PROGRESS IN HUMANITARIAN DEMINING:  TECHNICAL AND POLICY CHALLENGES


                   Richard L. Garwin and Jo L. Husbands

                  Prepared for the Xth Amaldi Conference
                        Paris, 20-22 November 1997


BACKGROUND

Introduction
     If the treaty banning anti-personnel landmines (APL) to be signed in
Ottawa in December took effect today with universal adherence, the global
landmine crisis would still confront us for many years.  "The burden
imposed by the proliferation and indiscriminate use of these weapons is
beyond calculation" (U.S. Department of State, p. v).  According to
international and U.S. government estimates, 80-120 million landmines
remain to be detected and cleared in over 60 countries.  Angola,
Afghanistan, and Cambodia are among the most afflicted.  Landmines maim or
kill an estimated 26,000 people a year, mostly innocent civilians, and the
survivors need extensive medical care and rehabilitation services
(Department of State, p. 1). (1)

     Landmines are remarkably durable, posing a threat years after the wars
for which they were laid have ended.  Most of the 23 million mines reported
by the Egyptian government, for example, date from the North Africa
campaigns of World War II, while others in the Sinai were laid in more
recent wars with Israel.  A number of Central and Eastern European states
also suffer from mines and unexploded ordnance left over from World War II.
Indochina still has million of mines and other unexploded ordnance left
from the 1960s and 1970s.

     The tens of millions of mines laid in recent years are largely the
results of internal conflicts, civil wars, or armed interventions.  In such
conflicts, the use of mines frequently shifts from traditional tactical
defensive roles to offensive, strategic uses "often aimed deliberately at
civilians in order to empty territory, destroy food sources, create refugee
flows, or simply spread terror" (Arms Project, p. 9).  Despite the
stipulations of international humanitarian law (which does not support the
use of landmines against civilians), few records are kept, with the result
that the mines' locations are almost completely unknown and unmarked. (2)
In addition to the hazards they pose during conflict, these "hidden
killers" represent a substantial barrier to economic recovery and the
return to normal life.

     On October 31st, U.S. Secretary of State Madeline Albright announced
that "the United States is calling for and will lead a global campaign, the
'Demining 2010 Initiative,' to eradicate all landmines which threaten
civilian populations by the year 2010" (U.S. Department of State, 1997a).
The Mission Statement for the Initiative reads:

     The objective is to accelerate global humanitarian demining efforts
     and our goal is to increase roughly by a factor of five -- to $1
     billion a year -- the public and private resources devoted worldwide
     to identifying and clearing landmines posing threats to civilians by
     the year 2010.  The Demining 2010 Initiative, through U.S. leadership,
     will seek to coordinate these efforts.  This goal will be achieved by
     bringing together donors, demining experts, and assistance recipients
     to make tangible commitments to expand substantially operational
     demining and related programs of assistance, to agree on mechanisms to
     enhance the exchange of demining information and demining technology,
     and to optimize the use of worldwide demining resources to achieve our
     goal of eradicating landmines by the year 2010 (U.S. Department of
     State, 1997a).

In order to achieve its goals, the Initiative will have to overcome
formidable technical and policy challenges.  This paper describes some of
those challenges and suggests some potential solutions to them.  The
scientific community, and the world's academies of science and science
societies, have an important potential role in achieving some of those
solutions.

Current Humanitarian Demining Methods
     Current methods of detecting and clearing landmines for humanitarian
purposes are primitive, dangerous, slow, and costly.

     Current technology is old technology.  "In many cases, the
technologies employed by equipment now in use, even by technologically
advanced armies, have changed little since World War 2" (Craib, p. 6).  As
discussed in greater detail below, the standard equipment used in demining
operations is a metal detector and a hand-held probe, although mechanical
clearance methods adapted from the military, such as ploughs, flails, and
rollers, may be used when funding and terrain permit.

     Clearing mines is dangerous work.  According to one estimate, mine
clearers are seriously injured or killed at a rate of one per 1-2,000 mines
cleared (Tsipis, p. 12).  Even in Kuwait, under relatively favorable
terrain conditions and with generous funding for the clearance contractors,
there are reports that more personnel were killed clearing mines than the
combat casualties suffered by the Coalition forces during the Gulf War.

