Artificial Intelligence to Boost Space-Probe Efficiency
Richard A. Lovett
for National Geographic News
|May 30, 2006|
In a shift some hail as a revolution in space technology, scientists are reprogramming existing space probes to make more decisions on their own.
Experts say artificial intelligence will help unmanned spacecraft work more efficiently and send better data back to Earth.
Rebecca Castaño of NASA's Jet Propulsion Laboratory (JPL) in Pasadena, California, says even the current Mars rovers Spirit and Opportunity will soon benefit from smarter software. Their upgrades are scheduled for this summer.
(See National Geographic magazine's 2005 Mars rover roundup.)
As part of their mission, the rovers' scan for atmospheric phenomena, such as dust devils (small sandy whirlwinds) or clouds.
But because the probes can't recognize these features on their own, they waste time beaming photos of little or no interest back to Earth. That will change in coming weeks, when a planned software upgrade will teach the rovers how to look for clouds and dust devils and instruct them to send back only the most useful images.
Similarly, scientists are working to provide greater autonomy to probes now orbiting the red planet.
These satellites can take many more photos than they can send homea bottleneck scientists lament.
"If we can analyze [that data] onboard for features of interest, it's easier to get more good stuff," Castaño said, speaking last week at a meeting of the American Geophysical Union (AGU) in Baltimore, Maryland.
Volcanoes, Floods, and Icebergs
Steve A. Chien, a JPL computer scientist and AGU panel speaker, said that the new approach, called onboard autonomy, is already in use on the Earth Observing-1 spacecraft, or EO-1.
The Earth-orbiting satellite is designed to capture high-resolution images of geological events, such as volcanic eruptions, floods, and the breakup of large polar ice sheets.
But the space probe can't study the entire planet in detail all at once.
In the past, engineers instructed the satellite where to train its instruments when important geological events were occurring.
Chien says the inefficient process carried high staffing costs and the prospect of delays.
But no longer.
For more than a year, the satellite has been programmed to make its own decisions.
If the unit spots a volcanic eruption, EO-1 reprograms its mission and zooms its cameras in for a closer look. Likewise, if the satellite detects a flood, EO-1 takes pictures as soon as possible.
The upgrade, Chien says, has increased the satellite's scientific contribution a hundredfold, while cutting operating costs by more than a million U.S. dollars a year.
Experts think the greatest returns from onboard autonomy, however, may come from future missions. In upcoming years, robotic probes may explore far-flung regions of the solar system, such as the area around Saturn's gigantic moon, Titan.
Earth-based scientists have already remotely visited Titan once, via the Huygens probe, which successfully parachuted to the surface in January 2005. (See "Huygens Sends Images of Titan" [January 2005].)
Ralph Lorenz, an assistant research scientist for the Titan mission based at the University of Arizona's Lunar and Planetary Laboratory in Tucson, spoke at the AGU meeting last week.
He believes Huygens barely scratched the surface of what can be learned from Titan.
Bypassing Mission Control
"We've had our taste buds tantalized by the Huygens probe," Lorenz said. "I think we're going to want to go back."
The scientist says that, unless NASA wants to send a fleet of landers, which is highly unlikely, the best approach is with a mobile probe, such as a balloon-like device that could navigate through Titan's dense atmosphere.
Given Titan's distance from Earth, it takes three hours to send a message between the Saturn moon and our planet.
That fact, Lorenz says, makes it more necessary than ever to empower future probes. In this way, robotic explorers can pick and choose relevant images and avoid sending useless data.
Lorenz compares such a tool to an efficient Web search engine like Google. "It's not a tool that's going to supplant scientists," he said. "It's a tool that allows scientists to do their jobs more effectively."
The researcher adds that the communication lag between Earth and Titan rules out the possibility of flying a probe from an Earth-based mission control center.
"You need the [probe to have] autonomy for orders like 'Try to head north,'" Lorenz said.
Lorenz says that given the rate of robotic advances, it's hard to determine what type of autonomous computing will be possible when a new Saturn probe is under development.
But while the details are difficult to predict, he said, "This sort of thing will be the key to making the most of the mission."
Free Email News Updates: Voted Best Online Newsletter, 2006 Codie Awards
Sign up for our Inside National Geographic newsletter. Every two weeks we'll send you our top stories and pictures (see sample).
|© 1996-2008 National Geographic Society. All rights reserved.|