WHEN the guns finally fell silent in Bosnia earlier this year, there was an
almost audible sigh of relief around the world. But the bloodshed is still not
over. Several times in recent months TV cameras have captured horrific images of
children maimed by land mines. Those images have reinforced the message being
presented to disarmament talks this week in Geneva that land mines should be
banned. Even if mines are useful during a war, their opponents argue that the
damage they do in the years after a conflict makes them unacceptable.
During the two weeks that the disarmament talks last, about 1000 people will
be killed or maimed by mines. Most will be ordinary people, and many of them
children. What鈥檚 more, that toll is unlikely to fall in the near future because
the world is littered with mines鈥攁t least 100 million in more than 60
countries, according to UN figures. And they are still being laid at the rate of
2.5 million a year. Mines can be cleared of course and in 1993 UN personnel
removed 80 000. But at this rate, even if a land mine ban came into force
tomorrow, it would still take more than 1000 years to clear what鈥檚 left.
So can science help? After all, researchers have developed sensors that peer
into the ground or sniff out the chemical signatures of explosives. Others can
identify aircraft, missiles and satellites by searching for telltale signals in
the electromagnetic spectrum. These sensors are used every day, in airports and
by the police and armed forces. Surely similar technologies could simplify mine
clearance?
Advertisement
The right combination
Indeed, scientists around the world are turning their attention to just this
question. Already, in laboratories and some field trials, researchers are
testing these sensors to see whether they can spot mines or even entire
minefields. But all the methods tried so far are plagued by the same problem.
Take them onto the site of an old battle, be it open fields or the remains of a
razed town, and none of them will spot every mine鈥攊n fact they don鈥檛 even
come close to a 100 per cent detection rate. Yet nothing less than this rate is
acceptable: a single mine missed means grievous injury or death for somebody.
And until these methods are reliable, the people who clear mines will not use
them. 鈥淚鈥檝e not seen anything that would convince me to rely on these new
technologies,鈥� says Bob Keeley, who heads the UN mine clearance programme in
Zagreb, Croatia. 鈥淎nd I鈥檓 going to take some convincing.鈥�
Most experts now agree that a reliable mine detector will need to make use of
several different sensors. And until the right combination is found, mine
clearance will remain a painfully slow process, says Paul Jefferson, an ex-Royal
Engineer with the British Army and a specialist in explosives and ordnance. His
experiences of clearing mines after the Soviet withdrawal from Afghanistan
illustrates the difficulties of conventional methods. While working for the HALO
Trust, a British charity devoted to mine clearance, Jefferson was assigned to
the town of Pol-e Khomri, 200 kilometres north of Kabul.
The hill he had to clear was once the site of a Soviet gun emplacement. It
commanded views of the flood plains on either side of the Qonduz river and had a
clear line of fire into the snowcapped mountains beyond. Surrounded only by a
flimsy wire fence, the camp appeared to have been poorly protected. But beyond
the fence, in the few places where grass did not cover the ground, the heads of
Russian POMZ fragmentation mines poked through like tiny pineapples, their trip
wires occasionally visible in the breeze. And where the wind had blown away the
earth, the small, flat heads of Russian PMN antipersonnel mines lay clearly
visible. The dirt track that led to the single entrance was the only safe route
through the minefield. Here, Jefferson taught his trade to 20 local people.
Mine clearance teams work in twos: one with a metal detector, the other with
a sharp, thin rod for probing the ground whenever a suspect object is found. It
is laborious, painstaking work. A team can expect to clear just 1 square metre
per hour. They have to go slowly because many common types of mine contain only
a little metal, and so can be overlooked even with a metal detector. These mines
are made of plastic, which is cheap and easy to mould, and only the trigger
mechanism is metal. Battlefields are full of shrapnel, old shell casings and
ration tins that cause false alarms, and every one must be treated as a
potential mine. To make matters worse, when a mine is discovered, the UN
recommends that it is blown up in situ, which can spread even more
shrapnel.
Signs of explosions
Detecting mines may be difficult but it is simple compared with mapping the
edges of a minefield鈥攖he first stage of any clearance operation. 鈥淭his is
extremely difficult,鈥� says Jefferson. 鈥淵ou start by chatting to the locals but
the vast majority don鈥檛 know what they鈥檙e talking about. They tell you that an
entire hill is mined but you know that the Soviets would never have bothered
with an area so big.鈥� Narrowing down the perimeters is often a matter of looking
for signs of explosions such as animal carcasses, and relying on military
judgement. 鈥淵ou look at the topography and for disused machine gun and artillery
emplacements and how they might have been protected.鈥�
Two years after clearing mines at Pol-e Khomri, Jefferson worked for the
British arms firm Royal Ordnance in Kuwait, clearing land mines in the wake of
the Gulf War. He was 鈥渢aken out鈥� by a Russian PMN antipersonnel mine and lost a
leg and his sight. The HALO Trust employs about 1000 people round the world. In
the past eight years, about 44 have been injured and eight killed.
