Dynamics and Controls Seminars

Abstracts 2003 -2004

* * * * * * * * * * * * *
Friday, May 28, 2004
10:00 a.m.
479 EBU-II

Shawn Liu
LabVIEW Real-Time Product Manager
National Instruments

LabVIEW Graphical Development: Tools for Interactive Controls Education and Research”

Come and see the new LabVIEW tools for system identification, control design, and simulation.  During this 2 hour seminar, we will describe the new LabVIEW Simulation Module, Control Design Toolkit, and System Identification Toolkit.  Learn how these new tools can be used with existing NI hardware and software such as NI data acquisition hardware, LabVIEW Real-Time, and LabVIEW FPGA.

 

* * * * * * * * * * * * *

Wednesday, May 19, 2004
11:00 a.m.
479 EBU-II

Jason L. Speyer
Mechanical and
Aerospace Engineering Department
University
of California, Los Angeles

“Characterization of LQG Differential Games with Different Information Patterns”

A discrete-time linear-quadratic-Gaussian (LQG) differential game is considered where the players have access to separate measurement histories. The particular problem that is solved is where one adversary has access to only noisy partial information of the state while the other makes a perfect measurement of the state vector. The system dynamics are assumed linear with additive process noise.  The solutions show a significant departure from previously published results. First, process noise is included in the dynamical system and a quadratic weighting in the state is included in the cost criterion. Secondly, the optimal strategies of both players are shown to be finite dimensional, not infinite dimensional as was originally thought. There is, therefore, no reason for the player with perfect measurement to have additional information to reconstruct the other player’s measurement. Thirdly, it is assumed that the perfect-measurement adversary's control matrix is in the range space of the other adversary's measurement matrix. Then, by a limit of the linear-exponential-Gaussian game solution to the LQG game solution, it is seen that the partial information player avoids reproducing an estimated version of his adversary's strategy. A key feature of this problem solution is the filter structure of the player with partial measurements, which prevents the player with perfect information from winning the game by injecting infinite white noise into the system.  Finally, the notion of a saddle point for deterministic games is extended to a saddle interval for games with uncertainty.

 

* * * * * * * * * * * * *

 

Wednesday, April 21, 2004
10:00 a.m.
479 EBU-II

Faryar Jabbari
Department of Mechanical and Aerospace Engineering
University of California, Irvine

“Disturbance Attenuation with Bounded Actuators:
Scheduled Output Feedback Controllers”

Actuator saturation is one of the oldest problems in control, which is -- again -- the subject of much attention. The motivating application for this work is earthquake engineering: a wide range of disturbance magnitudes that can cause damage; small signal regions that are inconsequential; inherent actuator limitations and the need for performance measures (since the structure is often open-loop stable) to evaluate the benefits of the control system.


A new class of output feedback controllers, for disturbance attenuation in linear systems with bounded actuators, will be discussed. The emphasis is on obtaining performance guarantees, both in small and large signal regions. The controller has a relatively generic structure and is scheduled according to the proximity of the compensator states to the origin, so that smaller states lead to larger gains, hence higher performance controllers.


The main results establish a scheduled controller with a parameter-dependent performance measure, with straight forward extensions to problems with rate limits, different performance measures and systems with certain class of nonlinearities (e.g., lpv cases). Linear splines are used to obtain solutions through standard LMI software. Examples --including an open-loop unstable system-- highlight the application of the results.

 

* * * * * * * * * * * * *

 

Friday, April 2, 2004
11:00 a.m.
479 EBU-II

Alison L. Marsden
Mechanical Engineering
Stanford University


“Suppression of vortex-shedding noise via derivative-free shape optimization”

The application of optimization methods to physically realistic fluid mechanics problems presents a number of challenges because of the computational cost involved in doing accurate simulations and the difficulty of obtaining gradient information.  In this work, derivative-free shape optimization is applied to minimize aerodynamic noise in the flow over an airfoil trailing-edge.  Reduction of trailing-edge noise is relevant to a number of engineering applications including airframe noise reduction and wind turbine, hydrofoil and rotorblade design.  Aeroacoustics problems such as this necessitate the use of modern computational techniques such as large-eddy simulation (LES) in order to capture a wide range of turbulence scales which are the source of broadband noise.  Substantial effort has been invested in development of accurate methods for aeroacoustic computations in recent years.  For these methods to be applied to engineering problems of greater interest, the focus must now move from simulation development to control applications.

