ECE 2007The Bradley Department of Electrical and Computer Engineering

Makingthe smallestthings visible

Carbon Nanotubes

Quantum

Dots

onSilica

Nanosperes

BraggOnionResonatorshell

One day, early in my college career, my Circuit Analysis professor, an

enthusiastic man prone to tangents, managed to turn some ho-hum

circuit analysis review into a half-hour explanation of the obstacles

faced in current microelectronics research.

“Electronic components are getting small enough that quantum

mechanical effects are beginning to dominate, and devices no longer behave

the way they’re supposed to,” he announced. Then he told us that in order to

continue producing smaller, faster computers at the current breakneck pace,

researchers must develop a technology that is completely new—probably

based on the same quantum mechanical effects that cause the current

technology to flounder.

Until then, I had thought of computer engineering as my career path

and physics as my hobby: never the twain will meet. Suddenly, here was

Dr. Hendricks telling me that they had not only met; they had met, married,

and started a family. “It will be a paradigm shift,” he continued, “and you’ll

be the ones to shape it.”

Today, I can’t remember if I shivered my way to class under two jackets

and a sweater or ambled, sweating under the hot sun. I can’t remember if it

was late morning and I was eager for lunch, or if it was late afternoon and I

was ready to go home. I can’t remember if I was a bright and eager freshman

or a self-assured and all-knowing sophomore. The details of that class and

that day are lost in murky memories of the Far Distant Past, but I can

remember with absolute clarity listening raptly to my professor’s words and

thinking to myself, “That’swhat I want to do.”

Captured by a tangent

My Linh Pham, CPE/Physics ’07

Minors: Microelectronics, Mathematics

Bradley Scholar

This report was produced with funds from the Harry Lynde Bradley Foundation.

Photography by Virginia TechVisual Communicationsunless noted otherwise.Josh Armstrong: pp. 1, 2, 3, 4,8, 9, 13, 14, 20, 21, 22, 26, 31Rick Griffiths: pp. 1, 2, 16, 40-44John McCormick: pp. 3, 36, 40-44Michael Kiernan: p. 27

Produced by The Right Word Communications, Blacksburg, Virginia

2 From the Department Head3 From the Advisory Board Chair

38 Ph.D. Degrees Awarded39 Patents Awarded40 Bradley Scholars42 Bradley Fellows45 Honors & Achievements46 Donors to ECE48 ECE Faculty49 ECE Advisory Board

4 Nanoprobing:making the smallest things visible

8 Tree of Trust: protecting embedded systems11 Tackling the mystery: the Mono virus14 Is an energy router in your future?16 Renaissance man of the acoustic waveforms40 Matt Carson ’98: life in the nascar fast lane

RESE

ARCH

FEATURES

ECE

PEOP

LE18 Biomedical Applications20 Computer Systems22 Networking24 Software and Machine Intelligence26 VLSI & Design Automation28 Communications30 Electromagnetics32 Electronics34 Power36 Systems & Controls

After three years of hiring, we havebecome one of the largest ECEdepartments in the country. Weare in the top 10 in terms of totalnumber of faculty and in a com-parable position in terms of both

undergraduate and graduate students.Since January of 2005, we have added

17 faculty members in ECE and are recruit-ing for three more this year. We currentlyhave a record high of 75 tenured and tenure-track faculty members in the department.

In 2006, eight more new faculty mem-bers joined us. They are Leyla Nazhandali,assistant professor in computer engineering;Khai Ngo, professor in power electronics;Scott Bailey, assistant professor in space sci-ences; Bob Clauer, professor in space sci-ences; Brent Ledvina, assistant professor inspace sciences; Ming Xu, assistant professorin power electronics; Jason Xuan, associateprofessor in imaging; and Yaling Yang, as-sistant professor in computer engineering.The eight have Ph.D.s from Caltech, Col-orado, Cornell, Illinois, Maryland, Michi-gan, UCLA, and Virginia Tech.

We have added three faculty membersin the space sciences area due to the NSFgrant to Wayne Scales and colleagues men-tioned in previous years. The Virginia TechIndustry/University Partnership for SpaceScience (IUPSS) has been formed and re-ceived initial support from VPT, Inc. ASpace Science Center at Virginia Tech en-

compassing: ground based analysis of GPSsignals, development of scientific instru-ments for sounding rockets and satellites,and leadership of major satellite programs isplanned.

Governor Timothy Kaine cut theribbon to open the new ECE MicronTechnology Semiconductor Processing Lab-oratory on October 4th. The lab is arenovation of a teaching laboratory thatopened in 2001. It has new and improvedcapabilities that will be used for bothteaching and research. Micron Technologymade a $750,000 gift that provided the finalfunds to complete the laboratory.

The Microelectronics, Optoelectronicsand Nano-technology (MicrON) group,which involves more than 10 facultymembers in four departments, will beconducting research in the new lab, and BobHendricks will be teaching ECE’s semi-conductor processing course.

Notable faculty awards include: ArunPhadke was one of four professors honoredwith the Docteur Honoris Causa di I’INPGrenoble; Daan Van Wyck received theIEEE power Electronics Society 2006 Dis-tinguished Service Award; Jason Lai wasnamed a Fellow of the IEEE; DushanBoroyevich was named the American Elec-tric Power Professor of ECE; and AnboWang was named the Clayton Ayer Profes-sor of ECE.

Our young faculty members continue

to garner national recognition in research:Tom Martin received a Presidential EarlyCareer Award for Scientists and Engineers(PECASE), and Patrick Schaumont andYong Xu received NSF CAREER Awards.With Martin, Schaumont, and Xu, we nowhave three faculty members who have wonPECASE awards and an even dozen CA-REER awardees. Congratulations to all!

My many thanks to Dan Sable, who isthe chair of the advisory board. Dan and theboard have been extremely helpful in sup-porting the initiative in space sciences.

We had a ceremonyMarch 9th to honorpast and present ECE department heads. Oilpaintings of many of the department headsare now on a stone wall inWhittemore 302.The painting of Samuel Pritchard (1892-1935) was rescued from Pritchard dormitoryduring a renovation. The more recent paint-ings are from photographs. Portraits of W.A.Murray 1935-56 and B.M. Widener (1956-58) are the only missing. If you have a goodphoto of either that you could part with fora while, we would appreciate your help incompleting the wall. Debbie Hallstead,Jaime De La Ree, and Kathy Atkins deserveall the credit for the ceremony and the wall.

James S. ThorpDepartment Head

FROM THE

Department Head

PERS

PECT

IVES

Headsof ECE

2

Samuel R. Pritchard1892-1936

George Powley1958-1964

William A. Blackwell1966-1981

Daniel B. Hodge1981-1989

It is an honor and a privilege to havebeen elected chairman of the ECEDe-partment Industrial Advisory Board(IAB). I have been associated with ECEfor almost 25 years in various capaci-ties including as a master’s student, a

Ph.D student, a research scientist, and thefounder and president of a VPT Inc., a Vir-ginia Tech spin-off company initially locatedin the Corporate Research Center and nowlocated in the Blacksburg Industrial Park.

The Virginia Tech ECE department hastruly outstanding undergraduate and grad-uate programs with excellent faculty world-renowned for its teaching and research. I canstill remember when I was a master’s studentlistening to a truly fascinating guest lectureby Professor Charles Bostian on satellitecommunications. After his researchers hadreceived anomalous signals from the satellitedishes on the top of Whittemore Hall, theybegan to develop new theories of RF wavepropagation only to later determine that thecause was a dead fly in the wave guide.

Virginia Tech has always been a leaderin the application of technology directly inthe classroom. It was not too long ago thatVirginia Tech became the first engineeringcollege in a land-grant university to requireall incoming students to possess a personalcomputer. The state-of-the-art model at thetime was an IBMPCwith a whopping 640Kof memory. This tradition continues withtoday’s application of wireless tablet PCs in

the classroom.ECE department excellence does not

stop at great teaching. During my associa-tion with Tech I have witnessed ECE de-velop internationally recognized researchprograms in power electronics, fiber opticsand sensors, wireless communications, mi-croelectronics, computers, and others. Thefuture challenges for the department are toestablish leadership positions in a variety ofemerging fields including nanomaterials,biomedical engineering, energy and the en-vironment, and space science.

As alumni andmembers of the IAB, it isin our own interest to see the ECE depart-ment prosper and gain higher reputation. AtVPT Inc., 100 percent of our engineers re-ceived a bachelor’s or master’s or Ph.D. fromVirginia Tech. At Orbital Sciences Corp.,which is also represented on the IAB, thereare about 100 engineers with a Tech degree.My vision for the IAB is to encourage an ac-tive role for the members in several ways:

1. Alumni and IAB member companiescan encourage practical experience by spon-soring undergraduate and graduate studentinterns. At VPT, we have had an active in-tern program for 12 years. This is a mutuallybeneficial relationship as VPT often hires thestudents full-time upon graduation.

2. Alumni and IAB member companiescan participate directly in the classroomthrough guest lectures. Relating practicalproblems encountered in industry to the

classroom theory goes a long way with stu-dent understanding of course material.

3. Alumni and IAB member companiescan sponsor research in areas of concern tothem. VPT has been active in promoting theemerging space science group in the ECEDepartment and is also a member of Centerfor Power Electronic Systems (CPES) Indus-try-University Partnership Program.

4. Alumni and IAB member companiescan take an active role in being advocates forthe ECE department at the college and uni-versity level. ECE has been a traditionalstrength of the university. It is critical that itcontinues to receive appropriate priority.

5. Finally, alumni and IAB membercompanies can support the ECE departmentthrough generous financial donations.

Jim Thorp has demonstrated remark-able leadership in focusing the department,and the IAB fully supports his efforts. Theuniversity is fortunate to have person of hisstature, integrity, caliber, and vision as theECE department head.

Great teaching, great research, greatstudents, great alumni, great leadership–what more can one ask?

Dan Sable (MS ’85, Ph.D. ’91)Chair, ECE Advisory Board

FROM THE

Advisory Board Chair

PERSPECTIVES

3

In March, ECE began displayingportraits of the departmentheads who have served since ECEwas established in 1892. We areseeking photos of W.A. Murray(1935-56) and M.M. Widener(1956-1958).

F. William Stephenson1990-1994

Leonard A. Ferrari1995-2000

Robert J. Trew2001-2002

James S. Thorp2004-present

NANO

PROB

ING

Yong Xu (left) is developing technology that employs a nanotube on theend of a fiber to achieve an optical imaging resolution of 1nm. His teamuses a laser imaging system in order to “see” and manipulate the nano-tubes. Graduate student Yaoshun Jia observes from the right.

Making the smallest things visible

4

Nanoprobe technology could achieveultimate optical imaging resolution,while helping explore one of the lastmysteries of quantum electrodynamics

Yong Xu, an assistant professor of electrical engineering, isdeveloping an active nanoprobe technology that could enablesignificant breakthroughs in optics, physics, and commu-

nications.Optically, his nanoprobe technology can break the diffraction

limit and achieve the ultimate optical imaging resolution of 1nanometer. Furthermore, he hopes to use the nanoprobes tomeasureone of the few remaining mysteries of quantum electrodynamics—vacuum field fluctuations. The probe could also be configured as asingle-photon emitter, enabling quantum cryptography and theultimate secure communications link.

Xu has received a National Science Foundation (NSF) FacultyEarly Career Development Program (CAREER) Award to supporthis efforts. CAREER grants are NSF’s most prestigious awards forcreative junior faculty members who are considered likely to becomeacademic leaders.

A tube and a dot

The potential of Xu’s technology seems inversely proportionalto its size—that of a single carbon nanotube. A carbon nanotube isa single-atom-thick sheet, shaped as a cylinder, that can have adiameter of about 3 nm. Its length can reach tens of microns. Likemany nanoscale molecules, carbon nanotubes have novel properties,including great strength and numerous intriguing electricalproperties.

While a nanotube will serve as the body of the probe, the tip ofan optical nanoprobe will be a single 1-3 nm quantum dot and thewhole assembly attached to an optical fiber. Also called nanocrystals,quantum dots are nanometer-scale pieces of semiconductor that canemit light (fluoresce). “We needed an emitter with a size as small aspossible,” Xu explains. “We also needed a fluorescent element thatis bright, stable, commercially available, and can cover a broadwavelength range.”

Xu expects it may take up to two years to perfect the fabricationtechniques and develop working prototypes of the appropriate sizeand strength. Once the fabrication is proven, his teamwill apply thenanoprobes in near-field imaging, vacuum-field imaging, and single-photon generation.

Improving the optical microscope

“Improving resolution is critical to the advance ofnanotechnology,” Xu says. “We need to be able to image optical

5

structures and observe physical phenomena at ever smaller sizes.”Most optical microscopes are restricted by the diffraction limit

and cannot produce optical images with resolutions better than a fewhundred nanometers. To overcome the diffraction limit, the currentapproach relies on a nanoscale fiber aperture to obtain higherimaging resolution, as in the case of near field scanning opticalmicroscopes (NSOM). However, even in the most advancedNSOMsystems, the imaging resolution is limited to about 50 nm resolution.In addition, NSOM technology relies on metal-coated tips, whichcreates a large perturbation of the original optical field, Xu explains.

Since the ultimate resolution limit of nanoprobe imaging isdetermined by the size of the fluorescent element, Xu’s nanoprobeshould enable optical imaging at a resolution of 1-3 nm, he says. Theprobe can generate images by scanning the position of the quantumdot and detecting the fluorescence into the silica fiber. “Using asingle carbon nanotube, the nanoprobe can achieve for the first timenon-perturbing imaging at this resolution,” he says.

Viewing a quantum mystery

With a nanoprobe of such small scale, Xu hopes to demonstratea 3-dimensional vacuum field imaging for the first time. Modernphysics has determined that a vacuum is not a complete void, as manypeople believe, he says.

“By definition, there is always something there, even in the

absence of all particles.We call this tiniest level of physical existenceof electromagnetic field ‘vacuum field fluctuation.’ Light, contraryto what youmight imagine, comes as waves of tiniest energy bundlescalled ‘photons.’ And a photon is emitted when an atom or amolecule jumps from a high energy level to a low level,” he explains.

“If there is no electromagnetic vacuum field, then they won’tjump. There is a perturbation that in essence tries to trick them intojumping from high to low. Our most precise theory of naturepredicts this and we know it must be there because otherwise ourdevices would not work.We have verified vacuum fields by countlessexperiments. We have a theory to calculate it, but we have not beenable to measure what is going on and how it is distributed,” he says.

Xu plans to analyze the decay of the quantum dot to measure avacuum field. “The decay of the light emitting process indicatesthere is a field. Bymeasuring how it decays or how quickly a photonis emitted, we can infer the distribution of the field.”

The nanoprobes in this application will have a tip of multiplequantum dots. In addition, an integral part of the 3-D vacuum fieldimaging is the use of silicon-based Bragg onion resonators he helpeddevelop while he was a postdoctoral scholar at Caltech.

New, single photon emitters

The third application of the nanoprobes is to convert them intosingle-photon emitters, which can be used in quantum cryptography

Colorized images of nanotechnologyFrom the left: Quantum dots (darkspots, ~3nm in diameter) on siliconnanospheres (red).

Center: An army of Bragg onion res-onators, similar to the ones Yong Xuplans to use to demonstrate and mapvacuum field fluctuations—which havebeen proven, but never mapped.

Far right: a close-up of the resonatorshows the layers of “shell,” whichmake it possible to isolate a vacuumfield fluctuation.

“We need to detect something so incrediblyweak and still isolate it from the noise.We spend a lot of time in a dark room.”

6

and quantum information processing.“In our approach, we use a nanoprobe to achieve optimal

coupling between a single quantum dot and a cavity, so that singlephotons can be triggered as needed,” he says. “We can engineer thecavity so that all the photons generated by the quantum dot will gointo the cavity. In the end, we can create a ‘push button’ device thatcan generate one, and only one, photon at a time, i.e., single photonon demand.”

A reliable, constant source of single photons will enable thedevelopment of quantum cryptography, where the security ofcommunications is guaranteed by one of the most basic laws ofnature. “A photon cannot be split in two, which is a wonderful giftof life.” Xu says. “A photon is either detected or not detected. It is asignal of one. It is either received or stolen by an eavesdropper.” Ifsidetracked by an eavesdropper, the recipient knows the data ormessage is compromised.

The trouble with tiny

Xu’s nanoprobe technology in imaging, computing,measurement, and communications, may find future use in physics,biology, chemistry, and even quantum teleportation, he suggests.However, in the near-term, his team faces challenges related to thesmall size of the technology.

“We deal with structures with dimensions of a few nanometers,so we must make sure everything is perfectly stable andnanostructures are precisely manipulated,” he explains. “We are alsoworking with one photon at a time. That is a very weak signal. Weneed to detect something so incredibly weak and still isolate it fromthe noise. We spend a lot of time in a dark room.”

Calling on an armyof onion resonators

For Yong Xu, 3-D vacuum field imaging is one of the mostimportant aspects of his research. “It is a great challenge andgreat fun to be able tomeasure something that has never been

measured before to the precision of a few nanometers,” he says.Not only would success reflect the first time vacuum field

imaging is demonstrated, but Xu’s workmay also lead to the cre-ation of a silicon laser, which is the most critical missing compo-nent in the race towards silicon-based integrated optics.

Xu’s team plans to demonstrate vacuum fields by usingBragg onion resonators he and his colleagues at Caltech and San-dia National Laboratories developed in 2004. Each resonator is ahollow cavity encased in a shell made of alternating layers ofhigh-index silicon and low index silicon dioxide (SiO2). Each res-onator has a “stem” opening so that a nanoprobe can be inserted.

The shell is a very thin 250 nm and each resonator is 10 to 15microns in diameter. Due to the large index contrast between sil-icon and SiO2, the shell can behave as a perfect conductor andprovide 100 percent reflection for any incident light. This can ef-fectively isolate the vacuum field in the hollow core from out-side field.

Xu’s team has developed a theoretical model and demon-strated that onion resonators can both enhance and inhibit spon-taneous emission by more than one order of magnitude. “Thistells us we can funnel most of the spontaneous emission into thedesired mode, which is important for both single photon sources,and for laser applications,” he says.

7

While hackers and viruses grab headlines incomputing security, Patrick Schaumont’steam is working to protect data from anewer, growing threat—a threat fromthe loss and theft of embeddedcomputers that store personal and

private information.“Today’s computers are so small that we can carry

them around and lose them,” he says. “Who among uscan claim never to have lost a credit card, a cell phone, aPDA, a laptop?”

