Cleaning Up Signals has Widespread Applications
by Marcia Goodrich
If pressed, most of us would
define the atmosphere as the air we breathe. Professor Mike Roggemann
sees the Earth’s nitrogen-oxygen envelope in a somewhat different
light.
“That atmosphere is one
big, thick, random lens,” he pronounced. “That’s why, when
it’s hot, the car down the road shimmers. We try to undo that.”
Roggemann is among a core of
Michigan Tech researchers who specialize in signal processing. Through
the use of algorithms and computers, they take distorted signals, be they
lasers or cell phone chat, and try to make them clear again.
Like any other type of light,
lasers can be broken up, or scattered, as they travel through the atmospheric
lens. This poses a problem for the military.
In the darkness of his lab,
Roggemann illustrates. “See this speckling?” he asks, pointing
to a glittering spot where a red laser is directed. “This makes it
very hard to kill something.”
If you want to knock a missile
out of the sky with a laser (“We call them direct energy weapons,”
Roggemann notes), it has to fire straight and true—no speckling allowed.
This is no small task when the laser has to travel through a big, thick,
random lens. Roggemann is trying to overcome the effects of atmospheric
turbulence, so a laser, whether is serves as a weapon, a sensor, or a
targeting mechanism, doesn’t get diverted or distorted, not to mention
speckly, on its intended journey.
“There’s darn little
you can do to solve this problem without adaptive optics,” Roggemann
notes.
Here’s how adaptive optics
works in the lab. On the laser’s journey to its target, it deflects
off a liquid crystal. The shape of this crystal is controlled by computer,
specifically, by a specialized computer code. A camera at the target takes
a picture of the laser beam at the moment of impact. “We compare
that with our theoretical projection of what happens,” Roggemann
said. The better the projection, the better the guidance system compensates
for the atmospheric distortion.
Roggemann notes that it’s
impossible to be perfectly accurate, since the atmosphere is in a constant
state of flux. But it is possible to be good enough.
His work is funded through
the multibillion-dollar Airborne Laser Project. “You put a laser
on a 747 to hit a theater ballistic missile in its boost phase,”
he explains. The goal is to eliminate a problem encountered during Desert
Storm. Iraqi Scud missiles were successfully blown apart in Israeli skies,
only to have the pieces rain down on the populace below. Intercept the
missile right after it’s launched, and you can shower the enemy with
its own weaponry.
“Imagine a minefield--you
don’t have to imagine too hard,” says Schulz, chair of the Department
of Electrical and Computer Engineering.
An estimated 110 million mines
have been planted in 64 countries around the world, and every year, they
reportedly kill about 10,000 civilians and maim another 20,000. Under
international law, military forces are supposed to record where mines
are laid, a requirement that is widely ignored.
Mines are unseen, and that’s
what makes them so dangerous. “One way to find them is to send in
tanks or people, and when one blows up, you’ve found a mine,”
Schulz said. The Army Night Vision Lab is sponsoring research to develop
a less-lethal method, using an unmanned, flying vehicle equipped with
multispectral sensors.
Mines may be invisible, but
they do leave clues as to their presence. The soil above them heats and
cools at a different rate than the surrounding dirt. If sensors can detect
those differences and identify a few mine sites, Schulz could then use
signal processing algorithms to make a very good guess as to where the
other mines might be.
“Somebody had to put those
mines out there,” Schulz notes. “They may be in a grid, or they
may be thrown out, but there will be a pattern.”
Once you know where the mines
are, you can begin to get rid of them. But at $300 to $1,000 apiece, safely
detonating mines is beyond the means of the residents of former war zones,
which don’t tend to be among the world’s most prosperous regions.
“The big goal is to find
minefields and then not go near them,” Schulz said.
While armies are the
primary sowers of minefields, it’s no surprise that the military
is funding research aimed at detecting and eventually eliminating them.
“The soldiers are the ones on the ground, the ones who get blown
up,” Schulz notes.
A lack of confidentiality is
one. “When you make a cell call, you’re actually making a broadcast,”
Tian notes. “Usually, nobody else hears it, but sometimes they do.
. . . You can used advanced signal processing techniques to prevent that.”
This accidental public speaking
happens because the system is trying to be efficient and piles as many
conversations into one frequency band as possible, assigning each one
a code so they don’t interfere with one other. But occasionally those
codes overlap, and you can hear another conversation in the background.
By minimizing the correlations between these codes, Tian has developed
a method to transmit your call in greater privacy, so neighbors, friends,
and strangers won’t overhear your true confessions or your grocery
list.
Clarity is another issue. Cell
phones are more vulnerable to “noise” and interference than
conventional telephones. The problem can be minimized by having multiple
antennae send and receive messages. But the antennae have to work together,
and signal processing algorithms direct them to winnow out all but the
sound of conversation.
And lastly, Tian is working
with a new technique called data fusion to help solve a major irritation
among cell phone users.
Cell phones are so named because
they operate in a cell around a base station. When you travel from one
cell to another, your phone’s signal is handed off to the next cell’s
base station, just as a relay runner hands a baton. Sometimes, however,
the baton drops.
“You have to break down
the connection with one base station and set up a connection with the
base station in another cell,” Tian said. “But if the new cell
has too many users, the mobile telephone switching center, which controls
the handoff, won’t establish a connection. You’ll be cut off.”
Using signal processing and
data fusion, the switching center can tell both stations to broadcast
your call, until you cross over into the next cell and your call is successfully
handed off. “In the boundary area between the cells, it can really
help,” Tian said.
“Communications performance
has to be above a certain threshold,” Tian notes, or consumers will
rebel.
Signal processing is taking
wireless communications well beyond that watermark. Cell phone performance
could soon be as trouble-free as a chat with your neighbor across the
backyard fence.
In
contrast, Tim Schulz is using signal processing to detect the deadly detritus
of war.
Signal
processing may be key to advanced military research, but it also has applications
a little closer to home. Gerry Tian, assistant professor of electrical
and computer engineering, is using advanced signal processing to improve
the wireless communications technology that is fundamental to cell phone
use. For as reliable as cell phones have become, they still have their
weaknesses.
