| |
|
Six one-day workshops and one two-day workshop have been selected
from the proposals. The two-day workshop will span June 2nd and 3rd.
Three one-day workshops are scheduled for Monday, June 2, with three
additional one-day workshops on Tuesday, June 3.
|
|
| |
C-1: Practical Techniques
in Control Engineering
|
| Monday, June 2, 2003, and Tuesday, June 3, 2003
|
- Dennis S. Bernstein; University of Michigan
- Carl R. Knospe; University of Virginia
|
This two-day short-course will provide a bridge between
recent developments in control theory and their practical
application in the laboratory and industry. Beginning with
an overview of fundamental tradeoffs and issues that affect
control-system performance, the course will systematically
cover topics in linear and nonlinear modeling, linear and
nonlinear controller synthesis, and fixed-gain and adaptive
tuning. In addition, controller implementation issues such
as saturation, quantization, and state constraints will be
discussed. The theoretical foundation of each topic will be
reviewed along with a discussion of practical ramifications
and limitations. The course is suitable for students, instructors,
and researchers who wish to obtain a broad perspective of
the control engineering enterprise as well as control engineers
from all industrial applications seeking a coherent, self-contained
overview of recent developments relevant to control practice.
|
SCHEDULE: Monday, June 2, 2003 |
8:30-10:30 DEFINING THE ISSUES AND CHALLENGES IN
CONTROL ENGINEERING
Course Overview
Control-System Design: Strategy, Physics, Architecture, and
Hardware
Fundamental Tradeoffs: Plant Properties, Hardware Constraints
and Performance Limitations
10:45-12:30 DEVELOPING LINEAR MODELS FOR CONTROL
Linear Modeling Issues: Representation, Analysis, and Limitations
Empirical Linear Modeling: Identification, Reconciliation,
and Validation
13:30-15:30 SYNTHESIZING LINEAR CONTROLLERS FOR PERFORMANCE
AND ROBUSTNESS
Performance and Robustness Metrics
Optimality-Driven Synthesis: H2 and Hinfinity Methods
Robust Control and Loop Shaping: Classical and Modern Methods
15:45-17:30 REDUCING MODEL DEPENDENCE IN CONTROLLER
SYNTHESIS
Minimal-Information Control: The Art and Science of PID Tuning
Adaptive Control: What Do You Need to Know? How Well Do You
Need to Know It?
Adaptive Disturbance Rejection with Applications to Noise
and Vibration Control
|
|
SCHEDULE: Tuesday, June 3, 2003
|
08:30-10:00 DEVELOPING NONLINEAR MODELS FOR CONTROL
Nonlinear Modeling Strategies: System Structure and White/Grey/Black
Box Models
Nonlinear Identification Methods: From Block-Structured to Neural
10:15-11:30 INEXACT APPROACHES TO NONLINEARITY
Treating Nonlinearity as Uncertainty: Absolute Stability
Treating Nonlinearity Approximately: Describing Functions and Statistical
Linearization
Treating Nonlinearity as Linearity: Gain Scheduling, LPV”¦s, and
Frozen Linear Methods
11:30-12:30 EXACT APPROACHES TO NONLINEARITY
Feedback Linearization
Backstepping
13:30-14:30 FACING THE REALITY OF CONSTRAINTS
Dealing with Amplitude and Rate Saturation: Antiwindup and Command-Modification
Methods
Dealing with State Constraints: Model Predictive Control
14:30-15:45 IMPLEMENTING REAL CONTROL SYSTEMS IN REAL HARDWARE
Control-Relevant Aspects of Sensors, Actuators, Amplifiers, and
Filters
Digital Implementation Issues: Aliasing, Phase Lag, Quantization,
and Bandwidth Limitations
Controlling Dangerous Plants: Stabilization in the Real World
16:00-17:30 PUTTING IT ALL TOGETHER
Design Examples and Case Studies
Wrap-up: Theoretical Gaps with Practical Implications and New Ideas
for Old and New Problems
|
 |
M-1: Process and Control
System Performance Monitoring and Trend Analysis CANCELLED
|
|
| |
M-2: Hybrid
Systems: Modeling, Control Design and Applications |
Monday, June 2, 2003
|
-
Manfred Morari; Automatic Control Laboratory, ETH-Zurich,
Switzerland.
