in the form ofsemantic knowledge. The system may employ themodel achieved to execute task in an way similarto the example shown and adapt the motion torobot own constraints in terms of physical limits andinterferences. The system integrates perceptionand action algorithms in order to autonomouslyextrapolate the context in which to operate. Itrobustly changes its behavior according to the


Tues

day,

31


environment evolution. The performances of thesystem have been verified through a series of tests.The tests run on the Kuka youBot platform and allthe tools and algorithms are integrated into WillowGarage “Robotic Operating System” (ROS).


Improving Stereo Vision Odometry for MobileRobots in Outdoor Dynamic Environments

Dan Pojar and Sergiu NedevschiTechnical University Cluj-Napoca, Cluj-Napoca, Romania

Keywords: Visual Odometry, Stereo Vision, Local-ization, Mobile Robot, Real-Time.

Abstract: This article presents a method forlocalization able to provide the pose in 3D usingstereo vision. The method offers a better andinexpensive alternative to classical localizationmethods such as wheel odometry or GPS. Using onlya calibrated stereo camera, the method integratesboth optical flow based motion computation andSURF features detector for stereo reconstructionand motion computation. Robustness is obtainedby finding correspondences using both featuredescriptors and RANSAC inlier selection for thereconstructed points. Least squares optimizationis used to obtain the final computed motion. Worldscale pose estimation is obtained by computingsuccessive motion vectors characterized throughtheir orientation and magnitude. The methodinvolves fast algorithms capable to function at realtime frequency. We present results supporting globalconsistency, localization performance and speed aswell as the robustness of the approach by testing it inunmodified, real life, very crowded outdoor dynamicenvironments.


Robust Control of Excavation Mobile Robot withDynamic Triangulation Vision

Alexander Gurko1, Wilmar Hernandez2, OlegSergiyenko3,1, Vera Tyrsa1, Juan Ivan Nieto

Hipólito4, Daniel Hernandez Balbuena4 and PaoloMercorelli5

1 Kharkov National Automobile and Highway University,Kharkov, Ukraine

2 Polytechnic University of Madrid, Madrid, Spain3 Engineering Institute of Autonomous University of Baja

California, Mexicali, Mexico4 Engineering Faculty of Autonomous University of Baja

California, Mexicali-Ensenada, Mexico5 Leuphana University, Lüneburg, Germany

Keywords: Excavation Robot, Robust Control,Game Approach, Multiple Identification.

Abstract: The problem of control system synthesisfor excavation works autonomous mobile robot onthe basis of the game approach is considered.Vision function and spatial orientation of the robotis realized by the dynamic triangulation laser visionsystem. It is assumed that the real state of the objectbelongs to the certain set of potential states in theform of polyhedron. Simulation results and functionalability analysis for the proposed control system areconcluded.


Supplier Selection based on Process Yield forLED Manufacturing Processes

Y. T. TaiKainan University, Taoyuan, Taiwan

Chien-Wei WuNational Taiwan University of Science and Technology,

Taipei, Taiwan

Keywords: Supplier Selection, Yield, Multiple Char-acteristics, LED.

Abstract: In today’s fierce competitive businessenvironment, it is very essential to work with rightsuppliers in the supply chain systems. Consequently,supplier selection problem is very important andhas received considerable attention. In the supplierselection problem, quality is the most popularcriterion. In the paper, we consider a supplier


Tuesday, 31


selection problem of comparing two suppliersand selecting the one that has a significantlyhigher process capability for the light emittingdiode (LED) assembly process with multiplecharacteristics. Testing hypotheses for the LEDassembly supplier selection are presented. Forpractitioners’ convenience, the corresponding criticalvalues for the supplier selection in LED assemblyprocess are tabulated. For illustration purpose, anapplication is presented.


An Integrated Model to Evaluate thePerformance of Solar PV Firms

Chun Yu Lin1, Amy H. I. Lee1,1,1, He-Yau Kang2

and Wen Hsin Lee11 Chung Hua University, Hsinchu, Taiwan

2 National Chin-Yi University of Technology, Taichung,Taiwan

Keywords: Analytic Hierarchy Process (AHP), DataEnvelopment Analysis (DEA), Fuzzy Set, Solar Cells.

Abstract: The use of renewable energy resourcesis being stressed in the 21st century due tothe depletion of fossil fuels and the increasingconsciousness about environmental degradation.Renewable energies, such as wind energy, fireenergy, hydropower energy, geothermal energy, solarenergy, biomass energy, ocean power and naturalgas, are treated as alternative means of meetingglobal energy demands. After Japan’s nuclear plantdisaster in March 2011, people are aware that agood renewable energy resource not only needs toproduce zero or little air pollutants and greenhousegases, it also needs to have a high safety standardto prevent the chances of hazards from happening.Solar energy is one of the most promising renewableenergy sources with an infinite sunlight resource andenvironmental sustainability. However, photovoltaicproducts currently still require a high productioncost with low conversion efficiency. In addition, thesolar industry has a rather versatile market cycle inresponse to economic conditions. Therefore, solarfirms need to strengthen their competitiveness inorder to survive and to acquire decent profits inthe market. This research proposes a performanceevaluation model by integrating data envelopmentanalysis (DEA) and analytic hierarchy process (AHP)to assess the business performance of the solarfirms. From the analysis, the firms can understandtheir current positions in the market and to knowhow they can improve their business. A case study

is performed on the crystalline silicon solar firms inTaiwan.


An Integrated Replemishment Model underDynamic Demand Conditions

He-Yau KangNational Chin-Yi University of Technology, Taichung,

Taiwan

Amy H. I. LeeChung Hua University, Hsinchu, Taiwan

Chun-Mei LaiFar East University, Tainan, Taiwan

Keywords: Lot-sizing, Mixed Integer Programming,Multi-objective Programming, Genetic Algorithm(GA), Safety Stock.

