2000 Events:

November 16, 2000: "Distributed Control of Air Traffic Systems" by Professor Claire Tomlin, Departments of Aeronautics and Astronautics and Electrical Engineering, Stanford University

Abstract: Today, with air traffic rates in the United States growing by 5% annually, and those across the Pacific Rim by more than 15% annually, the problem which poses the greatest challenge to air transportation in the United States and abroad is the design of a safe and efficient method to control this system of aircraft. Today's air traffic control system cannot even handle current traffic demands: ground holds and airborne delays in flights due to congestion in the skies have become so common that airlines automatically pad their flight times with built-in delays.

In this talk, I will discuss a proposed solution, called `free flight': in which each aircraft flies along an optimal route which ensures that the flight time is short, the fuel consumption is minimized, and inclement weather is avoided, and at the same time will maintain safe distance from other aircraft. I will present a model for such a system, and a method which incorporates dynamic game theory with computer aided verification techniques, to design control laws which guarantee the safety of the system. Finally, I will show how my students and I have applied this method to design control laws for the automated resolution of trajectory conflicts between multiple aircraft, and to the design of provably safe maneuvers for closely spaced parallel approaches.

bio: Claire Tomlin received the Ph.D. degree in Electrical Engineering from the University of California, Berkeley, in 1998. Since September 1998 she has been an Assistant Professor in the Department of Aeronautics and Astronautics at Stanford University, with a courtesy appointment in Electrical Engineering. She was a graduate fellow in the Division of Applied Sciences at Harvard University in 1994, and she has been a visiting researcher at NASA Ames Research Center during 1994-1998, at Honeywell Technology Center in 1997, and at the University of British Columbia in 1994. Claire Tomlin is a recipient of the Terman Fellowship, Stanford (1998), the Bernard Friedman Memorial Prize in Applied Mathematics, Berkeley (1998), the Zonta Amelia Earhart Awards for Aeronautics Research (1996-98), and the Natural Sciences and Engineering Research Council of Canada 1967 Scholarship (1992). Her research interests are in hybrid systems, nonlinear control systems, air traffic management, and flight vehicle dynamics and control.

Topical Papers in PDF format:
Computing Provably Safe Aircraft to Aircraft Spacing for Closely Spaced Parallel Approaches
A Game Theoretic Approach to Controller Design for Hybrid Systems
Level Set Methods for Computation in Hybrid Systems

September 21, 2000: "Flight Control System Design and Test for Unmanned Rotorcraft" by Mr. Chad R. Frost, NASA Ames Research Center

Abstract: Mr. Frost will discuss flight control system design methodologies for Unmanned Aerial Vehicles (UAVs), illustrated using flight test results from a recent helicopter development program. The presentation will focus on the techniques and advanced design tools used for aircraft stabilization and control, and explore their applicability to other areas of control system design.

bio: Mr. Frost is an Aerospace Engineer in the Army/NASA Rotorcraft Division's Flight Control and Cockpit Integration Branch, at NASA Ames Research Center, Moffett Field, California. His areas of research include design of flight control systems for manned and unmanned rotorcraft, control system optimization, flight vehicle modeling, and simulation.

Prior to joining NASA, Mr. Frost was Program Manager for Spacecraft Structures at Applied Aerospace Structures Corp., and at Preece Inc. as Senior Project Engineer. He received his B.S. and M.S. degrees in aeronautical engineering from California Polytechnic State University, San Luis Obispo, where he worked on the DaVinci human-powered helicopter, the first such vehicle to achieve hover.

June 8, 2000: "Digital Control Systems Design: Limitations and Optimization" by Dr. Wei-Min Lu, Symmetry Communications Systems

Abstract: From a layered point of view on control system design, even though the system-level design on the higher layers is crucial to the support of higher quality of service or performance of the overall systems, the component-level design (physical layer) is still very important in some applications, in particular if the component-level performance is the bottleneck of the overall system performance. This is exactly the case in computer hard disk drive (HDD) data storage systems.

