May 13, 1997

Hardier Breed of Antarctic and Lunar Explorers: Robots

By WARREN E. LEARY

ASHINGTON -- If a place like Antarctica was discovered on another planet, scientists from Earth undoubtedly would send hardy robot spacecraft to explore such an intriguing but desolate frozen world.

To operate in such a harsh environment, the robot explorer would not only have to be rugged, but also smart enough to fend for itself without immediate human direction if it ran into the unexpected.

Naturally, the place to test such robots would be Antarctica itself. And scientists and engineers from many disciplines have begun thinking about new ways to bring robots and Antarctica together. Under the proposals, Antarctica would serve simultaneously as the subject of robotic exploration and as a place to test robot technology that could operate on the Moon and beyond.

Plans are under way for using robots to gather meteorites that cluster in certain areas because of glacial flows. And researchers are discussing using robot probes to penetrate and explore a vast, ice-covered Antarctic lake as a prelude to a possible mission to an ice-covered moon of Jupiter that may have a subsurface ocean.

Antarctica is no stranger to robots. Its dry valleys served as a model for instruments sent to the cold, windy plains of Mars on Viking spacecraft in the 1970s. And in 1993, an eight-legged walking robot named Dante made an abortive attempt to creep into the crater of Mount Erebus, the continent's only active volcano. Dante was thwarted after going only 21 feet into the crater when a communication line broke.

"We need more experience operating robots remotely and we can use Antarctica as an analog or substitute for the lunar surface," said David B. Lavery, manager of the Telerobotics Research Program in the Space Science Office of the National Aeronautics and Space Administration. NASA is sponsoring the meteorite search project.

Since 1969, researchers from the United States and Japan have collected more than 16,000 meteorites in areas near the Allan Hills and the Yamato Mountains. This cache of meteorites has included scores of pieces blasted away from the Moon by impacts with asteroids and about a dozen samples that are believed to have originated on Mars.

Last August, NASA scientists announced that a potato-size rock found in Antarctica in 1984, designated Allan Hills 84001, originated on Mars and carried chemical and possible fossil evidence of primitive microscopic life.

The major problem in looking for these valuable samples is that Antarctica's inhospitable weather gives human searchers only six to eight weeks a year to search for them each summer, and occasional blizzards or logistical problems often cut into this time, researchers say.

This provides an opportunity for "meteorobot," the meteorite collector to be built by Carnegie Mellon University's Robotics Institute in Pittsburgh under a $5 million grant recently awarded by the space agency.

Dr. William L. Whittaker, the principal researcher on the project, said finding meteorites "is a profoundly difficult task for a machine," but one that would push the technology forward in order to master more complex extraterrestrial duties in the future. "Robots will be the exploration agents to other worlds and lots of basic training has to go on right here on Earth," Whittaker said.

Engineers envision a wheeled robot powered by an internal combustion engine that would slowly move an array of sensors in a detailed search pattern across the ice. The eyes of the robot would be high-resolution color cameras, perhaps with an infrared ability for seeing in the dark, that allow operators to move sensors close to objects of interest. Instruments would include a metal detector, a magnetometer for detecting the magnetic properties of metal-bearing rock and a spectrometer for examining the chemical composition of samples.

In initial operations, Whittaker said, a robot would cross a field of blue ice to detect surface meteorites, even those obscured from human sight by drifting snow, and perhaps some that are a few inches below the ice. As each is detected, onboard computers would record sensor data as well as position, depth, size and geometry, and the robot would plant a flag to mark the spot for human searchers to find later.

Project plans call for Whittaker and his colleagues to go to Antarctica in November to test sensors, power units, communication systems and other components. The next year, sensors are to be put on a robot and it would be put in the field under close human supervision. If all goes well, several robots will be put on the ice in late 1999 to look for meteorites.

Some veteran meteorite hunters are hopeful, but skeptical, of the robots' chances of success. "I don't know if TV cameras are good enough to do what the human eye and experience does," said Dr. William A. Cassidy of the University of Pittsburgh, who recently retired from field work after decades of searching Antarctica for meteorites. Members of a team led by Cassidy discovered the Allan Hills Mars rock that reportedly contains evidence of life.

"A big part of picking out a meteorite from other rocks scattered around is experience, just knowing what to look for," Cassidy said. "Still, it would be kind of fun to sit back in an office somewhere using a robot by remote control to hunt for meteorites."

To test the principle of having a robot in a desolate area operate either on its own or by remote control, the Carnegie Mellon researchers are to put a self-contained robot called Nomad out for a field trial this summer. In July, Nomad is to do a 125-mile trek across a Moon-like section of the Atacama desert in northern Chile. Using four-wheel drive and four-wheel steering, the 825-pound vehicle is to traverse a changing area of hills, craters, rocks and sand either on its own or while under control from a center in Pittsburgh. Using a special panoramic camera with a 360-degree field of view and four additional stereo cameras as its eyes, the machine is to creep along at about a foot per second, moving from one type of terrain to another.

Whittaker says Nomad is a generic prototype for both the meteorite-hunting robot and one that could operate on the Moon. Signals relayed by ground lines and satellites are delayed several seconds, simulating control delays that would occur between Earth and the Moon during a lunar mission, he said.

Another possible Antarctic robot mission would involve sending a twin-robot probe through 2.5 miles of ice permanently covering Lake Vostok, an immense body of water frozen off from the outside world for thousands of years. Interest in the project has been spurred by close-up pictures taken of Jupiter's moon Europa, which appears to be covered by deep layers of fractured ice that may have liquid water beneath which could support some form of life.

Researchers at NASA's Jet Propulsion Laboratory in California have held several workshops to bring together ideas for ice-penetrating robots that could be tested in Antarctica in advance of a possible mission to Europa. "It's not a formal proposal yet because we are still trying to figure out how to do it," said Joan C. Horvath of the laboratory's Advanced Concepts Office.

The initial idea is for a robot called a Cryobot that would give off heat to melt its way through ice. It would probably be heated by internal generators powered by radioactive isotopes. "The concept involves unspooling a cable from the Cryobot as it goes down, creating a thin layer of water around the probe as it sinks that refreezes as it passes through," Ms. Horvath says. The cable would be connected to a surface communications station for sending data back to Earth, perhaps relayed by a satellite orbiting Europa.

Once the Cryobot penetrated the ice and entered the body of water beneath, it would release a stowed robot called a hydrobot, which would move away from the penetrator to take pictures and look for signs of life. There is a possibility that an under-ice ocean on Europa might have volcanic vents like those found in the icy depths of oceans on Earth that could sustain unusual forms of life, experts said.

One of the main technical hurdles facing designers of ice-penetrating robots is reaching the depths without contaminating the water with microbes from the surface, intermediate ice or the probes themselves. "We want to answer the contamination question before sending probes into Lake Vostok or, later to Europa," Ms. Horvath said. "And we'd probably want to test the concept first in Greenland to be sure. Greenland would be a test for Antarctica and Antarctica would be a test for Europa."

---

Other Places of Interest on the Web

NASA Space Telerobotics Program

---