Blog     e-Newsletter       Resource Library      Directories      Webinars     Apps
AFCEA logo

Land Mine Detector Makes Waves

July 2004
By Cheryl Lilie

Researchers pick up good vibrations to locate buried mines.

Exciting a land mine may not sound like a good idea, but developers of the Seismic Landmine Detection System are doing just that. A group of researchers from the Georgia Tech Research Institute in Atlanta, Georgia, has developed a land mine detection system that sends seismic waves through a minefield, slightly moving the earth and items buried beneath. A noncontacting radar sensor measures the ground displacement to identify and locate plastic anti-personnel or antitank mines.

Waymond R. Scott Jr., professor, School of Electrical and Computer Engineering at Georgia Tech Research Institute (GTRI), and one of the principal investigators of the system, describes how the technology works as similar to dropping a pebble in a pond. “When the pebble hits the water, the wave will radiate out from that spot in little rings,” he explains. “If there was something right under the water’s surface, those waves would change as they go over that object.”

What happens in the ground is only slightly different, he says. A signal generator placed on the ground outside a minefield emits a seismic wave between 30 hertz and 1 kilohertz. The seismic wave, or Rayleigh wave, travels across the surface of the minefield, exciting the ground and causing it to move up and down approximately one micrometer. When the wave crosses over a land mine, it vibrates differently from that of the surrounding area. “There are very few items in the ground that have a hollow plastic case like a land mine,” Scott says. “When the wave excites the land mine, it creates a distinct resonance that you would not get from a rock or a piece of metal.”

Unlike water ripples, waves on the ground cannot be seen with the human eye, but a radar sensor instrument can detect the resonance. The sensor measures the displacements of the ground, and a computer converts the data into an animated color simulation of the wave as it propagates across the mine field. The system then processes the data with an imaging algorithm to create a single color image. A red screen image indicates a high probability that a mine is present, while black and blue mark lower probabilities.

Although the displacement above the mine is greater than the area around it, it is not the best detection cue. One unique physical property of a mine in the ground is that it will continue to resonate after the Rayleigh wave passes over it. “It’s analogous to tapping a glass,” Scott notes. “If you stop hitting it, you can clearly hear it continue to ring or resonate.”

Tests of the seismic land mine detector have been conducted at various sites. The results were as good in hard-packed clay roadbeds as they were in mostly sandy ground. Scott explains that, when pressure is applied to sand, it forces the particles to interlock and creates a stiffness. This process allows the wave to travel through the sandy soil at the same rate as through clay soil.

Results at a frozen field site were not as promising. “We wanted to test the system in the most extreme condition we could think of,” Scott shares. The hard ground changes the way the wave propagates, which in turn changes the character of the resonance, making it much more difficult to find the mine. “It barely worked at the frozen site. But that was a very extreme case, and we were happy to get anything to work out there,” he relates.

Scott notes that the greatest development of the seismic mine detector is its ability to differentiate between mines and other material in the ground. “Metal detectors or ground-based radar detectors are really good at finding mines,” he says. “The problem is that they are finding a lot of other stuff in the ground too.” Because a land mine produces a distinct resonance, the seismic system is more resistant to clutter and less prone to false detections, he contends.

However, some objects can fool the system. “If you have an item with a thin, light, flexible case—like a Coke can—it will resonate just like a land mine and can trick our system,” Scott says. “To distinguish between these two, we will have to physically find something that is different between the two that will show up on the radar, or to train algorithms to tell the difference.”

The research team also hopes to create algorithms to distinguish between various mines and other items. “What we have now is something that can tell us if something is resonating there, but we do not know if it is mine A, B or C or a Tupperware bowl,” he states.

One solution to this problem may be to combine the sensor with other land mine sensors currently in use. For example, a metal detector can distinguish a soda can from a mine because the aluminum found in cans has a distinct signature, whereas the seismic detector can eliminate false positives from buried metal objects. “It is very unlikely that any one system is going to stand alone. It’s a difficult problem to solve,” Scott admits.

In its current state, the seismic sensor does not work quickly enough to be used on the battlefield. “If time is not crucial, the seismic system can work well,” Scott says. “But if you need to clear a road at 50 kilometers per hour, it’s going to be extremely difficult, if not impossible.” Because the system is slow, its most likely military application would be as a confirmation sensor. “GPRs [ground-penetrating radars] can be really fast, but they have a lot of false alarms. Seismic sensors can look at those false alarms and determine if they are mines or clutter,” he says.

Researchers are working to improve the system’s speed. Scott says his team was surprised to learn that taking the tests out of the laboratory and into the field yielded faster results because the seismic noise was 20 to 30 decibels less in the field than in the laboratory. “Being in the middle of a big city and having large compressors in the room next door to the lab made the seismic noise very loud. The more noise, the slower the system,” he explains. “Actually, we completed a field test at the end of the runway on a military base, and it was quieter there than in the lab.” Scott relates that the results were 64 percent faster in the field than in the laboratory, and the data was just as good.

The system as it exists today is not ready for the soldier’s pocket. Researchers working on the project need to understand the physics of the mechanism before they can build a practical unit. “I would call the system as it exists right now a laboratory system,” says Scott. “We haven’t put a lot of effort into incorporating it into a handheld device, a robot or vehicle system.”

Another problem of making the seismic system more compact is having to present the data visually. Scott notes that a person operating the system would find it difficult to watch the screen for the results while conducting a scan. “Pilots are trained to do things like this, but you don’t want a mine detector that is so complicated that it takes a heavily trained person to operate it,” he explains.

As a result, developers have created a way to relay the resonance results using sound. Scott likens how the system produces audible results to tapping on a wall to find the stud. The resonance sounds are measured from the ground and fed through a headset or speakers. “You can clearly hear when it goes over a mine,” he says.

While the system is not near implementation, collaboration with the CyTerra Corporation holds possibilities for the future. The company is trying to incorporate the technology’s audio capability into a handheld system it is developing for the U.S. Army.

The Office of Naval Research, the Army Research Office and the Army Night Vision and Electronic Sensors Systems Directorate are sponsoring the research on the Seismic Land Mine Detection System. Funding for the research at GTRI started approximately seven years ago; however, Scott recalls beginning his portion of the research in the late 1980s. He acknowledges that there have been earlier attempts to do what his team is doing that date as far back as World War I.

“The idea has been around,” he says, “But the system could not be established because of technological issues. It was a while before a sensor that could sense the waves well could be developed.”