-
Stratos Pistikopoulos; Centre for Process Systems Engineering,
Imperial College London, UK.
Application presenters:
- Francesco Borrelli; Automatic Control Lab, ETH-Zurich,
Switzerland
- Vassilis Sakizlis; Centre for Process Systems Engineering,
Imperial College London, UK
- Francis J. Doyle III; Dept. of Chemical Engineering, University
of California, Santa Barbara, CA, USA
- Dr. Vipin Gopal; Honeywell, Minneapolis, USA
- Dr. Eduardo Gallestey; ABB, Corporate Research Ltd, Baden-Daettwil,
Switzerland
- Michael Fodor; Ford Motor Company, Dearborn, MI, USA
- Rainer Mobus; DaimlerChrysler AG, Stuttgart, Germany
|
Hybrid systems theory has matured from being an
exciting research field to the point of delivering results.
The aim of this workshop is twofold: to teach the use of systematic
methods for modeling and controlling systems which exhibit both
continuous and discrete dynamics and to show how such techniques
can be applied to a wide range of industrial applications.
Hybrid systems are heterogeneous systems that exhibit both continuous
and discrete dynamics. Hybrid systems can switch between many
operating modes where each mode is governed by its own characteristic
dynamical laws. Mode transitions are triggered by variables
crossing specific thresholds, by the lapse of certain time periods,
or by external inputs. Over the last few years they have attracted
much interest in industrial and academic circles because of
their wide applicability and the system theoretic challenges
they pose.
In the first part of the workshop we will teach the use of
systematic methods for modeling, control and optimization
of hybrid systems. We will focus on discrete time models and
will show how to recast them into a form that is suitable
for solving analysis and synthesis problems such as stability,
controllability and observability analysis, filter and controller
design, by using standard commercial optimization software.
We will show how to design optimal feedback control laws for
constrained linear dynamic systems and how this technique
can be extended to the general case of hybrid systems.
In the second part of the workshop, participants from various
industries will report how they have used these techniques
on a wide range of problems. These include the control of
combined cycle power plants, electrical drives, biomedical
systems, traction control of automobiles, electronic throttle
control and automatic steering in Driver Assistance systems.
To get an appreciation of the scope of the workshop and to
download relevant references go to http://control.ee.ethz.ch/~hybrid/
|
08:30-09:30 MODELING FRAMEWORK FOR HYBRID SYSTEMS;
THE HYSDEL COMPILER
09:30-10:30 TECHNIQUES AND SOFTWARE FOR MIXED INTEGER OPTIMIZATION
10:45-11:45 MULTI-PARAMETRIC OPTIMIZATION FOR PROBLEMS INVOLVING
CONTINUOUS AND MIXED INTEGER/CONTINUOUS VARIABLES
11:45-12:30 OPTIMAL CONTROL LAW DESIGN FOR CONSTRAINED LINEAR
SYSTEMS VIA PARAMETRIC PROGRAMMING
13:30-14:30 OPTIMAL CONTROL LAW DESIGN FOR CONSTRAINED HYBRID
SYSTEMS VIA PARAMETRIC PROGRAMMING
14:45-17:45 APPLICATIONS |
- Traction control of automobiles (Ford)
- Scheduling and planning of combined cycle power plants
(ABB)
- Direct Torque Control of electrical drives (ABB)
- Driver Assistance systems (Daimler-Chrysler)
- Electronic throttle control (Ford)
- Control of biomedical systems (ETH - Inselspital Bern)
- Polymer grade transition problem (UCSB - Basell Polyolefins)
|
| |
|
| |
M-3: Multiple
Models in Control & Signal Processing CANCELLED
|
|
| |
T-1: Modern
Anti-windup Synthesis
|
Tuesday, June 3, 2003
|
-
Andrew R. Teel, University of California, Santa Barbara,
USA
-
Gene L. Grimm, University of California, Santa Barbara,
USA
-
Luca Zaccarian, University of Rome, Tor Vergata, Italy
|
| The goal of this workshop is to explain and demonstrate
several constructive linear and nonlinear anti-windup techniques.
