(FD=Full Day Workshop, HD=Half a Day Workshop)
FD-1 Control, estimation, and optimization of interconnected systems: from
theory to industrial applications
Sunday, December 11, 8:30 - 17:30 Prado
Organizers: Mihailo R. Jovanovic (University of Minnesota) and Cédric Langbort (California Institute of Technology)
Additional Participants: Bassam Bamieh (Univ. of California at Santa Barbara) , Raffaello D’Andrea (Cornell University), Geir Dullerud, (Univ. of Illinois at Urbana-Champaign), Dimitry Gorinevsky (Honeywell Laboratories/Stanford University), Sanjay Lall ( Stanford University), Fernando Paganini (Univ. of California at Los Angeles) and Greg Stewart ( Honeywell/University of British Columbia).
Target Audience: The workshop is aimed at a broad audience of students, researchers, and industry professionals within the control community.
Summary: Large networks of interconnected dynamical systems are becoming prevalent in modern technological applications, as exemplified by the development of cross-directional control systems for paper machines, power distribution systems, automated highways, formations of unmanned aerial vehicles, and arrays of micro-cantilevers for massively parallel data storage, to name just a few. These applications are also of theoretical interest because they pose new challenges for analysis and control design.
BACK TO LIST
FD-2 Embedded Control Systems (CANCELLED *)
Organizers: Alfons Crespo (Universidad Politécnica de Valencia, Spain) and Pedro Albertos (Universidad Politécnica de Valencia, Spain)
Additional Participants: KarlErik Arzen (Lund University, Sweden), Anton Cervin (Lund University, Sweden) and Martin Torgren (Royal Institute of Technology in Stockholm, Sweden)
Target Audience: This workshop is addressed at a broad audience of young researchers in realtime and control, experts in Embedded Systems with interest in Control applications and experts in control with interest in embedded RT control solutions.
Summary: Currently, most automated control applications are implemented as embedded systems where the control and realtime aspects are fully connected. This strong interaction between Control and Embedded systems forces the need of researchers that can combine both fields.
BACK TO LIST
FD-3 State of the Art Computational Methods and Software for Computer-Aided Control Systems Design and Software (CANCELLED *)
Organizers: Biswa N.Datta (Northern Illinois University, USA)
Target Audience: The workshop will be of interests to the graduate students, researchers and practicing engineers working on a wide variety of control applications. The workshop will also help the instructors design a graduate level course on computer-aided control systems design and analysis.
Contact Information: http://www.math.niu.edu/~dattab/
Summary: The workshop will present lectures on the state-of-the-art computational methods and software on almost all aspects of control systems. The lectures will be organized to clearly explain the algorithms in a manner that is suitable for easy implementations, the important aspects of computer implementations will be clearly discussed, a clear and concise comparative study of one algorithms over the others for a given problem will be presented and recommendations will be made for the practicing engineers. Mathematical and computational jargon that seem to be distractive for most engineers to learn these techniques will be avoided.
BACK TO LIST
FD-4 Discrete-time nonlinear control systems (CANCELLED *)
Organizers: Eduardo Aranda-Bricaire (CINVESTAV, México)
Additional Participants: Claude Moog (IRCCyN, France) and Ülle Kotta (Tallinn University of Technology,Estonia)
Summary: Discrete-time systems arise in technology and applied mathematics in different ways. The study of such systems possesses interest both from theoretical and practical points of view. The workshop will present a sound theoretical framework which allows solving a broad class of control problems, e.g. feedback linearization, disturbance decoupling, control by output feedback, observer design, identification, realization, etc. Such framework is provided by the notions of differential forms and exterior differential systems. The workshop is based on a series of papers published by the organizers and coworkers over the past 12 years. It relates also to the difference algebraic framework originally introduced by Fliess.
BACK TO LIST
FD-5 Fault Detection and Diagnosis Based on Explicit Models and on Principal Components (CANCELLED *)
Organizers: Janos Gertler (George Mason University, USA) and S. Joe Qin (University of Texas, Austin, USA)
Summary: The detection and diagnosis of faults (malfunctions) in engineering systems is of paramount importance in modern technology. While limit checking is the classical approach in industry, utilizing a mathematical model of the plant allows for consistency checks. Such checks lead to residuals that are much more sensitive and specific to faults. Consistency checks have been developed in the control and aerospace community using explicit plant models (analytical redundancy). A great wealth of know-how has been accumulated in this framework, including methods of disturbance decoupling, statistical testing and residual enhancement for fault isolation. In the chemical and process industries, principal component analysis has been the method of choice. Principal components lead to an implicit model with a significantly reduced dimension. Principal component models also lead to residuals that have fundamentally the same properties and may be subjected to the same treatment as analytical redundancy residuals.