 

In this work, a tailored version of the surrogate management framework (SMF) (Booker et al., 1999) has been implemented to optimize noise reduction for an airfoil trailing-edge using several shape parameters with constraints.  The SMF method provides a robust and efficient alternative to gradient-based methods.  Using SMF, design space exploration is performed not with the expensive actual function but with an inexpensive surrogate function. The use of a polling step in the SMF guarantees convergence to a local minimum of the cost function on a mesh in the parameter space.  In the trailing-edge problem, constraints on lift and drag are enforced using the filter method of Audet and Dennis (2000). Within the framework of this method, a penalty function is systematically added to the surrogate model to aid in searching the design space.  Using this method, several interesting and unexpected optimal shapes have been identified, all of which resulted in significant reduction of trailing-edge noise (as much as 80%).  These shapes have provided motivation to study the physics of the flow, and in particular, the trade-off between noise reduction and loss of lift.

 

The results of this study demonstrate the successful application of shape optimization to a time-dependent complex flow problem, and validate the use of a novel adaptation of the SMF method with constraints.  The application of these cutting-edge optimization methods allows for optimization of a wide class of fluid mechanics design problems, including complex geometries, turbulent flows and unsteadiness.  Because of the portability of the SMF method, it can be coupled to turbulent flow solvers based on LES or unsteady RANS for high Reynolds number flows in future applications.

 

* * * * * * * * * * * * *

 

Thursday, March 18, 2004
11:00 a.m.
479 EBU-II

Izhak Bucher
Mechanical Engineering
Technion, Israel Institute of Technology
Haifa, Israel

“Tailoring the Dynamics of a Miniature (MEMS) Scanning Mirror”

In this talk, a new structure synthesis method will be presented for the dynamics of a MEMS scanning mirror. With this approach, the required operating forces to create a non-sinusoidal periodic motion can be reduced by 2-3 orders of magnitude. It will be shown that a vibrating structure can be tuned, in open loop to generate a triangular wave motion using a force that is several orders of magnitude smaller than a non-optimal closed loop controlled structure. The ability to create complicated geometries using standard MEMS manufacturing procedures, allows designers to embrace previously neglected inverse vibration theories.

 

Specifically, a new multi-degrees of freedom structure is designed to create a scanning mirror with which a direct on-retina, raster-scan video display can be developed. The ability to tailor the periodic motion of the scanning mirror to the optical requirements maximizes the uniformity of intensity distribution in a way that is currently not possible with electronic means.  A control scheme consisting of a phase lagging controller and suitable pre-filter are shown to complement the tailored structure in an optimal manner.

 

The links of the proposed approach to inverse vibration theory and to system and control theories will be described. A laboratory macro- as well as micro-sized demonstrators will be presented and discussed.  During the talk, the interdisciplinary nature of the presented device will be highlighted by describing the adaptive signal processing, finite-element modeling, topology optimization and the experimental procedure.

 

* * * * * * * * * * * * *

 

Tuesday, March 2, 2004
3:30 p.m.
479 EBU-II

Mark J. Balas
AES Department
University
of Colorado

“Dynamics and Control of Modern Horizontal Axis Wind Turbines”

Next generation wind turbines for electric power generation are larger and more flexible than past designs to promote greater energy capture and reduce the cost of energy.  It is here that the implementation of active feedback control is crucial to meet design objectives.  Not only must there be power regulation or optimization but also load mitigation to extend the life of the turbine.

 

Wind turbine dynamics modeling for active control is composed of three principal parts:

  • Aerodynamics and Inflow Behavior
  • Structural Dynamics
  • Feedback Control Algorithms

(A fourth area Power Electronics is dynamically so fast compared to the other three that it need not be considered in this discussion, but may be designed separately.)

 

Active control is a relatively new technology for wind turbines. Basic control theory used is periodic control and disturbance accommodation, but with a strong connection to the aerodynamics and structural dynamics of wind turbines.

