Dante Tours the Inferno

By MICHAEL D. LEMONICK

It wasn't until refrigerator-size boulders began hurtling down from above that the scientists sitting in an Anchorage, Alaska, control room started to get seriously worried. Until then the robot known as Dante II had successfully negotiated a steep, muddy descent and ambled unconcernedly through hot steam and poisonous gases. But even a 10-ft.-tall, 1,700-lb. automaton has its limits, and multiton chunks of rock moving at high speed were beyond Dante's. "That big one," said Carnegie Mellon University robotics expert John Bares, pointing nervously at a video screen after a rockslide, "would've wiped us out."

In the end, it was a misstep, not a rock, that toppled Dante, and only after the robot had completed its main mission: a detailed study of the crater floor 300 ft. below the rim of Alaska's active Mount Spurr volcano that included a 3-D survey of the hellish terrain and an analysis of gases issuing from belching vents. Among the significant results: the first maps of the crater's surface, normally hidden by outcroppings and haze. Dante also discovered scant sulfur dioxide and hydrogen sulfide in the noxious air, implying that the volcano, which erupted in 1992, will probably stay quiet for a while.

But important as this news was to volcano experts and the people of Anchorage, just 80 miles from Mount Spurr, the volcano study was perhaps the least noteworthy part of the robot's mission. Despite the final slipup, which toppled Dante and left it stranded on the steep mountain slope, the 10-day trek went a long way toward proving the potential of a technology that could let humans explore a wide range of sites too hazardous to visit in person -- other volcanoes, deep caves, the barren wastes of Antarctica, the ocean floor and even the surfaces of the moon and Mars. "The robot has performed like a champ," says David Lavery, manager of the Telerobotics Research Program at NASA, which paid for most of Dante II's $1.7 million development cost.

Dante II is the brainchild of Bares and William ("Red") Whittaker, the principal research scientist at Carnegie Mellon's robotics lab and a legend among robot designers. Whittaker helped design the machine that cleaned up the Three Mile Island reactor after its near meltdown in 1979, and he oversaw development of a system that will automatically inspect the heat-resistant tiles on NASA's space shuttles.

Dante is perhaps his most sophisticated product. Its vaguely spider-like aluminum body has eight legs, four of which are always on the ground; that provides maximum stability as the machine moves forward at a top speed of 3 ft. per minute, stepping lightly over obstacles up to 4 ft. high. Eight on- board video cameras enable scientists to view the terrain. Even more useful is a laser-ranging system -- a sort of light-based radar -- that makes 30,000 distance measurements every second and generates a virtual-reality computer image of the landscape. Says Bares: "It gives us a very complete picture of what's around us."

What makes Dante II truly revolutionary, however, is its four computers and their controlling software. Although the robot was connected by cable to a power generator and transmitter at the crater rim, which let the scientists direct it via a satellite hookup to the control room, Dante II can operate independently at times and did for nearly half the mission, negotiating its own path through the boulders.

That skill will be crucial if a Dante-like robot is sent to another world. On Mars, for example, says Lavery, contact would probably be limited to once a < day, and even then the enormous distances would result in a minimum 10-minute time lag in communications. Dante II is not quite smart enough for full autonomy, but considering that it took less than a year to design and build, it is remarkably close to self-sufficient. Says Lavery: "The consensus was, if we had another four or five months, we would have had that ability."

Another barrier to sending robots to the planets is weight: every pound you launch into space is expensive. At nearly a ton, Dante II would break the bank. Whittaker is already thinking about lighter models, though. And while NASA's Lavery cautions that Dante II is still "far from any sort of flight opportunity," he acknowledges that much of the technology used aboard Dante II will probably find its way into future space missions. In fact, NASA wants to launch a robot explorer toward Mars as early as 1996. And a private company working with Carnegie Mellon scientists hopes to send a Dante-like robot to the moon in 1997. The purely commercial purpose: to gather images for a gamelike, virtual-reality tour across the lunar surface.

In the meantime, Dante II-type robots should be in hot demand from earthbound volcanologists, 11 of whom have died exploring active craters in the past few years. As for Dante III, or whatever Whittaker calls the next generation, its task will be to spare humans from facing even greater dangers on other worlds.

With reporting by Patrick E. Cole/Anchorage and Mia Schmiedeskamp/Washington