Technology
ARL engineers analyze, research ways to improve mine-detecting radar technology
By Ric Kositzke, Managing Editor
Before World War II, countries began developing mines to counter the effectiveness of tanks, the trench-busting vehicles that altered how land wars were fought.
By the end of the war, roughly 300 million mines, each filled with powerful, lightweight trinitrotoluene (TNT), were used to destroy enemy tanks.
The biggest drawback of the mines, however, was that opposing forces could easily swoop in before they detonated and use them against the tanks of the very same army that positioned them there first. To counter this phenomenon, countries began encircling their anti-tank mines with anti-personnel mines–the type of mines still used today by poor countries.
Today, despite advances in mine-detecting technology, the United States government is still searching for more accurate methods in locating visible and buried enemy mines.
Andy Sullivan, Army Research Laboratory, Electronics Engineer Team Lead, and Traian Dogaru, ARL Electronics Engineer, hope their research will pave the way for advances in low- and high-frequency radar technology that will someday allow for wide-area surveillance, by land or air, of surface and buried targets.
“Landmines, in effect, are a psychological weapon,” said Sullivan. “Enemies use the threat of landmines to affect not only the morale of individual soldiers but also how the military operates on the battlefield.”
Over the years, electromagnetic induction (EMI) and magnetometers sensors have been developed by military experts to find and remove mines. According to Sullivan, the sensors are fairly effective; however, they usually need to be close to the ground to be effective. Because of the dramatic effect mines and minefields have on military operations, a wide-area surveillance tool is needed.
In previous anecdotal studies, low-frequency (<2 GHz), ultra-wideband (UWB) synthetic aperture radar (SAR) measurements have shown some potential to detect surface and buried targets over very large open areas in a high-standoff mode.
Likewise, in Sullivan and Dogaru’s work, they use model-based results to investigate the effectiveness of a low-frequency UWB radar system (<3 GHz) and an X-band SAR for detecting minefields. In particular, the engineers examined the scattering phenomenology of mine simulations on and under dirt lanes and in short-grass areas.
“Low-frequency radars are an emerging technology, which is why modeling and testing low-frequency radars is so important. In a model, I can perform many numerical experiments and simulations,” said Sullivan.
Sullivan, who said Dogaru and he have been running their models on the SGI O3K and IBM P3 platforms at the ARL MSRC, said, “The total computer hours used over the past couple of years has been approximately 100,000 hours. The breakout is about 75,000 on the IBM and 25,000 on the SGI.
A typical job might require anywhere from 10 to 90 processors, depending on the specific nature of the job. Some of the post-processing is done on our local desktops using Matlab, but we also use Tecplot, which is available on the SGI and IBM platforms for visualizing results.”
The data Sullivan and Dogaru analyzed was gathered by the Army’s Night Vision Electro-optics Sensors Directorate (NVESD), who have been conducting a series of tests at various locations around the country using a variety of airborne sensors, including a low-frequency UWB radar and an X-band SAR.
An aircraft equipped with radars created images of different mines placed in different clutter backgrounds. The aircraft flew past the areas several times, collecting images from a variety of angles into the target areas. That image data was then provided to the ARL for further processing and analysis.
With the data in hand, Sullivan and Dogaru began modeling surface plastic mines, flush-buried plastic mines and deeply buried mines to see whether low- and high-frequency radars could detect the mines. According to Sullivan, the dielectric contrast between the mine and the soil backgrounds leads to a very small radar cross section (RCS) for the buried cases. This case is considered to be one of the more challenging scenarios for low-frequency radar: pointing out that clutter is the real problem.
To estimate the surface clutter, the two engineers employed a finite difference time domain solver (FDTD) to model the backscatter from rough surfaces.
Their data was computed using more than 50 independent realizations of the rough surface.
In the field test, the two created a three-layer mixture model for anti-tank mines in the grass to be used for high-frequency (X-band) SA. An average permittivity constant can be derived for the grass layer using an appropriate mixture model.
It is based on the fact that the average or effective dielectric constant for the medium can be related to the dielectric constants of the individual components (in this case air and vegetation), their relative fractional volume, and their orientation and spatial distribution.
Despite his and Dogaru's extensive research, Sullivan admits that completely accurate mine detection using low- or high-frequency radars is still years away.
But he and Dogaru will continue analyzing and modeling surface and buried landmines, hoping that someday their research will lead to 100-percent accurate mine detection that will save many soldiers' lives.