BACK TO LIST
FD-6 Iterative Learning Control: Algebraic Analysis and Optimal Design (CANCELLED *)
Kevin L. Moore – Colorado
School of Mines and Jari Hätönen – University of Sheffield
YangQuan Chen (Utah
State University, USA), David H. Owens ( University of Sheffield , UK) and
Eric Rogers ( University of Southampton, UK)
Target Audience: The expected audience includes engineers, scientists, postgraduate students, and academics. The workshop will be self-contained so that it is suitable for systems and control researchers and practitioners who may not be familiar with the concept of ILC as well as to those with some background in the field.
Summary: The purpose of this workshop is to present a unified exposition of recent advances in ILC analysis and design, providing an integrated view of the presenters’ collaborative research on theoretical and experimental aspects of ILC research, two streams that have developed over several years into a systematic methodology. At the 2000 CDC in Sydney a tutorial introduction to ILC was given that concentrated on the fundamentals of ILC. The present workshop is aimed at a more advanced and systematic focus on the algebraic approach to ILC analysis and on norm-optimal design of ILC algorithms that has developed since the previous tutorial. Throughout the workshop the underlying theme will be on ILC as a mature design methodology with both significant demonstration of, and further potential for, actual implementations with clearly visible returns in terms of improved performance.
BACK TO LIST
FD-7 Model Predictive Control of Hybrid Systems
Sunday, December 11, 8:30 - 17:30 Nervión III
Organizers: Alberto Bemporad (univ. of Siena, Italy)
Additional Participants: Stefano Di Cairano (University of Siena, Italy) and Nicolo Giorgetti (University of Siena, Italy)
Target Audience: The workshop is intended for researchers and engineers that want to learn about the theory and practice of model predictive control (MPC) for linear and hybrid systems, from the basic MPC setup to advanced hybrid modeling and explicit representations of MPC in piecewise a.ne form via multiparametric programming.
Summary: Most of the control synthesis approaches developed in the last few years for hybrid systems involve the solution of optimal control problems. For continuous-time hybrid systems, researchers either studied necessary conditions for a trajectory to be optimal, or focused on the computation of optimal/suboptimal solutions by means of dynamic programming or the maximum principle. The hybrid optimal control problem becomes less complex when the dynamics is expressed in discrete-time or as discrete-events, as in general the main source of complexity becomes the combinatorial (yet .nite) number of possible switching sequences. By looking at the MPC problem as a multiparametric program where the control inputs are the optimization variables and the states and references are the parameters, the equivalent explicit form of the MPC control law can be computed algorithmically. Such a control law is piecewise a.ne, and consequently, MPC can be implemented as a look-up table of linear gains, drastically easing on-line computations and making the control code much simpler because no numerical optimization solver is involved.
During the workshop, the attendee will be introduced to a free software tool, the Hybrid Toolbox developed by the workshop’s proposers , which enables a direct application in Matlab of the techniques for modeling, simulating, and verifying the safety properties of hybrid systems, for designing MPC controllers for linear systems with constraints and hybrid systems, and for determining equivalent piecewise a.ne control functions that can be immediately prototyped on hardware. During the workshop a few case studies will be described that were performed in collaboration with Ford Research Laboratories (Dearborn, MI), to provide an industrial viewpoint and highlight the potentials and limitations on the proposed hybrid MPC technologies.
BACK TO LIST
The Behavioral Approach to Systems and Control:
Introduction and Recent Advances
Sunday, December 11, 8:30 - 17:30 Arenal III
Organizers: Jan Willem Polderman (University of Twente, the Netherlands) and Harry Trentelman (University of Groningen, the Netherlands)
Additional Participants: Paula Rocha (University of Aveiro, Portugal), Eva Zerz (University of Kasiers Lautern, Germany) and . Jan Willems (Catholic University of Leuven, Belgium)
Summary: Behavioral models are well suited to treat physical models where the signal flow direction implied by the usual input/output setting is awkward and physically artificial. For instance in in-terconnected systems input-to-output interconnection is seldom the way physical systems interact, witness the force equality implied in interconnection of mechanical components. Furthermore the behavioral framework is very convenient for studying dynamical systems as composed of smaller subsystems. Other areas where the behavioral approach has proven its suitability is in coding theory, and in composition of Discrete Event Systems through shared variables.
The aim of the workshop is twofold. Firstly, to present a self-contained introduction to the behavioral theory of systems. Secondly, to provide some of the recent advances in the behavioral approach. The introductory part covers the behavioral theory of linear systems. The second part is devoted to systems described by partial di erential equations and to dissipative systems. The audience that we have in mind are graduate students, researchers, and senior scientists.
BACK TO LIST
Developments in Control Performance Limitation Research: A Tale in the Network
Sunday, December 11, 8:30 - 17:30 Nervión II
Organizers: A. Pedro Aguiar (Univ. of California, Santa Barbara, USA), Jie Chen (Univ. of California, Riverside, USA), Rick Middleton (Univ. of Newcastle, Australia) and Li Qiu (Hong Kong University of Science and Technology, P.R. China)
Additional Participants: Andrzej Banaszuk (United Technology Research Center), Munzer Dahleh (MIT), James S. Freudenberg ( Univ. of Michigan), Nicola Elia (Iowa State University, ), Graham C. Goodwin (Univ. of Newcastle), Shinji Hara (Univ. of Tokyo), Joao Hespanha (Univ. of California, Santa Barbara), Iven Mareels (Univ. of Melbourne) and Sanjoy Mitter (MIT).
