Pittsburgh, PA
Saturday
December 3, 2022
    News           Sports           Lifestyle           Classifieds           About Us
Health & Science
 
Place an Ad
Running Calendar
Travel Getaways
Headlines by E-mail
Home >  Health & Science >  Environment Printer-friendly versionE-mail this story
CMU robotics pioneer sets sights on old mines

Wednesday, August 28, 2002

By Byron Spice, Post-Gazette Science Editor

In the mid-1980s, experimenters sent a six-wheeled robot called Terregator crawling through the U.S. Bureau of Mines' research mine in South Park Township, bouncing laser beams off the tunnel walls to make measurements for a map.

The ungainly machine was about the size of a desk, averse to water and forever tethered to a power supply, but it produced some amazingly precise maps of the old coal mine.

The Bureau of Mines is defunct and Terregator is now a museum piece, but those experiments by the Robotics Institute of Carnegie Mellon University have taken on new significance in the weeks since the Quecreek Mine accident, in which nine miners were trapped for three days by flooding from an adjacent, abandoned mine.

Inaccurate maps of the abandoned Saxman Mine played a key role in the Quecreek inundation, and that has mining experts puzzling once again over how they can do a better job of mapping the boundaries of long-closed, inaccessible mines.

This morning, the latest in mapmaking robots will be demonstrated for Sen. Arlen Specter, R-Pa., and state Sen. Tim Murphy, R-Upper St. Clair, during a briefing at the National Robotics Engineering Consortium in Lawrenceville.

That doesn't mean coal companies or government regulators are about to make robots the standard method for scouring the abandoned mines scattered throughout the country. There are other techniques, that can detect old mine workings and that might be improved through further research.

But some technologists say it is time to talk about using robots to enter the old mines and obtain measurements that would be too risky, if not impossible, for humans to make.

A robot like Terregator wouldn't be of much help, but Carnegie Mellon robotics pioneer William "Red" Whittaker says advances in automated mapping methods, power sources and computers have made it technically feasible to build a robot that's up to the task.

What Whittaker envisions is an amphibious robot compact enough to enter the mine through a borehole, "transform" into its operational configuration and use laser range finders or sonar arrays to map the mine as it swims or crawls through it.

An earlier Carnegie Mellon robot, called Houdini, was similarly designed to fold itself up so that it could enter hazardous waste storage tanks through existing access pipes and then reconfigure itself so it could move around and work inside the tank.

Whittaker, famed for designing robots to inspect the crippled Three Mile Island nuclear reactor, explore an active Alaskan volcano and search for meteorites in Antarctica, hopes to have a prototype assembled by Christmas.

A Penn State skeptic

Not everyone is as enthusiastic about robots as Whittaker.

Christopher Bise, head of the mining engineering program at Penn State University, doubts whether the abandoned mine problem requires such a high-tech solution. The idea of sending robots into every abandoned mine, he said, "is ridiculous."

Drilling test cores 20 feet or more in advance of the mining face, a standard practice when miners believe they are close to another mine, has worked well for years, Bise contended. And even if that practice is deemed inadequate -- the Quecreek miners didn't do core drilling because they didn't think they were close to the Saxman workings -- noninvasive geophysical analysis methods, such as seismic studies or electrical resistivity readings, should do the trick, he said.

Others are less pessimistic about robots.

Franklin Orr Jr., head of a National Research Council panel that studied the abandoned mine issue and dean of earth sciences at Stanford University, and panel member Don Steeples, a geophysicist at the University of Kansas, both agreed that robotic mapping is a technology worth exploring.

"We need to bear in mind that robotics is one of those things that's going to get cheaper and cheaper," Steeples said. Today's $250,000 robot, an extravagance for any mining operation, could well become tomorrow's $10,000 to $20,000 everyday tool, he suggested.