Abstract: This research develops an integratedreplenishment model considering supplier selection,procurement lot-sizing, quantity discounts andsafety stocks under dynamic demand conditions.The objectives of the model are to minimize totalcosts, which include ordering cost, purchase cost,transportation cost, shortage cost and holdingcost, and to maximize service level of the systemover the planning horizon. First, a multi-objectiveprogramming (MOP) model is proposed in the paper.Next, the model is transformed into a mixed integerprogramming (MIP) model based on the -constraintmethod. Then, the genetic algorithm (GA) modelis constructed to solve a large-scale optimizationproblem by finding a near-optimal solution. Anexample of a bike manufacturer is used to illustratethe practicality of the proposal model. The resultsdemonstrate that the proposed model is an effectiveand accurate tool for the integrated replenishmentand logistics management.


Tues

day,

31



A Study on a Decision Support Model forStrategic Alliance in Express Courier Service

Friska Natalia Ferdinand1, Ki Ho Chung2, HyunJeung Ko3 and Chang Seong Ko2

1 Pusan National University, Busan, Korea, Republic of2 Kyungsung University, Busan, Korea, Republic of

3 Kunsan National University, Gunsan, Korea, Republic of

Keywords: Strategic Alliance, Express CourierService, Multi-Objective Programming, GeneticAlgorithm.

Abstract: The market competition of expresscourier service has become severe so that anexpress delivery company is consistently underpressure to operate its service network as efficientas possible. In this regard, the strategic allianceamong small and medium companies can be a usefulway in order to maintain their competitiveness. Tocope with such challenges, this study proposes adecision support model to examine the feasibility ofmerging under-utilized courier service centers andcollaborating consolidation terminals with strategicalliances among them. The proposed models can beformulated as multi-objective programming modelsmaximizing the minimum expected profit increase ofeach participating company. A solution procedurebased on the maxmin criterion is developed by usinga genetic algorithm. The applicability and efficiencyof the proposed models is demonstrated throughillustrative numerical examples.


Production Planning and Control Model ofTechnology Migration for DRAM Industry

Ying-Mei TuChung Hua University, Hsin Chu, Taiwan

Keywords: Technology Migration, DRAM Industry,X-factor, Production Planning and Control.

Abstract: Due to product life cycle has beenshortened rapidly, it forces the product generationand technology should be enhanced quickly. Whentechnology generation change occurred, DRAMmanufacturers always used the past experiences to

handle the change process. However, the issuesare totally different and it made the companiessuffered many difficulties. In this work, a productionplanning and control model is developed. Theproduction planning focuses on CCR (CapacityConstraint Resources) to define the complete waferrelease schedule and apply X-factor to schedule theproduction processes during the migration period.Regarding to the shop floor control, there are twocontrol mechanisms to control and monitor themigration process, real time control and predictingcontrol. WIP status is the important factor to decidewhether the production planner needs to launchthe rescheduling module or not in the real timecontrol portion. Besides, a foresee function isperformed by predicting control portion which firingthe rescheduling module by the bias between theloading and capacity curves.


Development of Robust Learning Control andApplication to Motion Control

Meng-Shiun Tsai, Chung-Liang Yen andHong-Tzong Yau

National Chung-Cheng Unviersity, Chiayi, Taiwan

Keywords: Iterative Learning Control, Motion Con-trol, Nurbs Curves, Robust H∞ Control.

Abstract: In this paper, the error dynamic equationof the ILC algorithm is derived with consideration ofparameter uncertainties and noise. The H∞ framework is utilized using the derived error dynamics todesign the robust learning controller. The properlearning gain is designed based on an optimizationprocess to ensure that both tracking performance andconvergence condition can be achieved. Simulationand experiments are conducted to validate the robustlearning algorithm and the system is stable everunder high payload uncertainty.


Tuesday, 31



Statistical Models for Emotion Recognitionusing Facial Expression Analisys

Hernan F. Garcia1, Augusto E. Salazar1,2 andÁlvaro A. Orozco1

1 Universidad Tecnolgica de Pereira, Pereira, Colombia2 Universidad Nacional-Sede Manizales, Manizales,

Colombia

Keywords: Emotion Recognition, AppearanceModel, Facial Feature Detection.

Abstract: Emotion recognition is one of thelatest challenges in intelligent human/computercommunication. In this paper we present ourframework for emotional state classification based onEkman’s study and facial expression analysis. FacialAction Coding System (FACS) on facial featurestracking, based on Active Appearance Model ispresented for facial expression analysis of featuresextracted from the parametric model Candide3. Wedescribe the characterization methodology fromparametric model to obtain the best set points offacial feature which improve the emotion recognitionprocess. Also quantitatively evaluated the accuracyof both feature detection and estimation of theparameters associated with facial expressions,analyzing its robustness to variations in pose andlocal variations in the regions of interest. Then, ametodology of emotion characterization is introducedto perform the recognition. The experimental resultsshow that the proposed model can effectively detectthe different facial expressions. Also this approachoutperforms the conventional approaches for emotionrecognition obtaining high performance results inthe estimation of emotion present in a determinedsubject. The model used and characterizationmethodology showed high accuracy to detect theemotion type in 95.6% of the cases.


New Robust Controller Synthesis OptimizationMethodology under Six Sigma Constraint

J. Frechard1, D. Knittel1,2 and J. Renaud11 INSA, Strasbourg, France

2 University of Strasbourg, Strasbourg, France

Keywords: Six Sigma Controller Synthesis, DesignFor Six Sigma, Robust Optimization, Robust Control,Hinfinity Synthesis, Roll-to-roll Systems.

Abstract: Industrial systems usually containuncertain parameters. These uncertainties comefrom lack of knowledge of physical phenomena orfrom evaluation difficulties. The stochastic behaviourof these parameters have to be taken into accountduring the design process of industrial systems. Totackle with these uncertainties, the Design for Multi-objective Six Sigma method is used. This methodpermits to fix six sigma constraint and it is appliedon the synthesis of web tension controller for anindustrial large-scale roll-to-roll system. The key ideais to synthesize the controller in frequency domainby Hinfinity synthesis. Then the six sigma constraintis fixed on output web tension in time domain. Therobust controller synthesis is then compared withstandard Hinfinity synthesis approach. This paperpresents for the first time a controller synthesisincluding six sigma constraints.


Flexibility in Organic SystemsRemarks on Mechanisms for Adapting System

Goals at Runtime

Christian BeckerLeibniz Universitaet Hannover, Hannover, Germany

Jörg Hähner, Sven TomfordeAugsburg University, Augsburg, Germany

Keywords: Organic Computing, Flexibility, MachineLearning, Goal Exchange, Network Protocol Config-uration.

