Controls expert Miroslav Krstic holds record seven fellowships in technical and scientific societies
|Ronald Graham -- Click Here to visit JSOE Flickr|
San Diego, Calif., Nov. 20, 2017 -- This year is turning out to be a banner year for Miroslav Krstic, a controls expert at the University of California San Diego who also serves as the senior associate vice chancellor for research here on campus.
In June, he received the Ragazzini Education Award of the American Automatic Control Council. Then, in October, he received the Oldenburger Medal, a lifetime achievement award from the American Society of Mechanical Engineers, which is sometimes thought of as the Nobel Prize in the field of controls research. And in November, he found out he has been elected Fellow of the American Association for the Advancement of Science. He was recognized “for revolutionizing control of distributed parameter systems and extremum seeking, with applications to fluid flows, nuclear fusion, particle accelerators, batteries, and semiconductor manufacturing.”
This latest honor makes Krstic a fellow of seven different technical and scientific societies—a rare feat. That’s because his discipline, controls, transcends just one field of study, with applications to everything from oil drilling to batteries for electric vehicles.
“I chose controls in college because it is abstract, conceptual, and general—it applies to just about any problem where time plays a role, there are objectives to meet, and there are parts that interact,” Krstic said.
The principles of control theory are always the same: sensors take measurements of the physical environment and then the control system drives inputs to regulate and manage many different systems that interact with that environment. The Segway and cruise control in cars are the best known examples of control systems. Essentially, a cruise control system figures out how much gas your car needs in order to stay at a certain speed despite terrain changes, after measuring the current speed.
Krstic opened up a whole new area of research in control theory by reviving what are known as extremum seeking algorithms. His advances in that field make it possible to conduct chemical analysis of rocks on the Mars rover Curiosity. They also have helped achieve a 200-fold increase in area density in the microchips that run your smartphone, resulting in a multi-billion dollar impact for the semiconductor photolithography industry alone.
Krstic had proven himself to be an outstanding researcher from the get-go. When he started his Ph.D., his advisor at UC San Barbara, Professor Petar Kokotovic, told him about an open problem in the controls field. He solved it before classes started that quarter. It turned out that many top researchers in the field had been struggling, unsuccessfully, to do just that.
Success didn’t necessarily bring job offers in his field after he completed his Ph.D. in three years. “With highly theoretical results and still limited English, my record wasn’t translating into faculty job offers in electrical engineering departments to which I was applying,” Krstic told IEEE Control Systems Magazine. “But several mechanical engineering departments saw promise in my work in nonlinear and adaptive control, with applications involving fluids, and I made a permanent and fortunate academic transition into ME.”
Krstic was on faculty for two years at the University of Maryland, before joining UC San Diego in 1997. He currently holds the Daniel L. Alspach chair in Dynamic Systems and Control at the Jacobs School of Engineering at UC San Diego. He also is the founding director of the Cymer Center for Control Systems and Dynamics here on campus.
In addition to AAAS, Krstic is a fellow of the Institute of Electrical and Electronics Engineers, American Society of Mechanical Engineers, Society for Industrial and Applied Mathematics, International Federation of Automatic Control, Institution of Engineering and Technology and associate fellow of American Institute of Aeronautics and Astronautics.
When not running a large research group, teaching or taking care of administrative duties as senior associate vice chancellor for research, Krstic likes to read Greek and Roman classics and listen to classical music. His Ph.D. advisor, Kokotovic, turned him into a museum and ballet aficionado. Krstic’s passion for playing the electric guitar has not left him since his middle school days and he now owns an impressive collection of instruments, including several 1970s Gibson Flying Vs, custom shop Strats and classic Marshalls.
Extremum seeking, or “hill climbing,” is the problem of tuning inputs of a system, in real time and without a mathematical model of the system available, to reach optimal operation. Imagine sitting at a mixing console of a sound system in a music auditorium, faced with dozens of knobs to turn in order to optimize sound quality. Extremum seeking automates such an overwhelming “knob-tweaking” process.
The first extremum seeking algorithms were invented before WWII but were abandoned in the 1960s because their creators were not able to guarantee the method’s stability. Krstic resurrected extremum seeking around the year 2000 by providing stability proofs for the algorithms and introducing new applications like those for autonomous vehicles performing a search in environments where GPS doesn’t work. Extremum seeking has grown to be one of the most active areas in control engineering research and, as a result, Krstic is currently one of the top three controls authors worldwide in terms of the annual citation rate.
Krstic also was able to develop algorithms that solve the long-standing problem of delays in a non-linear control system—that is a system where input and output are not proportional.
Researchers in the field of non-linear controls had built into their algorithms the assumption that there are no delays in sensing, computation, actuation, or communication over a network. But of course delays are inevitable.
In the mid-2000s, Krstic’s work on the topic of control of fluid flows to reduce aerodynamic drag led to an elegant solution to this delay conundrum. He adapted his flow control techniques to compensation of large delays in nonlinear control systems. He developed a method that takes any controller designed for a system without a delay and makes it work in the presence of a delay of arbitrary length, even when the delay varies with time or depends on the system's state (such as the solid-liquid interface in extruders for 3D printing), and even when the delay value is unknown.
Krstic’s control designs were recently employed on a 700-meter-deep oil drilling system by the Norwegian Statoil company to regulate violent oscillations of the fluids in the drilling system using only a valve at the top as the input.
For his Oldenburger Lecture, Krstic chose a topic that may seem completely unrelated to oil drilling but is mathematically nearly identical—control of congested traffic dynamics on long freeway sections using inputs only at the on-ramp of the freeway.
In congested, slow traffic, perturbations travel fast in the upstream direction (opposite from the direction of cars) because, in bumper-to-bumper traffic, drivers respond rapidly (and even anxiously) to the driver in front of them braking. Krstic’s research exploits this human behavior to suppress stop-and-go oscillations in congested traffic by varying the duration of green and red lights at on-ramps in response to the measured traffic speed at the same location on the freeway.
Other recipients of the Oldenburger media at the Jacobs School of Engineering include mechanical and aerospace engineering professor Robert Bitmead, in 2014.