Medical records, high-resolution pictures of pen-drawn signatures, and codes to unlock electronic car

locks are just some of the examples of information nowstored on smart cards, key fobs, and other embeddedcomputers. The problem in keeping the informationsafe, Schaumont says, is that conventional computersecurity and cryptography focuses on protecting dataduring transmission, but once thieves have possession ofthe computer, transmission is not an issue and the data isvulnerable.

The solution is to carefully design security into thehardware and software of a computer, but engineerscommonly add protection only as an afterthought.Schaumont is working to develop a methodology forembedded systems engineers to follow in designing bothTR

EEOF

TRUS

T

Pengyan Yu (left) and Eric Simpson demonstrate SAM (secure authenticated mode). Unautho-rized devices see only the noise on the right, while the authorized device sees the magic word.

8

GROWING THE TREE OF TRUST STARTS BY PROTECTING THE ROOT-OF-TRUST

INEM

BEDD

EDSY

STEM

S

hardware and software for security.His team is seeking answers to questions, including: How can

we systematically deal with private information processing in aportable computer? How can we implement cost-effectiveencryption mechanisms that are energy-efficient and secure? Howcan we store secrets in a reliable way in a portable system?

“Securing secrets is not part of the typical engineer’s toolbox,”he says. “We want to change that.”

Schaumont has received funding for his effort from a NationalScience Foundation (NSF) Faculty Career Development Program(CAREER) award. The CAREER grant is worth $400,000 over fiveyears and is NSF’s most prestigious award for junior facultymembers. Schaumont joined ECE in 2005 as an assistant professor.

Greatest hazards are side channel attacks

“As in any kind of security, a secure embedded system is only assafe as its weakest link,” Schaumont says. Once thieves have

possession of an embedded computer, secrets can be easily probedout, and the thief doesn’t need to know the password to do that. Forexample, a cryptographic chip can often be analyzed with a sidechannel attack by measurement of only its power consumptionpattern.

“You don’t need to open the chip to do this,” Schaumontexplains. “It’s a systematic problem, much like a communicationsproblem. You can filter out the noise and detect the code you seek.”He was on a team in the Secure Embedded Systems Group at theUniversity of California at Los Angeles (UCLA) that demonstratedsuch a side-channel attack on a standard encryption chip. “We couldbreak the code in just three minutes by measuring the power,” hesays.

Side-channel attacks have been recognized as a security issueonly in the past decade, according to Schaumont. “Most people arefamiliar with software attacks and physical attacks, but not side-channel attacks,” he says. (Continued on p. 10.)

Patrick Schaumont receivedan NSF CAREER Awardfor his work on protectingembedded systems fromside-channel attacks.

“We could break the code in just threeminutes by measuring the power.”

9

Software attacks often arrive via virus or spyware and involve acomplicated process that requires expert knowledge, Schaumontexplains. “Logical attacks require brains, but they are cheap. All anattacker needs is patience.” Classical security software is designed tothwart these attacks.

Physical attacks, often used in reverse engineering, requiredisassembly and are costly. “Physical attacks require expensive,specialized equipment, such as microscopes and chemicals to getdown to the chip layer,” he says. “It’s expensive, but it’s not hard.Using a systematic approach, it can be done.”

Side-channel attacks are ideal for embedded systems from anattackers perspective. “You don’t need to open a chip and don’t needexpensive hardware. Plus, it is systematic.” Protecting against theseattacks is very challenging, Schaumont says. “There are manyabstraction layers—all of them exposed. Current countermeasuresare all point solutions, implemented only in software or hardware.They offer no guarantees for the overall system.”

Hardware/Software Codesign

The solution to designing secure embedded systems is to payattention to all abstraction layers, and this requires hardware/software codesign, he says. “The flexibility of software supportscomplex crypto-algorithms. On the other hand, hardware canprovide side-channel countermeasures that are hard to implementwith software, such as constant-power execution. A combination ofthe two will enable flexible and secure system design.”

Tree of Trust

Schaumont wants to develop a systematic approach to side-channel-resistant embedded system design. “Security is hard tooptimize, hard to quantify,” he explains. He is proposing asystematic approach to side-channel-resistant embedded systemdesign he calls “The Tree of Trust.”

With the Tree of Trust, a designer can systematically partitiona system into a secure-critical and a non-critical part. A designerneeds to be able to systematically shield those secrets. Secrets go intothe critical part. A completely isolated secret is meaningless,however. All these protected secrets are only useful if they caninteract. The Tree of Trust makes sure that secrets are integrated intothe system with a secure interface.

“You are trying to defend your root of trust,” he says,explaining that the root of trust is the element that is implicitlytrusted, such as the key of an encryption algorithm. “Implementingthe root of trust is always expensive, so the smaller you can makethis, the cheaper your system is. Indeed, the non-trusted elementscan use standard components or software.”

The embedded system contains several abstraction levels,including the protocol, the algorithm, the architecture, hardware,and circuit levels. “If you partition the system correctly at everylevel, you will obtain a very small root-of-trust which can beprotected with cost-effective countermeasures across thehardware/software boundaries,” he explains.

Building With Secrets

His goal, he says, is “to come up with a way in which peoplecan systematically deal with the design of secrets when designing asystem…We are trying to find a canned sequence of operations thata designer of secure hardware and software could use.”

His team is applying the Tree of Trust concept to differentembedded-system implementations. One ongoing applicationinvolves protecting content using “Secure AuthenticatedMode,” orSAM for short. Using SAM, a video message can be created that canbe displayed only on a unique and single device. Another project isthe development of protection mechanisms for DSP software insensor nodes that remain active in the field for several years. “Thesensor nodes contain highly advanced signal processing to gather

information on activities in theirenvironment. The software representsan important amount of intellectualproperty,” he says.

Targeting the Engineers

A critical part of Schaumont’s planis educating the engineers who designembedded systems. He has involvedundergraduate students in the research

project and introduced a new undergraduate course last fall calledHardware/Software Codesign. He is also working on a team that isdeveloping a freely available CD-ROM for undergraduates thatincludes a codesign environment and tools.

“Hardware/software codesign has typically been a graduate-level topic, however we want our undergraduates to be able tocompete on a global scale and be capable of designing complexembedded systems,” he says. Schaumont plans to develop anadditional course at the graduate level focusing specifically on secureembedded systems.

“We want engineers to develop a sense about how to buildsecure embedded systems and to understand what is breakable andwhat is not.”

“Securing secrets is not part ofthe typical engineer’s toolbox.We want to change that.”

10

TACKLINGTHE MYSTERY:

THEMONOVIRUS

Many people knowinfectiousmononucleosis(mono) as the dis-

ease of fevers, sorethroats, and fatigue thatcan rob students of entiresemesters. Medical re-searchers know mono asthe disease caused by theEpstein-Barr virus (EBV)—one of the oldest, mostcommon human viruses.

By age 45, about 95 per-cent of all adults havebeen infected, accordingthe U.S. National Centerfor Infectious Diseases.EBV is a tenacious virusthat can maintain a life-long, persistent infectionin healthy people. Re-searchers know it ‘hidesout’ in healthy people,but what triggers it to gofrom ‘sleeper agent’ to aninfection that spreads toother individuals is stillunknown. (Continued onp. 12.)

ECE’s Paul Plassmann—along with Mark Jones and a team ofgraduate students—is working with researchers from TuftsUniversity to understand the dynamics of EBV; how it interacts

with human cells and organs over time. Using supercomputers, theyare developing a simulation that researchers hope to use to modelthe behavior of the virus at the cellular level in a human body.

“We just can’t watch cells interacting over time in a humanbody,” Plassmann says. Given the impossibility of observing thevirus infecting people, a simulation testbed is the only solution forsuch situations, he says, adding that biologists have coined a term forit. “They call experiments in living tissue ‘in vivo,’ experiments inthe laboratory ‘in vitro,’ and now, experiments using computer sim-ulations, ‘in silico.’”

The mystery

Viruses typically target certain kinds of cells and particular or-gans, he explains. “EBV hangs out in the tonsils.” Specifically, EBVtargets naïve B-cells, which are part of the body’s adaptive immuneresponse system. An EBV-infected naïve B-cell then goes throughthe “germinal center process” and fools the body’s immune systeminto thinking that the infected cell is real memory B-cell. “It thenhides out in the adaptive immune system’s memory compartment.For unknown reasons, these infected B-cells occasionally are acti-vated and become ‘lytic.’ In the lytic phase, the virus reproduces,destroying the original cell and producing free virus, which can in-fect other B-cells or leave in the saliva to infect other people.”

After the initial acute phase of the infection, EBV survives inthe body at a constant rate of about 5 in a million cells. The mys-tery, Plassmann says, is how EBV cells remain at a constant popula-tion under the radar of the body’s immune system. “Biologists donot know how long an infected B-cell lives,” he explains. “It couldbe as short as a week. At some point, the cell has to reproduce or golytic and infect other cells.”When people develop mono, as much as50 percent of their memory B-cells become infected and destroyedby the immune system, causing the familiar symptoms.

Simulating extraordinary complexity

Simulating a live biological process is extraordinarily complex.The simulation models cells, viruses, and other biological objects asagents and simulates their motion and interaction through the solu-

tion of stochastic differential equations. Each agent has an internalstate, which can bemodeled as a finite state machine or ultimately bychemical pathways that will involve hundreds of states and thousandsof transitions. The real computational challenge, however, is withthe number of cells. “We are modeling hundreds of millions to bil-lions of cells,” Plassmann explains. “The time step for the simula-tions is on the order of minutes, but we want to see what happensover a span of months or years.” The simulations easily get into theteraflop to petaflop range (greater than 1012 to 1015 floating pointoperations per second), which is the edge of what is possible withtoday’s fastest machines, he says.

Such simulations typically require supercomputers, or parallelcomputers, he says. Plassmann uses machines with between 200 and

400 processors. An interestingnew development in computerengineering is that number ofcores per chip is rising quickly,“so we will have to develop thealgorithms and software that cantake advantage of these mas-sively multi-core architectures,”he says.

Another complicating factor is that the reactions and dynamicsinvolve different time scales, creating a multi-scale problem. “Thereis a slower time scale where we model the motion and interactionbetween cells,” Plassmann explains. “However, where wemodel thechemical pathways inside a cell, things happen on a much faster timescale. The trick is to have the simulation move at the longer timescale of the cells, while accurately representing what is happening atthe shorter time scales,” he says.

Borrowing from combustion

The team is using some ideas from their experience in combus-tion modeling. In combustion, the fluid mechanics happens at theslow time scale (millisecond time scale) and the chemistry happens atthe fast time scale (nanoseconds). “The observation is that similarchemistry calculations are done repeatedly. So, rather than redoingthem, we store the results of these calculations in a ‘scientific data-base’ and approximate subsequent chemistry calculations basedqueries to this database. This approach has enabled speedups in com-bustion simulations of 100 to 1000 times, Plassmann says.

While Plassmann enjoys the challenge of multi-scale simula-tions, he says the most fun is working with people in other fields onpredictive capabilities. “In my work on EBV, I’m working with ex-perimentalists. Their experiments look at understanding the specificsof particular cellular transitions, interactions, or properties. How-ever, the simulation is the ‘whole picture’ and gives an experimen-tal biologist a chance to see how these specific properties affect the

“The more we can do with simulation,the more we can help to find ways

to improve people’s health.”

12

Paul Plassmann is developing simulations to help solve the mystery of the latent infec-tion of the Epstein-Barr virus that causes mono. The screen behind him, and the imageon page 11, show a cross section of a simulation of a germinal center—an active areaof the tonsils used by the adaptive immune system to develop memory B-cells.

complete biological system. In the predictive mode, a biologist cannow ask the question, ‘if I can experimentally change a specific in-teraction (by, for example the use of a drug) howwill that affect theentire system?’”

The give-and-take between biologists and engineers is fun, hesays. “There are many things we discover that they don’t know, butare necessary for an accurate simulation. For example, how long doesa memory B-cell live?Where does it divide? They don’t know. Andyou can’t track a memory B-cell around the body in vivo. Youmighthave a picture of the germinal center, but that’s just a snapshot.” Ger-minal centers are areas of activity that develop dynamically after theimmune system activates B-cells. And ultimately one has to be ableto understand these dynamics to get to the bottom of latent EBV in-fection.

“Together, we design experiments to get the information we

need. Then we see how the results map to what they already know…The neat thing about this is once we figure out how the germinalcenter works, we have a model that can be used for predictivepurposes. There is no way they can get that without oursimulations.”

The team is discussing similar models for other biological sys-tems, he says, such as modeling influenza infections of the lungs.They are in discussions with one group about developingmodels forH5N1, the bird-flu virus. Another potential project involves mod-eling the T-cell response to dioxin in the environment.

Plassmann says the hope is to develop a general simulationframework that can make big contributions to research in systemsbiology. In biology, for example, “the more we can do with simu-lation, the more we can help to find ways to improve people’shealth.”

13

Imagine a device that automatically detects demand forpower and delivers processed electricity in the required form(AC or DC) at the correct voltage and frequency. Energyrouters, analogous to the familiar data and communicationsrouters, could improve reliability, efficiency, and safety atevery level of the electrical power system.Ships, aircraft, and land vehicles could be made smaller,

lighter, and more fuel-efficient. Businesses, factories, and evenhomes could better use greener andmore efficient energy sourcesand loads, such as different equipment and appliances with vari-able-speed motors. The power grid could employ intelligent en-ergy routers to quickly isolate problems while still providingnecessary power.

Energy routers are still a concept, and their true realizations

may be years in the future, but they are a dream of many powerelectronics researchers today. According to Fei (Fred) Wang, anassociate professor of electrical engineering, at the heart of theconcept are the electronic power distribution systems promotedby the Center for Power Electronics (CPES), an NSF EngineeringResearch Center led by Virginia Tech.

“It’s a long-term goal, but in the end, your power electronicconverter will be an energy router,” he says.

Smaller, lighter high-power

Wang’s interest in the concept evolved from his work inhigh-power electronics—developing electronic circuits to con-vert and control electricity in systems that range from tens of kWto megawatts of power.

IS AN

ENERGY ROUTERin your future

Luis Arnedo (left)and Fei (Fred)

Wang in the HighPower Laboratoryat the Center for

Power ElectronicsSystems. Behindthem is a “plug-

and-play” modularconverter and be-

side them is aconverter system

testbed.

14

?

The team led by him and his colleagues at CPES is currently fo-cused on improving the power systems of jets, ships, land vehicles,mini-grids, and even oil rigs. The team is not only engineering high-density power converters, but also designing the architecture andconcepts that will help control and operate the high-power electri-cal systems of the future.

In high-power applications, the motivation for developinghigher density power converters springs from a drive toward greaterfuel efficiency, a desire to use less materials, and a need for increasedusable space,Wang says. For example, the team is workingwith Boe-ing on an advanced motor drive system for potential aircraft use.“The goal is that it be lightweight andcompact,” he says. “To make thesesystems smaller, wemust use advancedtechnology, like better semiconductorand magnetic materials, better capaci-tors, better cooling and packaging,new circuits, and new controllers.”

Part of the solution was usingadvanced silicon carbide (SiC)switches that are not yet commerciallyavailable. “SiC can operate at 300°C,whereas silicon can only work up to150°C. That is a big difference and means we need less cooling anda smaller heat sink.” SiC also has lower loss and allows very fastswitching enabling the design to reduce other components, such ascapacitors and inductors, he says. “Every component we eliminateequals a savings in space and weight.”

CPES is working on a variety of projects with other aircraftfirms, including Rolls Royce, General Electric (GE), and SAFRAN.Much of the focus involves technology for the more-electric aircraft.The aircraft retain their fuel engines, but from that point onward,the goal is for all the subsystems to be electrical. Power density andreliability are key requirements for such a system.

The U.S. Navy has similar goals for its ships. “The Office ofNaval Research (ONR) has been funding research on the all-electricship, for, among other things, the reduced size and weight,” Wangsays. “We are looking at everything after the generating plant. Thismeans we are trying to get rid of bulky, inefficient mechanical andhydraulic equipment. To do that, however, we need a family of highdensity power converters as large as 100MW.” Because the goals arealso to reduce size and weight, the CPES team is collaborating withindustry and national labs and studying the potential use of SiC de-vices for these converters, he says.

Transforming the entire system

Developing the individual electronic components—no matterhow large—is only part of the story, Wang says. “When we put all

these converters together, how canwemake sure theywork togetherin an optimal way?”

Conventional high-power systems employ generators, motors,transformers, and mechanical devices to process electricity, he says.“Now, we are talking about systems with a high penetration of elec-tronic power converters. Power converters have much faster dy-namics than traditional equipment. They are also highly nonlinear.This is good and bad.”

Electronic converters are intelligent and fast, but they generatea lot of noise. This can result in poor power quality and electro-magnetic noise interference. “We are now studying these system is-

sues. We need to understand theinteractions for systems to work in anoptimal way,” Wang says. Althoughthe research is still at an early stage, heis already talking about “getting rid ofthe bulky stuff,” like mechanical pro-tection devices and breakers.

CPES has conducted power ar-chitecture studies for server, data cen-ter, aircraft and ship power systems,and is currently investigating a powersystem for offshore oil drilling. “Tra-

ditionally, these oil rigs use pneumatic power, but they want to useelectric. How do you design that?”

Another project, for the NSF, involves designing a distributedgeneration microgrid for electric utilities. “Today the grid is slowand based on electromechanical inertia. There is so much inertia thatit doesn’t crash easily,” Wang explains. “Once we have electronicconverters, they will operate so fast that we will no longer have in-ertia in the system. We will maximize utilization, but we need tomake sure that while working fast, our system can also react fast.”

Whether for offshore drilling, vehicles, or power grids, thequestions can be as basic as whether to use AC or DC power, and atwhat voltage, he says. “The problem is that everybody is hand wav-ing, but nobody has an analytical tool.” CPES wants to develop afundamental tool that uses scientific evaluation and provides engi-neers with the ability tomodel, analyze, and control the power in themost efficient manner possible.

Once we have a raw form of electricity, how do we convert itto what we need?Wang and his CPES colleagues believe that futuresystems will be electronic power distribution systems, consisting ofthree types of power converters: source converters tied to the gen-erating source, distribution converters for power delivery, and loadconverters for the dedicated use. The ultimate answers to gettingthe right form of power when neededmay transform the whole sys-tem. One of those changes may be energy routers. Imagine lifewithout wall warts …

HIG

HPOW

EREN

GIN

EERING

15

“The problem isthat everybody ishand waving, butnobody has ananalytical tool.”