Abstract: Within the last decade, technical systemsthat are capable of self-adaptation at runtimeemerged as challenging approach to cope with theincreasing complexity and interconnectedness intoday’s development and management processes.One major aspect of these systems is their ability


Tues

day,

31


to learn appropriate responses for all kinds ofpossibly occurring situations. Learning requires agoal function given by the user – which is subject tomodifications at runtime. In order to allow for flexiblemanipulations of goals within the system’s operationperiod, the learning component must be able tokeep knowledge in order to respond to varyinggoals quickly. This paper describes attempts toimplementing flexible learning in rule-based systems.First results show that efficient approaches arepossible even in real-world applications.

Paper 18011:30 - 13:00 Room VelazquezParallel Session 6a - Intelligent Control Systemsand Optimization

Distributed Control System for Crystal Growth

A. E. Kokh, V. A. Vlezko and K. A. KokhInstitute of Geology and Mineralogy, Novosibirsk, Russian

Federation

Keywords: Crystal Growth Control System, LBOCrystal, Controller, Load-commutator, Thermal FieldSymmetry.

Abstract: Distributed system for control over thecrystal growth process is presented. The mainadvantages of the system are its low cost, ability torecover after power failure, an application of standardISaGRAF software environment and available low-power PC-Controller. One of the option of the systemis an ability to control symmetry and dynamics ofthe heat filed. This option is the key factor for theprogress in growth of nonlinear optical LBO crystal.


Multi-Agent System for Adaptation of DistributedControl System

Dariusz Choinski and Michał SenikSilesian University of Technology, Gliwice, Poland

Keywords: DCS, Multi-Agent Systems, Ontology,Java, .Net, JADE, FIPA, OPC, XML, NHibernate, Hy-brid Systems, Concurrent Programming, KnowledgeSharing, Learning.

Abstract: A solution based on Multi-Agent Systemsproperties has been proposed. The presentedstructure is designed to Distributed Control Systembehaviour abstraction and encapsulation of thetechnical characteristics of its individual elementssuch as OPC (OLE for process control) servers. An

ontology facilitating the creation of user interface forMulti-Agent System environment has been proposed.This ontology is based on a set of concepts andsymbols understandable for the operator and theknowledge defining the hierarchical structure ofobject. Presented solution is not only a conceptionbut it is a real, cross platform implementation basedon the both Java and .Net programming platform. Itpractically shows how new programming solutions,tools and methodologies can be integrated andreused to solve real life, practical automation systemproblems.


An Optimization Method for Training GeneralizedHidden Markov Model based on Generalized

Jensen Inequality

Y. M. Hu1, F. Y. Xie1,2, B. Wu1, Y. Cheng1, G. F.Jia1, Y. Wang3 and M. Y. Li1

1 Huazhong University of Science and Technology,Wuhan, China

2 East China Jiaotong University, Nanchang, China3 Georgia Institute of Technology, Atlanta, U.S.A.

Keywords: Generalized Hidden Markov Model,Generalized Jensen Inequality, Generalized Baum-Welch Algorithm.

Abstract: Recently a generalized hidden Markovmodel (GHMM) was proposed for solving theproblems of aleatory uncertainty and epistemicuncertainty in engineering application. In GHMM,the aleraory uncertainty is derived by the probabilitymeasure while epistemic uncertainty is modelledby the generalized interval. Given any finiteobservation sequence as training data, the problemof training GHMM is often encountered. In thispaper, an optimization method for training GHMM,as a generalization of Baum-Welch algorithm, isproposed. The generalized convex and concavefunctions based on the generalized interval areproposed for inferring the generalized Jenseninequality. With generalized Baum-Welch’s auxiliaryfunction and generalized Jensen inequality, similarto the multiple observations training, the GHMMparameters are estimated and updated by thelower and the bound observation sequences.A set of training equations and re-estimatedformulas have been derived by optimizing theobjective function. Similar to multiple observations(expectation maximization) EM algorithm, thismethod guarantees the local maximum of the lowerand the upper bound and hence the convergence ofthe GHMM training process.


Tuesday, 31



Constrained Predictive Control of MIMO SystemApplication to a Two Link Manipulator

Joanna ZietkiewiczPoznan University of Technology, Poznan, Poland

Keywords: Predictive Control, Constraints, MIMOSystems, LQ Control.

Abstract: In the paper application of constrainedpredictive control to multi input, multi output systemis presented. The method is based on feedbacklinearization and LQ control. Constraints of thesystem are implemented by interpolation of referencetrajectory. Finding solution is a compromise betweenthe unconstrained LQ control and a constrainedfeasible control and is executed by minimization ofone variable. The application of the method to a twolink manipulator is used to present advantages andlimitations of the algorithm.


An Autonomous BipedConcept and Design

Peter JakubikAalto University, Espoo, Finland

Keywords: Biped, Biped Design, Dynamic Walking,Engineering Design.

Abstract: This paper argues for a new approach inthe mechanical design principle for the humanoidwalkers. Applying linear electric direct drivemotors the biped mechanism is able to behaveas dynamically highly reactive walker admissible toexploiting its own natural dynamics. Based on this,a whole new concept of an anthropomorphic walkerprototype is described including the interaction of thedesign and algorithmic aspects of the motion control.


Kinematic Analysis of Lower MobilityCooperative Arms by Screw Theory

Philip Long, Wisama Khalil and Stéphane CaroÉcole Centrale de Nantes, Nantes, France

Keywords: Cooperative Manipulators, HumanoidRobots, Screw Theory.

Abstract: This paper studies the kinematic modelingand analysis of a system with two cooperativemanipulators, working together on a common task.The task is defined as the transportation of an objectin space. The cooperative system is the dual armedhumanoid Nao robot, where the serial architectureof each arm has five degrees of freedom. Themobility of the closed loop system is analyzed andthe nature of the possible motion explored. The serialsingular configurations of the system are considered.Furthermore the parallel singularities of the closedloop system associated with each actuation schemeare analyzed.