In this talk, we will consider digital control system design as part of the component-level system design. Our goal is to examine the limitations of the digital system design to reduce the impact of system uncertainty on system performance and develop design techniques to derive optimal solutions. We will use HDD servo control system design as an example throughout the discussion. We will particularly examine the feedback and feedforward approaches widely used in the disturbance attenuation for practical digital control system design. We will formulate optimal digital design problem with probabilistic noise description equivalently as deterministic H_infinity control problems, and develop algorithms to obtain suboptimal solutions.

bio: Wei-Min Lu (S'88 - M'90) received the B.Sc. degree with honors in engineering from Tsinghua University, Beijing, China in 1986, and the M.Sc. and Ph.D. degrees in electrical engineering and mathematics from California Institute of Technology, Pasadena, CA, in 1990 and 1995, respectively.

He is now with Symmetry Communications in San Jose, CA as a Senior Member of Technical Staff responsible for system architecture and performance for wireless data communications. Before he joined Symmetry, he was with IBM Advanced Magnetic Recording Laboratory as an advisory engineer to work on performance analysis and control architecture design for computer data storage systems. He also previously worked at NASA Jet Propulsion Laboratory as a Member of Technical Staff responsible for the development of guidance and control techniques for space vehicles. His technical interests include computer networks and wireless data communications, as well as theory, computation, and industrial application of control sciences.

May 11, 2000: "Control of Systems with Large Spatially Distributed Actuator and Sensor Arrays" by Dimitry Gorinevsky, Ph.D., Honeywell Technology Center

Abstract: The talk will discuss control architectures, applications, and algorithms associated with spatially distributed systems. In such systems, control and measurement arrays are used to influence processes evolving in time and in space. Examples include distributed flow control, pressure and temperature uniformity control in jet engines, web manufacturing processes, active wavefront control in adaptive reflectors, and iterative control of batch processes. New emerging applications are related to multi-zone thermal processing, MEMS, smart structures and distributed control of vibrations and acoustic fields. Development of practical control approaches for such systems is focused on localized control implemented with distributed computing elements. Control design and analysis for array systems requires considering dynamics stability and control performance in time and in spatial coordinates. The spatial dynamics are "non-causal" and defy "normal" controls design intuition. The presentation will attempt to give the listeners the flavor of the developments and approaches in this area without being too technical.

bio: Dr. Dimitry Gorinevsky is a Senior Staff Scientist in Controls and Navigation Department at Honeywell Technology Center (HTC) working in Cupertino, CA. He received his M.S. from Moscow Institute of Physics and Technology (¡®Phystech¡¯), Ph.D. from Moscow State University and held research and industrial positions in Moscow, Munich, Toronto, and Vancouver. Before joining HTC in 1999, he was with Honeywell-Measurex, a consultant for the Canadian Space Agency, and an Adjunct Professor at the University of British Columbia. At Honeywell-Measurex, Dr. Gorinevsky led the development of advanced control technology for industrial paper machine systems with hundreds of spatially distributed measurements and actuators. His other industrial and application experience and interests include process control and monitoring, spacecraft control, robotics, biomechanics, engine control, aerodynamical flow control, turbine systems, adaptive optics, and optical communication networks. He has authored a book published in English and Russian, more than 100 reviewed technical papers and book chapters and has several patents. He is currently active in IEEE CSS committees and boards.

April 13, 2000: "Control Techniques and Challenges for Hard Disk Drives" by Dr. Matthew T. White, IBM Almaden Research Center

Abstract: The area density of hard disk drives (HDDs) is increasing at the astonishing rate of 100% per year, while the price per megabyte drops about 50% per year. The task of the HDD servo engineer is to provide increasingly fast and accurate control systems to support reading and writing of these shrinking bits while maintaining competitive costs. The effects of these industrial trends will be discussed, as well as the differences in performance required by the high-end server, desktop, mobile, and audio-visual markets.

