University of Wisconsin-Madison College of Engineering
UW-Madison College of Engineering banner 

Electrical and Computer Engineering

Field programmable gate array

Field programmable gate arrays (FPGAs) are computation and routing resources controlled with an underlying layer of static random access memory (SRAM). (50K JPG)

Katherine Compton

Assistant Professor Katherine Compton (17K JPG)

The ultimate handheld is just a reconfiguration away

The allure of powerful handheld computing devices for communication, organization, recreation and more is such that consumers practically need to wear a tool belt to hold all of the pods, phones, players and digital assistants. But a better answer to carrying more computing power in smaller spaces will more likely be found in devices called field programmable gate arrays (FPGAs.)

FPGAs are pieces of hardware that can be reconfigured and reassigned to perform various tasks. They achieve higher performance than software and more flexibility in the hardware. It's essentially the ability to redesign a circuit after it's installed in the device. FPGAs can become the graphics processor at one point or an MP3 decoder, or satellite data processor or encryption processor at another point, all while increasing computing speed and using less power.

But before all of these tasks can be combined into one reconfigurable device, someone has to figure out how these FPGAs will relate to the host processor and operating system. What happens when multiple programs want to use the same reconfigurable piece of hardware at the same time?

Assistant Professor Katherine Compton is working toward better understanding and control of FPGAs so that a host processor can make efficient decisions dynamically at runtime regarding whether a program should rely on the processor or reconfigure the specialized hardware. Answers to these complicated questions ultimately could mean electronics products will more quickly get to market and remain useful longer, since a device that goes out of date could be rebuilt by installing new logic.

Reinventing the laser

Lasers are modern marvels with myriad applications to daily life — supermarket scanners, DVD players and communications infrastructure, to name a few. But the future holds even more promise for these amazing devices as researchers seek to build them on the atomic scale.

Philip Dunham Reed Professor Dan Botez, McFarland-Bascom Professor Franco Cerrina and Milton J. and A. Maude Shoemaker Professor of Chemical and Biological Engineering Thomas Kuech are working toward a continuous wave, direct current, room temperature, infrared semiconductor laser based on quantum boxes. The success of their project would be a dream come true in areas of application as diverse as medical diagnostics, environmental monitoring and military countermeasures, and much more.

Researchers have created quantum-scale lasers operating in pulsed mode at room temperature, but none can operate continuously without cooling to cryogenic temperatures. A continuous-wave, mid-IR laser operating at room temperature would be portable and powerful — a revolution in lasers.

Already the College of Engineering team has demonstrated, for the first time, emission at room temperature, in the mid-infrared wavelength range (3-5 microns) from quantum-well devices. The novel structure is the subject of a patent application filed by the Wisconsin Alumni Research Foundation.

Enhancing system stability through better modeling

Professor B. Ross Barmish is an expert at finding answers for the kinds of problems that seemingly have no answers. In many areas of science and engineering designers are stuck with bad models of systems. The models have variables and tolerances that range between 20 and 30 percent and yet designers have to engineer methods to control those systems. Consider, for example, an airplane: The flight control and landing systems have to perform without fail through wide variations in temperature, pressure, wind speed and other possibilities. In other words, the systems have to be stable and robust.

In the case of stability, Barmish pursues an approach called set-valued frequency response methods. He says these kinds of problems can often be boiled down to a simple math equation about polynomials and their roots. But if the polynomial has variations in its coefficients because of the physical uncertainty in the system, one is left with a very fundamental problem about roots of polynomials. Barmish searches for robust results— those that are where they need to be, no matter how the parameters of the system change the roots. The answers to these kinds of problems are applicable over a wide range of areas—so much so, that Barmish’s approach has become a standard toolset for those seeking to control the seemingly uncontrollable.


University of
Wisconsin-Madison College of Engineering logo

Copyright 2004 The Board of Regents of the University of Wisconsin System
Date last modified: Thursday, 17-Feb-2005 14:09:29 CST
Date created: 17-Feb-2005