David Gagnon (EE ’07) came to Virginia Tech in 2002 in-tending to major in electrical engineering. But his di-verse interests kept complicating his academic life. Hewill graduate this May with a double major in EE andmusic and a minor in mathematics, with a load of re-search projects on his resume. He’s also principal cellist

with the New River Valley Symphony.For graduate school, Gagnon may add an-

other subject area to the mix. He’s being courtedby two graduate programs—architecture at Rens-selaer Polytechnic Institute and physical acousticsat the University of Texas at Austin.

“I love learning,” he says. “A lot of topics in-trigue me.”

Gagnon added the mathematics minor because heenjoys it. He added music because he loves it and hasplayed with the New River Valley Symphony for fiveyears. An honors student, Gagnon started pulling his in-terests together with the Horton Honors Scholarship,which financed a trip to study the acoustics of Europeanconcert halls. Attending performances in Salzburg, Austria’s2,179-seat Festspielhaus during Salzburg Festival trained hisear in a way that performance never had, Gagnon says.“When I’m on stage, I don’t get to hear a lot of what goes

on in the concert hall,” he says. “In Europe, I compared venues

RENAISSANCE MANof the acoustic

waveforms

David Gagnon (EE ’07) is also a music major and has amathematics minor. He is interested in music acoustics—analyzing and characterizing the sounds of music.

and noticed how the dimensions of the building effect sound clarityand echo.”

Now he’s thinking about a career designing concert halls andachieving good acoustics for various types of performances. His sen-ior honors thesis also involves music—analyzing and characterizingsound waves of music. He’s investigating how the sound changes asit bounces off materials of different absorptive values and howechoes show up in the audio signal. He’s also interested in charac-terizing music waveforms so music genres can be automatically cat-alogued and retrieved by computer.

“Our brains are so complex; it’s a challenging task to have thecomputer listen to notes and figure out what type of music it is,” hesays. “The beat histogram, for instance, can give a lot of informa-tion about the type of song. Classical music can be speedy or slow,and the beat is not strong; it produces a blurry histogram. A technosong, on the other hand, has a strong rhythm.”

Gagnon is also looking at the timbre of different musical in-struments—the tonal character and changes over time—and ana-lyzing the waveform. In one test, he will play his cello in a studiowith a friend on a euphonium, which has a different range, to de-termine whether his characterization program is only good for oneinstrument.

His engineering experience will have a major role in the studyas well. Gagnon is using MATLAB to analyze the wave file of themusic.

Gagnon is familiar with the data acquisition involved in researchprojects; he has participated in two separate REU (Research Expe-rience for Undergrads) programs under the auspices of the NationalScience Foundation. In the first, at Central Florida University, heworked with ultra fast lasers. Gagnon spent last summer at PennState, where he was involved with testing the conductivity of“graphene” thin carbon films, which have “cool, unusual proper-ties” and the ability to conduct electricity well.

“I’ve just picked things that interested me duringmy college ca-reer,” he says. “I have a lot of things I want to learn. That means I gofrom musical theory class over to the engineering building for an-tenna design. Electromagnetic waves are similar to those in acoustics.These things are very exciting to me.”

—By Su Clauson-Wicker

Gagnon spent the summerof 2005 visiting and sketchingthe acoustic environments ofEurope’s major concert halls.

Understandingovarian cancerdrug resistance

ECE researchers are collaborating with researchers from JohnsHopkins University School of Medicine to develop an inte-grated systems biology approach to study ovarian cancer drug

resistance. Ovarian cancer is the fifth leading cause of cancer deathamong women in the United States and up to 80 percent of patientsdiagnosed at advanced stages die within five years.

In many cases, the patients initially respond to chemotherapybut later relapse with recurrent tumors that are resistant to the ther-apy. The Johns Hopkins researchers have produced a cell line modelfor which they can reversibly turn on and off the cell’s drug resist-ance.

Stimulating this simple biological system with relevant inputsand observing the dynamic response of the protein levels within thecells will provide data that the ECE teamwill use to build models ofthe signaling and regulatory networks that are responsible for drugresistance. These models will then be validated experimentally andultimately used to more precisely target drugs to improve the long-term survival of patients.

The Computational Bioinformatics & Bio-imaging Laboratory,in collaboration with the medical team at theWake Forest Uni-versity School of Medicine, is working on a genome-wide as-

sociation study to identify an individual’s risk of cardiovasculardisease from the joint effects of multiple genetic factors.

Although there are a number of studies underway to collect ge-netic data from individuals with cardiovascular diseases, the analysisof these massive data sets to identify the key risk factors and theirinteractions is a critical but unsolved problem. The staggering num-ber of possible combinations of factors that must be considered, cou-pled with the probable nonlinear nature of these interactions, makes

this problem immensely complex.Using recent advances in machine learning, such as support-vec-

tor machines, conditional mutual information, and cluster comput-ing, Joseph Wang and his colleagues expect to make significantadvances over conventional techniques that rely on linear modelsand simple interaction terms. Ultimately, this work will enhance theability to predict, treat, and prevent cardiovascular disease.

The ability to identify individuals with a genetic predispositionfor disease will allow for more effective preventive interventions, in-cluding modifying environmental exposures that have been shownto interact with the genetic predisposition to increased disease risk.

BIOMEDICALAPPLICATIONS

FACULTY

Masoud AgahPeter AthanasWilliam BaumannAmy BellLouis BeexGary BrownMichael BuehrerRichard ClausWilliam DavisMark Jones

G.Q. LuTom MartinKathleen MeehanPaul PlassmannT.-C. PoonJoseph TrontAnbo WangYue (Joseph) WangChris Wyatt

The key risk factors in cardiovascular disease

Ovarian cancer cells dividing

CO

UR

TESY

CA

NC

ERM

EDIA

CEN

TR

E

18

Does changing white matter in the aging brain affect cognitivefunction? Do the effects of alcohol abuse on the brain beginearly in the abuse process? Does childhood exposure to stim-

ulants such as methylphenidate (Ritalin) predispose adolescents todeveloping substance abuse disorders?

These questions are being investigated using image analysis toolsdeveloped by researchers in the Bioimaging SystemsLaboratory. The laboratory is a joint effort betweenECE and the Virginia Tech/Wake Forest School ofBiomedical Engineering and Sciences to acceleratethe use of imaging and image analysis in biomedi-cine. Much of the work is done in collaborationwith clinical and basic science researchers at theWake Forest School of Medicine and the Virginia-MarylandRegional College of VeterinaryMedicine.

The nerve fibers that transmit signals long dis-tances within the brain constitute the white matter.As adults age, they can develop patchy white matterlesions, which are referred to as leukoaraiosis (LA).These lesions appear as hyperintensities—or higher contrast spots—in radiologic imaging. LA lesions create a problem for researchersstudying how the brain changes with age, as current image registra-tionmethods do not perform effectively in regions where tissue con-trast has changed.

ECE’s Chris Wyatt is developing a new registration approachthat can accommodate contrast changes and determine whether theLA changes in white matter correlate with changes in cognitive func-

tion. The project is funded by UCLA’s Center for ComputationalBiology (CCB) and the U.S. National Institutes of Health (NIH).

Investigations into the effects of alcohol and stimulants on thebrain are underway in conjunction with research atWake Forest; alsofunded by the NIH. The ECE team will apply cutting-edge imageanalysis techniques used in humans to primates, such as macaques or

Rhesus monkeys.The alcohol abuse study seeks to understand

whether the effects of abuse begin early in theprocess. Most of what is known of alcohol’s effectson the brain is based on studies of individuals whohave abused for a long time. Human studies are alsocomplicated by factors including non-alcohol drugabuse, poor nutritional states and other medical con-ditions.

Although most studies of children and adultswith ADHD have suggested that medication withmethylphenidate or amphetamine has either no ef-fect or can be protective from substance abuse, it is

impossible to distinguish between the effects of the stimulants andthe ADHD itself, according to Linda Porrino, professor of physiol-ogy and pharmacology atWake Forest. “Given that these stimulantscan produce dramatic changes in the brain’s dopamine system, thequestion of long-term adverse consequences still remains,” she says.The studies on non-human models are designed specifically to an-swer whether or not stimulant exposure can predispose adolescentsto developing substance abuse disorders.

The successful sequencing of the genome for humans and otherorganisms has provided a “parts list” for these cells, but howthese parts function and interact is largely unknown. Under-

standing these networks in human cells will help us better under-stand the nature of many diseases and point the way tomore effectivemedical interventions. In microbes, this knowledge can aid in engi-neering designer bugs that can help with waste remediation or en-ergy production.

Deducing the structure and dynamics ofthe biochemical networks that control life atthe cellular level is of the major goals of sys-tems biology. From an engineering perspec-tive, this work falls into the area known assystem identification.

What makes system identification of bi-ological systems so difficult is the sheer com-plexity of the systems, the fact that they arehighly nonlinear, and the difficulty of pro-viding inputs and measuring outputs of the

WINDOWS to the brain

Applying system identification to biologysystem. Researchers cannot just hook up a function generator andan oscilloscope, as they might to analyze a radio circuit.

ECE’sWilliam Baumann, is working in collaboration with JeanPeccoud of the Virginia BioInformatics Institute, John Tyson of bi-ology and JosephWang (ECE) onmethods to identify these systems.To understand the basics of this problem, they are measuring the re-sponses of synthetic gene networks engineered in bacteria. In this

case, where the network can be built up andmeasured step-by-step, it is possible to verifyour techniques and understand the limits ofour models. At the next level of complexity,they are looking at stochastic modeling of thecell cycle in yeast and the incorporation ofnew data from measurements of protein lev-els in single-cell organisms. At the highestlevel of complexity, they are investigating theuse of gene and protein expression data tomodel networks of genes involved in diseasessuch as breast cancer andmuscular dystrophy.

Image estimate of density

and direction of white mat-

ter connections in a brain.

The probability over time of an HIV re-sponse to a transcription stimulation.

19

Ever since the introduction of Field-Programmable Gate Array(FPGA) technology in the mid 80’s, people have imagined com-puters that could evolve and adapt to changing conditions with-

out requiring external assistance. That dream has come closer toreality as ECE researchers have demonstrated the first known in-stance of a system changing its own hardware in a non-trivial fash-ion while continuing to run.

A team in the Configurable Computing Laboratory (CCM) hasdemonstrated an embedded system that is able to interactively im-plement new circuits inside itself or remove or reconnect existingcircuits. The systemwas developed byNeil Steiner, a Bradley Fellowand Ph.D. candidate, working with his advisor, Peter Athanas.

By folding low-level design tools into a demonstration system,the team has made it possible for computers to participate in theirown design. “If our system finds defects within itself, it can simplyavoid using the damaged resources,” Steiner said. “This capability ismuch more cost-effective than throwing away large, but imperfectchips.”

Internal modeling and design tools also allow the system tofunction at a higher level of abstraction than previous FPGA sys-tems. Instead of requiring a configuration bitstream that supports a

specific FPGA model, the proof-of-concept system can accept cir-cuit netlists in the widely popular EDIF format. These circuits arethen placed, routed, configured, and connected to other circuits in-side the system, without interrupting its operation. The same EDIFnetlists can be used for any device in the same FPGA family, whichensures that designs remain more portable.

The current research falls within a proposed roadmap for au-tonomous computing systems—systems that are able to functionmore independently and assume responsibility for their own re-sources and operation, according to Steiner. The next step is to in-clude a synthesizer within the system, so that it can accept behavioralcircuit descriptions in high-level languages like VHDL or Verilog.“The most ambitious levels on the autonomy roadmap describe sys-tems that can adapt to changing conditions and even learn from ex-perience, a prospect that is not as far-fetched as it sounds,” he said.

Steiner credits his Bradley Fellowship and support fromXilinx,the leading FPGAmanufacturer, in enabling his research to push theenvelope to show what can be accomplished with current technol-ogy. “FPGAs have taken on important and well-deserved roles inmodern systems, but we have shown that the early dreams are still vi-able,” he said.

COMPUTERSYSTEMS

FACULTY

Lynn AbbottPeter AthanasVirgilio CentenoMohamed EltoweissyMichael HsiaoChao HuangMark JonesTom MartinLeyla NazhandaliJung-Min Park

Cameron PattersonJoAnn PaulPaul PlassmannBinoy RavindranPatrick SchaumontSandeep ShuklaJoseph TrontChris WyattYaling Yang

FPGAs can finally adapt without external control

20

ECE’s Configurable Computing Laboratory, directed by PeterAthanas, is part of a new NSF Center for High-PerformanceReconfigurable Computing (CHREC, pronounced “shreck”).The new center, headed by the University of Florida, includes

The George Washington University, Virginia Tech, and BrighamYoungUniversity. CHREC is the nation’s first multidisciplinary re-

search center in reconfig-urable high-performancecomputing and is part of theNSF Industry /University

Cooperative Research Centers (I/CURC) program.Including the university partners, about two dozen organiza-

tions comprise CHREC, ranging from the Air Force Research Lab-oratory to the Aerospace Corporation. Shawn Bohner of computerscience is the Virginia Tech lead and co-director of the Center. Join-ing Bohner and Athanas are computer science faculty membersWu-Chun Feng and Francis Quek.

ACpE team is creating middleware that allows software applica-tions to easily use a dynamically loaded library of functionsimplemented in FPGA hardware. The “Wires on Demand”

technology will enable greater use of FPGAs in software-definedradio (SDR) applications.

“FPGAs are ideally suited to the DSP requirements of softwareradios,” said Cameron Patterson, who, with Peter Athanas, serves asa faculty member on the project. “However, it is difficult to changeSDRwaveforms at run time,” he said. In current reconfigurable ap-proaches, the inter-module communication structure usually remainsfixed, which in software-defined radio applications, limits logic androuting resource efficiency.

Reconfigurable applications development is daunting for soft-ware radio designers, largely because inter-module communicationrequires low-level physical design, he said. “The designers must learnmore about the FPGA architecture and tools than they ever want to

know. Run-time reconfigurable application design would be mucheasier if module communication circuitry was automatically syn-thesized.” Routing should be optimized between modules as muchas within a module and avoid the use of buses, on-chip networks andcrossbar switches with their significant latency, area and power over-heads, he said.

With the team’s approach, choosing the module’s coordinatesand completing connections to other modules are run-time opera-tions. A library of relocatable partial bitstreams is created, which isused by the run-time system to complete application requests for in-stancing and connecting modules.

Bradley Fellow alumniMark Bucciero and Jonathan Graf, bothof Luna Innovations, are on the project, along with graduate stu-dents Kipp Bowen, Tim Dunham, Matt Shelburne, Jorge Suris, andBradley Fellow Justin Rice. The project is sponsored by the AirForce Research Laboratory.

ECE joins new NSFcenter for reconfig-urable computing

Wires on Demand: AIDING THE USE OF FPGAS IN SDR

e-Rug:The e-Textile Laboratory is incorpo-rating electronics into large surface areas, such asrugs and draperies. The rug prototype shown herewas made with readily available electronic yarns and isbeing used as a testbed for tracking human gait. Therug lights up when the sensors are activated and haspotential uses such as tracking people in low-visibilitysituations or directing traffic during evacuations.

e-Textile LaboratoryContacts: Mark Jones,Tom Martinwww.ccm.ece.vt.edu/etextiles/

Configurable ComputingLaboratoryDirector: Peter Athanaswww.ccm.ece.vt.edu

Real Time SystemsLaboratoryDirector: Binoy Ravindranwww.real-time.ece.vt.edu

21

According to research in ECE’s Laboratory for Advanced Net-working (LAN), the medium access control (MAC) protocolsof state-of-the-art wireless sensor networks (WSN) are sus-

ceptible to denial-of-sleep attacks that can reduce network lifetimefrom years to days.

“Like other networks, sensor networks are vulnerable to mali-cious attacks,” said Ph.D. candidate Lt. Col. David Raymond (U.S.Army), who conducted the study with his advisor, Scott Midkiff.“However, their limited resources make these devices particularlyvulnerable to denial-of-sleep attacks.”

The sensor platforms typically use 8-bit processors, less than 8kb of RAM, and 100 kb of program memory. “Not only are theymore sensitive, but because oftheir hardware simplicity, defensemechanisms designed for tradi-tional networks are not feasible,”he said. The team identified fea-tures of WSN MAC protocolsthat increase vulnerability to de-nial-of-sleep attacks, then ana-

NETWORKING

FACULTY

Luiz DaSilvaMohamed EltoweissyThomas HouYao LiangAllen MacKenzieScott MidkiffAmitabh MishraLeyla Nazhandali

Jung-Min ParkBinoy RavindranYaling Yang

Wireless sensor nets vulnerable to denial-of-sleeplyzed the impact of attacks tailored to these vulnerabilities.

“With full protocol knowledge and an ability to penetrate link-layer encryption, all WSN MAC protocols examined are suscepti-ble to a full domination attack, which reduces network lifetime tothe minimum possible by maximizing the power consumption ofthe nodes’ radio subsystem,” he said. “Even without the ability topenetrate encryption, subtle attacks can be launched that reduce net-work lifetime by orders of magnitude.”

The team is working to design and implement mechanisms toprevent or mitigate the impact of these attacks. As WSNs grow lessexpensive, the potential for widespread use in applications fromhealth monitoring to military sensing continues to rise, he said.

Motes—test platforms for low-power, limited resource systems—in the Laboratoryfor Advanced Networking (LAN) are used to test network vulnerabilities.

Complex Network andSystem Research GroupDirector: Thomas Houwww.ece.vt.edu/thou/CNSR

Laboratory for AdvancedNetworkingDirector: Luiz A. DaSilvawww.irean.vt.edu/lan

22

Advanced Research inInformation Assuranceand SecurityDirector: Jung-Min Parkwww.arias.ece.vt.edu

Researchers in electrical and computer engineering aredesigning routing protocols for airborne networks, atechnology that provides reliable communication be-

tween aircraft, unmanned aerial vehicles (UAVs), and otherairborne assets.

Airborne networks arise naturally in military scenar-ios, search-and-rescue operations, and even potentially incivil aviation. Ensuring that airborne platforms can com-municate effectively without necessarily relying on groundrelays is a major objective in this area.

Luiz DaSilva and graduate student Bo Fu are workingwith a team from SCA Technica, Inc. as part of an AirForce Research Lab SBIR project. The team is developingmechanisms that will overcome routing challenges in net-works of high-speed nodes through the use of locationawareness and disruption tolerant network techniques. In-spired by the idea of mesh networks, wireless networks ofground nodes that sometimes cover an entire city. They areproposing to opportunistically create a mesh in the sky,with underlying robust mechanisms that address the par-ticular challenges of airborne platforms.

InNovember, Virginia Tech is launching a first-of-its kind com-petition for students and researchers in mobile ad hoc net-working. Called the MANIAC Challenge (Mobile Ad Hoc

Networking Interoperability and Cooperation), the competitionwill be held in conjunction with IEEE’s Globecom 2007, Novem-ber 25-26 in Washington, D.C.