 

This talk will be an overview of the fundamentals of wind turbine dynamics and control.

 

* * * * * * * * * * * * *

Friday, February 27, 2004
10:30 a.m.
479 EBU-II

Jeff S. Shamma
Mechanical and
Aerospace Engineering
University
of California, Los Angeles

Multiagent Repeated Games and Convergence to Nash Equilibria

Consider a scenario in which multiple decision making agents repeatedly play a matrix game and adjust their strategies according to observations of each other's actions. The game is noncooperative in that each agent may have its own objective/utility function, and these objectives are not shared among agents. A central issue is whether agent strategies will converge to a Nash equilibrium. Prior work shows how convergence to a Nash equilibrium in this setting may or may not occur. This talk presents new strategic update mechanisms that can lead to convergent behavior in previously nonconvergent cases, such as the Shapley and Jordan counterexamples, through the use of fundamental feedback control concepts.

 

 

* * * * * * * * * * * * *

 

Monday, February 23, 2004
11:00 a.m.
584 EBU-II

W. Paul Malcolm
Haskayne School
of Business
University
of Calgary
School
of Applied Mathematics
University
of Adelaide, Australia

“A General Smoothing Theorem for Discrete-Time Gauss-Markov Jump Linear Systems”

In this seminar we extend the new state and mode estimation algorithms computed by Professors Robert J Elliott and Francois Dufour.  The algorithm developed by Elliott and Dufour is distinct from extant methods, such as the so called Interacting Multiple Model algorithm (IMM) and sequential Monte Carlo methods, in that it is precise; that is, their algorithm is based upon well defined approximations of the "exact" hybrid filter.

 

To compute a smoothing algorithm, we exploit a duality between forwards and backwards (dual) dynamics.  The natural framework to exploit this duality is the method of reference probability, whereby one chooses to work under a new, or 'reference' probability measure. Under this new measure, both the state process and the observation process are independently and identically distributed with Gaussian statistics, however, the Markov chain, whose state value fully determines the system dynamics, remains unchanged.  Since our smoothing algorithm depends upon the filter, we start by giving a review of jump Markov system filter developed by Elliott and Dufour. This filter has been shown to significantly out perform the IMM in conventional object tracking scenarios and the more challenging bearings only tracking problem.  The new smoothing algorithm we compute is general and can be readily configured into the standard special cases of fixed point, fixed lag and fixed interval smoothers.

 

* * * * * * * * * * * * *

Friday, February 13, 2004
2:00 p.m.
584 EBU-II

Frank Allgower
Institute for Systems Theory in Engineering
University of Stuttgart,
Germany
(Currently on sabbatical leave at UCSB)

“Nonlinear Model Predictive Control: From Theory to Applications”

In the past decade model predictive control (MPC), also referred to as receding horizon control, or moving horizon control, has become a preferred control strategy for a large number of industrial processes. The main reasons for this popularity include the ability to explicitly handle constraints and to consider multivariable processes with potentially many manipulated and controlled variables. In this presentation we will give an overview over the area of model predictive control with special emphasis on nonlinear model predictive control. After a brief discussion of the history and impact of MPC we will discuss recent results regarding system theoretic properties like stability, robustness, output feedback and performance of the closed loop.  With a number of applications we will demonstrate that by using specially tailored optimization methods even large problems, having hundreds of states, can be controlled efficiently using NMPC methods.

 

* * * * * * * * * * * * *

 

Friday, February 13, 2004
10:30 a.m.
479 EBU-II

Kenny Breuer
Division of
Engineering
Brown University


“Challenges, Successes and Failures in the Control of Turbulence”

We will present some recent results from experiments on the control of turbulent shear flows using two approaches - (i) electromagnetic forcing and (ii) feed forward control based on shear stress measured at the wall. Both approaches have been demonstrated to reduce the intensity of turbulent fluctuations and the turbulent skin friction, although the cost of implementing the control far exceeds the power saved.  Details on the changes in the flow fields will be presented.  Issues regarding the challenges with the physical implementation of turbulence control systems will be addressed as well as well as some of goals and interim results of new experiments currently underway.