"Anti-windup synthesis" denotes the design of suitable
filters to be introduced on top of an existing controller for
a plant subject to input saturation, under the assumption that
the controller has been designed to guarantee stability and
desirable performance properties when interconnected to the
plant without input saturation. The two key properties of such
an augmented system are 1) the unconstrained trajectories are
fully preserved as long as the saturation limits are not exceeded;
2) upon activation of the saturation nonlinearity, stability
is retained and the unconstrained performance is recovered as
much as possible by the augmented system.
Within this workshop, modern anti-windup techniques will
be overviewed and useful recipes will be provided for their
application to practical industrial problems, ranging from
the simplest ones that apply to asymptotically stable linear
systems, to the more complicated ones, which apply to nonlinear
systems in general. Handouts will be distributed to the participants
of the workshop
|
SCHEDULE: Tuesday, June 3, 2003
|
08:30-09:15 ANTI-WINDUP: NOTATION AND PRELIMINARIES
Illustration of the windup effect through examples
Formalization of the anti-windup problem
Description of some simple compensation schemes and illustration
of their limits
Introduction to modern constructions
Introduction to LMIs
09:15-10:45 LINEAR ANTI-WINDUP
DESIGN VIA LMIs: THE STATIC CASE
The linear anti-windup scheme for linear plants
External and full-authority anti-windup
Performance measures
Feasibility conditions for static anti-windup design
Static anti-windup algorithms
Examples
11:00 -12:30 LINEAR ANTI-WINDUP DESIGN
VIA LMIs: THE DYNAMIC CASE
Dynamic linear anti-windup schemes for linear plants
Feasibility conditions for general anti-windup
Plant-order anti-windup algorithms
Design for guaranteed robustness
Examples
13:30-15:00 NONLINEAR ANTI-WINDUP
FOR ASYMPTOTICALLY STABLE LINEAR SYSTEMS
Nonlinear anti-windup design via RHOC, continuous-time and discrete-time
cases
Nonlinear anti-windup via scheduled linear compensation
Examples
15:00-16:00 ANTI-WINDUP DESIGN FOR
MARGINALLY UNSTABLE SYSTEMS
Linear anti-windup algorithms for marginally stable plants,
continuous time and discrete time
Examples
Nonlinear anti-windup algorithms for marginally unstable plants,
continuous and discrete time
Examples
16:15-17:15 ANTI-WINDUP DESIGN FOR
EXPONENTIALLY UNSTABLE SYSTEMS
Nonlinear anti-windup design for exponentially unstable plants
Examples
17:15-18:15 ADVANCED ANTI-WINDUP CONSTRUCTIONS
Nonlinear anti-windup design for Euler-Lagrange systems
Application to robotic manipulators
Examples
Anti-windup for nonlinear systems: design guidelines
Anti-windup for systems with rate saturation
|
|
| |
T-2: Stochastic
Search and Optimization
|
Tuesday, June 3, 2003
|
-
James C. Spall; Applied Physics Laboratory, The Johns
Hopkins University
-
Stacy D. Hill; Applied Physics Laboratory, The Johns
Hopkins University
|
| This workshop will review search and optimization
algorithms and analysis techniques that are widely used in the
control systems community. The emphasis will be on general principles,
comparative analysis of algorithm performance, and issues relevant
to practical implementation. Participants will be directed to
the textbook or other appropriate literature for details associated
with the implementation and/or the underlying theory.
This workshop is an introduction to stochastic search and
optimization, as oriented to systems and control problems.
Stochastic search and optimization plays an increasing role
in the analysis and control of modern systems as a way of
coping with inherent system noise and with providing algorithms
that are relatively insensitive to modeling uncertainty. Methods
for stochastic search and optimization are used throughout
virtually all aspects of control theory and practice. To name
a few areas, these include: decision aiding, system identification,
flight control for aircraft, simulation-based optimization
for discrete-event systems, performance analysis of communication
networks, control and scheduling of complex manufacturing
processes, and computer-based personnel training.