Competing teams will come together to form a large (esti-mated 30-50 node) ad hoc network. The organizers will create andsend traffic to each team. Teamswill be judged based on howmuchof the traffic destined to themmakes it through the network, how

Battery-sensingintrusion systemfor mobile computers

Using a dynamic threshold calculation algorithm, a CpE teamhas developed an attack-detection system for mobile comput-ers, which alerts users when battery power changes are detected

on handheld wireless devices, such as PDAs and smart phones.Hosts enabled with the battery-sensing intrusion protection sys-

tems (B-SIPS) are employed as sensors in a wireless network andform the basis of the intrusion-detection system (IDS). “When asmall mobile device is kept in a high activity state for extended pe-riods of time, the battery resources are depleted faster than normal,decreasing its charge life,” explained Timothy Buennemeyer, a Ph.D.candidate on the project. “Our system not only alerts the user, but

it also provides security administratorswith a nontraditional detection capabil-ity that is scalable and complementaryin a network environment with existingcommercial and open system IDSs.” Ir-regular and attack activity is detectedand reported to the intrusion detectionengine for correlation with existing sig-natures and further investigation.

An analytic model was developedto examine the smart battery character-istics to support the theoretical intru-sion-detection limits and capabilities of

B-SIPS. The dynamic polling rate algorithm allowed the smart bat-tery to gauge the network’s illicit attack density and adjust its pollingrate to more efficiently detect attacks and conserve battery charge.

The team is currently developing a device-profiling and attackmatching capability, and a usability study is being conducted. Buen-nemeyer is working on the project with fellow graduate studentTheresa Nelson, Professor Joe Tront, and Randy Marchany, direc-tor of Virginia Tech’s IT Security Laboratory.

1st-ever ad hoc networking contest set

23

little energy they consume in forwarding traffic, and a subjectiveevaluation of the quality of their solution’s design. To get theirtraffic across the network, each team must rely on other teams’willingness to forward traffic for them, according to Luiz DaSilva,who with Allen MacKenzie is organizing the project with fund-ing from the NSF.

The two-in-one effort is aimed at meeting educational and re-search goals of improving network throughput, deepening under-standing of overall network behavior, and motivating students inthe field. For more information, see www.maniacchallenge.org.

AIRBORNE NETWORKSCreating a mesh in the sky

TIM

BUEN

NEM

EYER

U.S.

Air

Force

Technology developed in ECE laboratories is serving as thefoundation of a new effort to improve flood and drought fore-casting in the Ohio River Basin.“Floods and droughts are twomajor natural hazards in the Ohio

River Basin, and they have major impacts on the region’s agricul-ture, industries, commercial navigation, and residential communi-ties,” said Yao Liang, an assistant ECE professor at the AdvancedResearch Institute (ARI) and principal investigator (PI) on the ef-fort. “There are data andmodels available from different sources anddifferent systems, that, if integrated, can significantly improve fore-casting accuracy and help disaster management.”

The integration effort stretches across universities and govern-ment organizations, including the National Weather Service (co-PIThomas Adams is a Virginia Tech alumnus), George Mason Uni-versity, the University of Pitts-burgh, and NASA GoddardEarth Sciences Data Informationand Services Center.

The team is developing asystem to integrate soil moisturedata from NASA satellites andNASA-NOAA land surface mod-els, and a spatial data assimilation

framework recently developed at the University of Pittsburgh intothe National Weather Service River Forecast System. Surface soilmoisture data from multiple satellites in conjunction with theNASA-NOAAmodels and the data assimilation framework will sig-nificantly improve the evapotranspiration rate calculation, whichplays a critical role in the river forecast system. The integration willbe achieved by extending the hydrological integrated data environ-ment (HIDE) developed by Liang and Nimmy Ravindran.

“Our ultimate goal is to enable the National Weather Serviceto seamlessly avail itself of the soil moisture data, that was previ-ously unavailable,” Liang said. The project results are expected to beextendable to the national level, via the adoption of the system byriver forecast centers at both national and regional levels. The$860,499 project is funded by NASA.

SOFTWARE& MACHINEINTELLIGENCE

FACULTY

Lynn AbbottRobert BroadwaterMark JonesPushkin KachrooYao LiangCameron PattersonPaul PlassmannBinoy RavindranSandeep Shukla

Yue WangChris WyattJason Xuan

Hydro-informatics for floods, droughts

BioImaging SystemsLaboratoryDirector: Chris Wyattwww.bsl.ece.vt.edu

Computational Bioinfor-matics & Bioimaging LabDirector: Yue Wangwww.cbil.ece.vt.edu

Computer VisionLaboratoryDirector: Lynn Abbotthttp://vision.ece.vt.edu/

24

Forecasting of flooding and droughts will be improved (right) by integrating remote sensinginformation (left) with land surface models through the Virginia Tech HIDE data environment (center).

Most people think of fingerprints, faces, and gait and iris pat-terns when they hear of automatic human recognition sys-tems. ECE researchers, however, are investigating using ears

for this biometric.“Ear-based recognition is of particular interest because it is non-

invasive and because it is not affected by environmental factors suchas mood, health, and clothing,” according toMohamed Saleh, a stu-dent in Virginia Tech’s graduate program in the Middle East andNorth Africa (VT-MENA).

“The appearance of the outer ear is relatively unaffected byaging, making it better suited for long-term identification whencompared to other non-invasive techniques, such as face recogni-tion,” he explains.

Surprisingly, ears have a different appearance for every individ-ual—even for “identical” twins. Ear recognition has not receivedmuch attention, he says, but other research teams had indicated that

it may be at least comparable to face recognition, at about 71 percentaccuracy. Saleh and his advisor, Lynn Abbott, recently appliedimage-based classifiers and feature-extraction methods to a smalldataset of ear images with strongly positive results.

The team used six small (50x40 pixels) grayscale im-ages for 17 individuals, each image showing different lumi-nance and orientation. Comparing seven different imagerecognition techniques, they achieved accuracies rangingfrom 76 percent to 94 percent.

“We believe ear recognition has significant potential,”Saleh says. Further research needs to be done in the area,using larger datasets, and considering problems of interfer-ing hair, clothing, jewelry, eyeglasses and other artifacts, he says. Forfuture research, the team is considering “multimodal” approachesfor recognition that involve ears and faces simultaneously.

Many applications today would benefit from an ability to pickout human faces automatically among other elements in acolor image. These include systems that perform face recog-

nition, systems that track faces for videoconferencing or human-computer interaction, and surveillance systems. Unfortunately, bothface detection and face recognition are very difficult computationalproblems, although the human visual sys-temmakes it seem easy. The challenges in-clude variations in illumination, size andorientation changes, occlusion, and skincolor.

An ECE team has developed a newtechnique that detects faces based on vari-ations of skin tone, as well as shapes ofskin-colored portions of an image. Thesystem does not explicitly attempt to lo-cate face-related features such as noses and eyes. The system relies onan unusual combination of two techniques, known as the DiscreteCosine Transform and a type of artificial neural network known asa Self-Organizing Map.

The technique has dramatically lower computational require-ments than more conventional techniques and is well suited for low-cost, real-time hardware implementation, according to Abdallah S.Abdallah, a doctoral student in the VT-MENA program. (VT-MENA is Virginia Tech’s graduate program in the Middle East andNorth Africa.) Abdallah is working with ECE’s Lynn Abbott andMohamad Abou El-Nasr of the Arab Academy for Science and Tech-nology in Alexandria, Egypt.

Abdallah notes that “even though different people have differ-ent skin colors, studies have demonstrated that intensity, rather than

chrominance (color) is the main distinguishing characteristic.” Usingsegmentation based on skin color eliminates the need for multireso-lution image pyramids, which are used in more traditional face-de-tection standard methods. “Most of the computational load inprevious techniques comes from the large sizes of the feature vec-tors being used,” Abdallah says.

Given an intensity representation (agrayscale, or “black-and-white” image),the technique relies on steps of regionanalysis, feature extraction, and patternrecognition. If faces are present in theimage, a self-organizing map (which hasbeen trained in advance) attempts to lo-cate each one.

For benchmarking and testing, theteam compiled a new image database,

named the VT-AAST image database, consisting of 286 images, con-taining 1027 faces. The database is currently available on-line fornon-commercial use at http://filebox.vt.edu/users/yarab/VT-AASTDatabase.htm

“Several large image databases are widely available for humanface recognition,” Abdallah says, “but relatively few are available forface detection.” The main difference is that most face-recognitiondatabases contain images that closely resemble “mug-shot” photo-graphs, with the subject facing the camera and centered in the image.More general face-detection systemsmust deal with individuals whodo not face the camera, who may be at various distances from thecamera, andmay be partially occluded. Each image in the VT-AASTdatabase is available in four formats, including the original color andimages segmented by skin color.

Finding Faces at Any Angle

Your ears can give you away

25

The Virginia Tech VLSI for Telecommunications (VTVT) grouphas developed a prototype system to monitor the health of in-frastructures, from buildings and railroads to spacecraft. The

prototype, which is the first all-digital signal processing system, re-duces power consumption by 80 percent of that of its predecessorsand paves a way for a self-contained monitoring system.

“We want to be able to diagnose damage to infrastructures andverify the repair efficacy,” said ECE’s Dong Ha, director of theVTVT group. “The single most important aspect of developing aself-contained monitoring system is low-power dissipation. The sys-tem should operate on energy harvested from the ambient such assolar, thermal, or vibration.We need to squeeze out every drop to re-duce the overall power dissipation ranging from DSP algorithms tothe interface between a sensor and a digital signal processing chip.”

Ha explained that, in typical monitoring systems, sinusoidaltones are used to excite the structure, then the responses in the formof electrical impedance are measured and processed. His team, in-cluding graduate students Jina Kim (CPE) and Ben Grisso (ME), re-duced power consumption by using digital pulse trains rather thansinusoidal tones and by observing only the polarity of the responsesignals. Using digital signals on both the structural excitation andthe sensing allowed the team to eliminate both power-hungry digi-tal-to-analog and analog-to-digital converters. It also simplified sig-nal processing of the response signals, Ha said.

The interdisciplinary team from the VTVT lab andmechanicalengineers from the Center for Intelligent Material Systems andStructures is supported by NSF on the effort. VTVT is part of theCenter for Embedded Systems for Critical Applications (CESCA).

VLSI & DESIGNAUTOMATION

FACULTY

Peter AthanasDong HaMichael HsiaoChao HuangTom MartinPatrick SchaumontSandeep ShuklaJoseph Tront

VT VLSI for Telecommuni-cations (VTVT)Director:Dong Hawww.ee.vt.edu/~ha/research/research.html

PROACTIVEDirector: Michael Hsiaowww.proactive.vt.edu

FERMATDirector: Sandeep Shuklahttp://fermat.ece.vt.edu

Center for EmbeddedSystems in CriticalApplications (CESCA)Director: Dong Hawww.cesca.centers.vt.edu

All-digital monitorfor structural health

Jina Kim demonstratesthe size and functionof an all-digital healthmonitor for structures.

26

While scientists and engineers view the promise of greaterperformance, efficiency, and smaller size from nano-scaleelectronics, a group of computer engineers see the chal-

lenges of higher manufacturing defect rates, more frequent faultincidence and greater susceptibility to variational effects. They areinvestigating nano computing: how to build complex processors orapplication-specific computing architectures with nano scale com-ponents and materials that may be fault-prone or defective.

“People are building quantum dot cellular automata, molec-ular switches and fabrics and nano tube-based transistors/inter-connects and discovering there are issues with reliability, quality,and performance,” says Sandeep Shukla, an assistant professor ofcomputer engineering. “Silicon devices reaching 45nm and beyondare experiencing variational effects, which lead to unpredictabletiming, capacitance, resistance and other factors. Manufacturingtechniques are imprecise at that scale and we are seeing high defectrates. Moreover, with very thin oxide layers, these devices are sus-

ceptible to atmospheric radia-tion,” he explains.

“How can we efficientlymap logic onto an array of quan-tum dot cellular automata so thelogic uses minimum resources,bypasses defective cells andworks correctly in the presenceof radiation that might inadver-

tently change the states of the quantum cells involved?” he asks.Molecular fabrics introduce issues as well. “Many researchers

believe that in the near future, hybrid structures will dominatecomputing. How do we interface CMOS with nanofabric; microstructures with nanostructures? There will be more nanowires inthe memory than micro addressing lines and, hence, a problemwith interfacing,” he adds.

Nanotube-based fabrics will also affect software and may re-quire spatial programming rather than temporal view of pro-gramming today, he says. Also probabilistic computing willpossibly replace deterministic computing of today. “How do youwrite code that can harness the massive parallelism afforded by bil-lions of switches?”

Some of the technologies under development are speculativeand may not be feasible, but the industry must be prepared to dealwith the issues involved, he says. The issue is growing in impor-tance as evidenced by the growing number of workshops studyingnano computing.

This spring, Shukla co-chaired a workshop on nano comput-ing at Europe’s biggest conference onDesign Automation and Test(DATE 07). He is also organizing a workshop on nano electronicsat the NSF in September to help NSF assess the importance of in-vesting in the area. Last May, Shukla organized a workshop for re-searchers fromVirginia sponsored byNSF, Virginia Tech, VirginiaCommonwealth University, and the IEEE Joint Chapter on Com-puter/Industrial Electronics and Control.

Computer engineering researchers, led by Michael Hsiao andJung-Min Park are applying their verification, testing, andnetwork security expertise to a system that can help protect

the online privacy of children.The system involves using a trusted third-party server, where

parents register and allow different levels of access. The computerengineering team is developing the prototype and verifying thatthe overall system is secure, in order to protect children’s onlineprivacy. They are working with researchers in the business collegewho are providing legal and focus group input and assessing thesystem with users.

The verification challenges in the system spring from the verylarge state spaces involved, said Hsiao. “One of the key steps is rep-resentation of the system during the state space traversal to avoidstate space explosion.” The team is developing methods that areless vulnerable tomemory explosion, such as automatic test patterngeneration and satisfiability solvers.

Applying CPE skills to protect children’s privacy

Nano electronics bring mega design challenges

Secure EmbeddedSystems GroupDirector:Patrick Schaumontwww.ece.vt.edu/schaum/research.html

27

An ECE research team is investigating UWB-based wireless sen-sor network technology to boost the safety of mining and im-prove rescue operations during disasters. “Mining continues to

be one of the most dangerous professions due to the possibility ofmine accidents,” said Haris Volos, a graduate student on the project.

“We hope to develop the technology that would provide robustwireless notification and coordination under normal operations andduring disasters, plus enable rescue teams to track the movements ofminers and locate them if they are disabled and unable tocommunicate.”The team’s goal is a wireless sensor networkspecifically designed for a mine environment. UWBwas selected asthe base technology due to its ability to simultaneously provide both

communications and position location data. Initial efforts involvecharacterizing UWB propagation in mines. “Relatively littleinformation exists on how electromagnetic signals—particularlyUWB signals—propagate in a mine environment,” Volos said.Measurement work has begun at the Kimballton limestone mine inwestern Virginia, where the team took bothUWB and narrow-bandpath loss measurements in tunnels in addition to grid measurementsin a room-and-pillar environment.

Volos is working with fellow graduate student Chris Headley;Chris Anderson, a post-doctoral associate; ECE faculty membersMichael Buehrer and Claudio da Silva; and Antonio Nieto of min-ing and minerals engineering.

COMMUNICATIONS

FACULTY

WirelessCharles BostianMichael BuehrerClaudio da SilvaAllen MacKenzieTim PrattWilliam TranterAmir Zaghloul

Fiber CommIra Jacobs

DSPLouis BeexAmy BellLamine MiliJeffrey Reed

NetworksLuiz DaSilvaThomas HouYao LiangScott Midkiff

Amitabh MishraJung-Min ParkYaling Yang

VLSI CircuitsPeter AthanasDong HaMichael Hsiao

OpticsT.-C. Poon

PropagationGary BrownWilliam DavisSteven EllingsonSanjay RamanSedki RiadAhmadSafaai-Jazi

Wireless@Virginia Tech is one of the largest wireless researchgroups in the United States. The center brings together 27faculty members and more than 100 graduate students, in

efforts ranging from circuits to networks. W@VT incorporateseight centers, groups, and laboratories, including the Center forWireless Telecommunications, theMobile and Portable Radio Re-search Group (MPRG) and the Virginia Tech Antenna Group.

W@VT one of largest wireless groups

Adapting UWB wireless to mine safety

From left: Chris Anderson (Ph.D. ’06), Chris Headley and Haris Volos preparepropagation measurements at the Kimballton limestone mine.

Wireless@Virginia TechDirector: Jeffrey Reedwww.wireless.vt.edu

Digital Signal Processing& Communications LabDirector: Amy Bellwww.ece.vt.edu/fac_support/DSPCL

28

Ateam led by Amir Zaghloul, an ECE professor, is developing alow-power system for randomly located sensors to communi-cate with satellites. The sensors can be dropped—and even

camouflaged—in uncontrolled or unfriendly environments, mak-ing the technique appropriate for disaster situations as well as home-land security, counter-terrorism and combat.

Each sensor in the system operates independently andmay relaythe same information, which creates a built-in redundancy. “Dam-age or discovery of the any of the sensor package nodes will not af-fect the transfer of the required information,” he says.

The sensors use scanning antennas with hemispherical coverageand use small phased arrays with relatively high gain and narrowbeamwidth. The array beam is continuously scanned within the ex-pected angular range of the satellites, with a pre-set dwell time ateach scanning position. The coded sensor data is transmitted in allscanning directions.

A geostationary (GEO) or a moving low-earth-orbiting (LEO)satellite then receives the data when the array beam is in the corre-sponding direction. “This data acquisition by the satellite is per-formed without the need for satellite tracking by the sensor,”Zaghloul explains. Low-gain hemispherical-beam antennas can alsobe used for lower frequency operation.

The low-power sensor and antenna are housed in a miniaturepackage along with a sensor processor to translate the sensor infor-mation to the RF package. “Initial calculations indicate that, at an in-formation rate of 100 bps, a 3x3 cm phased array of the sensors canlink to a LEO satellite at 1400 km altitude with 36 W of RF powerat 30 GHZ, with a 9 dB link margin,” he says. The low-power sys-tem is in the milliwatt range for medium data rates.