     Demining is slow.  In 1996, 3,000 workers cleared 12 square kilometers
of Cambodia with currently available technology (Morrison and Tsipis,
p.40).  According to one UN estimate, it would take 1,100 years to clear
all the APL around the world using current technology (Rouhi, p. 16).
Meanwhile the problem is getting worse: in 1995, while an estimated 250,000
APL were cleared, an estimated 2.5 million new mines were laid. (3)

     Demining is expensive and cost-ineffective.  The costs of clearance
are out of proportion to the costs of the mines.  The most common types of
antipersonnel mines in use cost between $3.00 and $15.00, while clearing
them can cost between $300-$1,000 per mine (Joint Research Centre, p. 10).
In 1993 U.N.-sponsored demining teams cleared 80,000 landmines at a cost of
$24 million, and the cost/clearance ratios have not improved. Clearing the
12 square kilometers of Cambodia cited above cost $8 million.  The United
States has invested $153 million in demining programs since 1993, which
have resulted in the removal of 1.5 million mines (U.S. Department of
State, 1997c). (4)  The UN estimate cited above puts the cost of clearing
all the mined areas of the world using current methods at $33 billion
(Rouhi, p. 16).

     Donor countries and international aid agencies cannot increase funding
for demining to the level necessary to clear millions of mines with current
methods.  The new U.S. Demining 2010 Initiative aspires to a five-fold
increase in current resources, but $13 billion ($1 billion per year through
2010) is only about 1/3 of the UN estimate of what it would cost to remove
all the landmines currently in the ground and likely to be laid by 2010.
The answer must come from significant improvements in the productivity of
current approaches and from development of new technologies.  Secretary
Albright acknowledged the problem, asserting that "We need to intensify
research into better methods of demining -- for in this era of
technological miracles, the most common tool we have for detecting land
mines is still a stick attached to a person's arm" (Associated Press,
1997). (5)

Improving Demining Methods:  The Technical Challenges
     Improving detection and clearance methods is a formidable technical
challenge.  The humanitarian demining problem is characterized by an
enormous variability in the nature of explosive ordnance to be removed, and
in the type of terrain and vegetation.  In addition, the mines are
infesting some of the world's poorest countries, where the indigenous
personnel available to undertake demining may lack technical skills and
experience.

     The terrain to be cleared includes everything from jungle to deserts
to mountainsides and every kind of climate.  What works in land normally
sown with field crops will not work in a tea plantation, and what works in
a tea plantation is unlikely to be suitable for a rice paddy.  And
techniques that work in areas afflicted with tripwire-actuated mines may
not be necessary for areas known to be without them.  In some cases
substantial benefit would accrue from demining of paths of ample width for
people and animals in order to allow access to water, farms, and supplies.

     The variety of mines being used is enormous, including many with very
small amounts of metal.  A U.S. Army database made available to the United
Nations, for example, contains profiles of 750 different types.  APL laid
on or just under the surface may have as little as 30 g of high explosive
(HE), like the U.S. M-14 or the Italian VS-50.  The M-16 "bounding" mine
contains 500 g of HE and leaps into the air before exploding, with a lethal
radius of some 30 m.  While surface APL may have only a fraction of a gram
of metal (or in rare instances none at all), their purpose is to maim and
not to kill.  The larger mines achieve their much larger destructive range
by creating and projecting with great force metal pellets, as in the case
of the common hand grenade and are either positioned on stakes, hung from
trees, or are buried and bound into the air before bursting.

     Humanitarian demining is complicated by the fact that land that has
been unused for several years in most portions of the world will be covered
with substantial vegetation that makes it impossible to see the ground or
to move the normal hand-held metal detector freely above the ground.  And
the overall land clearance by burning or shearing that may be suitable for
farmland would sacrifice a tea plantation.
     With the passage of time, APL become covered with soil even if
initially laid on the surface.  If tripwire mines are present, demining
might begin with the tedious process of casting snag lines repeatedly into
the area and drawing them back from a safe distance.  This typically works
either on slack-line or taut-line mines.  If there are no tripwires, then
demining can begin with shearing of a border of the region to be demined,
with due care not to exert more than a few pounds of force on any small
region of ground.

Improving Demining Technologies:  Some Early Lessons
     The landmine crisis has evoked enormous interest and since the early
1990s numerous studies and conferences have examined ways to improve on
current practice and technology. From all this activity, four early
lessons/conclusions are emerging that appear to reflect a degree of
relative consensus among the experts.

