SAN FRANCISCO NASA may mimic the ant in its future efforts to explore the belt of asteroids that lies between Mars and Jupiter, a resource-rich region that astronauts might tap as humans move out in the solar system.

A small group working at NASAs Goddard Space Flight Center has proposed launching a massive colony of miniature spacecraft using social insect-inspired artificial intelligence to spend several years prospecting among perhaps 1,000 of the space rocks.

"The idea is to have a totally autonomous swarm you can send out to explore multiple bodies," said Steven Curtis, a scientist at the Greenbelt, Maryland, NASA field center.

Called ANTS thats "Autonomous Nano Technology Swarm" the fleet of buggy spacecraft would cruise independently to the Asteroid Belt. Each futuristic probe would hoist its own solar sail to capture the minute pressure of the suns rays and push it along on its journey.

The tiny probes, each weighing about 2.2 pounds (1 kilogram), would then fan out among the hundreds of thousands of asteroids at least 0.62 mile (1 kilometer) in diameter or larger.

"You build a bunch of them and just throw them out there," said Curtis, who recently presented the concept at the fall meeting of the American Geophysical Union.

Scientists look to the Asteroid Belt for multiple reasons that include their potential as threats to Earth and as sources of raw materials. For example, the asteroids contain metals otherwise prohibitively expensive to launch into space from Earth that could be employed in the further exploration of the cosmos.

Even a cursory but up-close survey of the Asteroid Belt, and the elemental composition and types of the bodies found there, would allow astrogeologists to pinpoint which individual rocks would merit mining.

"You could rapidly determine which are the most important mineral resources in terms of future exploitation," said Donald Yeomans, an asteroid and comet expert at NASAs Jet Propulsion Laboratory (JPL).

The ANTS mission would not launch before 2020 at the earliest, and only then from human-occupied space stations parked in Lagrangian orbits, or L-points, where the gravitational forces of Earth and sun are perfectly balanced, Curtis said.

The swarm would take three or more years to travel to the main belt of asteroids. Once there, 100 ruler and messenger "ants" would look on, guiding operations, as the 900 or so worker probes did the bulk of the work. Only a small number of messengers would then make the return trip to their space-based safe harbor, ferrying with them the data acquired during the mission.

Each worker would carry a single instrument, whether a magnetometer, gamma ray sensor or some other tool, to perform a specific task.

"Basically, [it's] everything youd need to characterize the asteroids from remote sensing," Curtis said of the complement of instruments. Instead of a single, large spacecraft though, those functions would be "smeared" across the 1,000 tiny probes.

That large number builds redundancy into the system -- lose 20 percent of the spacecraft and the mission can still succeed. It also lowers the cost of a mission, since NASA could churn the probes out on an assembly line.

Most importantly, the scheme would permit individual probes of modest means to collectively perform Herculean efforts just as occurs in the realm of the ant and other social insects.

"Insects do a lot with what they have," Curtis said.

Indeed, scientists are increasingly looking at ants and other creatures in designing robotic networks that mimic the collective, emergent intelligence in the behavior found in places like the colony of insect critters in your backyard.

"The ant colony as a whole, some people refer to it as a superorganism, because any one individual isnt that intelligent. But if you view the ant colony as an organism itself, it does intelligent things," said Tucker Balch, a research scientist at Carnegie Mellon University who models ant behavior for adoption in future robotic systems. "The lesson is, you can build a complex, almost intelligent, system with many inexpensive and possibly disposable parts."

In the case of the ANTS proposal, the probes would perform their tasks individually, but at the same time would swap what theyve learned back and forth in a way that makes the collective behave Borglike as would a single, larger spacecraft.

Elsewhere within NASA, scientists are developing similar concepts for application in planetary exploration. At JPL, for instance, the Sensor Webs Project seeks to develop an independent network of wireless sensor pods that could be deployed to monitor and explore a limitless range of environments.

Like space ants, a sensor webs individual components would communicate among themselves, allowing for the diffused yet orchestrated exploration of dynamic regions like the northern polar cap of Mars.

"The thing with sensor webs is you really want the emergent behavior to resemble conscious thought for the whole being rather than the individual," said Kevin Delin, leader of the JPL project. "The important thing is there is a global purpose to their behavior, rather than just saying, Were all cooperating."