So long as mine clearance relies on such primitive methods, injury and death
are inevitable. Making the task safer and more efficient is one of the jobs of a
30-strong team of computer scientists, engineers and armed forces personnel who
work in the Countermine Warfare Project (CWP) at the Defence Research Agency
near Chertsey. They are developing a new generation of mine detectors for the
armed forces.
One of its sensors is a camera that produces images with millimetre waves,
which fall between infrared and microwaves in the electromagnetic spectrum. The
sensor relies on the fact that at everyday temperatures all objects emit
electromagnetic radiation. The strength of that emission depends on the object鈥檚
temperature and its composition. For example, human flesh is a good emitter
while metals are not. Poor electrical conductors, such as plastics and
explosives, fall somewhere in between.
Last September, researchers from the CWP took the millimetre-wave camera to a
testing ground in Wales to see if they could use it to spot mines laid on the
ground. Finding these is more difficult than it sounds: they are designed to
blend in with their background and are easy to miss, especially when visibility
is poor. But when viewed through the millimetre-wave camera, metal mines show up
as dark objects against the relatively bright background of a wet forest road,
even at night. Plastic mines are more difficult to spot because in the world of
millimetre waves they look similar to wet soil and grass.
Nevertheless, the team is pleased with the results. They show that the camera
mounted on a vehicle may be capable of spotting suspect objects on the road. For
the moment, however, the camera is nowhere near ready for operational use. Such
a system would need to produce images in real time, but at the moment it takes
several minutes to scan an area to build up a single image. 鈥淚t might be useful
for mine clearance operations where images could be built up over hours or
days,鈥� says the man in charge of the CWP who cannot be named for security
reasons.
Another disadvantage of millimetre-wave imaging is that the amount of
radiation that objects emit in this region of the spectrum is relatively small:
most energy is emitted as infrared radiation. Unlike millimetre waves, infrared
radiation can penetrate soil so it can be used to find mines that are buried
close to the surface. But infrared radiation has its own drawbacks.
鈥淚R sensors detect the difference between the temperature of a mine and the
surrounding area,鈥� says the CWP team leader. Mines can be spotted because metal
and plastic heat up and cool down at different rates from the surrounding soil.
On a warm day, for example, a metal object is hotter than the soil, but at night
it is cooler. 鈥淏ut at the beginning and end of the day the temperatures match.
During this crossover period you can鈥檛 see a thing,鈥� he says. The CWP prefers
millimetre-wave imaging because it does not suffer from this problem.
Another major challenge is to design an algorithm that will pick out the
electromagnetic signature of a mine from all the background clutter. At present,
computer programs designed to analyse infrared, radar or millimetre-wave images
can still miss mines that are visible to the naked eye and identify lumps of
grass as potential mines. 鈥淚f you used this in Bosnia, you wouldn鈥檛 be able to
move forward because of false alarms,鈥� says a team member.
Going underground
One way to reduce the number of false alarms would be to screen an area with
the millimetre-wave camera, for example, and use another sensor for a more
detailed examination. One of the candidates for the in-depth analysis is
ultrawide band radar, which is capable of distinguishing plastic and metal mines
from rocks. As a rule of thumb, the lower the frequency of the radio waves, the
better they penetrate the ground. By including low frequencies in the spectrum
of the ultrawide band radar, the technique can detect mines buried in most soils
and may even reveal the type of mine.
Ultrawide band radar transmits a short pulse of radio waves into the ground
and analyses the reflection. 鈥淎 wide band pulse acts like a hammer,鈥� says one of
the CWP team. 鈥淲hen it hits an object, it makes it ring with radio waves, like
an electromagnetic bell.鈥� The frequency of this radiation depends on the size,
shape and material of the object. By building up a database of the way different
mines ring, the CWP researchers hope to be able to identify each one.
This work is in its early stages. The CWP鈥檚 radar is mounted 18 metres up on
the roof of a hangar and to build up an image the team drags mines along a track
on the ground at a slow walking pace. 鈥淥f course, in practice the radar would be
moving, mounted on top of a vehicle or aircraft,鈥� says the team leader. The
results take two or three minutes to emerge. However, building a useful database
of mine signatures will be difficult because a mine gives a different radar
response depending on its orientation in the earth.
