A self-directed robot working in the driest spot on Earth is helping a team of scientists in the search for life? on Mars.

Robots destined for work in space get their start in Pittsburgh on the campus of Carnegie Mellon University. The Field Robotics High Bay is where scientists using an array of components thick cables, computer boards, wires and vehicle parts build robots with a mission.

"There's a robot over there that was designed to work with the space shuttle," explains guide David Wettergreen, a research scientist at Carnegie Mellon University's Robotics Institute. "There is a robot across the room that has been to the Antarctic, I think three times now searching for meteorites. And on the far end there is a group of three robots that work together to do assembly tasks for a space truss or a structure that would be assembled in space."

David Wettergreen heads the field robotics team working on the technology for a smart machine that will work, someday soon, on Mars. The three-year project, funded by the U.S. space agency, NASA, began six months ago in the Atacama Desert in northern Chile, the driest spot on earth.

"There is no vegetation at all," said Mr. Wettergreen. "So there are no bushes, no cactus or sagebrush blowing around. It is just exposed soils and rocks."

"That's right," he continued when asked if they are going there looking for life. "The life that we are looking for is microscopic. We are looking for bacteria. And in fact in some areas of the desert there are microscopic organisms, some lichen that survive in the rocks. So we are studying where those organisms live and where they don't live, and then trying to understand why [do they live] here and not there." David Wettergreen says the Atacama is an ideal test laboratory for the Mars mission. Its aridity, soil composition and extreme ultraviolet radiation are analogous to conditions the Viking Lander spacecraft actually found when it set down on a rocky Martian plain 30 years ago.

"So, if we can understand better what is happening in the Atacama and measure that in a way that we can relate it to what we know about Mars then we can formulate ways to investigate Mars or better understand information we are getting from Mars," he said.

The team built a four-wheeled, solar-powered autonomous vehicle named Hyperion. The rover senses the position of the sun so it can keep its batteries charged, and it will be able to regulate its activities so it has enough power to complete its mission.

Experiments on board the rover also test its ability to move into unknown terrain using cameras and internal sensors. "What we are trying to do is develop a rover that can traverse long distances and as it makes that traverse take specific measurements about the geology and biology of the terrain," continued Mr. Wettergreen. "So, it is taking pictures, recording spectra, which tell us about the mineralogical and the chemical compositions of soils and it is also looking for indications of life."

Alan Waggoner, the director of the Molecular Biosensor and Imaging Center at Carnegie Mellon University College of Science, is in charge of the life-detection instruments mounted on the rover.

"Our approach was to think about what molecules life on Mars might contain," he said. "When earth formed in the solar system four and one half billion years ago, life evolved on earth quite quickly. And a lot of people think that the environment on Mars might have been very similar, in which case they would have had life very early on. Well if there were places on Mars where there is water and enough warmth to maintain that life, it may still exist on Mars."

One approach to finding that life is to shine specific wavelengths of light known to be absorbed by chlorophyll molecules in plant cells and to look for the fluorescent signals they would emit. Another is to use fluorescent dyes that bind with the building blocks of life - DNA, proteins, carbohydrates, and lipids.

Alan Waggoner built an imaging microscope to do the job. "The imaging system would look for evidence of photosynthetic system florescence," explained Mr. Waggoner. "Plus, the rover would spritz or spray the ground with a solution with the four different florescent dyes. It [would] wait a certain period for the dyes to incubate with the bacteria or bio-films, and then it would image those regions that had been sprayed as well.

"Hopefully we could do this at a low resolution initially and see if there are any obvious signals coming, [and then] microscopically be able to look in crevices of rocks and in much smaller areas for evidence of very small patched colonies or patches of microorganisms," he added.

After the first tests in the Atacama Desert in April, Hyperion was brought back to the Carnegie Mellon University Field Robotics High Bay for an overhaul.

These look like the guts of a robot.

"They certainly are," said Mr. Wettergreen. "Right now we've taken out a number of components that we will be reworking and replacing and we are developing new wheels for the robot to better operate in the type of soft soils that we are encountering, and we are changing the steering mechanism."

Although Hyperion proved that it could outdistance previous planetary rovers, David Wettergreen says that in 2004 the goal is to have it travel 50 kilometers on its own, up from 20 kilometers in 2003.

And by the end of next year, the science team plans to be ready for a full-dress simulation setting the rover loose in the Atacama Desert as if it were exploring Mars.