Not that Li has aspirations to kick up the red dust himself.

Instead, as a professor of geodetic science at Ohio State University, he's working to develop the navigational tools to guide robotic rovers across that alien world in 2003 and beyond.

Li was alerted to NASA's rover problem in 1997 when he watched television images of the rover Sojourner bumping into rocks during the Mars Pathfinder mission.

"One thing which caught my eye (watching Sojourner) was the navigation," he recalled.

The rover would approach a big rock, stop, then try to climb up its surface, he said. If the rover had flipped over, its solar panel would have been pointed toward the ground, and that would have put it out of action because it was solar powered.

Li called experts at NASA's Jet Propulsion Laboratory in California to see whether there might be a better way to guide the rovers, especially as they move farther from their spacecraft.

Sojourner ventured out only about 33 feet based on preset course headings and crude estimates of distances to rocks.

But rovers during Mars landing missions this decade will wander miles from their mother ships. That, Li said, will require a hundred-fold improvement in the accuracy of systems used with Sojourner.

Travelers normally collect their maps before they arrive, and Martian rovers would be no exception.

Li has proposed that rover maps be made from a series of camera images of the landing area snapped from the spacecraft as it descends. Another option is map making from photographs taken by satellites orbiting Mars.

The aerial views would be compared with pictures taken on the ground from a camera mounted on a mast atop the rover.

The resulting computer-generated maps should be able to locate obstacles and terrain features to an accuracy of about an inch close to the spacecraft and about 30 feet 10 miles away from the lander, Li said.

Li and his team, postdoctoral student Kaichang Di and graduate student Fengliang Xu, have a four-year $440,000 grant to develop the computer software for the mapping strategy.

Tests of the system have been completed in California, with a helicopter substituting for the descending spacecraft and a dried-up lake bed substituting for Mars.

A NASA manager said the agency plans to use the OSU navigational software for rovers during Mars missions in 2003 and 2007.

In 2003, the rover will roam more than a half mile from its landing craft. In 2007, the range could be extended to more than 6 miles.

"We are fairly enthusiastic about what he's doing," said Larry Matthies, supervisor of the Machine Vision group at the Jet Propulsion Laboratory. "It's not the only game in town, but we expect to use it."

The goal, Matthies said, is to be able to direct the rover to interesting terrain features photographed either by the descending spacecraft or by the orbiting satellites. 

Navigation also could be critical in future rock-recovery missions if the rover has to return its cargo to the spacecraft for the trip back to Earth, he said.

The unexplained disappearance of the Mars Polar Lander as it descended to the Red Planet in 1999 was a setback for Li's project because he had hoped to use images taken from that mission.

But Li said he was encouraged by the ability of the NEAR-Shoemaker spacecraft to take a series of photographs as it lowered itself onto the Eros asteroid last month.

Landing on an asteroid and landing on Mars are different, but the OSU team hopes to use asteroid images to perfect its system, Li said.

dlore@dispatch.com

Caption: Ron Li, a geodetic scientist at Ohio State University, is helping NASA.