Low-power satellite/sensor systemcan operate with randomly located packages

An ECE research team has been awarded a $400,000 Cyber Trustgrant from the National Science Foundation (NSF) to improvethe security of software-defined radio (SDR) technology.Although software radio technology promises to alleviate the

spectrum shortage problem and improve spectrum utilization, itraises new security issues, according to Jung-Min Park, principal in-vestigator, and co-principal investigators Thomas Hou and JeffreyReed. “The software may be vulnerable to failure andmalicious tam-pering,” Park said.

“Changes to conventional ASIC (application specific integratedcircuit) devices requires technical skills and specialized equipment,which makes unauthorized changes very difficult,” he said. WithSDR, user services or RF parameters can be reconfigured via thesoftware. SDR research and standardization efforts focus on cryp-tosystems securing the download process and preventing tamperingof downloaded software.

“That is the first line of defense,” Park said. “We are exploringother measures to secure SDR devices from malicious hackers andmilitary opponents seeking tactical advantages.”

The team is investigating security vulnerabilities in the physicaland MAC layers of SDR networks.

Securing SDR devices from hackers, opponentsIncumbent users

Spuriousinter–cellbeacons

33 – 100Km

802.22 BS1

CPE3

CPE2

CPE1 802.22 BS2

Incumbenttransmitter

CMAC

WRAN 1 WRAN 2

DSSreport

Beacon report

Spurious Beacon attack: Amalicious CPE sends spuriousinter-cell beacons to neighboring802.22 entities.

SECURITYTHREATSIN

COGNITIVERADIO

NETWORKS

29

ELECTRO-MAGNETICS

Bob Clauer wants to understand the global electrical current sys-tems that are driven by the solar wind/magnetosphere inter-action. He is currently involved in establishing an ionospheric

HF radar at Virginia Tech to measure ionospheric plasma convec-tion—measurements that can be used to deduce the electric field ofthe ionosphere.

He is also developing an autonomous magnetic field measure-ment platform for use in remote regions of the Antarctic. In De-

cember (the Antarctic summer), the system will be redeployed fromthe South Pole where it has been tested for the past year, onto theAntarctic Plateau. The system will be the start of a magnetometerchain along the 40-degree magnetic meridian that will be magneti-cally conjugate (reciprocal) to the Greenland west coast chain ofmagnetometers. Using data from both hemispheres, it will be possi-ble to begin to map the global electrical current systems that aredriven by the interaction of solar wind with the magnetosphere.

FACULTY

Scott BaileyCharles BostianGary BrownC. Robert ClauerRichard ClausWilliam DavisSteven EllingsonLouis Guido

Ira JacobsBrent LedvinaKathleen MeehanHardus OdendaalT.-C. PoonRanjay RamanSedki RiadAhmad Safaai-Jazi

Wayne ScalesAnbo WangYong XuAmir Zaghloul

ECE’s Scott Bailey is serving as deputy principal investigator forNASA’s AIM mission, which launches in April. AIM is a two-year mission to study polar mesospheric clouds, which form an

icy membrane 50 miles above the ground at the edge of space. Theclouds, when viewed from the ground, are seen as very bright cloudsreflecting the sun at twilight and are called noctilucent clouds.

The clouds have been seen for more than a century in springand summer at high latitudes. The mission’s primary goal is to ex-plain why the clouds form and discover what is causing them to ap-

pear more frequently and at lower latitudes.The AIM satellite is 55 inches tall, 43 inches wide and weighs

430 pounds. The satellite carries three instruments: Cloud Imagingand Particle Size (CIPS); Solar Occultation For Ice Experiment(SOFIE), which measures the thermal and chemical environment in-side the clouds; and the Cosmic Dust Experiment (CDE), whichmeasures the input of cosmic dust into the atmosphere.

James M. Russell III of Hampton University serves as the pro-ject’s principal investigator.

AIM satellite to study noctilucent clouds

Mapping the systems driven by solar winds

Scott Bailey is deputy principal investigator of NASA’s AIM mission. Right: At South Pole Station—testing a prototype ofan instrument that will help map the electrical current systems that are driven by solar wind and the magnetosphere.

30

NA

SA

ETA telescope coming near Tech

In 2006, an ECE/physics research team, including Steven Ellingsonand Cameron Patterson, began operation of a low-cost, direct-sampling radio telescope in a rural mountainous region of North

Carolina. Called the ETA (Eight-meter-wavelength Transient Array),the telescope was designed for unattended operation to monitor theskies for single, dispersed radio pulses associated with prompt emis-sion from gamma ray bursts and exploding primordial black holes.The telescope analyzes signals from 24 dipole antennas using anarrangement of 28 FPGA boards and four PCs interconnected withcustom I/O adapter boards. Thanks to additional funding from thephysics department, a second, portable instrument is planned nearthe Tech campus to improve performance using “anticoincidence”—a technique in which simultaneous detections are required at bothsites as a means to rule out local radio frequency interference.

Sapphire temperature sensorstands up to harsh environment

Aminiature fiber optic sensing probe has proven its mettle in theharsh industrial environment of a coal gasifier at Tampa Elec-tric’s Polk Power Station. The sapphire-wafer-based fiber optic

temperature sensor was installed in late May 2006 and continuedoperation for sevenmonths—far exceeding the project objective of45 days. During that time, all thermocouples installed in the gasifierwere replaced.

In a coal gasifier, most of the fuel is not burned, but chemi-cally broken down by temperature and pressure in a limited oxygenenvironment, producing syngas. The syngas is cleaned and used asfuel in a combustion turbine. Accurate, reliable temperature meas-urements are critical: operating too high will reduce conversion ef-ficiency and shorten the refractory life, while operating too lowwill cause the molten slag to become viscous, plugging the tap-hole.

The sapphire temperature sensor, developed by the Center forPhotonics Technology, provided continuous, real-time tempera-ture data as high as 1392°C. The research program is funded by theNational Energy Technology Laboratory of the U.S. Departmentof Energy. A second field test is underway in 2007.

High-res, 3D imagingfor asphalt aggregates

Work is underway at the Center for Photonics Technology todevelop a new laser imaging system for small rocks and sandparticles. Sponsored by the National Cooperative Highway

Research Program (NCHRP), the three-year project is geared to-wards improving the analytical models of rocks and small particlesin asphalt aggregates.

A special mounting system holds particles ranging from 50mmto 75µm in diameter, allowing the user to position them for view-ing at any angle. The object is illuminated by a series of interfer-ence fringes, generated by a red visible laser and an end-polishedfiber optic coupler. These fringes, when reflected off the face of theobject, contain quantitative information about the surface heightprofile. The image itself is collected with a high-resolution digitalcamera. The camera uses a thermoelectric cooling device to coolthe CCD pixel array for extremely low noise operation. This lownoise results in increased measurement resolution. The system isprojected to achieve better than 5µm resolution in all three imagingdirections.

Center for PhotonicsTechnologyDirector: Anbo Wangwww.ee.vt.edu/~photonics/

Fiber & Electro-Optics Re-search CenterDirector: Richard Clauswww.ee.vt.edu/~feorc/

Dusty Plasma LaboratoryDirector: Wayne Scales

Time Domain & RF Meas-urement LaboratoryDirector: Ahmad Safaai-Jaziwww.ee.vt.edu/~tdl

ElectroMagneticInteractions LaboratoryDirector:Gary Brownwww.ee.vt.edu/~randem/emil/emil

Virginia Tech AntennaGroupDirector: William Davishttp://antenna.ece.vt.edu

Measuring an ozone destroyer

ECE researchers are studying high altitude nitric oxide, which isa catalytic destroyer of ozone, in the first-ever effort to meas-ure the chemical in the polar night. Scott Bailey is collaborat-

ing with Chris Hall of aerospace and ocean engineering on the$1.4-million NASA project, which will launch in 2010. The instru-ment will consist of a very large telescope and will view a bright UVstar in occultation. Bailey and Hall are working with researchersfrom the University of Colorado on the project.

Bradley Fellow Evan Lally sets astone on a prototype imagingsystem for asphalt aggregates.

31

ELECTRONICS

FACULTY

Masoud AgahDushan BoroyevichRick ClausWilliam DavisLouis GuidoRobert HendricksChao HuangJason LaiFred C. Lee

G.Q. LuKathleen MeehanKhai NgoHardus OdendaalRanjay RamanAnbo WangFred WangMing Xu

Sedki RiadAmir ZaghloulKwa-Sur TamKrishnan RamuDouglas LindnerJoseph TrontYong Xu

ECE opened its new Micron Technology Semiconductor Pro-cessing Laboratory this fall. The Class 1,000/10,000 clean-roomlaboratory is the latest of the new facilities operated by Virginia

Tech’s multidisciplinaryMicroelectronics, Optoelectronics andNan-otechnology (MicrON) Group.

The laboratory occupies 1800 square feet in Whittemore Halland builds on a teaching laboratory that opened in 2001. New or im-proved capabilities include state-of-the-art photolithography, wetprocessing, dry etching (RIE/DRIE), chemical vapor deposition, andmetallization. This gives Virginia Tech researchers on-site capabili-ties to produce cutting-edge prototypes for nano-structured biolog-ical and chemical sensors, organic and molecular nanoelectronics,

solid-state lighting devices, microelectromechanical systems (MEMS)for sensing and communications, microfluidics for on-chip biologi-cal assays and cooling of high-power electronic circuits, and ad-vanced chip-level packaging strategies.

The research and teaching laboratory is named for MicronTechnology, in recognition of the firm’s $750,000 gift that providedthe final funds that allowed the laboratory’s completion. The clean-room joins several other of Tech’s MicrONGroup facilities: a metal-organic chemical vapor depositions (MOCVD) laboratory, anelectronic/optoelectronic materials laboratory, and a device charac-terization laboratory. An electronic packaging laboratory is alsoavailable.

Semiconductor Processing Lab Dedicated

Future EnergyElectronics CenterDirector: Jason Laiwww.feec.ece.vt.edu

Center for PowerElectronics SystemsDirector: Fred C. Leewww.cpes.vt.edu

Microelectronics, Opto-electronics and Nanotech-nology Group (MicrON)Director: Louis Guidowww.micron.ece.vt.edu

Motion Control SystemsResearch GroupDirector: Krishnan Ramu

32

Bradley Fellow Mark Lehne is investigating new analog-to-digital conversion-circuit architectures that improvea broadband wireless transceiver’s ability to efficientlydetect information in the presence of strong interfer-ence. The goal is better overall performance whileallowing for higher data rates and more intelligenttransceivers. In order to coexist with legacy narrow-band transceivers, high-speed analog-to-digital con-verters must allow the receiver to operate with sufficientfidelity to detect target signals while identifying andrejecting narrow-band interference.

A “black box”system of modeling

Researchers in the Center for Power Electronics (CPES) have de-veloped a “black-box” modeling approach for designing newpower electronics-based electric energy distribution systems.

The new method diverges from the conventional “white-box” ap-proach, which requires detailed knowledge of the converter struc-ture and parameters. However, the growing use of power electronicsin distribution systems has complicated design: the size and com-plexity of future systems alone make white-box modeling difficult.Another complication is that converters are commercially availablefrom numerous different vendors, who provide little or no infor-mation about the internal structure of their products.

The CPES approach is based entirely on the terminal charac-teristics of power converters, using a simple set of small-signal meas-urements to construct the low-frequency model without requiringany knowledge of the converter structure or any of its internal pa-rameters. Called the black-box modular-terminal-behavioral mod-eling methodology, the model captures the information withoutneglecting parasitics or simplifying the converter internal operation,as is common with conventional reduced-order models. The CPESapproach is modular, enabling computations efficient analysis of allelectrical phenomena in the system.

Power electronics researchers from around the world are collab-orating on how to focus efforts to ensure that power electron-ics helps industries and countries globally to reduce energy

consumption and environmental impact.Power electronics has emerged as an enabling technology in this

global effort.

As one of the largest power electronics research groups in theworld, CPES has taken the lead on gathering academic and indus-trial researchers to develop a roadmap for development. CPES hasorganized two annual roadmap workshops, the most recent duringOctober 2006. The roadmap effort is funded by an NSF supple-mental grant.

To facilitate communications in remote locations, CPES re-searchers have developed a self-contained, 3 kW, autonomoushybrid power system for data communications in remote loca-

tions. The testbed system serves as a demonstration of an electricpower system for future self-sustained, zero-emission applications.The scaled-down generic electronic power distribution system is rep-resentative of systems that could be used for future homes, data cen-ters, hybrid electric cars, aircraft, and ships.

The testbed incorporates representative renewable sources,power converter-based loads, and a grid-interfaced controllable dis-

tribution network, in order to evaluate system-level impacts of CPEStechnologies.

As implemented in the CPES laboratory, the testbed includes asolar photovoltaic source simulator and AC grid connection as en-ergy sources, lead-acid batteries, a battery charger, an AC transferswitch, and DC-AC inverter, andmixed AC-DC power distributionwith computer-controlled measurement and supervision as the en-ergy management system.

The system is designed to provide uninterrupted energy sup-ply to the data servers and air-conditioning electrical loads.

Testbed developed for self-sustained,zero-emissions power system

Mapping power electronics development

33

High-speed analog-to-digital converters

Higher energy costs and recent blackouts in North America andother systems around the world have spawned an increased em-phasis on protection systems.Wide area measurement concepts

and devices developed at Virginia Tech are being applied worldwideto reduce catastrophic failures of the power grids, limit the regionsaffected by such events, and to increase the speed of restoration aftersuch an event.

In one project, researchers are using real-time wide area meas-urements of the California power system to determine optimumprotection policies and settings for critically located relaying sys-

tems. The goal of the three-year development and demonstrationproject is to improve protection system supervision by making itadaptive to the prevailing state of the system.

Towards this goal, the team is developing techniques for adap-tive adjustment of dependability and security, potential load en-croachment alarm systems, andmore-intelligent out-of-step relayingfunctions.

The team will determine key locations on the California gridwhere an insecure relaying operation during stressed system condi-tions would be detrimental to the viability of the power system. Theteamwill also develop a method for using real-time wide area meas-urement data and the existing protection system data to determinewhich of the relay characteristics are in danger of being encroachedupon during a catastrophic event. Appropriate countermeasures willbe developed for those vulnerabilities.

Finally, the team will develop a technique to use wide-areameasurements to improve out-of-step decision-making at key loca-tions where out-of-step blocking and tripping functions are used.Out-of-step relays are traditionally set based on transient stabilitystudies performed for assumed base case and contingency conditions.However, in practice, the power system and the actual complex se-quence of events differ from the study cases, and the protection set-tings in use are not appropriate to the changed system conditions.

POWER

FACULTY

Robert BroadwaterVirgilio CentenoJaime de la ReeYilu LiuLamine MiliSaifur RahmanKwa-Sur TamJames Thorp

Improving California’s power system

34

Dawei Fan (left) andJerel Cullis test thedynamic responseof several commer-cial PMUs.

Virginia Tech’s wide area measurement technologies are gainingpopularity worldwide, with researchers making presentationslast year in Taiwan, China, India, Sweden and France. ECE

teams are working with the Power Grid in India and the NationalSystemOperation Center in Brazil to aid in developing optimal widearea measurement systems based on the phasor measurement tech-nology (PMU) developed at Virginia Tech. The Indian PMU systemis being designed to improve power system monitoring and to con-trol their rapidly growing system. The Brazilian system seeks tomonitor and control their system to take full advantage of hydrogeneration and reduce the effect of regional weather on energy gen-eration. ECE researchers are working to determine the optimal place-ment and deployment schedule to obtain the greatest benefit fromthe measurement system at every stage of implementation.

The increased demand, availability, and applications of wide area measurement devices have created the need for proper testing andstandardization. Virginia Tech power engineering researchers are collaborating with the National Institute of Standards and Testing(NIST) to develop procedures and testing devices that will allow researchers, users, and manufacturers to document, evaluate and

compare the dynamic response of the different wide area measurement devices on the market. In addition to the NIST work, graduate stu-dents fromVirginia Tech andOtto-Von-Guericke University in Germany have collaborated on efforts to test industrial prototypes of widearea measurement devices.

PMUs go globalDesigning the distributionsystem of the future

Virginia Tech, Southern California Edison andKEMA, Inc. are working to extend the use ofsynchronized measurements beyond conven-

tional transmission system applications to enhance dis-tribution system reliability in a project sponsored bythe U.S. Department of Energy. The team seeks toimprove fault localization, prediction, isolation, andservice restoration capabilities. Moreover, system se-curity will be enhanced through the use of distributedintelligence to execute system islanding and servicerestoration. The team is also investigating the use ofsynchronized measurement of single phase and har-monic components for the detection of incipient faultsin distribution protection.

35

Center for PowerEngineeringDirector: Yilu Liu

Center for Energy & theGlobal EnvironmentDirector: Saifur Rahman

DEVICE CHARACTERIZATION

SYSTEMS& CONTROLSNo-cable elevatoruses VT switchedreluctance motordrive technology

ECE researchers have built a 1:10 scale prototype of a cable-lesselevator that would cost less than one-fifth the price of con-ventional elevator technology. The energy-efficient elevator

technology has the potential to make home elevators more afford-able, improve ship-board transportation, and even allow more thanone cage in a shaft.

The elevator prototype was designed for weapons elevators inbattleships and was funded primarily by the Naval PostgraduateSchool. The elevator uses an electric propulsion system based on theswitched-reluctance linear actuator first developed in ECE labora-tories for maglev transportation.

The elevator uses no cables and is propelled by a system com-prising of stators along the shaft and translators on the elevator cage.It was designed so that increasing the lift force does not require morestators, just translators.

The control equipment is carried at the sides of the cage, in-stead of at the top, which is in the plenum space between floors inconventional elevator technology. “Carrying the control equipmenton the sides of cage means the elevator does not need as much headroom as traditional elevators,” said Krishnan Ramu, director of theMotion Control Systems Research Group that developed the sys-tem. “This means that tall buildings can fit more usable floor space.It also means that in home elevator systems, even a short, 7-feet tallbasement can accommodate an elevator.”

Putting the control equipment on the cage also enables multi-ple cages in one shaft, according to Ramu. The shaft could be a cir-cular system in a large building or ship, where the unoccupied cagestravel horizontally above the top floor and beneath the bottom floor.

The project has takenmore than four years from concept to pro-totype. “The control systemwas the most excruciating,” Ramu said.“We had to make sure we could accelerate and decelerate smoothlyin all circumstances and that, in case of power interruption, the ele-vator would glide smoothly without harm to occupants.”

FACULTY

William BaumannA.A. (Louis) BeexPushkin KachrooDouglas LindnerKrishnan RamuDaniel StilwellChris Wyatt

A 1:10 scale prototype of an elevator thatcan operate inexpensively in smaller spacesuses Virginia Tech’s switched reluctancemotor drive technology.