An overview of the HDD mechanics and electronics will be presented. The typical HDD controller uses mode switching, with one mode optimized for maintaining the position of the head over a data track while reading or writing and another optimized to move between tracks as quickly as possible. The position signal used for feedback control is generated from information encoded on the disk, and tradeoffs between sampling frequency, performance and loss of disk space for customer information must be considered when designing the servo system.

The challenges of the HDD control system have made them an active area for industrial and academic research. A review of recent work will be discussed, including the use of a secondary MEMs or piezo actuator to augment the single voice coil motor currently used for positioning of the read/write head.

bio: Matthew White received a B.S. from Michigan State University in 1990 and a Ph.D. from the University of California Berkeley in 1997, both in mechanical engineering. His Ph.D. dissertation focused on control techniques for increased disturbance rejection in HDDs. Since graduation from UC Berkeley, he has worked in the Storage Systems and Technology group at IBM's Almaden Research Center in San Jose on a variety of servo and mechanics projects for data storage devices.

March 9, 2000: "Computational Tools for Nonlinear Control" by Drs. P. K. Menon, V. H. L. Cheng, and L. S. Crawford, Optimal Synthesis, Inc.

Abstract: Most real-world control problems are nonlinear. Control system development for these problems have traditionally been based on Taylor series linearized system dynamics in conjunction with linear control techniques. Since the system dynamics will behave differently in different parts of the state space, several controllers will have to be designed and ¨¬scheduled? with respect to the operating conditions to yield acceptable control system performance. In many situations, coming up with a satisfactory controller schedule can consume far more human resources than the linearization-linear system design tasks. Moreover, the stability of the resulting system cannot be guaranteed.

Over the past two decades, several nonlinear control system design methods have emerged in the literature. Some of these methods have reached a sufficient level of maturity to permit their application in real-world problems. Nonlinear control methods use the system dynamics to develop their structure. No linearization or gain scheduling is involved in their implementation. These aspects free up the designer to focus on the control system design aspects of the problem, enabling the development of high-performance controllers.

This talk will discuss practical nonlinear control methodologies, and will present the development of computational tools that enable transparent implementation of these methods. Illustrative examples will be presented. The presentation will conclude with a software demo.

bio: Dr. Menon has been involved in the development of nonlinear control systems for aircraft, rotorcraft, missiles, launch vehicles, spacecraft, medical devices and robots. In the past, he has received support in these research areas from the NASA, Navy, Air Force, Army and the industry. He has published extensively in professional journals and has presented papers at various national and international conferences.

Dr. Menon has been the president and chief scientist at Optimal Synthesis Inc since 1992. His experience includes 16 years as a research scientist in the aerospace industry, 8 years at universities as a faculty member and 3 years with NASA as a visiting scientist. He has taught advanced courses in Control, Flight Vehicle Dynamics, Guidance and Control, Machine Vision, and Digital Signal Processing. He has directed Ph. D dissertations and numerous graduate Projects. Dr. Menon has lead short courses on automatic control at national conferences.

Dr. Menon is a member of the IEEE, AHS, Sigma Gamma Tau, and is an Associate Fellow of the AIAA. He is the recipient of research awards from NASA and the IEEE. Dr. Menon is a reviewer for over 10 archival journals on automatic control and signal processing, and has served as an Associate Editor of the AIAA Journal of Guidance, Control and Dynamics.

** Dr. V.H.L. Cheng and Dr. L.S. Crawford will Assist Dr. Menon in this presentation. **

February 10, 2000: "Helicopter Flight Control and Handling Qualities" by Dr. Jeff Schroeder, NASA Ames Research Center

Abstract: Helicopters are described by some as mechanical nightmares. Even those offering a more generous description will approach their control with great caution. If a pilot has modest objectives and flies in clear weather with good visibility, he or she can supply their own stability and control with sufficient training and experience. However, when strict handling qualities requirements are imposed in inclement weather, pilots must rely on a flight control system to accomplish their objectives satisfactorily.