Virtual Mechanism Approach for Dual-armManipulation

Nejc Likar1,2, Bojan Nemec2 and Leon Žlajpah21 Jozef Stefan International Postgraduate School,

Ljubljana, Slovenia2 Jozef Stefan Institute, Ljubljana, Slovenia

Keywords: Dual-arm Manipulation, Object Manipu-lation, Virtual Mechanism.

Abstract: We propose a novel control approachfor cooperative dual-arm object manipulation. Ourscheme has three typical features: (1) the twoarms with the object together form a new kinematicchain, where the base of the second arm is theend-effector of the new robot; (2) the object betweenthe robots is defined as a virtual mechanism,therefore manipulating the object is accomplished bycontrolling the virtual mechanism; (3) the proposedscheme allows cooperative dual-arm systemsperforming a task while moving on mobile platforms.The proposed algorithm is verified with experimentson a dual-arm system with Kuka LWR robots, andsimulations with 2 different robots: Kuka LWR on afixed support and Mitsubishi PA10 robot on a mobileplatform Nomad XR400.


Tues

day,

31



Two-arm Robot Teleoperation using aMulti-touch Tangible User Interface

Andreas Angerer, Andreas Bareth, Alwin Hoffmann,Andreas Schierl, Michael Vistein and Wolfgang Reif

University of Augsburg, Augsburg, Germany

Keywords: Software Architecture, Teleoperation.

Abstract: Teleoperation of robots can be interestingin various scenarios: operation in hazardousenvironments, medical surgery in the presence ofradiation or, in general, remote control of robotsin (partially) unknown environments. Crucial toteleoperation systems is an intuitive interface forthe remote operator to ensure straightforward andprecise control of the operated system. Advancedhardware devices and powerful software frameworksallow for quick development of such interfaces. Wepresent a case study implementing a novel two-armrobot teleoperation interface using multi-touch andtangible user interface concepts. The focus lies onthe implementation of this system, using an objectoriented robot programming framework and theMicrosoft Surface as a user interface platform.


Motion Detection and Velocity Estimation forObstacle Avoidance using 3D Point Clouds

Sobers L. X. Francis, Sreenatha G. Anavatti andMatthew Garratt

University of New South Wales@ADFA, Canberra,Australia

Keywords: Scene Flow, Lucas/Kanade,Horn/Schunck, 3D ,PMD Camera, Motion De-tection, Velocity Estimation.

Abstract: This paper proposes a novel threedimensional (3D) velocity estimation method byusing differential flow techniques for the dynamicpath planning of Autonomous Ground Vehicles(AGV) in a cluttered environment. We providea frame work for the computation of dense andnon rigid 3D flow vectors from the range data,obtained from the time-of-flight camera. CombinedLucas/Kanade and Horn/schunck approach is usedto estimate the velocity of the dynamic obstacles.The trajectory of the dynamic obstacle is predictedfrom the direction of the 3D flow field and theestimated velocity. By experiments, the utility of theapproach is demonstrated with the results.


The Modular Behavioral Environment forHumanoids and other Robots (MoBeE)

Mikhail Frank, Jürgen Leitner, Marijn Stollenga,Simon Harding, Alexander Förster and Jürgen

SchmidhuberDalle Molle Institute for Artificial Intelligence (IDSIA),

Manno-Lugano, Switzerland

Università della Svizzera Italiana, Lugano, Switzerland

Scuola Universitaria Professionale della Svizzera Italiana,Manno-Lugano, Switzerland

Keywords: Robotics, Modelling, Simulation, Archi-tecture, Framework, Humanoid, Adaptive RoadmapPlanning, Machine Learning, Cooperative Robots,Shared Workspace, Autonomous Adaptive Behavior,Unstructured Environment.

Abstract: To produce even the simplest human-like behaviors, a humanoid robot must be ableto see, act, and react, within a tightly integratedbehavioral control system. Although there existsa rich body of literature in Computer Vision, PathPlanning, and Feedback Control, wherein manycritical subproblems are addressed individually,most demonstrable behaviors for humanoid robotsdo not effectively integrate elements from allthree disciplines. Consequently, tasks that seemtrivial to us humans, such as pick-and-place in anunstructured environment, remain far beyond thestate-of-the-art in experimental robotics. We viewthis primarily as a software engineering problem, andhave therefore developed MoBeE, a novel behavioralframework for humanoids and other complex robots,which integrates elements from vision, planning,and control, facilitating the synthesis of autonomous,adaptive behaviors. We communicate the efficacy ofMoBeE through several demonstrative experiments.We first develop Adaptive Roadmap Planning byintegrating a reactive feedback controller into aroadmap planner. Then, an industrial manipulatorteaches a humanoid to localize objects as the tworobots operate autonomously in a shared workspace.Finally, an integrated vision, planning, control systemis applied to a real-world reaching task using thehumanoid robot.


Tuesday, 31



Robot-Dog – Human Interaction in Urban Searchand Rescue Scenarios

Improving the Efficiency of Rescue Teams inHazardous Environments

Anna Bosch, Xavier Cufi, Josep Ll. De la Rosa andAlbert Figueras

University of Girona, Girona, Spain

Keywords: Robot-Dog – Human Interaction, Au-tonomous Robot, Cognitive Systems, ImprovingEfficiency of Search and Rescue Teams, TrainedDogs.

Abstract: After a natural urban disaster the interiorof the rubble is often where the majority of victims arelocated. Mortality rates increases and peaks after 48hours, so it is of major importance to have fast andeffective search and rescue teams. Nowadays, therescue and exploration teams normally use dogs asa companion to find victims. Trained dogs are veryhelpful in these situations since their high mobility,speed and detection capacity. However they needconstant instructions and supervision, they can be indanger in some situations and they are not able tocollect precise data from the environment. Instead oftrying to build competing devices, the COMPANIONSproject looks at cooperation between natural andartificial creatures and in particular robots and dogs.This is rather new ground for research, where allthe dog shortcomings can be complemented withautonomous robots with cognitive abilities able tocooperate with dogs and humans in search andrescue environments. The aim of the project is toanalyse how a team of agents (robots-dogs-humansin this case) can cooperate and interact duringsearch and rescue. Research will be towards anew rescue scenario composed that will allow: (i) toempower the best characteristics of all the involvedagents and to minimize the worst ones; (ii) providethe fundamental tools for enabling these three agentsto work in a cooperative and efficient way in rescuemissions; and (iii) and to lengthen the human-doglink by allowing the exploration combining mobilerobots and trained dogs with more distant and saferhuman intervention in the dangerous rescue scenes.The main challenge will be the dog-robot interaction:to give visual cognitive and reasoning abilities to therobot in order to let him autonomously interact andcooperate with the dog according its behaviour andthe environment conditions; and to specifically traina dog to correctly accept and interact the robot (incharge of an expert dog training company).