     1)  To use U.S. slang, there is no "silver bullet," that is, no single
approach or technology will emerge to offer the solution to the landmine
crisis.  Many experts stress the need for a "tool kit" that would offer a
variety of equipment that could be combined in different ways in different
situations.

     2)  There are at least two critical problems within the rubric of
"detection."  The first is the need to discriminate landmines from all the
other metal that may be in the ground. When mines are laid in or near
former battlefields, there may be millions of fragments that will set off a
metal detector.  One study reports that false alarm rates can be as high as
1,000-to-1 (Tsipis, p.11).  But each must be treated seriously and
investigated as though it were a potential mine, at a high cost in time and
stress on personnel.

     A second key detection problem is the need to know where mines are not
present, so demining efforts are not wasted and recovery and rebuilding can
begin in the safe areas.

     3)  Traditional military countermine techniques and equipment are not
directly applicable to humanitarian demining, largely because the standards
for successful clearance are different.  During battle, the goal is quick
breaching of a path for troops to pass through the minefield.  And it is
adequate to clear a small portion of the field where the attack is focused.
Furthermore, in conventional conflict relatively high personnel losses are
accepted.

     Once the conflict is over, the humanitarian task becomes to clear as
many mines as possible with as high a level of confidence as possible so
that civilians may return to normal life in safety.  The reasons for the
current UN standard of a rate of clearance over 99.5% are understandable,
but this puts heavy demands on potential technologies. (6)

     Traditional countermine R&D programs may nonetheless offer valuable
spin-offs for demining.  In addition, the peacemaking and peacekeeping
situations in which many militaries are finding themselves today pose
demining challenges similar to humanitarian situations.  The roads and
large areas that must be cleared in Bosnia, for example, require efforts
more like humanitarian mine clearance tasks than traditional breaching.

     4)  Sustained, systematic engagement is needed among organizations
engaged in demining, military technologists charged with R&D programs,
humanitarian/NGO organizations, and the broader scientific community that
could be a source of both technical assessments and new ideas.
Communication between the military technology "producers" and the
humanitarian demining "consumers" is still relatively new and tentative,
however, and far from systematic. (7)

     One critical aspect of this engagement is the need for feedback from
those doing demining to ensure that R&D requirements are based on genuine
operational needs and grounded in real world conditions.  The engagement
should also include R&D managers charged with developing for other purposes
potentially relevant technologies.  One of the most obvious examples is the
need to develop better means to detect explosives in terrorist bombs, which
has fostered a wide variety of efforts in a range of government agencies.
As discussed below, some of the technologies are potentially very
interesting, but a "real world" perspective will be critical to assessing
the possibilities of taking advantage of possible spin-offs.

     Before moving on to a discussion of some candidate technologies, it is
worth noting that there are no reliable estimates of the amount of money
currently being invested in R&D to improve humanitarian demining.  There is
no central source of information nor any organization that has undertaken
to collect these data.  Information from the private sector would
understandably be difficult to collect or estimate, but information on
government investment is also elusive.  In the United States, the
Department of Defense is officially charged with coordinating the R&D
efforts of its organizations, as well as other government agencies, and is
a member of the Interagency Working Group on Humanitarian Demining. All one
can say with confidence, however, is that the budget for the humanitarian
demining research, development, testing, and evaluation (RDT&E) program
housed in the U.S. Army Communications and Electronics Command, Night
Vision Electronic Sensors Directorate at Ft. Belvoir, Virginia (the one
program clearly devoted to the task, was $10 million in Fiscal Year 1995
(FY95), $3 million in FY96, $14.7 million in FY97, and $21 million in FY98,
which began on October 1st. (8)
Data on investment by individual European countries or the European
Community is also relatively hard to acquire.


THE TECHNICAL CHALLENGES:  SOME CANDIDATE TECHNOLOGIES

     This section discusses a number of the new technologies or
improvements in old technologies that offer promise of improving the
detection and clearance of landmines.  One of the authors, Richard Garwin,
participated in the JASON and MIT (Tsipis, 1996) efforts cited in the
References.