Summary: The worshop addresses constraints, limitations, and tradeoffs in feedback control design and implementation. The proposal is motivated by recent intense activity in this fundamental area of research and its broad applications to control over communication networks, nonlinear control, and control applications. The workshop will bring together leading researchers to make tutorial presentations on emerging problems and new results, with an emphasis placed on control over communication networks and nonlinear tracking illustrated by signi.cant applications. We believe that the theme and the topics to be presented in the workshop are of major interest and will attract a wide audience.
BACK TO LIST
HD-1 Dynamics and Control of Smart Structures (CANCELLED *)
Organizers: S. O. Reza Moheimani (Univ. of Newcastl, Australia) and Andrew Fleming (Univ. of Newcastle, Australia)
Target Audience: It is expected that this workshop will be of great interest to a systems and control audience.
Summary: The field of smart structures is an emerging engineering field that has captured the attention of many engineering professionals and academics in recent years. A smart structure incorporates "smart material" sensors and actuators, electronic signal processing, and advanced control systems to produce appropriate actuator response for particular sensor inputs. Smart materials are materials that possess adaptive capabilities to external stimuli. Electro-strictive polymers, piezoelectric materials, electro-rheological and magneto-rheological fluids, shape memory alloys and magneto-strictive materials fall within this category. The purpose of this workshop is to introduce this emerging field to a systems and control audience. Smart Structures is an important field of research with great potential for developing new technologies across many industries. Given the important role of feedback control algorithms in optimizing the performance of smart structural systems, control engineers can make significant contributions to the development of this field.
BACK TO LIST
HD-2 Advanced Control in Siderurgy (CANCELLED *)
Organizers: Nicolás de Abajo (Knowledge Innovation Research Centre Aceralia Arcelor Group)
Summary: The area of control in steel production is one of the most dynamic issues for the minimization of production costs and improving the facilities yield. The high competitiveness of the steel market and the continuous increase of the client quality requirements are translated in a permanent investment in new control systems both hardware and software. Even upstream (Blast furnace and Steel Making) production where traditionally the control was not considered a precise tool in terms of quality (safety is critical), it’s now considered a priority due to the reduction of the time cycle for production. Downstream production where quality is a key factor of success, the precision of the control system is critical and the stability is a major topic for increasing the reliability of the facilities. Some of the most advanced control techniques are tested and used in the steel production like State Process Control and ANN. The purpose of this workshop is to introduce this topic to systems and control audience, motivating the launching of R&D projects and applied research demonstrations in the steel production field. The transferability of these projects is enormous and the top level requirements in terms of real time and reliability should challenge the control community. All the areas related with control in the steel production are open with special emphasis in sensors data reconciliation, control algorithms, learning feedback loops and hard real time solutions not only applied to an specific facility but even factory wide.
BACK TO LIST
HD-3 Identification of Hybrid Systems
Sunday, December 11, 8:30 - 12:30 Triana
Organizers: Aleksandar Juloski (Eindhoven University of Technology) and Giancarlo Ferrari-Trecate (Politecnico di Milano)
Additional Participants: Rene Vidal (Johns Hopkins University, USA) and Simone Paoletti (Univ. of Siena, Italy)
Target Audience: The workshop is thought for graduate students and researchers with a background in the fields of hybrid systems or system identification. The material covered at the workshop will present state of the art in hybrid identification as well as open research problems in this area. The workshop will also focus on several practical case studies and a discussion on merits and drawbacks of the presented methods. This may be of interest to researchers in other areas and practitioners who want to use identification techniques to obtain hybrid models in practical applications. A minimal prior knowledge in the fields of linear identification and hybrid systems is assumed.
Summary: Due to their ubiquity and many potential applications, hybrid systems attracted a lot of attention in the control community during the last few years. Numerous results on modelling, analysis, verification and control synthesis appeared in the literature. However, most of the theoretical developments hinge on the assumption that the accurate quantitative hybrid model of the process at hand is readily available. In some situations it is possible to obtain such a model starting from first principles. However, in most practical situations first principles modelling is too complicated or even impossible to use, and the model needs to be identified on the basis of experimental data. Methods geared specifically towards the identification of models with a hybrid structure areof very recent date, and to date, several methods exist which have been successfully applied in practical situations. This workshop will focus on the following ones: a) Clustering-based algorithms, b) Bounded error algorithms, c) Bayesian algorithms, and d) Algebraic algorithms. The comparison and several practical applications will be also presented.
* If you are registered in a cancelled workshop you can either register in another Workshop of the same duration (just e-mail registration co-chair Carlos Bordons) or you can have the money refunded. You can also register in another Workshop if you wish (go to PaperPlaza). If we do not hear from you by Nov 14th we will refund the money.