Adaptive Data Update Management in SensorNetworks

C. M. KrishnaUniversity of Massachusetts at Amherst, Amherst, U.S.A.

Keywords: Sensor Networks, Energy-aware Com-puting, Energy Harvesting, Accuracy Balancing.

Abstract: Transmitting messages is by far the mostenergy-intensive thing that most sensors do. Weconsider the problem of a sensor which regularlysenses some parameter in its operating environment.Based on the value it knows to have been estimatedat the base station (or other central informationcollation station) for that parameter, the actualsensed value, its remaining energy levels, and otherquantities such as the time-to-go in the mission (iflimited) or the anticipated energy inflow (if energyharvesting is used), it decides whether that sensedvalue is worth transmitting. We present heuristics tomake this decision and evaluate their performance.


Hyper-elastic Pressure SensorsTemperature Dependence of Piezoresistivity of

Polyisoprene – Nanostructured CarbonComposite

Juris Zavickis, Maris Knite, Artis Linarts andRaimonds Orlovs

Riga Technical University, Riga, Latvia

Keywords: Temperature Dependence, Piezoresis-tivity, Polyisoprene, Carbon Black, Composite.

Abstract: Our previous efforts revealedpolyisoprene-nanostructured carbon compositeas prospective sensitive material for elaboration ofentirely hyper-elastic piezoresistive pressure sensorelement. In this article we investigate the temperaturedependence on initial electrical resistivity as well aspiezoresistive properties of such material and self-elaborated hyper-elastic pressure sensing element.Certain conclusions about the effect of temperatureon electroconductive structure and piezoresistivityare made.


Tues

day,

31



On the Design of Change-driven Data-flowAlgorithms and Architectures for High-speed

Motion Analysis

Jose A. Boluda, Pedro Zuccarello, Fernando Pardoand Francisco Vegara

Universitat de València, Burjassot, Spain

Keywords: High Speed Motion Analysis, BioinspiredVisual Sensing, Data-flow Architectures.

Abstract: Motion analysis is a computationallydemanding task due to the large amount of datainvolved as well as the complexity of the implicatedalgorithms. In this position paper we present someideas about data-flow architectures for processingvisual information. Selective Change Driven (SCD)is based on a CMOS sensor which delivers, orderedby the absolute magnitude of its change, only thepixels that have changed after the last time they wereread-out. As a natural step, a processing architecturebased on processing pixels in a data-flow method,instead of processing complete frames, is presented.A data-flow FPGA-based architecture is appointed indeveloping such concepts.


Optimal Control Strategy of NG Piston Engine asa DG Unit Obtained by an Utilization of Artificial

Neural Network

Jaroslaw Milewski, Lukasz Szablowski, Jerzy Kutaand Wojciech Bujalski

Warsaw University of Technology, Warsaw, Poland

Keywords: Artificial Neural Network, Control Strat-egy, Distributed Generation, Internal CombustionEngine.

Abstract: The paper presents a control strategyconcept of a piston engine fueled by Natural Gasas a DG unit obtained by using an Artificial NeuralNetwork. The control strategy is based on severalfactors and directs the operation of the unit in thecontext of changes occurring in the market, whiletaking into account the operating characteristicsof the unit. The control strategy is defined by anobjective function: for example, work at maximumprofit, maximum service life, etc. The results ofsimulations of the piston engine as a DG unit atchosen loads are presented. Daily changes in the

prices of fuel and electricity are factored into thesimulations.


Neuro-fuzzy Sliding Mode Control for a Two LinkFlexible Robot

Khedoudja Kherraz1, Mustapha Hamerlain2 andNouara Achour1

1 University of Sciences and Technology U.S.T.H.B,Algiers, Algeria

2 Center of Development of Advanced Technologies(CDTA, Algiers, Algeria

Keywords: Robot Manipulator, Flexible Link, SlidingMode, Neuro Fuzzy, Vibration Control, Chattering,Trajectory Control, Hybrid Control, Super TwistingAlgorithm.

Abstract: In most robotic applications, trajectorytracking control and vibration suppression in flexiblelink manipulator is a recurring problem, due to theunknown nonlinearities and strong coupling oftencaused by the presence of flexibility in the links.In order to solve this problem, a new sliding modecontroller using neural networks and fuzzy logic ispresented in this paper. The stability of the proposedcontroller is proved with the Lyapunov functionmethod. The neural network is used to compensatethe highly nonlinear system uncertainties. Thefuzzy logic is used to eliminate the chattering effectcaused by the robust conventional sliding modecontrol. The effectiveness of this control system willbe compared to the performance obtained with asecond order sliding mode control which is the supertwisting algorithm. Comparative simulations show thesuperiority of the proposed controller regarding thesecond order sliding mode controller and confirm itsrobustness with bounded disturbance and its abilityto suppress the flexible link manipulator vibrations.


Tuesday, 31



Multi-device Power-savingAn Investigation in Energy Consumption

Optimisation

Jean-Marc Andreoli and Guillaume BouchardXerox Research Centre Europe, Grenoble, France

Keywords: Energy Consumption Modelling,Sequential Decision Processes, Exact Multi-dimensional Optimisation.

Abstract: We are interested here in the problemof optimising the energy consumption of a set ofservice offering devices. Our target is a printingdevice infrastructure as typically found in a mediumor large office, where all the devices are connected tothe network, and clients can submit jobs to individualdevices through that network. We formulate a costoptimisation problem to find a trade-off between theenergy consumption of the infrastructure and the costof allocating jobs to devices. The latter cost resultsfrom the potential gap between the expectation of theclients and the quality of the service delivered by thedevices. We present a model of some of the typicalconstraints occurring in such a system. We thenpresent a method to solve the optimisation problemunder these constraints, and conclude the paper withsome experimental results.


