Page top
Skip navigation
UW-Madison crest

College Directory
College Directory


Content begins

Electrical and Computer Engineering

Patrick Cheung, Pam Limpiti, Professor Barry Van Veen, Andrew Bolstad and Matt Rebholtz

Professor Barry Van Veen (center) with students (from left) Patrick Cheung, Pam Limpiti, Andrew Bolstad and Matt Rebholtz. (large image)

Turning down the noise:
helps researchers ‘listen’
to the brain

To study how regions of the brain communicate, neuroscientists often use a technique called electroencephalography (EEG), which reads electrical activity in the brain through sensors on the scalp.

However, the skull and the scalp blur these EEG readings. In addition, a multitude of signals from “background” processes make it difficult to pinpoint electrical activity corresponding to specific tasks. “It’s like standing outside a crowded party and trying to sort out individual conversations,” says Professor Barry Van Veen.

Illustration of EEG brain study

Van Veen and his students use signal-processing techniques to filter out that noise and enable them to study how one area of the brain influences another (inset graphic). “The brain is active all the time,” he says. “It’s in the midst of that background noise that you have to identify a specific set of connections associated with a task.”

One research paradigm is working memory, a type of task-oriented short-term memory. For example, working memory allows a person to remember a phone number long enough to dial it, or to remember a series of notes or pattern of shapes long enough to repeat it. Neuroscientists hypothesize that several regions of the brain are connected in working memory tasks. Van Veen and his students use their signal-processing techniques to identify electrical connections from EEG data and determine how they change under different conditions, such as task difficulty or recall accuracy.

The group also is interested in how connectivity in the brain changes between waking and sleep, and more complicated activity such as language processing.

Van Veen is hopeful that, as the research progresses, his methods will provide some insight into the workings of the brain and lead to better understanding for treatment of medical conditions like epilepsy or schizophrenia.

Biometric encryption
guards your electronic identity

For Assistant Professor Stark Draper, identity theft means much more than a lost credit card number. Draper is developing a novel type of encryption for biometric data—such information as fingerprint and iris scans, DNA profiles and the like. His research may add an extra layer of protection to your identity.

Now, for example, if your credit card information is stolen or the credit card company’s database is broken into, the company simply issues a new card with a new number. “But since you’ve only got ten fingers, if that happens to your fingerprint, you’ve got a problem,” says Draper.

He studies a type of data encryption that prevents burglars from retrieving original biometrics from the data stored in a database or security program, while allowing the legitimate owner of the biometric data to verify his or her identity.

For example, a user could set up a fingerprint-scan lock on his or her computer. Using Draper’s “secure biometrics” approach, the computer would be able to verify the user’s fingerprint based on stored data. “But if someone broke into your computer and looked at the data stored there, they couldn’t replicate your fingerprint,” says Draper. “That’s different from how current biometric systems work. Right now they just store your biometric in a recognizable form.”

The underlying ideas have applications beyond biometrics. The ideas also can be used in wireless ad-hoc networks to derive encryption keys from natural phenomena in an eavesdropper-proof manner, and even form the basis for reliable communication across the backbone of the Internet.

Power struggle:
Advocating for energy consumers

Too-high electricity bills can leave consumers wondering if their energy providers are cheating them. According to Associate Professor Bernard Lesieutre, they just might be.

Lesieutre and his research group are trying to determine whether electricity suppliers can manipulate the markets to their advantage.

Current guidelines overseeing energy markets are based on financial models and regulations for market share; however, those models don’t take into account the physical limitations of the energy grid. Power lines have a limited capacity for how much electricity they can carry before they become congested, and too much power can physically warp the cables. Even an exceptionally hot day could reduce the amount of energy the lines can tolerate.

Because the lines cannot carry any more power, when conditions create congestion, competitors might not be able to supply power to where it’s needed. As a result, one power company might become the only provider in an area for a time. “If they know or can guess that, they can raise their prices to make more money,” explains Lesieutre. “They know their electricity is no longer substitutable. People can’t get their energy from somewhere else because the grid is overloaded.”

Based on sensitivity analyses, Lesieutre’s group has determined that such inflation is possible. While there are regulations for substantial manipulation, current measures only apply to instances where prices increase by 300 percent or more.

“Our concern is this high threshold. It doesn’t detect a lot of times when rates are noncompetitive,” says Lesieutre. “Our research is to come up with something with a much finer resolution than that.”

Having identified scenarios with potential for market manipulation, the group’s next step is to develop measures to determine when companies are taking advantage of those scenarios. Ultimately, Lesieutre hopes to prevent this market manipulation.

Back to topEnd of page