This workshop introduces the fundamental issues in stochastic
search and optimization with special emphasis on cases where
classical deterministic techniques (linear and nonlinear programming,
etc.) do not apply. These cases include many important practical
problems, which will be discussed throughout the course (e.g.,
neural network training, simulation-based optimization, target
tracking, nonlinear control, image processing, discrete-event
systems, experimental design, etc.)
|
SCHEDULE: Tuesday, June 3, 2003
|
08:30-09:15 BACKGROUND
Basic issues in search and optimization.
Stochastic vs. deterministic methods.
No Free Lunch Theorems.
Summary of classical methods of optimization and their limitations.
Brief review of necessary multivariate analysis, matrix algebra,
probability, and statistics.
09:15-10:15 DIRECT
SEARCH TECHNIQUE
Introduction to direct random search.
Monte Carlo methods.
Nonlinear simplex (Nelder-Mead algorithms).
10:30-12:00
LEAST-SQUARES-TYPE METHODS
Recursive methods for linear systems.
Recursive least squares (RLS).
Least mean squares (LMS).
13:15-15:00 STOCHASTIC APPROXIMATION FOR LINEAR AND NONLINEAR
SYSTEMS
Root-finding and gradient-based stochastic approximation (Robbins-Monro).
Gradient-free stochastic approximation: finite-difference (FDSA)
and simultaneous perturbation (SPSA) methods.
15:15-16:45 METHODS MOTIVATED BY PHYSICAL PROCESSES
Simulated annealing and related methods.
Evolutionary computation and genetic algorithms.
16:45-17:00 WRAP-UP AND DISCUSSION
|
|
| |
T-3: Automated
Multivariable System Identification: Basic Principles with
Control and Monitoring Applications |
Tuesday, June 3, 2003
|
-
Wallace E. Larimore; Adaptics, Inc.
-
Dale E. Seborg; University of California, Santa Barbara
|
| Over the past several years, computational methods
and software have been developed to reliably identify system
dynamics from input/output data with optimal statistical accuracy.
These automatic methods apply to a very general class of linear
systems including multi-input/multi-output, with state and measurement
noise disturbances, unknown feedback, unknown state order, and
possibly unstable or highly resonant dynamics. However, existing
methods for high accuracy identification such as Box/Jenkins
and prediction error methods are problematic in that they are
both computationally unreliable and involve a tedious toolbox
approach requiring graduate level training. The automatic methods
presented in this workshop are fundamentally different and involve
direct determination of the system states, i.e., system rank,
using stable singular value decomposition (SVD) computations
and optimal rank selection based on canonical variate analysis
(CVA). CVA is related to partial least squares (PLS), principal
component analysis (PCA), and subspace system identification
methods.
The concepts are presented in a direct first principles way
that is appropriate for advanced undergraduate and graduate
curriculum so that automated system identification can be
made much more accessible to those in most need of using it.
This advance in system identification has major implications
for analysis, system monitoring, and design and implementation
of control systems for many applications including aerospace,
automotive, and chemical and industrial processes. For process
monitoring, CVA provides powerful new methods for the detection
and analysis of changes in colinear multivariable processes.
For control system implementation, CVA provides high accuracy
multivariable models even in the presence of unknown state
feedback. CVA provides a starting point for robust control
design since accuracy confidence bands on the identified model
are automatically provided. Since the computation is completely
automatic, it is ideal for online and adaptive control applications.
|
SCHEDULE: Tuesday, June 3, 2003
|
08:30-09:00 OVERVIEW OF SYTSEM IDENTIFICATION
METHODS
09:00-10:00 MODEL ORDER AND STATE SELETION USING STATISTICAL
CRITERIA
10:00-11:00 MODEL PARAMETER ESTIMATION AND FILTERING
11:00-12:00 COMPARISON OF ALTERNATE SYSTEM IDENTIFICATION
APPROACHES
13:30-14:30 IDENTIFICATION WITH UNKNOWN DELAYS AND FEEDBACK
14:30-15:30 PROCESS MONITORING USING CVA
15:30-16:30 PROCESSING MONITORING APPLICATIONS
16:30-17:30 IDENTIFICATION AND CONTROL APPLICATIONS
|
|
| |
|
|
|