Implementation of Cognitive Chips in MachiningError Attenuation

Maki K. RashidSultan Qaboos University, Muscat, Oman

Keywords: Cognitive Chips, Piezoelectric Switching,Vibration Attenuation, Smart Structure, MachiningErrors.

Abstract: Machining is a complex process thatrequires a high degree of precision with tightgeometrical tolerance and surface finish. Those areconfronted by the existence of vibration in the turningmachine tool. Overcoming a micro level vibrationof a cutting tool using smart materials can save oldmachines and enhance development in designingnew generations of machine tools. Using smartmaterials to resolve such problems represent oneof the challenges in this area. As a continuationfrom previous work for the transient solution for a

tool tip displacement using pulse width modulation(PWM) technique that was implemented for smartmaterial activation to compensate for radial disturbingcutting forces. A Fuzzy algorithm is developed tocontrol the actuator voltage level to improve dynamicperformance. Such technique together with the finiteelement method as dynamic model proved a greatsuccessfulness. To implement such results in reallife industrial system we may use chips that mimichuman brain as developed recently by IBM which isintelligent to learn through incidents, find patterns,generate ideas and understand the outcomes toreduce tool vibration error.


Nonlinear Model Validation of an InternalCombustion Engine with Few Sensors

Agustín Trobiani, Alvaro Agüero, Andrés García andOsvaldo Agamennnoni

Universidad Nacional del Sur, Bahía Blanca, Argentina

Keywords: ECU, Nonlinear Model, ParameterIdentification, Combustion Engine, Optimal Control,Attractor Reconstruction.

Abstract: This paper presents a nonlinear modelingof a small combustion engine and its validation.The salient characteristic of the procedure is thefewness on the amount of sensors, in fact usingonly three commercial sensors available for othercombustion motors, an error less than five percentwas obtained during the identification process.As another important point, the control of thecombustion engine replacing the original carburetorfor a commercial injector using the same hardwareutilized for identification (this time reprogrammed)made possible the control of the motor in closed loopforming in this way a cheap ECU.


Modeling and Active Vibration Control of aSmart Structure

Nader Ghareeb and Rüdiger SchmidtRWTH Aachen University of Technology, Aachen,

Germany

Keywords: Super Element Model, Lyapunov Func-tion Controller, Strain Rate Feedback Controller.


Tues

day,

31


Abstract: Active vibration control of flexiblestructures has gained much attention in the lastdecades. The major components of any activevibration control system are the mechanicalstructure suseptible to vibration, sensor(s) toperceive it, actuator(s) to counteract the influenceof disturbances causing vibration and finally,the controller responsible for the generation ofappropriate control signals. In this paper, a Lyapunovfunction controller and a strain rate feedbackcontroller (SRF) are used to attentuate the vibrationsof a cantilivered smart beam excited by its firsteigenmode. A super element (SE) with a finitenumber of degrees of freedom (DOF) is derived fromthe modified and damped finite element (FE) model.The state-space (SS) equations of the same modelare also extracted. Controllers are applied to theSE and SS models and results are presented andcompared.


Performance Shaping through Cost Cumulantsand Neural Networks-based Series Expansion

Bei Kang, Chukwuemeka Aduba and Chang-HeeWon

Temple University, Philadelphia, U.S.A.

Keywords: Statistical Optimal Control, CumulantMinimization, Neural Networks, Cost Cumulants,Performance Shaping.

Abstract: The performance shaping method isaddressed as a statistical optimal control problem.In statistical control, we shape the distribution ofthe cost function by minimizing n-th order costcumulants. The n-th cost cumulant, Hamilton-Jacobi-Bellman (HJB) equation is derived as thenecessary condition for the optimality. The proposedmethod provides an approach to control a higherorder cost cumulant for stochastic systems, andgeneralizes the traditional linear-quadratic-Gaussianand Risk-Sensitive control methods. This allows thecost performance shaping via the cost cumulants.Moreover, the solution of general n-th cost cumulantcontrol is provided by numerically solving the HJBequations using neural network method. The resultsof this paper are demonstrated through a satelliteattitude control example.


Robot Phase Entrainment on Quadruped CPGController

Vitor Matos and Cristina P. SantosUniversity of Minho, Guimarães, Portugal

Keywords: Central Pattern Generator, Quadrupedlocomotion, Locomotion Entrainment.

Abstract: Central Pattern Generators are used inseveral kinds of robot locomotion, from swimmingand flying, to bipeds, quadrupeds and hexapods. Itis thought that this approach can yield better resultsin dynamical and natural environments. In thiswork we expand a previous quadruped locomotioncontroller and propose a method to couple thestep cycle phase onto the locomotor CPG of aquadruped robot, creating a feedback pathway tocoordinate the phases of each leg to the phase ofthe step cycle. This approach is tested in a simulatedquadruped robot and the performed locomotion isevaluated. Results demonstrate that the proposedphase coupling synchronizes the swing step phaseof ipsilateral legs to the respective step phase of thecycle and show an improvement in stability of theperformed walk gait.


A Mobile Service Robot for IndustrialApplications

Luca Lattanzi, Giacomo Angione, Cristina CristalliAEA srl, Loccioni Group, Ancona, Italy

Florian Weisshardt, Georg Arbeiter, Birgit GrafFraunhofer IPA, Stuttgart, Germany

Keywords: Robot Design, Development and Con-trol, Mobile Robots and Intelligent AutonomousSystems, Autonomous Agents, Vision, Recognitionand Reconstruction, Service Robotics.

Abstract: This paper addresses the challenge ofintroducing mobile robots in industrial applications,where changes in the working environment anddiversification of tasks require flexibility, adaptabilityand in some cases basic reasoning capabilities.Classical industrial robots hardly permit to meetthese requirements, so a new concept of servicerobots facing challenging industrial productionsystem needs is proposed. The realization of suchan autonomous agent is illustrated and describedin details, focusing on mobility, environmental


Tuesday, 31


perception and manipulation capabilities. The resultis a mobile service robot able to face changeableconditions as well as unexpected situations anddifferent kinds of manipulation tasks in industrialenvironments. In this paper an implementationdedicated to household appliances production isdescribed, but the results achieved can be easilyextended to many industrial sectors, goods andelectromechanical components where high levels offlexibility and autonomy are needed.