Demining by Detonation without Prior Detection
     The tined roller:  A roller with curved spring-steel tines spaced at
axial distance of 1 cm can be used to exert a load of some 30 kg on each
tine (3000 kg/m width) and can mimic walking over rather uneven terrain.  A
tine roller with long stiff fingers was tested in 1995 at Fort Belvoir and
"was effective in rice paddies, soft ground, and mud."  Assuming an
effective inelasticity of 0.1, the energy required to propel the roller per
square meter is 3 kilojoules (3 kJ).  To move at 3 km/hr (1 m/s) would thus
require 3 kw or about 4 hp (perhaps a 10 hp engine).

     The Dervish:  The Dervish may be imagined as three radial axles of
length typically 2 or 4 m, radiating from a central vertical spindle
temporarily pressed into the ground (Salter and Gibson, 1997).  Each of the
axles ends in a steel wheel perhaps 70 cm diameter and about 2 cm thick,
self-propelled by a hydraulic motor in its hub.  The Dervish thus treads on
a circular path 2-cm wide with sufficient footprint force to detonate APL.
Much effort has gone into the demonstration that the system survives
detonation of normal APL and is readily repaired if it encounters a large
mass of explosive.

     A mechanism at the spindle can be contrived to move the center of
rotation of the three axles along a desired direction, with an advance of
about 3 cm per revolution of the Dervish.  In this way the Dervish can
proceed along a path, while scouring it with nearly overlapping circles.
Alternatively, a sensing mechanism at the central spindle could be used
with commandable rotation of the steel wheels to provide the same advance,
but without putting any force on the spindle that touches the ground and
serves essentially as a reference for the short-time small advance of the
Dervish.  It could process 5 square meters per minute of suitable terrain.

     The Dervish or a fleet of Dervishes can be commanded to follow a
straight or complicated path either by differential GPS signals with
respect to a local GPS receiver within a kilometer or so, or by means of a
kind of multi-frequency (or single-frequency time-shared) Decca type
system.  Finally, and perhaps most desirable, two hydraulic steering rams
and hinged axles can change the orientation of the wheels, not only to
control the fine advance but to allow rapid straight-line motion of the
Dervish for quick traversal of fields, following a path, etc.

Finding and Defeating Buried APL
     As described in the reports by JASON (1996) and Tsipis (1996),
humanitarian demining ordinarily uses a "metal detector" like that used for
detecting old coins.  For instance the Shiebel Model MMD-100 Handi Mini
Mine Detector easily detects a fraction of a gram of not necessarily
magnetic metal in soil and thus all common APL close to the surface.  But
it also detects the enormous numbers of metal fragments in inhabited areas
or battlefields.  A single 155-mm shell bursts into 3,000 fragments of a
gram or more in mass. So it is essential to be able to mark any candidate
APL location, and to locate it as accurately as possible, in order to
minimize the region that must be explored.

     Discrimination from metallic clutter is usually done by manual probing
using a pointed metal or fiberglass rod typically some 5 mm diameter and
200 mm long.  Some 5 to 20 minutes is spent probing the earth at a small
angle to the horizontal, with the deminer prone and wearing a face mask.
If no solid object that might be a mine is found, the deminer moves on to
something else, or more usually digs up the piece of metal.

     The enhanced metal detector:  A metal detector may be 30 cm in
diameter and can thus readily locate an isolated metal fragment  or mine to
an accuracy of 5 cm or so.  But the sweep rate of a metal detector is
determined only by its transverse dimension, and if it were only 5 cm long,
it would allow location accuracy to about 1 cm (in one dimension).  A cross
sweep would then locate in the other direction, so that an isolated piece
of metal could be bounded to an uncertainty region of about 1 cm.

     Furthermore, a small piece of metal gives a signal whose width at the
surface of the ground is similar to the depth below the detector, so a
measurement of the width of the signal as the surface is scanned gives an
indication of depth.

     The enhanced metal detector should be able to signal if there is only
a single small piece of metal within the sweep region, and if so should
guide the operator to appropriate sweeps and modes to determine if its
horizontal position to within 1 cm or so, and its depth to within about 20%
accuracy.  It needs also to aid the operator in marking the spot clearly
and unambiguously, and perhaps even numbering and coding it.

     The air knife:  The "air knife" is a long tube with a thin radial slit
near its tip from which issues a jet of air driven by about 15 atm
pressure.  The idea is for the particles of soil to be levigated and then
carried away by the stream of air without introducing large forces in the
vicinity.