36

Naval UnderwaterWarfare Center in Newport, RI. The teamwentfrom concept to first successful demonstration in just 10 months.

The high-speed AUV presented a number of unusual designchallenges. It is 50 percent heavier than the water it displaces. If it isnot being actively controlled, the AUV sinks quickly to the bottom.

The most unusual feature of the high-speed AUV is that it doesnot possess a vertical separationbetween its center of mass and itscenter of buoyancy. Almost allsubmersibles are designed so thatthe heavy parts are near the bot-tom and the lighter parts are nearthe top. This means it naturallytends to stay upright. The high-speed AUV, however, has no pas-sive stability. The teamdeveloped and is patenting amethod of actively stabilizing theroll motion.

When the high-speed AUVis not in high-speed flight, it tran-sitions into nose down hover. Itcan then transition back intohigh-speed flight.

Avery small, high-speed autonomous underwater vehicle (AUV)has been developed by ECE’s Dan Stilwell and Wayne Neu ofaerospace and ocean engineering. The high-speed AUV is ap-

proximately 39 inches long and 3 inches in diameter. It is capable ofspeeds greater than 15 knots; most AUVs operate at between 2-3knots. The AUV was developed as a “crash project” directed by the

Controls researchers are developing a new rifle system technol-ogy to enable the shooter to maintain better aim control underergonomic disturbances.The technology is expected to improve shot accuracy and dis-

persion under combat stress and to extend the effective range of smallarms rifles, according to Douglas Lindner, the Virginia Tech lead onthe project.

Called the INertially STAbilized Rifle (INSTAR), the system’sgoal is to compensate for these physiological effects and improve theoverall accuracy of the soldier by stabilizing the gun barrel relativeto the stock taking out the human induced disturbances.

INSTAR operates similar to a video camera with a simple feed-back control system that rejects any tracking commands at lower fre-quencies and compensates for the jitter disturbances due to breathing,heart rate, and muscle movement, etc., at higher frequencies.

INSTAR features a specially designed gun barrel suspension,compact high energy density piezoceramic actuators to effect gunbarrel motion, inertial sensors to measure gun barrel motion andpower, control, and signal conditioning electronics integrated intothe multifunctional gun stock.

Lindner is collaborating with Chris LaWigna of Techno-Sci-ences, Inc. and Diann Brei of the University of Michigan.

Mini, high-speed AUV exceeds 15 knots

Anti-jitter technology for better rifle aim

37

Autonomous Systems &Controls LaboratoryDirector: Dan Stilwellwww.ascl.ece.vt.edu

Digital Dignal ProcessingResearch LaboratoryDirector: A.A. (Louis) Beexwww.ee.vt.edu/~dsprl/

Intelligent Control GroupDirector: Pushkin Kachroo

Motion Control SystemsResearch GroupDirector: Krishnan Ramu

Adams, William JosephDecentralized Trust-BasedAccess Control for DynamicCollaborative EnvironmentsCommittee Chairs: Davis, N.J.Midkiff, S.F.

Al-Bandakji, MohammadRachadModeling and Analysis ofPhotonic Crystal WaveguidesCommittee Chair: Safaai-Jazi, A.

Al-Ghadhban, Samir NaserMulti-Layered Space FrequencyTime CodesCommittee Chairs: Buehrer R.M.;Woerner, B.D.

August, Nathaniel J.Medium Access Control inImpulse-Based Ultra WidebandAd Hoc and Sensor NetworksCommittee Chair: Ha, D.S.

Bae, Kyung KyoonAnalytical Framework for thePerformance Analysis ofMultiple Antenna SystemsCommittee Chair: Annamalai, A.;Tranter, W.H.

Brownfield, MichaelIgnatiusEnergy-Efficient WirelessSensor Network MAC ProtocolCommittee Chair: Davis, N.J.;Midkiff, S.F.

Chandrasekar, KameshwarSearch-Space Aware LearningTechniques for UnboundedModel Checking and PathDelay TestingCommittee Chair: Hsiao, M.S.

Chung, Woo CheolThe Dual Use of Power Distri-bution Networks for Data Com-munications in High SpeedIntegrated CircuitsCommittee Chair: Ha, D.S.

Dietze, Kai P.Blind Identification of MIMOSystems: Signal Modulationand Channel EstimationCommittee Chair: Stutzman, W.L.

Goel, Nitin KumarDevelopment of “Core-Suction”Technique for Fabrication ofHighly Doped Fibers for OpticalAmplification and Characteriza-tion of Optical Fibers for RamanAmplificationCommittee Chairs: Pickrell, G.R.;Stolen, R.H.

Gong, Michelle XiaohongImproving the Capacity of Wire-less Ad Hoc Networks throughMultiple Channel Operation:Design Principles and ProtocolsCommittee Chair: Midkiff, S. F.

Hadjichristofi, GeorgeCostaA Framework for Providing Re-dundancy and Robustness inKey Management for IPsecSecurity Associations in aMobile-Ad Hoc EnvironmentCommittee Chair: Davis, N.J.

Hall, Kristopher JosephThwarting Network StealthWorms in Computer Networksthrough Biological EpidemiologyCommittee Chairs: Davis, N.J.;Abbott, A.L.

Han, MingTheoretical and ExperimentalStudy of Low-Finesse ExtrinsicFabry-Perot InterferometricFiber Optic SensorsCommittee Chair: Wang, A.

Huang, ZhengyuQuasi-Distributed IntrinsicFabry-Perot InterferometricFiber Sensor for Temperatureand Strain SensingCommittee Chair: Wang, A.

Joshi, Gaurav GaurangUltra-Wideband Channel Model-ing using Singularity ExpansionMethodCommittee Chair: Stutzman, W.L.

Kim, Jong-KookInvestigation of High-NonlinearityGlass Fibers for PotentialApplications in UltrafastNonlinear Fiber DevicesCommitteeChairs: Safaai-Jazi, A.;Stolen R.H.

Lee, Hyung-JinDigital CMOS Design for UltraWideband CommunicationSystems: from Circuit-Level LowNoise AmplifierImplementation to a System-Level ArchitectureCommittee Chair: Ha, D.S.

Leu, Ching-ShanImproved Forward Topologiesfor DC-DC Applications withBuilt-In Input FilterCommittee Chair: Lee, F.C.

Liu, QianModular Approach for Charac-terizing and Modeling Con-ducted EMI Emissions in PowerConvertersCommittee Chairs: Boroyevich, D.;Wang, F.

Liu, WenduoAlternative Strutures forIntegrated ElectromagneticPassivesCommittee Chair: Van Wyk, J.D.

Nguyen, Anh Minh NgocHigh-Quality Detection inHeavy-Traffic Avionic Communi-cation System Using Interfer-ence Cancellation TechniquesCommittee Chair: Zaghloul, A.I.

Papenfuss, Cory M.Wideband Active Vibration Con-trol Synthesis and Implementa-tion on Uncertain ResonantStructuresCommittee Chair: Baumann, W.T.

Qiu, YangHigh-Frequency Modeling andAnalyses for Buck and Multi-phase Buck ConvertersCommittee Chair: Lee, F.C.

Subramanian, AnbumaniLayer Extraction and ImageCompositing Using a Moving-Aperture LensCommittee Chair: Abbott, A. L.

Sulistyo, Jos B.High Speed Circuit DesignBased on a Hybrid of Conven-tional and Wave PipeliningCommittee Chair: Ha, D. S.

Syal, Manan RohiniStatic Learning for Problems inVLSI Test and VerificationCommittee Chair: Hsiao, M. S.

Tcheslavski, Gleb V.Coherence and Phase Syn-chrony Analysis of Electroen-cephalogramCommittee Chair: Beex, A. A.

Tebben, Daniel JamesLimitations and Improvementof Subcarrier Multiplexed Sys-tems over Optical FiberCommittee Chair: Jacobs, I.

Teotia, SeemantInfluence of the Number of De-grees of Freedom on the Ca-pacity of Incoherent OpticalFiber Communication SystemsCommittee Chair: Jacobs, I.

Tsai, Shu-Jen StevenStudy of Global Power SystemFrequency Behavior Based onSimulations and FNET Meas-urementsCommittee Chair: Liu, Y.

Varma, Krishnaraj M.Fast Split Arithmetic EncoderArchitectures and PerceptualCoding Methods for EnhancedJPEG2000 PerformanceCommittee Chair: Bell, A. E.

Wang, ShuoCharacterization and Cancella-tion of High-Frequency Para-sitics for EMI Filters and NoiseSeparators in Power Electron-ics ApplicationsCommittee Chair: Lee, F. C.

PH.D. DEGREES AWARDED

38

Wang, ZhiyongIonic Self-Assembled MultilayersAdsorbed on Long Period FiberGratings for Use as BiosensorsCommittee Chair: Stolen, R.H.

Wu, HaisangEnergy-Efficient, Utility AccrualReal-Time SchedulingCommittee Chair: Ravindran, B.

Xu, ChunchunHigh Accuracy Real-Time GPSSynchronized Frequency Meas-urement Device for Wide-AreaPower Grid MonitoringCommittee Chair: Liu, Y.

Xu, JunchengHigh Temperature High Band-width Fiber Optic Pressure Sen-sorsCommittee Chair: Wang, A.

Xu, YangSynthesis and Characterizationof Silica Coated CdSe/CdSCore/Shell Quantum DotsCommittee Chair: Meehan, K.

Yin, JianHigh Temperature SiC Embed-ded Chip Module (ECM) withDouble-sided MetallizationStructureCommittee Chair: Van Wyk, J.D.

Zeng, DongsongPulse Shaping Filter Design andInterference Analysis in UWBCommunication SystemsCommittee Chairs: Zaghloul, A.I.;Annamalai, A.

Zhang, XinFully Distributed Control and ItsAnalog IC Design for ScalableMultiphase Voltage RegulatorsCommittee Chairs: Huang, A.Q.;Thorp, J.S.

Zhang, YanMiniature Fiber-Optic Multicav-ity Fabry-Perot InterferometricBiosensorCommittee Chair: Wang, A.

Zhang, ZhiyeSintering Micro-Scale andNanoscale Silver Paste forPower Semiconductor DeviceAttachmentCommittee Chair: Lu, G.Q.

Zhong, ZhianPower Systems FrequencyDynamic Monitoring SystemDesign and ApplicationsCommittee Chairs: Liu, Y.;Centeno, V.

Zhou, XigenElectrical, Magnetic, ThermalModeling and Analysis of a5000A Solid-State SwitchModule and Its Application asa DC Circuit BreakerCommittee Chairs: Huang, A.Q.;Wang, F.

Zhu, HuiyuNew Multi-Phase Diode Recti-fier Average Models for ACand DC Power System StudiesCommittee Chair: Linder, D.K.

Zhu, NingPlanar Metallization FailureModes in Integrated PowerElectronics ModulesCommittee Chair: Van Wyk, J.D.

“Planar Wideband Antennas,”S-Y Suh and W. L. Stutzman

“Topologies for Multiple EnergySources,”J.S. Lai

“Apparatus and Method thatPrevent Flux Reversal in theStator Back Material of a Two-Phase SRM (TPSRM),” K. Ramu

“Radial-Axial ElectromagneticFlux Electric Motor, CoaxialElectromagnetic Flux ElectricMotor, and Rotor for Same,”K. Ramu

“System to Generate and Con-trol Levitation, Propulsion andGuidance of Linear SwitchedReluctance Machines (LSRMs),”K. Ramu

“Quasi-Resonant DC-DC Con-verters with Reduced BodyDiode Loss,” M. Xu, F.C. Lee,Y. Qiu

“Self-Driven Circuit for Synchro-nous Rectifier DC/DC Con-verter,” M. Xy, F.C. Lee, Y. Ren,D. Sterk

“Two-Stage Voltage RegulatorsWith Adjustable IntermediateBus Voltage, Adjustable Switch-ing Frequency, and AdjustableNumber of Active Phases,”J. Wei, M. Xu, F.C. Lee

“Adaptive Bus Voltage Position-ing for Two-Stage Voltage Reg-ulators,” J. Wei, M. Xu, F.C. Lee

“EMI Filter and Frequency Fil-ters Having Capacitor with In-ductance Cancellation Loop,”S. Wang, F.C. Lee, W. Odendaal

“Buck Converter with High Effi-ciency Gate Driver ProvidingExtremely Short Dead Time,”Y. Ren, F.C. Lee

“Optical Fiber Pressure and Ac-celeration Sensor Fabricatedon a Fiber Endface,” Y. Zhu,X. Wang, J. Xu, A. Wang

“Optical Fiber Sensors forHarsh Environments,” J. Xu,A. Wang

“Optical Fiber Sensors Basedon Pressure Induced TemporalPeriodic Variations in Refrac-tive Index,” A. Wang

“Self-Compensating Fiber OpticFlow Sensor Having an End ofFiber Optic Element and Reflec-tive Surface Within a Tube,”W. Peng, B. Qi, A. Wang

“Method and Apparatus for Ac-cessing Memory Using EthernetPackets,” P. Athanas, H. Green,T. Brooks, K. Paar,P. McFall

“Smaller Aperture Antenna forMultiple Spot Beam Satellites,”A. Zaghloul, O. Kilic,A. Williams.

PATENTSAWARDED

39

BenjaminA. BeasleyEE/Music ’09Kernersville,NCJ.B. WestScholarship;NationalMerit Scholar-

ship; Robert C. Byrd Scholar-ship; Sam Walton CommunityScholarship; University Sym-phonic Wind Ensemble; Cham-ber Winds; Horn Ensemble;Co-op: Analog DevicesWhy ECE: It’s the ideal combi-nation of math, science, andtinkering. It manages to beboth abstract and practical in avery appealing way.

Ross Ben-jamin ClayCPE ’09Raleigh, NCMinor: Eco-nomics;Dean's Schol-arship;Aikido; Work

experience: IT consultant,WebSphere Application ServerSystem tester for IBMSummer plans: A Departmentof Energy SULI internship atLos Alamos National Labora-tory working with researcherson projects related to wirelesscommunications, streamingmedia optimization and virtualnetworks.

BrittanyCloreGE/CPE ’10Fairfax, VASociety ofWomen Engi-neers; Hypa-tiaengineering

community; intramural volley-ballCareer aspirations: Find waysto make computers availableto everyoneMost memorable experience:Being part of the Hypatia pro-gram....We do everything...from homework to laser tag.

ThomasAlanCooperEE ’10Oak Ridge,TNGalileoengineeringcommunity;

Student Engineers’ CouncilPublicity Committee; intramuralsoccer; Projects: SustainableEnergy Design Project, Con-temporary Issue Research Proj-ectWhy ECE: I have always likedworking with computers,elec-trical systems, and radios.

David C.CravenCPE ’08Winston-Salem, NCPratt Engi-neeringScholarship;Academic of-

ficer for HKN honor societyMost memorable experience:Being part of the honors com-munity in Hillcrest has been atremendous experience, I’vehad the opportunity to bemore exposed to fields outsidemy own.

It takes more than an expertdriver like Tony Stewart and afast Chevrolet Monte Carlo to

win a NASCAR race. Anothertype of race is run in the shop,where designers, machinists, and—yes, engineers—work againstthe clock and competitors to de-

sign, build, test, and install new parts.Matthew Carson (EE ’98, ME ’01), engineer on the Joe Gibbs

Racing team, often has only a few days to figure out what wentwrong in last weekend’s race and redesign it or fix it for a race theupcoming weekend.

“Luckily, I don’t need to go to the tracks to test a part,” saysCarson. “We can test it on our equipment. Ideally, we won’t have tobuild it to test it—that’s the project I’m working on when I’m not

putting out fires.”The role of the data provided by an engineer is crucial to the

success of Joe Gibbs Racing. Since Carson began there five yearsago, he’s partied at NYC’s Waldorf Astoria, celebrating the team’s2002 and 2005 Nextel NASCAR championships of driver TonyStewart. Many think the team’s advanced standing has everything todo with its advanced technology. Joe Gibbs Racing has ramped upits digital environment, especially its use of digital product devel-opment and testing.

Working virtually first and then later in the shop, engineersstripped excess metal from upper components and applied it belowthe centerline of axles to improve handling. Now Carson is work-ing on evaluating new engine configurations virtually. “I’m alwaystrying to make components lighter and stiffer,” he says.

Carson has written an analysis tool for valve train testing; he’screated programs in Visual Basic to keep track of overall testing.

BRADLEYSCHOLARS2006-2007

40

The BradleyScholarships werefounded by Marion

Bradley Via in honorof her father,

Harry Lynde Bradley

Matt Carson ’98:

Courtesy

of

Joe

Gib

bs

Racin

g

DanielMichaelHagerCPE ’08High Point, NCNationalMerit Finalist;Robert ByrdScholar; Soci-

ety of Automotive EngineersScholar; Work experience:Worked for Lockheed Martin onthe Information TechnologyAgency contractCareer aspirations: NASA Mis-sion Specialist AstronautWhy ECE: My Digital Designand Microprocessor Designcourses have confirmed that Imade the right choice.

EdwardJonesEE ’07Richmond, VAPresident andco-founder ofTech’s Engi-neers With-out Borders;

Tau Beta PiResearch: Honors thesis onpower electronics with the Fu-ture Energy Electronics CenterWork experience: Departmentof Energy: I traveled aroundthe country to meet withpower and power electronicsexperts and discuss the DOE'sfuel cell program

AdamShankCPE ’07Stuarts Draft,VAUniversityHonors Pro-gram; fenc-ing

Co-op: Software engineer/ IBMMost memorable experience:Current independent study: weare using Lego Mindstorms ro-bots linked to PDAs and writ-ing a group communicationslibrary for use with swarmtechnology.

Jacob Sim-monsCPE ’08Smithfield,VAHillcrest Hon-ors Commu-nity;intramural

softball; intramural soccerWhy ECE: ECE is always a chal-lenge to create and innovate,rather than simply relying onthe research and methods ofothers. It is a field thatchanges people's lives withevery new project.

JerryTowlerEE ’08Greer, SCStudent Tech-nology Coun-cil evaluatinghardware andsoftware for

use in the engineering curric-ulaCareer aspirations: Consider-ing patent law—its application,interpretation, and abuse.Summer plans: Working atDRT, Inc. to design automatedtesting software for industrialand military signal processinghardware.

Zachary LaCelleCPE ’09Lansing, NYUniversityHonors Pro-gram; NSCS;soccer

Career aspirations: Research AIand roboticsMy most memorable aspect ofTech experience has been boththe digital logic and analog cir-cuits classes.Why ECE: electronics have al-ways fascinated me and I wantto learn exactly how they workand how to engineer them

My LinhPhamCPE/Physics’07Annandale,VAMinors: Micro-electronics,Mathematics

Micron Scholar; Tau Beta Pi; PhiBeta Kappa; Sigma Pi Sigma;choir; swing dancingResearch: Spin Hall effect insemiconductors (physics) andhalf-metallic behavior of ferro-magnetic CrSe nanocrystalsWhy Virginia Tech: VirginiaTech told me that if I providedthe desire and motivation, theywould provide the means.