Modeling and Visualizing Individual and GlobalTrends of a Multi-agent System

Vinicius Nonnenmacher, Marta Becker Villamil andLuiz Paulo Luna de Oliveira

Universidade do Vale do Rio dos Sinos, São Leopoldo,Brazil

Keywords: Simulation, Social groups.

Abstract: This paper proposes a new model forreal-time visualization of the social dynamics as aresultant of individual changes due to their mutualinteractions. The model allows the dynamicalvisualization of both, the individual characteristicschanges as well as the resultant system trends asa whole. As an application, we investigated the roleof individual degrees of influence and the number ofagents in the global choice of a population betweentwo antagonistic options.


A Generative Traversability Model for MonocularRobot Self-guidance

Michael Sapienza and Kenneth CamilleriUniversity of Malta, Msida, Malta

Keywords: Traversability Detection, AutonomousRobotics, Self-guidance.

Abstract: In order for robots to be integratedinto human active spaces and perform usefultasks, they must be capable of discriminatingbetween traversable surfaces and obstacle regionsin their surrounding environment. In this work,a principled semi-supervised (EM) framework ispresented for the detection of traversable imageregions for use on a low-cost monocular mobilerobot. We propose a novel generative model forthe occurrence of traversability cues, which are a

measure of dissimilarity between safe-window andimage superpixel features. Our classification resultson both indoor and outdoor images sequencesdemonstrate its generality and adaptability tomultiple environments through the online learningof an exponential mixture model. We show thatthis appearance-based vision framework is robustand can quickly and accurately estimate theprobabilistic traversability of an image using notemporal information. Moreover, the reduction insafe-window size as compared to the state-of-the-artenables a self-guided monocular robot to roam incloser proximity of obstacles.


3D Dense Reconstruction from Two-FrameStereo Correspondences

Costantino Scozzafava1 and Giuseppe Zuffanti2,11 Erxa S.r.l., Turin, Italy

2 Lagrange Project - CRT Foundation - ERXA SRL,TURIN, Italy

Keywords: Computer Vision, Stereovision, DisparityMap, 3D Reconstruction, Local Area-based PixelMatching.

Abstract: In this paper a novel approach to theproblem of 3D reconstruction from stereovision isproposed. The analyzed solution is based on theidentification of suitable metrics of image pixels ableto measure their (dis)similarity using rectangular,fixed-size, local support windows. Every localsupport window, centred in a given pixel, is weightedbased on colour proximity and colour segmentationof surrounding pixels. Then a metrics for dissimilaritycomputation between potentially homologous pixelsacross the images is applied using an aggregation(convolution) of raw costs together a measure ofcolour affinities. Finally the resulting measures areused to build two disparity maps (left-to-right andright-to-left) and the maps fused in a final one.


Tues

day,

31



Object Recognition based on a Simplified PCNN

Yuli Chen1,2, Yide Ma1, Dong Hwan Kim2 andSung-Kee Park2

1 Lanzhou University, Lanzhou, China2 Korea Institute of Science and Technology, Seoul,

Korea, Republic of

Keywords: Simplified Pulse Coupled NeuralNetwork (SPCNN), Image Segmentation, ObjectRecognition, Region-based Matching.

Abstract: The aim of the paper is to propose aregion-based object recognition method to identifyobjects from complex real-world scenes. Theproposed method firstly performs a colour imagesegmentation by a simplified pulse coupled neuralnetwork (SPCNN) model, and the parameters ofthe SPCNN are automatically set by our previouslyproposed parameter setting method. Subsequently,the proposed method performs a region-basedmatching between a model object image and atest image. A large number of object recognitionexperiments have proved that the proposed methodis robust against the variations in translation, rotation,scale and illumination, even under partial occlusionand highly clutter backgrounds. Also it shows a goodperformance in identifying less-textured objects,which significantly outperforms most feature-basedmethods.


Integrating Kinect Depth Data with a StochasticObject Classification Framework for Forestry

Robots

Mostafa Pordel, Thomas Hellström and AhmadOstovar

Umeå University, Umeå, Sweden

Keywords: Image Processing, Depth Integration.

Abstract: In this paper we study the integration of adepth sensor and an RGB camera for a stochasticclassification system for forestry robots. The imagesare classified as bush, tree, stone and human andare expected to come from a robot working in forestenvironment. A set of features is extracted fromlabeled images to train a number of stochasticclassifiers. The outputs of the classifiers are thencombined in a meta-classifier to produce the finalresult. The results show that using depth informationin addition to the RGB results in higher classification

performance.


Wave Vibration Analysis of Classical Multi-storyPlanar Frames

C. MeiThe University of Michigan - Dearborn, Dearborn, U.S.A.

Keywords: Wave Vibration, Multi-story Frame,Bending Vibration, Longitudinal Vibration.

Abstract: This paper concerns free vibrationanalysis of in-plane vibrations in classical multi-storyplanar frame structures. An exact analytical solutionis obtained using wave vibration approach. Thecoupling effects between bending and longitudinalvibrations in frames are taken into account. Classicalbeam theories are applied in modeling the flexuraland longitudinal vibrations. Reflection matrices at“sliding” and “rolling” boundaries, as well as reflectionand transmission matrices at the “L” and “T” joints areobtained. Numerical examples are presented alongwith comparisons to results available in literature.


On the Bus Priority Dilemma: Modelling andAnalysis of Congested Traffic Network using

Coloured Petri Nets and (Min,+) Algebra

Hamza Boukhentiche, Abdeljalil Abbas-turki andAbdellah El Moudni

SeT-EA3317UTBM, Belfort Cedex, France

Keywords: The Bus Priority Dilemma, Petri Nets,Dioid Algebra, Resource Sharing, Modeling andAnalysis.

Abstract: In this paper, a mathematical approachfor modeling and analysis the bus priority system isintroduced. The main objective is to highlight the buspriority dilemma in a congested traffic network. Thestudy of the bus priority needs to consider the twokinds of vehicles that move on the same segment.Based on coloured Petri net and (min, +) algebra, anew proposition is introduced to consider mutuallythe two vehicle kinds in a given hybrid segment.Afterwards and for analysis, we propose two newconstraints that model the delay and the waitingtime of a vehicle in this exotic algebra. Thus, anequation which relates the bus delay and the traffic


Tuesday, 31


lights switching using (min,+) algebra is formulated.In order to provide us with interesting performanceevaluation, an example is worked out to highlight thisdilemma.