As well as investigating ways to detect individual mines, the CWP team is
developing a technique for spotting entire minefields, using synthetic aperture
radar. The SAR is mounted on a 鈥減latform鈥� such as an aircraft or satellite and
sends out a steady stream of radio pulses. It then analyses the reflections it
receives from objects on the ground. As with normal radar, the time it takes the
reflection to arrive reveals an object鈥檚 distance from the platform. What makes
SAR different from a standard radar is that by combining the data from repeated
pulses, received as the platform changes position, computer programs can create
two-dimensional, photograph-like, images of an area.
In May last year, the group flew an aircraft-mounted SAR over a pattern of
antitank mines laid on the ground. Some of these mines sat on a concrete runway
while the rest lay in thick grass. From an altitude of 7.5 kilometres, the mines
on the concrete showed up clearly on the SAR image, but those in the grass were
less obvious. Nevertheless, using these data, the team mapped the perimeter of
the minefield.
Searching for patterns
In future, the researchers aim to improve the SAR detection technique using
pattern recognition. Because mine-laying teams have to work methodically for
their own safety, they often place their mines in regular formations. A computer
program might be able to pick out these patterns on a radar image. Some mines
are scattered from the air, but even in these cases they may have a density
pattern that can be searched for, says the head of the CWP. 鈥淏ut you鈥檝e got to
be able to see at least 70 per cent of the mines and distinguish them from
battlefield rubbish.鈥� Again, the CWP is unable to produce the SAR images in real
time and the frequencies used by the CWP cannot penetrate thick foliage or
earth.
In the US, scientists are developing a system that they hope will advance
mine detection still further. At the US Army Research Laboratory in Adelphi,
Maryland, engineers have combined ultrawide band radar with synthetic aperture
techniques in a single system. The result is an airborne radar that can, at
least in principle, map buried minefields. The difficulty, as ever, is
distinguishing the mines from the background. 鈥淥bviously, metal mines show up
better than plastic mines,鈥� says John Miller, the engineer who led the SAR
development team. Spotting plastic mines in some soils, particularly sand, can
be very difficult because their responses to radio waves are so similar.
In order to differentiate better between mines and their background, the
American researchers are studying how radiation propagates through different
soils. 鈥淲e need to know how much energy is lost at the surface, how much
distortion occurs through the soil and how this varies with factors like
moisture content,鈥� says Miller. Working with geologists across the US, the team
is collecting soil samples with a view to building a database of how radio waves
travel through different soils, and creating a model to describe the
differences. 鈥淭he idea is that eventually we can fill gaps in our database using
the model.鈥� This information could be used to adapt the SAR鈥檚 performance to the
soil type.
The American and CWP groups are two of dozens of teams working on new
techniques for mine detection. All the groups approached by New
杏吧原创 agree that real improvements will be possible only by using a
combination of several types of sensor. 鈥淪omething that is hidden from one
sensor may not be hidden from another,鈥� says Miller.
Sensor fusion is still in its infancy when it comes to mine detection, and
agreeing which sensors to combine is unlikely to be straightforward. The head of
the CWP believes that the best option will be conventional metal detection,
ultrawide band radar and a technique for spotting the chemical signatures of
explosives known as nuclear quadrupole resonance (NQR). When a nuclear
quadrupole鈥攁 nucleus with a nonspherical distribution of charges鈥攊s
excited by radio waves, it radiates electromagnetic radiation at a
characteristic frequency. Carbon transmits at a different frequency from
nitrogen, for example. And since explosives generally have a high nitrogen
content, they can be found by looking for the electromagnetic signature of
nitrogen. 鈥淭he problem is getting each technology to the same level of
development so that they all work together,鈥� says the CWP. 鈥淣QR is still at an
early stage of development.鈥�
Sensor fusion
Alois Sieber, a physicist at the European Commission鈥檚 Joint Research Centre
at Ispra in Italy, is not impressed by NQR. He has coordinated a series of
international workshops on mine detection and last year published a report
outlining the most promising techniques. He believes that metal detection and
ultrawide band radar are hugely promising, but that NQR will only be an option
in the long term. One of the biggest problems is that the method needs huge
magnetic fields which have to be generated outdoors with portable power
supplies.
His favoured combination is metal detection, ultrawide band radar and
infrared imaging. The problems seen with infrared sensing at dawn and dusk are
only important when building mine detectors for the military, he says, because
the sensors must be reliable in a battle, whatever the time of day. 鈥淚nfrared
sensing is particularly good for surface laid mines,鈥� he says.