 

* * * * * * * * * * * * *

 

Tuesday, February 3, 2004
10:30 a.m.
479 EBU-II

Dr. Ann Marie Sastry
Department of Mechanical Engineering
Department of Biomedical Engineering
The
University of Michigan

“Cluster, Connections and Transport: Use of Percolation Concepts in Modeling Intracellular Ion Migration”

 

The exact connections between initiators of events within cells, cascades of chemical reactions, and human disease, including such physical events as stroke, cardiovascular disease, endocrine disease, and abnormal cell growth, are currently unknown. While there are established correlations between these signaling events and clinical outcomes, the tools required to determine causal mechanisms are not yet available. And, further identification of the ionic initiators of important cellular phenomena is unlikely to yield greater understanding, absent realistic modeling of the complex, stochastic interactions among the key signaling elements of the cell: divalent cations, including calcium and zinc. Ultimately, use of nanoscale, intracellular sensors, combined with intensive modeling of their location, affinity, and probable detection rates, will afford this needed understanding.

 

We are using percolation concepts, in conjunction with models for delivery of sensors to cells, to accomplish this. Traditionally, work in understanding intracellular cation transport has involved wholly continuum descriptions, both experimentally and in modeling. These were needed to first understand the approximate concentrations of the key ionic species for signaling in the cell. Recently, however, we have seen development of the ability to track single ion binding in the cell through optical sensing, and the simultaneous ability to efficiently model very large numbers of both particles and their interactions, through Monte Carlo methods. Indeed, the cytosol is comprised of a fine network of filamentary structures, which form a complex landscape for ions and the proteins that ferry them into and out of the cell. The presence of these filamentary, intracellular “highways,” which comprise cytoskeletal structure, dramatically alters local transport rates of key divalent cations. Tracking important ionic signaling agents through physiological and disease states using sensors requires detailed statistical modeling not only of sensor deployment and transport through this complex domain, but also their cluster statistics.

 

Protein-based sensors, however, are of complex shape. And, the mathematical toolkit of percolation solutions, in development since the early part of the last century, has contained no analytic approximations for noncircular or nonspherical particles, until recently. We have extended previous analytic solutions for circles and spheres in order to determine percolation points in fields of generalized ellipses and ellipsoids, which readily approximate a very wide class of particle shapes, including fibers. We have also studied 2D versus 3D percolation, to assess the importance of particle placement in thin domains, and compared analytic approximations of percolation of particles in a finite or infinite field with Monte Carlo simulations. In this talk, solution of these problems and their applications will be discussed, and simulations of delivery of sensors to cells, including models for clustering of sensors around targeted areas within the cells, will be presented. Our ultimate aim is to quantify the low-frequency biometal binding so critical in cell signaling, and apply these models to various physiological and disease states in mammalian cells.

 

* * * * * * * * * * * * *

Thursday, December 4, 2003
11:00 a.m.
479 EBU-II

Prof. Michael J. Grimble
Industrial Control Centre
University of Strathclyde
Glasgow, Scotland

“Nonlinear Predictive Control for Industrial Applications”

 

There are no commercial predictive controllers for non-linear systems that handle nonlinearities in a rigorous theoretical manner. There are empirical fixes to deal with nonlinear behavior but this suggests significant improvements may be possible in say process control by using improved solutions. A new approach to nonlinear predictive control will be introduced based upon an extension of the linear system techniques of LQGPC and GPC design. The new NLQGPC and NGPC methods are based upon so-called state dependent state-space models for nonlinear processes.  This approach is rather different to the more common linearization around a desired trajectory design methods.  Stability is ensured by introducing a "satisficing" type of projection.  A range of different industrial and aerospace applications will be presented.

 

* * * * * * * * * * * * *

Wednesday, December 3, 2003
11:00 a.m.
479 EBU-II

Dr. Peter Dower
Department of Electrical & Electronic Engineering
The
University of Melbourne

“Analysis of input to state stability for discrete time nonlinear systems via dynamic programming”

 

This seminar presents novel analysis results for input to state stability (ISS) that utilize dynamic programming techniques to characterize minimal ISS gains and transient bounds. These characterizations naturally lead to computable necessary and sufficient conditions for ISS. Our results make a natural connection between ISS and optimization problems in nonlinear dissipative systems theory, including L-2 gain analysis and nonlinear H-infinity theory.