Steeples' own expertise is in seismic analysis, in which explosives, air guns or hammers are used to generate ground vibrations and the resulting seismic waves are then recorded. Careful interpretation of those seismic signals can sometimes reveal underground mines.

It's a technique that's proven helpful in studying abandoned coal mines under Pittsburg, Kan., which have caused subsidence problems.

"But we've got a whole lot of things working in our favor there," Steeples emphasized. The topology is flat, the geology is simple and the mines are only 20 to 100 feet below the surface, he said.

Tricky geology

Appalachia's hills and valleys, complex geology and much deeper mines make it much more difficult -- and expensive -- to detect old mine workings, Steeples said. Seismic waves are harder to interpret, for instance, when the surface is uneven and the underground rock layers themselves can be jumbled.

Finding the right mix of methods is crucial, he said. It's entirely possible, for instance, to locate an abandoned mine simply by drilling deep holes in a 3-foot grid. But that's prohibitively expensive and environmentally hazardous.

Ground-penetrating radar is only good for 10 or 20 feet. Seismic methods are only accurate to depths of 50 feet, though it also is possible to do seismic readings from inside the mine itself. Microgravity detectors, which look for variations in mass density, are better for finding air-filled voids than water-filled voids. Measurements of electrical potential -- the electrical resistance of rock -- are better at detecting water than air.

Further development might make all of these techniques easier to use, cheaper and more accurate, he added.

Though it has been years since the Carnegie Mellon Terregator experiments and related mapping experiments in a limestone mine, the creation of maps is a technology that is at the basis of most mobile robots, such as the driverless autonomous vehicle program at the Robotics Institute.

In particular, Whittaker said, a technology called SLAM -- for simultaneous localization and mapping -- has proven to be one of the great advances of the last five years. The SLAM program is able to estimate the position of walls, roofs and other landmarks around the robot and the position of the robot itself at the same time.

The earlier mine mapping experiments were performed in dry tunnels, but automated mapping can be performed -- and might even be better suited to -- underwater tunnels, said William Stone, an engineer and diver who has been called the Jacques Cousteau of caves.

Rescue robots?

The day after the Quecreek inundation, representatives of the National Cave Rescue Commission consulted with Stone about the possibility of human divers reaching the nine trapped miners. A rescue probably could have been attempted if the distance to the miners had been less than 1,000 feet, he said. But he told the rescue group that swimming up to 8,000 feet in water with zero visibility and filled with mining equipment was too hazardous for human divers.

A traverse that's too risky for a human, however, might be just the sort of mission appropriate for a robot, said Stone, a North Allegheny High School graduate who heads a robotics group at the National Institute of Standards and Technology in Gaithersburg, Md., as well as a firm called Stone AeroSpace. In addition to mapping abandoned mines, an amphibious robot might be able to deliver supplies or establish communications with trapped miners.

Stone, Whittaker and the Southwest Research Institute have proposed building a fully autonomous robot called DEPTHX that would be used to map Zacaton, a sinkhole in northeastern Mexico that is more than 1,080 feet deep and thought to be the world's deepest underwater cave.

The team has sought funding from NASA, with the intent of demonstrating technologies for mapping and searching for signs of life in Europa, the moon of Jupiter that may have a deep layer of water under its surface ice. If the project is approved, Whittaker's group would build the SLAM-based mapper, while Stone would provide the robot.

Whittaker said he also will need some sort of financial support to build his prototype of a mine mapper. He nevertheless plans to move ahead with the project, and perhaps also make it a focus of a fall class, while he searches for groups willing to finance it. People can always find plenty of reasons not to tackle such a project, so the only way to get it done is to just start doing it, he reasons.

"You can't hold your breath waiting for someone to say, 'Here's what we should do,' " he said.


Byron Spice can be reached at bspice@post-gazette.com or 412-263-1578.

Back to top Back to top E-mail this story E-mail this story
Search | Contact Us |  Site Map | Terms of Use |  Privacy Policy |  Advertise | Help |  Corrections