For the helicopter control designer and analyst, the plant presents some challenges. Near hover, helicopters are typically unstable and sometimes have inherent dynamics that result in several large outputs for each input. Unfortunately, developing a feasible control scheme for this multi-input/multi-output problem is complicated by inadequacies of the analytical model. Some responses are even predicted to have the wrong mathematical sign. Finally, the spinning rotor adds several modes that can be destabilized by the unwary. This presentation discusses the above issues and common practices associated with rotorcraft flight control.

bio: Jeff Schroeder received a B.S. and a M.S. from Purdue University and a Ph.D. from Stanford University in Aeronautics and Astronautics. Currently, he is the Deputy Chief of the Flight Control and Cockpit Integration Branch at NASA Ames, and he also serves on the part-time faculty at San Jose State University. His research interests include simulator motion fidelity requirements, handling qualities, and pilot-vehicle-display design.

January 13, 2000: "Adaptive Inverse Control" by Prof. Bernard Widrow, Dept. of Electrical Engineering, Stanford University

Abstract: At present, the control of a dynamic system (the "plant") is generally done by means of feedback. Dr. Widrow proposes in his latest book "Adaptive Inverse Control" an alternative approach that uses adaptive filtering to achieve feedforward control. Precision is attained because of the feedback incorporated in the adaptive filtering.

The control of the plant dynamic response is treated separately, without compromise, from the optimal control of plant disturbance. All of the required operations are based on adaptive filtering techniques. Following the proposed methodology, knowledge of adaptive signal processing allows one to go deeply into the field of adaptive control.

In order for the adaptive inverse control to work, the plant must be stable. If the plant is not stable, then conventional feedback methods should be used to stabilize it. Generally, the form of feedback is not critical and would not need to be optimized. If the plant is stable to begin with, no feedback would be required.

Adaptive inverse control places an adaptive filter whose transfer function converges to the inverse or reciprocal of that plant in cascade with it. If the plant is minimum-phase, an inverse is easily obtained. If the plant is nonminimum-phase, a delayed inverse can be obtained. The delay in the inverse results in a delay in overall system response, but this inevitable with a nonminmum-phase plant. The basic idea can be used to implement "model-reference control" by adapting the cascaded filter to cause the overall system response to match a preselected model response.

Disturbance in a minimum-phase or nonminimum-phase plant, can be optimally controlled by a special circuit that obtains the disturbance of the plant output, filters it, and feeds it back into the plant input. The circuit works in such a way that the feedback does not alter the dynamic plant response. So disturbance control and control of dynamic response can be accomplished independently. The same ideas work for MIMO systems as well as SISO systems.

Control of nonlinear plants is an important subject that raises significant issues. Since a nonlinear plant does not have a transfer function, how would it have an inverse? By cascading a nonlinear adaptive filter based on neural networks with the nonlinear plant, the filter can learn to drive the plant as if it were the plant's inverse. This works surprisingly well for a range of training and operating signals. Control of dynamic response and plant disturbance can be done. Examples and demonstrations will be presented.

These methods show a great promise, but much more work needs to be done to characterize system responses and to establish optimality of disturbance control. This is very much an open area for research.

bio: Dr. Bernard Widrow is a Professor of Electrical Engineering at Stanford University. His fields of teaching and research are signal processing, neural networks, pattern recognition, adaptive filtering, and adaptive control systems. Before coming to Stanford in 1959, he taught at MIT. He did undergraduate and graduate training there, and received the Doctor of Science Degree at MIT in 1956.

Dr. Widrow is author of two books: "Adaptive Signal Processing", and "Adaptive Inverse Control", both published by Prentice-Hall. A third book, entitled "Quantization Noise" will soon be completed. Each is the first kind, established new fields of research and engineering that are being pursued worldwide by students, faculty, and practicing engineers.