A Pittsburgh robotics company, already competing for the Google Lunar X Prize, is making future plans for robots -- hopefully its own -- to build NASA's first lunar launch pad and landing site.

Astrobotic Technology Inc., an Oakland-based company led by renowned Carnegie Mellon University roboticist William "Red" Whittaker, has announced its strategy to use robots to build a moon pad at one of the moon's poles.

With technical assistance from CMU's Robotics Institute, Astrobotic presented its strategy Friday during a NASA Lunar Surface Systems conference co-sponsored by the U.S. Chamber of Commerce and its Space Enterprise Council in Washington, D.C.

The pad is scheduled for completion in 2020.

The key challenge is protecting people, their habitats and equipment from moon dust, which will be sent flying at sandblasting speeds during moon landings and takeoffs.

Astrobotic's plan calls for two robots, and possibly a third for standby, to work around the clock for almost six months to build a berm to prevent dust or regolith from traveling at high speeds in low gravity and sandblasting the outpost.

A second option calls for robots to construct a rock-lined launch pad so landings and blastoffs do not disturb dust. But that option depends on whether enough rocks exist within a reasonable distance from the moon's poles.

"For efficient cargo transfer, the landing site needs to be close to the outpost's crew quarters and laboratories," Dr. Whittaker said. "Each rocket landing and takeoff, however, will accelerate lunar grit outwards from the pad. With no atmosphere to slow it down, the dry soil would sandblast the outpost."

Robots already on the drawing board will shovel moon dirt or gather rocks into bins, then transport the payload for use in building a berm or a pad.

If NASA chooses the first option, lawn-mower-sized robotic rovers weighing 330 pounds each could build an 8.5-foot-high berm in a 160-foot semicircle in fewer than six months. That project would require the robots to move 2.6 million pounds of lunar dirt.

According to the second option, small robots would comb the lunar soil for rocks to be used to pave a durable grit-free landing pad, said John Kohut, Astrobotic's chief executive officer. "This might reduce the need to build protective berms."

Before such decisions are made, robotic scouting missions are necessary to gather more details. Researchers also must determine how much force, and how much energy, is needed to dig lunar soil.

"We need to get more data about lunar soil at the polar areas -- how easy it is to scrape up and collect and how abundant rocks are at the poles," David Gump, Astrobotic president, said. "Right now, we and NASA are making some pretty gross assumptions."

Moon dust -- a fine, spiky, self-cohesive dust formed by micrometeoric impacts on the moon -- represents the bane of lunar exploration. Human landings already have proven that moon dust readily penetrates and could threaten operation of equipment.

"The biggest technical challenge that we will face is protecting ourselves against the very fine moon dust," Mr. Gump said. "You must seal your motors, your axles and any joints."

Astrobotic and CMU officials, with their longstanding experience in building robots for difficult terrains, are interested in producing the robots for the NASA project once more information is gathered.

"We are positioning ourselves to be the first robotic picks and shovels to build NASA's first lunar base," Mr. Gump said.

Details of the study and lunar imagery are available at www.astrobotictech.com.

Dr. Whittaker also is leading Astrobotic in developing its first lunar robot, which has been undergoing field trials for months. The company's goal is to win the $20 million Google Lunar X prize by sending a robot to the Apollo 11 landing site and transmitting high-definition video back to Earth. That project is scheduled for liftoff in December 2010.

To mark the 40th anniversary of the Apollo 11 moon walk in July, Astrobotic plans a series of news conferences to demonstrate its prototype lunar robots.