BOSTON — In the search for better tools to detect land mines, scientists are taking a closer look at the nose of man's best friend.
Designing sensors that approximate a dog's ability to smell explosives is one of several approaches researchers in the United States are pursuing to help uncover what the International Red Cross estimates to be 120 million mines worldwide.
From simple hand-held antennas to ferret out nonmetalic mines to sophisticated X-ray imagers, these emerging technologies are aimed at helping mine-removal teams with their Herculean task.
According to the United Nations, if today's approaches - inching through a minefield, gently probing with a bayonets or metal probes, or using trained dogs - were the only ones available, global demining efforts would cost $33 billion and take 1,100 years.
The Defense Advance Research Projects Agency alone reportedly is spending $25 million over three years to advance the sensor and computing technologies needed to mimic a dog's sense of smell. The Defense Department's Humanitarian Demining Technology Development Pro- gram is spending nearly $15 million this year on new technology and on finding ways to apply existing technologies to demining.
Much of the work is funded by the Pentagon, which is looking for faster, more effective ways to clear mines. But if the technologies prove effective, researchers say, they also have the potential to cut the staggering cost and millennial time scale currently projected for global demining efforts.
'We can do better'
"We can do better than a bayonet," insists Joe Mack, defense programs coordinator for the physics division of the Los Alamos (N.M.) National Laboratory. "Ideally, you should be able to come in and quickly look at an area. If you detect something unusual, you should try to image it. Then if it's got the right shape, you pay attention to it."
Ground-penetrating radar is one tool for surveying an area for objects buried just below the surface. But it has its limits, Dr. Mack notes. For example, its images get blurry if the soil is moist. He is working on an alternative - a hand-held antenna that can sense changes in electrical properties between soil and a mine's casing. He came across the device in a mine, where it was used to measure the thickness of ore veins.
"It's worked in controlled experiments," he says, adding that it also can form a crude image as it's swept back and forth over an object. He says he and his team hope to have a prototype built within the next nine months.
Once an "anomaly" is detected, the next step is to try to image it. Within its limitations, radar can be used to build a 3-D image of a buried object, Mack says.
Like a dog's nose
Even with such a 3-D image, however, the object can't be classified as a mine until explosives are detected. To reach that goal, scientists in several labs are trying to build sensors that, like a dog, can zero in on specific odors.
One approach, being developed by David Walt and John Kauer at Tufts University in Medford, Mass., uses fluorescent dyes attached to the ends of tiny optical fibers. Each dye reacts differently when exposed to the vapor given off by explosives or by bacteria as they break down explosives.
Bundled in clusters of 20, the dye-tipped fibers form a unique pattern of color and brightness when they are exposed to different compounds. Using neural networks, computers can be taught to read these patterns to identify different kinds of explosives.
Duplicating the sensitivity of a dog's nose is a key challenge. The concentrations of volatile materials given off by mines are extremely low, says Dr. Walt, a chemistry professor at Tufts, who adds that he and a colleague are working on boosting the sensitivity of their sensors. Once they do, they face another hurdle: copying a dog's ability to find a particular odor amid a sea of smells. "It's one thing to capture a smell in a lab," Walt says, "but will it work against a complex background environment that may have fuel spills, wildflowers, and other competing fragrances?"
A new technique for imaging may reduce the need for an independent "sniff" test, according to its developer, Alan Jacobs, a professor of nuclear and radiological engineering at the University of Florida in Gainesville. His team has developed a device that uses X-rays to image mines. It shines a beam of X-rays into the ground, then detects the smaller, scattered field of X-rays that return. This approach can reveal both the shape of the object and spot air pockets unique to mine fuses.
In the end, however, "every method has its pros and cons," says Mack, so advances in mine detection are likely to come using a variety of sensors. The bottom line, he says, is producing something that works in the field.
"I have a son in the Army. He's in Bosnia," Mack explains. "He says: 'All you guys at the national labs ever do is design physics tools. Can I drop this thing in the mud and will it still work?' That's an important reality check."