 

* * * * * * * * * * * * *

Tuesday, November 18, 2003
11:00 a.m.
479 EBU-II

Michel Gevers
Center
for Systems Engineering and Applied Mechanics
(CESAME)

Iterative Feedback Tuning: from the bare essentials to some recent developments

 

Iterative Feedback Tuning (IFT) is by now a classical model-free controller tuning method, since it was developed as far back as 1994.  However, the methodology is still undergoing developments and extensions. This talk will present the basics of the methodology, some industrial applications, and some extensions and recent developments. In particular the talk will focus on the use of IFT for the tuning of controllers with minimum settling time.

 

* * * * * * * * * * * * *

Wednesday, November 12, 2003
11:00 a.m.
584 EBU-II

Michael Overton
Courant Institute of Mathematical
Sciences
New York University
, New York
, NY

Optimizing Stability and Controllability via Pseudospectra

 

A matrix is stable if its eigenvalues are in the left half of the complex plane.  More practical stability measures include the pseudospectral abscissa (maximum real part of the pseudospectrum) and the distance to instability (minimum norm perturbation required to make a stable matrix unstable).  Likewise, the classical definition of controllability is not as useful as a measure of the distance to uncontrollability.

 

Matrices often arise in applications as parameter dependent.  Optimization of stability or controllability measures over parameters is challenging because the objective functions are nonsmooth and nonconvex.  We solve such optimization problems, locally at least, via a novel method based on gradient sampling.  One of our stability optimization examples is a difficult problem from the control literature: finding stable low-order controllers for a model of a Boeing 767 at a flutter condition. We also give a controllability optimization example and explain its connection with an interesting open question: how many connected components are possible for pseudospectra of rectangular matrices?

 

Joint work with James V. Burke, University of Washington, Seattle, WA and Adrian S. Lewis, Simon Fraser University, Burnaby, BC, Canada.

 

 

* * * * * * * * * * * * *

Thursday, October 30, 2003
10:00 a.m.
479 EBU-II

Dr. D. Hrovat
Corporate Technical Specialist
Ford Research Laboratory
Dearborn, Michigan

“Automotive Control Systems and Related Developments”

The talk will summarize some of the developments in automotive dynamic systems modeling and control, with an emphasis on advanced and R&D efforts in vehicle dynamics and control at Ford, and especially at the Ford Research Laboratory in Dearborn.  This will include the following:

 

  • Design and evaluation of (advanced) Traction Control;
  • Active safety and Interactive Vehicle Dynamics and Control: present efforts and future potential developments and enhancements including industry-first Roll Stability Control, (first implemented in the new Volvo XC-90 SUV, which received numerous awards, including the prestigious 2003 Motor Trend SUV-of-the-Year award);
  • Tools and methods for evaluation of various (advanced) vehicle control / active safety concepts including the autonomous steering robot driver vehicle setup.

 

As time permits, the presentation will also give a brief overview of a number of other automotive control applications and important related activities.

 

* * * * * * * * * * * * *

Friday, October 17, 2003
10:30 a.m.
479 EBU-II

I. D. Landau
Laboratoire d’Automatique de Grenoble
France

“Adaptive rejection of unknown disturbances - 
Application to active vibration control”

 

The basic approaches for the rejection of unknown disturbances will be reviewed in the first part of the talk.  The methodology for feedback adaptive control of active vibration systems in the presence of unknown narrow band disturbances will be then discussed. A direct adaptive control scheme based on the internal model principle and the use of the Youla-Kucera parameterization is proposed. This approach is comparatively evaluated with respect to an indirect adaptive control scheme based on the estimation of the disturbance model. The comparative evaluation of the two approaches is done in real time on an active suspension system.

 

The slides for the talk can be downloaded from the following web site page:

http://www-lag.ensieg.inpg.fr/landau/bookIC/index_FR_Transparents_Tutorials.htm

 

 

* * * * * * * * * * * * *

For information: Sophia Bligh at (858) 822-1269