GFIS: Genetic Fuzzy Inference System forSpeech Recognition

Washington Luis Santos Silva and Ginalber Luiz deOliveira Serra

Federal Institute of Education, Science and Technology,São Luis, Brazil

Keywords: Recognition Speech, Fuzzy Systems,Optimization, Genetic Algorithm, Discrete CosineTransform.

Abstract: The concept of fuzzy sets and fuzzylogic is widely used to propose of several methodsapplied to systems modeling, classification andpattern recognition problem. This paper proposesa genetic-fuzzy recognition system for speechrecognition. In addition to pre-processing, with mel-cepstral coefficients, the Discrete Cosine Transform(DCT) is used to generate a two-dimensional timematrix for each pattern to be recognized. A geneticalgorithms is used to optimize a Mamdani fuzzyinference system in order to obtain the best modelfor final recognition. The speech recognition systemused in this paper was named Genetic FuzzyInference System for Speech Recognition (GFIS).Experimental results for speech recognition appliedto brazilian language show the efficiency of theproposed methodology compared to methodologieswidely used and cited in the literature.


Control of Variable Parametric Actuator inBuilding Temperature Regulation

Todor IonkovTechnical University of Sofia, Bulgaria, Sofia, Bulgaria

Keywords: Building Management Control Systems,HVAC, Nonlinear Controller, Actuator.

Abstract: Building management control systemsprovide occupants with comfortable temperature,humidity, etc. in different environments. Severalcontrol strategies have been introduced to control

the temperature regulation, where an operatingscheduling is pre-defined. In this context, standardPI control algorithms are adequate for the controlof Heating, Ventilation and Air Conditioning (HVAC)systems. Due to the specificity of object dynamic,the uses of standard algorithms for regulation lead tofluctuations in angle actuator position. Achieving ofenergy saving control in these systems is correlatedwith minimisation of these actuator fluctuations.A novel nonlinear controller is proposed for highperformance parametric variable motion controlsystems in actuators for Building TemperatureRegulation. This control utilizes Lyapunov functionto guarantee the convergence of the valve positiontracking error from all possible initial conditions.Simulation results and performance comparisonconfirm that the proposed control scheme offerssignificantly improved performance in terms ofthe trajectory tracking ability to time-varyingreference input, system control bandwidth, androbustness against external disturbances. The addedintegrator improves the system’s robustness againstmodelling uncertainties and external disturbance,thus improving steady-state control accuracy.


Productivity Improvement through LayoutRedesign

A Lean Approach Case Study

Wa-Muzemba Anselm TshibanguUniversity of Maryland Baltimore County, Maryland, U.S.A.

Steve BerlinskiState University of New York /College of Technology at

Alfred, New York, U.S.A.

Keywords: Lean, Cell Manufacturing, Layout De-sign, Continuous Improvement.

Abstract: This paper explains how a small-sizecompany located in the western New York regionhas used a lean approach to redesign its operationallayout and eliminate unnecessary transportationmoves to deal with financial turmoil and surviveglobal competition. Although this is a true realworld case study, the company’s name is referredto as West City Vacuum Forming Inc. (WCVF Inc.)throughout the paper to protect the company privacyas they are still in business today. A couple of yearsago when WCVF Inc. lost two of its major customersto competition and technology changes, the companysales were tragically impacted and severe measuresof cost reductions were urgently needed in order tomaintain the company in business. WCVF Inc. then


Tues

day,

31


decided to retrench and re-evaluate its manufacturingpractices. As a result of this self assessment effort,drastic measures including a significant downsizingof the workforce and a consolidation of the spacefloor were addressed to save the company fromfinancial turmoil. This paper specifically analyzes thefloor consolidation aspect because it resulted into anew configuration of operational layout that improvedWCVF Inc. operations, productivity, and materialflow by eliminating unnecessary transportationactivities. Other benefits recorded include annualoperational costs saving of approximating $50,000and a reduction in cycle times in the order of 4.8 daysfor some products, representing a cut of 5%.


The Optimal Gasoline Blending into RomanianRefinery

Case Study

Cristian Patrascioiu and Nicolae NicoletaPetroleum-Gas University of Ploiesti, Ploiesti, Romania

Keywords: Gasoline Blending, Mix Formula, LinearOptimization, Objective Function, Restrictions, MAT-LAB Optimization, Industrial Optimization Software.

Abstract: The paper presents the research intoRomanian industrial in line gasoline blending. Thepaper is structured in three parts. First part presentsan industrial bending system of a Romanian refinery.The second part contents a research of the blendingmathematical model for petroleum products. The lastpart is a case study, regarding an industrial problemof the gasoline blending in a Romanian refinery.The study has reported similitude and differencesbetween the optimal solution calculated by BlendOptimization and Supervisory System (BOSS) andthe optimal solution obtained by using the author’smodel.

16:00 - 17:00 Room PlenaryOptimization in Design of Automated MachiningSystemsKeynote Speaker: Alexandre Dolgui

Optimization in Design of Automated MachiningSystems

Alexandre DolguiEcole des Mines de Saint-Etienne, Saint Etienne, France

Abstract: The aim is to develop advancesoptimization techniques from the perspective ofcombinatorial design of automated manufacturing

systems. The suggested methodology is basedupon advanced line balancing methods and processplanning and equipment selection algorithms. Themain results are based on exact mathematicalprogramming methods and their intelligent couplingwith heuristics and metaheuristics. A decision-aidsystem based on these methods will be presented.This system is employed for the design of massproduction dedicated machining lines as well as forreconfigurable manufacturing systems.

Closing Session17:00 - 17:15 Room Plenary


Tuesday, 31

NOTES ICINCO 2012


ICINCO 2012 NOTES


NOTES ICINCO 2012


ICINCO 2012 NOTES


Documents

ICINCO 2012 Final Program and Book of Abstracts · ICINCO 2012 Final Program and Book of Abstracts 9th International Conference on Informatics in Control, Automation and Robotics