Matthew W.WelchEE ’09Richmond,KYNational Soci-ety of Colle-giateScholars; Stu-

dent Engineers Council; TaoBeta Pi; WUVT Assistant Engi-neer; University Honors Pro-gram; IEEE; learning Chineseand SpanishWhy ECE: I have always beeninterested in how electricityworks and how it can be ap-plied to almost everything.

Carson routinely evaluates engine parts from Gibbs’ vendors. Insome cases, he can test the parts to higher tolerances than the sup-pliers can.

“Sometimes I find a substandard part or design,” he says. “Imight want to redesign and have our shop make the part. Some ofour vendors sell to all the teams and, unless it’s a safety issue, we’lljust fix the problem ourselves and keep it a secret. Competitive ad-vantage is important.”

Cost is not an issue in the racing world, Carson says. “A fuelpump cable failed on Tony Stewart’s car at Bristol. Tony could havewon the race, so that cable breaking cost us about $100,000.”

Each car has a 780-hp motor that costs more than $50,000 tomake—and it is designed to last one race. The team goes through atleast 280 engines, each designed for 500 miles, in one season, and itmakes many parts for the engines in its modern machine shop. Car-son, who worked for General Motors, loves the excitement and fastpace of the racing industry. “It’s fun. It’s something different every

day, not like the corporate worldwhere you work on one part forone car and it may not come outfor four years. Here we see per-formance on the weekend.”

Carson enjoys tearing theengine apart and doing analysesafter a race. The thrill of the raceis not just vicarious with him; Carson has been racing in Sports CarClub of America autocross events since his college days, most re-cently in a Camaro. He also holds a commercial pilot’s license and isworking toward his flight instructor’s rating this year.

But Carson is spending more time at home these days. He andhis wife, Desha, have a year-old son, Micah Matthew, who likes tospend time with his dad.

—By Su Clauson-Wicker

41

LIFE IN THE NASCAR FAST LANE

Matt Carson ’98, as pilot

Des

ha

Car

son

Mark W.BaldwinBSEE ’93,MSEE ’05,Virginia TechAdvisor:Yilu LiuResearch:Power system

dynamic response to line andgenerator outages. Primaryfocus is the use of bulk powersystem eigenproperties to deter-mine event type; location andassess overall power systemcondition. Also investigatinggenerator rotor train torsionalresponse to pulsating loads,transformer condition assess-ment, FACTS/ESSS applicationsin power system oscillationdamping.

Aric D.BlumerB.S. Engineer-ing Science /Physics, ’92,Bob Jones;MSCPE, ’94,ClemsonAdvisor:

Cameron PattersonResearch: The acceleration ofdigital circuit simulation usingprocess migration betweenhardware and software—mov-ing busy processes into thehardware to be run in paralleland moving idle processes outof the hardware. Results showthat for an example circuit de-scription with 35 processes,the simulation can run 25times faster.

Evan LallyBSEE ’03,MSEE ’06,Virginia TechAdvisor:Anbo WangResearch:High-resolu-tion particle

imaging system for asphalt ag-gregates, plus preliminary in-vestigation into the productionof single-crystal sapphirestructures for pressure sens-ing. Studying the adaptation ofthe LHPG method of growingsingle crystal sapphire fibersto product more advancedstructures (see p. 31).

Robert M.GardnerBSEE ’03,MSEE ’05,Virginia TechAdvisor:Yilu LiuResearch:Wide area

monitoring and control oflarge-scale, complex networks,such as power systems. He isworking on frequency dataanalysis and conditioning,power system event detection,and power system event loca-tion, using the FNET wide areafrequency monitoring networkdeveloped at Tech.

DanielFriendBSEE, MSEE’98, BrighamYoung Uni-versityAdvisor: AllenMacKenzieResearch:

Distributed cognitive networks,which can sense their environ-ment, reason using observa-tions and experiences, enactchanges, and learn from prioractions. Developing architec-tures and analyzing distributedoptimization and the trade-offbetween partial informationand optimality.

JoAnn M. Adams (BSEE ’94)Co-owner, Big Fish DesignCenterville, Va.

Robert J. Adams (Ph.D. ’98)Assistant Professor, ECEUniversity of Kentucky

J. Shawn Addington (Ph.D. ’96)Department Head, ECEVirginia Military Institute

Sarah S. Airey (BSCPE ’01)

Christopher R. Anderson(BSEE ’99) Post–DoctoralAssociate; Virginia Tech

Matthew Anderson (BSCPE ’04)

Nathaniel August (Ph.D. '06)

Carrie Ellen Aust (BSCPE ’98)

William Barnhart (BSEE ’00,MSEE ’02)

Brian L. Berg (Ph.D. ’01)

Ray A. Bittner, Jr (Ph.D. ’97)

Kirsten Brown (BSEE ’94)Special Assistant to the CEOMicroStrategy Inc.,Alexandria, Va.

Steve Bucca (BSEE ’87, MSEE ’90)

Mark Bucciero (BSCPE ’01,MSCPE ’04) Argon ST, Inc.; Fair-fax, Va.

R. Michael Buehrer (Ph.D. ’96)Assistant ProfessorVirginia Tech

Charles F. Bunting (Ph.D. ’94)Associate ProfessorOklahoma State University

Carey Buxton (Ph.D. ’02)

Scott C. Cappiello (BSCPE ’94)Senior Director of Program Man-agement; MicroStrategy, Inc.,Carlsbad, Calif.

J. Matthew Carson (BSEE ’98)Joe Gibbs RacingHuntersville, N.C.

Ricky T. Castles (BSCPE ’03)Ph.D. Student; Virginia Tech

Eric Caswell (Ph.D. ’02)

Kevin Cooley (BSEE ’02)IAS Corporation

Cas Dalton (BSCPE ’03)

Phillip Danner (BSCPE ’91)

Bradley A. Davis (Ph.D. ’01)

Daniel Davis (BSEE ’03)

Scott Davis (BSCPE ’00)

Brian M. Donlan (MSEE ’05)

Joel A. Donohue (MSEE ’94)

Thomas Drayer (Ph.D. ’97)

Bradley D. Duncan (Ph.D. ’91)Professor, ECEUniversity of Dayton

Gregory D. Durgin (Ph.D. ’00)Assistant ProfessorGeorgia Tech; won 2006 NSFCAREER award

W. Ashley Eanes (BSEE ’95)Duke Energy, Burlington, N.C.

Richard B. Ertel (Ph.D. ’00)

Brian F. Flanagan (BSEE ’97,MSEE ’98)

Kevin P. Flanagan (BSCPE ’00,MSCPE ’01)

Todd Fleming (BSEE ’94,MSEE ’96)

Ryan J. Fong (BSCPE ’01,MSCPE ’04)

Jayda B. Freibert (BSEE ’98)

Bradley H. Gale (BSEE ’97)

BRADLEYFELLOWS2006-2007

42

ALUM

NI

More than 75 graduatestudents have been

supported by the BradleyFellowship program sinceit was instituted in 1988

Mark LehneBSEE ’94, Seat-tle Pacific Uni-versity; MSME’98, MSEE ’00,Oregon StateUniversityAdvisor:Sanjay Raman

Research: Investigating newanalog-to-digital conversioncircuit architectures that im-prove a broadband wirelesstransceiver’s ability to effi-ciently detect information in thepresence of strong interference.The goal is better overall per-formance while allowing forhigher data rates and more in-telligent transceivers (see p.33).

AndrewLoveBSCPE ’05,University ofVirginiaAdvisor:Thomas Mar-tine-Textiles re-

search involving garmenttracking and shape sensing. Auser’s garments incorporatemicroprocessors and sensors,which allow the garments tomonitor their acceleration andorientation. Additional tech-nologies are currently beinglooked at that will allow for theclothing to detect their shapeand position.

KeithMcKenzieBSEE ’01,MSEE ’04,University ofTennesseeAdvisor:Yilu LiuThe utiliza-

tion of wind energy with thepower grid, including develop-ing a model for wind turbineand associated power electron-ics interface. Issues beingstudied include the impacts ofutilizing different configura-tions of energy storage, issuesinvolved in connecting windturbines to the grid, and lowvoltage ride-through capabilityof a wind turbine.

Justin RiceBSEE/BSCPE’02, MSEE ’04,University ofFloridaAdvisor:Cameron Pat-tersonResearch:

Develop tools to allow dynamicrun-time reconfiguration ofFPGA devices. He seeks to cre-ate a design environment thatenables engineers to take ad-vantage of the benefits offeredby run-time reconfigurationwhile hiding the complexitiesinvolved (see p. 21).

David C.ReuschBSEE ’04,MSEE ’06,Virginia TechAdvisor:Fred C. LeeResearch:Developing

improved voltage regulatorsfor future computer micro-processors that will achievehigher efficiency and higherpower density than current in-dustry designs. He is currentlystudying high frequencyDC/DC conversion at the 5-20MHz switching frequency.

Parrish E.RalstonBSEE ’06, Vir-ginia TechAdvisor:KathleenMeehanResearch:Synthesis

and characterization of II-VInanoparticles. Room tempera-ture and pressure techniqueshave been developed to pro-duce CrSe and CoSe nanoparti-cles. Such nanoparticles havepotential in the biomedicalfield, where they can be usedas tags and as a contrastingagent in imaging.

Daniel J. Gillespie (BSCPE ’95)

Brian Gold (BSEE ’01)Ph.D. Student; Carnegie Mellon

Jonathan Graf (BSCPE ’02,MSCPE ’04) Luna Innovations,Blacksburg, Va.

Timothy Gredler (BSCPE ’03)

Christopher Robert Griger(BSCPE ’02)

Alex Hanisch (BSCPE/Math ’03)Joint Warfare Analysis Ctr., DoD

Abigail Harrison (BSCPE ’04)

Jennifer J. Hastings (BSEE ’96)

Dwayne A. Hawbaker (MSEE’91)Senior Staff Engineer; JohnsHopkins Applied Physics Lab

Matt C. Helton (BSEE ’01)Resident EE; Coal GasificationPlant; Eastman Chemical Co.;Kingsport, Tenn.

Benjamin E. Henty (MSEE ’01)

Jason Hess (MSEE ’99)

H. Erik Hia (MSCPE ’01)Software EngineerHatteras Networks, N.C.

James Hicks (Ph.D. ’03)

Hugh E. Hockett (BSCPE ’03)

Janie Hodges (BSCPE ’01)

Spencer Hoke (BSCPE ’03)

Russell T. Holbrook (BSCPE ’03)

Andrew HollingsworthBSCPE ’02

Ryan Hurrell (BSEE ’03)

John Todd Hutson (BSEE ’93)

Madiha Jafri (BSCPE ’03)

Daniel A. Johnson (MSEE ’01)

Adam Kania (BSEE ’01)

David A. Kapp (Ph.D. ’96)

Dimosthenis C. Katsis (Ph.D.’03)

David Kleppinger (BSCPE ’04)

Paul A. Kline (Ph.D. ’97)

Gregory Kozick (BSCPE ’03)

William B. Kuhn (Ph.D. ’95)Professor, EECEKansas State University,

Jeffery D. Laster (Ph.D. ’97)Product Specialist, Analog, MixedSignal, and RF IC; MentorGraphics

Charles Lepple (BSEE ’99,MSEE ’04)

Jason Lewis (BSEE ’99)

Joseph C. Liberti (Ph.D. ’95)

Zion Lo (BSEE ’94)Sr. SoftwareEngineer/ArchitectIQNavigator, Colorado

Daniel L. Lough (Ph.D. ’01)

Cheryl Duty Martin (BSEE ’95)

Stephanie Martin (BSEE ’04)

Michael Mattern (BSEE ’02)

Christopher Allen Maxey (BSCPE ’02)

Eric J. Mayfield (BSEE ’97)

Patrick McDougle (BSEE ’03)

Brian J. McGiverin (BSCPE ’96)

John T. McHenry (Ph.D. ’93)Senior Electrical Engineer,U.S. Department of Defense

David McKinstry (MSEE ’03)

Garrett Mears (BSCPE ’00)Head of DevelopmentOpen Vantage Ltd.,London

Vinodh Menon (BSCPE ’02)

Michael Mera (BSEE ’03)

“The BradleyFellowshiphas given methe opportu-nity to workon leadingedge stuffthat doesn’tyet have com-mercial value,and thus fewother sourcesof funding...”—Neil Steiner

43

Douglas R.SterkBSEE ’00,MSEE ’03,Virginia TechAdvisor:Fred. C. LeeResearch:Voltage regu-

lator modules for advanced mi-croprocessors targeting IntelCorp.’s 2010 specifications. Heand David Reusch are improv-ing the packaging and thermaldesign to make use of ad-vanced circuit design tech-niques and magnetic layouttechniques, and selecting anoptimal silicon design forserver applications.

RebeccaSheltonBSE EE,BA EnglishWriting ’06,University ofTennessee atChattanooga

Advisor: William BaumannResearch: Mathematical model-ing and parameter estimationof small gene networks, espe-cially man-made networks cre-ated by synthetic biology. Sheis currently modeling the acti-vation switch of the HIV virus.Such models will aid in design-ing synthetic gene networksand understanding naturalgene networks.

Neil J.SteinerB.A. ’98Wheaton;BSEE ’98 Illi-nois Instituteof Technology;MSEE ’02Virginia Tech

Advisor: Peter AthanasResearch: Developing infra-structure and proof-of-conceptfor autonomous computingsystems that take increasedownership for their resourcesand operation. Developed thefirst known occurrence of asystem changing its own hard-ware in a non-trivial fashionwhile in operation (see p. 20).

Juan E.SurisBSEE ’96,Puerto Rico;MCSPE, ’98Northwestern;MS Statistics,’99, Chicago

Advisor: Luiz DaSilvaResearch: Game theoretic ap-proach to opportunistic spec-trum sharing. Researchers areseeking to improve efficientuse of the limited spectrum.Also exploring cooperationamong users to achieve fairand efficient spectrum access.

Ethan SwintMSECE ’02,Baylor; MSME’05, Universityof Texas,AustinAdvisor:Krishnan RamuResearch:

Power electronics and switchedreluctance drives, with an em-phasis on reducing drive com-plexity and cost, including thedevelopment of new control al-gorithms. SR drives are ex-pected be more efficient andreliable than traditional motorsand generators, with lower cost.

ThomasRondeauBSEE ’03,MSEE ’06,Virginia TechAdvisor:Charles Bost-ianResearch:

Developing cognitive radio sys-tems that intelligently optimizewireless communications de-vices by appropriately tuningsoftware “knobs” to improveperformance and service. He isspecifically working on the AIalgorithms used to create theintelligent optimizationprocesses.

Carl Minton (MSCPE ’99)

John Morton (MSEE ’98)

Stephen Nash (BSCPE ’03)

Troy Nergaard (MSEE ’00)Tesla Motors; San Carlos, Calif.

Michael H. Newkirk (Ph.D. ’94)Applied Physics LaboratoryJohns Hopkins University

Paul Erik Nguyen (BSCPE ’98,MSCPE ’99)

J. Eric Nuckols (BSEE ’97, MSEE’99); Associate Principle EmbeddedSoftware Engineer; ITT AdvancedEngineering & Sciences

Neal Patwari (BSEE ’97, MSEE ’99)

Joseph Allen Payne (BSEE ’00)

W. Bruce Puckett (MSEE ’00)

Yaron Rachlin (BSEE ’00)

BRADLEYFELLOWSCONTINUED

44

Christian J. Reiser (Ph.D. ’05)

Steve Richmond (MSEE ’01)

Jamie Riggins (BSEE/BSCPE '04)Graduate Student, Virginia Tech

Pablo Max Robert (Ph.D. ’03)

Thomas W. Rondeau (BSEE ’03)Ph.D. Student; Virginia Tech

Thomas M. Rose (MSEE ’96)Engineer/Scientist IVBoeing, University City, Mo.

Jon Scalera (MSCPE ’01)

Amy Schneider (BSCPE ’03)

Steven Schulz (MSEE ’91)

David C. Schroder (BSEE ’05)

Jeff Scruggs (MSEE ’99)

Kashan Shaikh (BSCPE ’02)Ph.D. StudentUniversity of Illinois

Raymond A. Sharp (BSEE ’02)

Roger Skidmore (Ph.D. ’03)

Jeff Smidler (BSEE ’98)

Amanda (Martin) Staley(BSEE ’99, MSEE ’01)

Graham Stead (BSCPE ’93)

Douglas R. Sterk (MSEE ’03)Ph.D. Student, Virginia Tech

Scott Stern (BSEE ’93)

Samuel S. Stone (BSCPE ’03)

Anne (Palmore) Stublen(BSEE ’91) Newark, Del.

Seema Sud (Ph.D. ’02)

David Tarnoff (MSEE ’91)Assistant ProfessorEast Tennessee State University

Daniel Tebben (Ph.D. '06)

Rose Trepkowski (MSEE ’04)

Christian Twaddle (BSCPE ’01)

Matthew C. Valenti (Ph.D. ’99)Associate ProfessorWest Virginia University

Wesley Wade (BSEE ’93)

Kristin Weary (BSEE ’03)

Michael L. Webber (MSEE ’03)

Jason Wienke (BSEE ’02)

Thomas Williams (BSEE ’00)

William J. Worek (BSCPE ’99)

Kai Xu (BSEE ’95)

Jason Jon Yoho (Ph.D. ’01)

Gregory A. Zvonar (MSEE ’91)

ALUM

NI

Arun Phadke was one of four professorshonored with the “Docteur Honoris Causa diI’INP Grenoble.”

Tom Martin was honored as a recipient ofthe Presidential Early Career Award for Scien-tists and Engineers (PECASE).

Patrick Schaumont received an NSF CAREERaward for “Hardware/Software Codesign for Se-cure Embedded Systems: Methods and Educa-tion.”

Yong Xu received an NSF CAREER award for“Single Carbon Nanotube Based Nano-OpticalImaging.”

Jason Lai was named an IEEE Fellow.

Amy Bell was awarded the IEEE OutstandingStudent Branch Counselor award.

Daan Van Wyk received the IEEE Power Elec-tronics Society 2006 Distinguished ServiceAward.

Fred Lee was the keynote speaker at the 10thInternational Conference on Optimization ofElectrical and Electronic Equipment (OPTIM’06), May 2006.