Sieber also argues that creating one mine detector that will work anywhere is
simply not feasible. Factors such as surface vegetation, soil type and even
humidity will influence the choice of radar frequency that will be needed for
penetrating the ground. In turn, different frequencies need different sized
radar antennas. Sieber plans to simplify these problems by building a mine
detector for just one part of the world鈥擝osnia. He has requested a grant
of 50 million Ecus (拢41 million) from the European Union to coordinate
development of a prototype by 1998.
Such a system will have to work well to convince mine clearers such as Keeley
of its worth. 鈥淚 won鈥檛 be using any new technology until I鈥檝e seen the scientist
who developed it in the middle of a minefield using it,鈥� says Keeley. Sieber had
better be sure of his ground.
* * *
The science of killing
AS part of their work, scientists at the Countermine Warfare Project at
Britain鈥檚 Defence Research Agency in Chertsey have catalogued 500 different
types of mine. Despite this wide variety, they all kill and injure in one of two
ways.
The first is with a small explosion. These 鈥渂last mines鈥� are designed to be
triggered by the pressure of a descending foot, and the scale of damage depends
on the amount of explosive. As a rule of thumb, says Paul Jefferson, a former
Royal Engineer and mine clearance expert, about 50 grams is enough to remove a
foot, while 300 grams will destroy a leg. More than 500 grams will kill.
But the injuries caused by the shock wave from a blast mine are rarely
confined to a single limb and often extend to the genitals and abdomen, says the
International Committee of the Red Cross. Often the blast forces dirt into the
wound which makes it difficult to clean and susceptible to infection. When this
happens, a surgeon鈥檚 only course of action is to amputate the limb above the
wound.
The second type of mine is designed to spray a large number of high-velocity
fragments towards its victims. One example is the Claymore, an antipersonnel
mine designed in the early 1950s. It is shaped like a warped lunch box and sits
on four legs. The plastic box is packed with plastic explosive and embedded in
its convex face are 700 steel balls, each weighing about 150 grammes. When
triggered, the mine sprays these projectiles in a 60掳 鈥渒illing zone鈥� that
extends 50 metres in front of the mine.
A macabre branch of science can even tell a manufacturer how big pieces of
shrapnel need to be to kill hapless soldiers or civilians who trigger a mine.
The science of wound ballistics has grown since the Second World War with
studies of the effects of fragments of different size, mass and velocity on
human flesh and bone simulants, such as anaesthetised animals and blocks of
gelatin. The mathematics of wound ballistics predicts an optimum energy for
killing with a fast-moving fragment.
One of the most active teams in this area was the Wound Ballistic Research
Group at Princeton University in New Jersey. The results of its research during
the 1950s make gruesome reading. As a projectile rips through flesh at several
hundred metres per second, it creates a cavity that expands and contracts
several times before collapsing to leave a permanent wound. This process sends a
shock wave through the body that can break bones, damage nerves and rupture
internal organs far away from the entrance and exit wounds. The size of this
cavity depends on the kinetic energy of the fragment, calculated according to
the formula 陆mv2, where m is mass and v is velocity.
Clearly, for maximum impact, increasing the velocity of a projectile is far more
important than its mass.
An equally important factor is the proportion of the projectile鈥檚 energy that
is imparted to the target. If it lodges in the body, a projectile has passed on
all its energy. But the destructive energy is wasted if the projectile passes
through the target without slowing significantly. The Princeton group found that
projectiles with a large cross section and small mass decelerated most
efficiently inside the body. The team even developed equations that relate these
quantities to factors such as the density of flesh and the coefficient of drag
while passing through the body.
How likely are fragments to kill? Two researchers at Princeton, Howard
McMillen and J. Gregg, set out to answer this question by applying the formulas
for penetration to different parts of the human body. They found that the
probability of inflicting a fatal injury increases up to a maximum level.
However, this maximum kill rate is possible only with relatively large amounts
of explosive.
So, instead, McMillen and Gregg went back and looked at the effects of
fragments with lower kinetic energies, propelled by smaller amounts of
explosives. They found the relationship between a fragment鈥檚 energy and the
probability that it will kill is not linear. As the energy rises, the
probability increases rapidly at first until a critical point, after which it
rises more slowly. McMillen and Gregg suggested that explosives could be best
used to give fragments this critical point energy. Beyond this level, they
argued, any increase in explosive would result in only a small increase in
killing potential.
According to McMillen and Gregg鈥檚 calculations, the way to squeeze the
optimum killing efficiency from the small amount of explosive in a Claymore mine
would be to reduce the mass of the steel balls to roughly 15 grams鈥攁 tenth
of their present mass.
Further reading: The Technology of Killing by Eric Prokosch, Zed
Books