Dushan Boroyevich was named the Ameri-can Electric Power Professor of ECE.

Anbo Wang was named the Clayton Ayre Pro-fessor of ECE.

Jim Thorp was named Power Engineering Ed-ucator of the Year and also was recognized bythe U.S. National Committee of CIGRE as anAttwood Associate.

Krishnan Ramu was an invited keynotespeaker at the International Conference on In-dustrial Technology, December 2006.

Amy Bell is serving as associ-ate editor of IEEE Signal Pro-cessing Letters and IEEE SignalProcessing Magazine.

Michael Buehrer is associateeditor of IEEE Transactions onWireless Communications,IEEE Transactions on Vehicu-lar Technologies, and IEEETransactions on Signal Pro-cessing.

Luiz DaSilva is associate edi-tor of the IEEE Communica-tions Letters.

Thomas Hou is editor of theIEEE Transactions of WirelessCommunications, associateeditor for IEEE Transactions ofVehicular Technology, editorof ACM/Springer Wireless Net-works, and editor of ElsevierAd Hoc Networks.

Michael Hsiao served asguest editor of ACM Transac-tions on Design Automation ofElectronic Systems.

Lamine Mili is co-editor of theInternational Journal of Criti-cal Infrastructures.

T.-C. Poon is associate editorof the International Journal ofOptomechatronics and topicaleditor of Applied Optics.

Sanjay Raman is serving asassociate editor of IEEE Trans-actions on Microwave Theoryand Techniques.

Saifur Rahman served asguest editor of the Proceed-ings of IEEE.

Sandeep Shukla is associateeditor for IEEE Design & Test,IEEE Transactions of IndustrialInformatics, and co-guest edi-tor of a special issue on “Elec-tronic System Level Design” ofIEEE Design and Test maga-zine.

Amy Bell is serving as chair of the Public Awareness Committee ofthe IEEE Educational Activities Board.

Ira Jacobs serves on the Federal Communications TechnologicalAdvisory Council as a special government employee.

Scott Midkiff is serving as program director for the National Sci-ence Foundation, Division of Electrical, Communications and CyberSystems.

Krishnan Ramu served as a distinguished lecturer for the IEEE In-dustrial Electronics Society.

Saifur Rahman is serving as an IEEE Distinguished Lecturer.

Peter Athanas served as co-chair of Dynamically Reconfig-urable Architectures, April2006.

Thomas Hou chaired theFirst IEEE Workshop on Net-working Technologies for Soft-ware Defined Radio Networks,September 2006.

Michael Hsiao served asgeneral chair of the IEEE Inter-national High Level Design Val-idation and Test.

Fred Lee was general chair ofthe International Power Elec-tronics and Motion ControlConference, August 2006.

Allen MacKenzie and LuizDaSilva co-chaired the 1stWorkshop of Game Theory forNetworks (GameNets), whichwill expand into a full-sizedconference in 2008.

Tom Martin was program co-chair of the 2006 InternationalSymposium on Wearable Com-puters, sponsored by the IEEEComputer Society.

Krishnan Ramu served asGeneral Co-Chair of IEEE Indus-trial Electronics Society’s Inter-national Conference onIndustrial Technology, Decem-ber 2006.

ConferenceChairs

HONORS & ACHIEVEMENTSHonors & Awards Editorships

National &InternationalService

45

Christopher R. Anderson ’99Donald W. Armistead ’87Shawn B. Arner ’91John W. Ayres ’75William D. Barnhart ’00Andrew Beach ’80T. Wayne Belvin ’83Robert F. Bruce ’65Ryan L. Bunch ’99Ernest D. Crack ’66William C. Decker, Jr. ’63Steve O. Dixon ’64Kevin A. Dunetz ’89Stephen P. Earwood ’93Lewis S. Farinholt ’74Amy R. Flax ’86Robert W. Freund ’71Victor Fung ’91

Benjamin M. Grady ’00James T. Griffin ’70Jeffrey G. Hadley ’89Paul S. Hamer ’72Elliott L. Harper ’69Edgar D. Harras ’67Karen M. Heine ’00Kurt M. Hinds ’89John G. Hoecker ’73Hsien L. Huang ’68Timothy P. Huggins ’06Ryan E. Hurrell ’03John T. Hutson ’93Donald W. Kahl ’85Ashok N. Katti ’80Joel E. Keys, Jr. ’87Melissa B. Koch ’90Grady J. Koch ’91

Joseph P. Kurian ’94Ting-Pui Lai ’97Robert E. Lambert ’80William K. Lamp ’78John K. Loftis, Jr. ’77Carl S. Mills ’99James C. Mills ’97George A. Mizusawa ’94James C. Morizio ’82Aaron M. Moyer ’80Michael L. Oatts ’79Jeffrey T. Olson ’81Robert C. Patterson, III ’91David T. Perdue ’75Robert J. Phipps ’00Milton A. Pilcher ’38Don M. Powers ’59Bindiganavele A. Prasad ’74

Carlos V. Roberts ’70Daniel M. Sable ’85James F. Schooler ’85Noel N. Schulz ’88Edwin F. Sharpe ’49K. Douglass Smith ’69David W. Snodgrass ’62Michael E. Sockell ’81James R. Sowers, Sr. ’79Robert W. Sundeen, Jr. ’78Brian T. Unterwagner ’98Frederick E. White ’66Donna E. Wiltsey ’91Barry M. Wood ’72Ning Yang ’95

ABB, Inc.AMD Matching Gift ProgramBoeing Matching Gift

CompanyCaterpillar FoundationChevronTexaco Matching

Gift ProgramChoice Hotels InternationalFoundationComcast CableDominion FoundationElectrical Distribution

Design Inc.Genscape, Inc.

IBM InternationalFoundation

ITT Industries, Inc.International Foundation

for TelemeteringKEMA Inc.KEMA-TDC Inc.MathWorks Inc. (The)Micron Technology

Foundation, Inc.Micron Technology Inc.Microsoft CorporationRaytheon CompanyRonald A. Williams, Ltd.

SANS InstituteSTMicroelectronicsSamsung Electronics

Company Ltd.Science Applications

InternationalCorporation

Siemens Matching GiftsSouthern Company

Services, Inc.Tektronix, Inc.Texas Instruments Inc.The Scholarship Foundation-

Lockheed MartinMatching Gifts

The Walt Disney CompanyFoundation

Tyco ElectronicsCorporation

Tyco ElectronicsFoundation

United Space AllianceFoundation

Verizon FoundationVia-Bradley College of

Engineering FoundationWachovia Foundation, Inc.William B. Webber Trust

Maria E. ArnerDiane AyresStephen R. and

Rosemary P. ColeJames A. and

Kathleen M. EderHoward D. Frisch

Stanley HondaHui-Lein P. HuangIra and Irene JacobsSudha N. KattiJames M. and

Mary P. LazarekDavid M. and Luwanna S. Lofe

Michael L. MilesLarry L. Miller, Jr.Dorla S. PowersSudha S. PrasadLinda RobertsKirk H. SchulzVirginia H. Sharpe

Dorothy SnodgrassHarold (Hal) H. and

Tracey B. Speed, IIIJames A. and

Carla V. Wilding

Corporate Donors

ECE Alumni who contributed to ECE

ECE Friends who contributed to ECE

DONORS TO ECE During Fiscal Year 2006

46

CPES Partners

Principal Members PlusABB, Inc. - Corporate

ResearchAnalog DevicesC&D Technologies, Inc.Delta ElectronicsFreescale SemiconductorFSP Group, Inc.GE Global ResearchHipro Electronics Co., Ltd.Infineon TechnologiesInternational RectifierIntersil CorporationLinear TechnologyMKS Instruments, Inc.National Semiconductor

CorporationNXP SemiconductorsPrimarionRenesas Technology

CorporationRolls-RoyceTexas InstrumentsThe Boeing Company

Principal MembersCrane Aerospace/ELDEC

CorporationDRS Power and Control

Technologies, Inc.Eltek Energy ASEmerson Network PowerGroupe SAFRANRockwell Automation -

Allen-Bradley

Associate MembersCrown InternationalDensei-Lambda KKIntel CorporationITT Power SolutionsJohnson ControlsKEMA T&D Power

L-3 Communications,Power Paragon Inc.

Nippon Chemi-Con Corp.Northrop GrummanPanasonic Electric Works

Laboratory of America, Inc.Schneider Toshiba Inverter

EuropeShindengen Electric

Mfg. Co., Ltd.Toyota Motor Corporation

Affiliate MembersABB Inc., Drives & Power

Products Div.Agilent TechnologiesAnsoft CorporationAO Smith CorporationAstronicsBAE SYSTEMS Controls, Inc.Ballard Power Systems, Inc.Caterpillar, Inc.Cree, Inc.Curtis InstrumentsDanfoss Electronic DrivesEaton Corporation -

Innovation CenterElectronic ConceptsEmerson Motor CompanyEngineous SoftwareFlomericsFord Motor CompanyFuji Electric Advanced

Technology Co., Ltd.GE EnergyGM Advanced Technology

CenterGrundfos A/SHamilton SundstrandHoganas ABHoneywell International

CorporationIngersoll-Rand CompanyJapan Research InstituteKohler Company

LEM USA, Inc.Magnetek, Inc.Miller Electric CompanyMPC Products CorporationMTS Systems CorporationNissan Motor Co., Ltd.Oak Ridge National

LaboratoryOshkosh Truck CorporationOtis Elevator CompanyPacific Scientific EKDPhoenix IntegrationPower Efficiency

CorporationRemy, Inc.Rietschle ThomasRockwellAutomation/Dodge/

Reliance ElectricS&C Electric CompanySchneider Electric - Square DToshiba International

CorporationTrane CompanyTransim TechnologyUnico, Inc.United Technologies

Research CenterVPT, Inc.Whirlpool CorporationYaskawa Electric America, Inc.

Wireless @VTAffiliates

Analog DevicesArmy Research OfficeDigital Receiver

Technology Inc.DRS TechnologiesFreescale SemiconductorGeneral Dynamics

C4I SystemsHuawei Technologies

CompanyL3 CommunicationMotorola Inc.Qualcomm Inc.Raytheon IISSamsung Telecommunications

AmericaSony Ericsson Mobile

CommunicationsTexas Instruments

Although every effort has been made to ensure the accuracyof this report, we acknowledge that errorsmay have occurred.If your name was omitted or listed incorrectly, please acceptour sincere apologies and send corrections to the Office ofUniversity Development at (540) 231-2801.

Corporate and Industrial Affiliates

When you hear from the College Annual Fund this fall, make your gift to ECE

47

A. Lynn AbbottAssociate ProfessorIllinois ’89

Masoud AgahAssistant ProfessorMichigan ’05

Peter AthanasProfessor; Brown ’92

Scott BaileyAssistant ProfessorColorado ’95

William T. BaumannAssociate ProfessorJohns Hopkins ’85

A. A. (Louis) BeexProfessor; Colorado State ’79

Amy E. BellAssociate ProfessorMichigan ’97

Dushan BoroyevichAmerican Electric PowerProfessor; Virginia Tech ’86

Charles W. BostianAlumni DistinguishedProfessor; NC State ’67

Robert P. BroadwaterProfessor; Virginia Tech ’77

Gary S. BrownBradley DistinguishedProfessor of ElectromagneticsIllinois ’67

Michael BuehrerAssociate Professor;Virginia Tech ’96

Virgilio CentenoAssistant ProfessorVirginia Tech ’95

C. Robert ClauerProfessor; UCLA ’80

Richard O. ClausLewis A. Hester ProfessorJohns Hopkins ’77

Claudio da SilvaAssistant ProfessorUniversity of CaliforniaSan Diego ’05

Luiz DaSilvaAssociate ProfessorKansas ’98

William A. DavisProfessor; Illinois ’74

Jaime de la ReeAssociate Professor & Assis-tant Department HeadPittsburgh ’84

Steven EllingsonAssistant ProfessorOhio State ’00

Mohamed EltoweissyAssociate ProfessorOld Dominion ’93

Louis GuidoAssociate ProfessorIllinois ’89

Dong S. HaProfessor; Iowa ’86

Robert HendricksProfessor; Cornell ’64

Thomas HouAssistant ProfessorPolytechnic Univ. ’98

Michael HsiaoProfessor; Illinois ’97

Chao HuangAssistant ProfessorPrinceton ’05

Ira JacobsProfessor; Purdue ’55

Mark JonesAssociate ProfessorDuke ’90

Pushkin KachrooAssociate ProfessorUC Berkeley ’93

Jason LaiProfessor; Tennessee ’89

Brent LedvinaAssistant ProfessorCornell ’03

Fred C. LeeUniversity DistinguishedProfessor; Duke ’74

Yao LiangAssistant ProfessorClemson ’97

Douglas LindnerAssociate ProfessorIllinois ’80

Yilu LiuProfessor; Ohio State ’89

G. Q. LuProfessor; Harvard ’90

Allen MacKenzieAssistant ProfessorCornell ’03

Tom MartinAssociate ProfessorCarnegie Mellon ’99

Kathleen MeehanAssistant ProfessorIllinois ’85

Scott F. MidkiffProfessor; Duke ’85

Lamine MiliProfessor; Liege ’87

Amitabh MishraAssociate ProfessorMcGill ’85

Leyla NazhandaliAssistant ProfessorMichigan ’05

Khai D.T. NgoProfessor; Caltech ’84

Hardus OdendaalAssistant ProfessorRand Afrikaans ’97

Jung-Min ParkAssistant ProfessorPurdue ’03

Cameron PattersonAssociate ProfessorCalgary ’92

JoAnn PaulAssociate ProfessorPittsburgh ’94

Paul PlassmannProfessor; Cornell ’90

T.C. PoonProfessor; Iowa ’82

Timothy PrattProfessor; Birmingham ’68

Saifur RahmanJoseph Loring ProfessorVirginia Tech ’78

Sanjay RamanAssociate ProfessorMichigan ’97

Krishnan RamuProfessor; Concordia ’82

Binoy RavindranAssociate ProfessorUT Arlington ’98

Jeffrey H. ReedWillis G. Worcester ProfessorUC Davis ’87

Sedki RiadProfessor; Toledo ’76

Ahmad Safaai-JaziProfessor; McGill ’78

Wayne ScalesProfessor; Cornell ’88

Patrick SchaumontAssistant ProfessorUCLA ’04

Sandeep ShuklaAssistant ProfessorSUNY Albany ’97

Daniel StilwellAssistant ProfessorJohns Hopkins ’99

Kwa-Sur TamAssociate ProfessorWisconsin ’85

James S. ThorpDepartment HeadHugh P. and Ethel C. KellyProfessor; Cornell ’62

William H. TranterBradley Professor ofElectrical EngineeringAlabama ’70

Joseph G. TrontProfessor; SUNY Buffalo ’78

ECE FACULTY

48

Anbo WangClayton Ayre ProfessorDalian ’89

Fred WangAssociate ProfessorUSC ’90

Yue (Joseph) WangProfessor; Maryland ’95

Chris WyattAssistant ProfessorWake Forest School ofMedicine ’02

Ming XuAssociate ProfessorZhejiang; ’97

Yong XuAssistant ProfessorCaltech ’95

Jason XuanAssociate ProfessorMaryland ’97

Yaling YangAssistant ProfessorIllinois ’05

Amir ZaghloulProfessor; Waterloo ’73

Research Faculty& Instructors

Carl B. DietrichResearch Assistant Professor

Leslie K. PendletonInstructor

Arun G. PhadkeUniversity DistinguishedProfessor Emeritus

Gerald ReidInstructor

Warren L. StutzmanThomas Phillips Professor(Emeritus)

Michael ThompsonInstructor

Jason ThweattInstructor

Andrew Beach (’80, ’85)Principal ASIC Design EngineerRaytheon Corporation

Miroslav BegovicProfessorSchool of Electrical andComputer EngineeringGeorgia Institute of Technology

Rick BrisbinVice President, SoutheastRegion, L-3 Communications,Titan Group

Gautam ChandraVice PresidentWGL Holdings, Inc. andWashington Gas Work

Christopher DeMarcoProfessorElectrical and ComputerEngineeringUniversity of Wisconsin-Madison

Shivon Dosky (’89)Managing PartnerCollegiate Inn Properties

Scott DoylePartnerSteptoe & Johnson, LLP

Truls Henriksen (’90)Consultant

Michael Keeton (’71)Senior Program DirectorOrbital Science Corporation

Robert E. Lambert (’80)Senior Vice PresidentWorldwide Technology StrategyThe Walt Disney Company

Terrence C. LeslieUniversity and Academic Rela-tions ManagerMicron Technology, Inc.

Christopher G. Magnella (’84)Manufacturing ManagerMotorola

Gino ManzoDirector, MicroelectronicsTechnology and ProductsManassas Site ExecutiveBAE Systems

Dave Marsell (’88)Vice President of EngineeringPressure Systems, Inc.

Michael Newkirk (’88, ’90, ’94)Principal Professional StaffThe Johns Hopkins UniversityApplied Physics Laboratory

Dan Sable (’85, ’91)President and CEOVPT, Inc.

Karl Salnoske (’76)Chief Information OfficerSchering-Plough

Wayne Snodgrass (’61)Retired, Vice PresidentNorthrop Grumman

James Stewart (’83)Vice President &General ManagerImaging Systems DivisionAvago Technologies

C. Hyde Tucker (’56)Retired, President and CEOBell Atlantic International

Timothy J. Winter (’81)Director, Air and MissileDefense DepartmentSystems Development &Technology DivisionNorthrop Grumman Corpora-tion–Electronic Systems Sector

Paul K. Wong (’80)President and CEOProsoft

Virginia Tech does not discriminate against employees, students,or applicants on the basis of race, color, sex, disability, age, vet-eran status, national origin, religion, or political affiliation. Anyonehaving questions concerning discrimination should contact theEqual Opportunity/Affirmative Action Office.

FacultyBooks

Published2006

Bob HendricksJohn Wiley & Sons

Michael BuehrerMorgan-Claypool

William Tranter,Desmond P. Taylor,Rodger E. Ziemer,Nicholas F. Maxem-chuk,Jon W. MarkIEEE Press andJohn Wiley & Sons

T.-C. PoonSpringer

T.-C. Poon, T. KimWorld Scientific

ECEADVISORY BOARD

www.ece.vt.edu

The Bradley Departmentof Electrical & Computer EngineeringVirginia Polytechnic Institute and State University302 Whittemore HallBlacksburg, VA 24061

Non-Profit Org.U.S. Postage

PAIDBlacksburg, VA 24060

Permit No. 28

Modeling the growth

ofa

viru

sHydroInformatics

Canc

ercelldivision