     As with a snow blower (for those with experience in appropriate
climates), the air knife works well on a small scale, where the particles
do not need to be moved very far in order to expose the putative mine.  On
the other hand, the source of air requires power, and that means capital
investment and a supply of gasoline, and the actual compressor has uses
outside the demining community; as with all of the other materials
involved, in the demining environment the compressor and hoses need to be
protected against theft.

     This kind of problem has been solved in the field, since demining is a
respected profession, and the demining teams themselves must trust one
another implicitly if the job is to get done.

     Nevertheless, many choices need to be made in order to optimize the
performance of the air knife and of the system as a whole.  Because
deminers work far apart in order to minimize the damage if a mine explodes,
fairly long hoses need to be used to feed the air knives.  And there is not
a lot of experience with the air knife in the field.

Smart Prod
     Detection of characteristic acoustic resonances:  Since the canonical
approach to demining is the detection and location of a putative mine, and
then the careful probing with a metallic or fiberglass prod at a small
angle to the horizontal, while the deminer lies prone behind a face shield,
an improved prod which would provide immediate "mine-no mine" information
would be very welcome.  One would advance the prod carefully toward the
object, and obtain a signal which was "mine", "not mine", or "equivocal."
If the first, the mine could be detonated in place or marked for
detonation; if the second, the "not mine" could be quickly excavated so
that it no longer could be mistaken for a mine, or it could be durably
marked in place.  Only in the case of "equivocal" would routine prodding,
excavation, etc. be necessary.

     Although there are several characteristics that might be used to
discriminate mines, one of the approaches that has been picked up by Paul
Horowitz and Jonathan Wolff at Harvard University uses acoustic excitation
(Wolff, 1997).  In principle the idea is to drive the mine case by a prod
at a frequency swept through the audio range, while another prod against
the mine case, connected to an accelerometer or other detection of acoustic
signal, is used to feed an amplifier and correlation device.  The
excitation spectrum (transfer function) of the object from the "acoustic
input" to the "acoustic output" is correlated against similar transfer
functions for candidate mines, and if there is a robust match, the object
is a mine.

     Similarly, pebbles, pieces of wood, and metallic scrap need to be
investigated in the field environment, and those templates stored for cross
correlation.

     A "receiver operating characteristic" in the radar sense is generated
by a variable threshold, so that if the threshold for calling something a
particular kind of mine is high, very few false positives will be called.
However, real mines may then escape into the equivocal or the "not mine"
category.  At low thresholds, everything would be called a mine.

     In actuality, with several mine profiles, the ROC rises to 95%
probability of detection at a false alarm probability of something like 4%
(300 trials on 3 mines and 200 trials on 3 neutral objects).

     Work is underway to see whether a single probe can be used with pulsed
excitation to avoid construction and assembly complications of the
two-probe system, and to refine the decision procedure.

     Thermal diffusivity probe:  Remote-reading thermometers exist which
consist of a quartz optical fiber viewed at the instrument end by a
detector sensitive to thermal infrared. If the fiber of any reasonable
length is pressed against or brought to the neighborhood of some object,
the amount of infrared viewed by the fiber is characteristic of the
temperature of the object, and its emissivity.

     For mine characterization, a probe containing the fiber would be
pressed against the putative mine, and a laser diode would be used to
introduce a pulse of few microsecond duration, that would raise the surface
temperature of the object illuminated by the fiber.  The ir sensor would
then watch the decline of this surface temperature with time, the speed of
which is dependent upon the thermal diffusivity.

     Since stone feels cold to the touch, while the plastic or wood of
which mines are made is "warm", there is evidently a very big difference in
thermal diffusivity between stone and plastic.  It remains to be seen
whether objects in the ground environment such as wood, metal fasteners,
and the like can be reliably discriminated from common mines either by the
sole use of thermal excitation and observation, or whether this is a useful
adjunct to other elements of a smart probe.

Unexploded Ordnance in Laos
     The problem in Laos left over from the late 1960s and early 1970s is
not landmines since there was very little ground combat in Laos, but
unexploded "cluster bomblets."  U.S. forces were not permitted to operate
in Laos in support of U.S. activities in South Vietnam, but hundreds of
thousands of aircraft missions were flown, many of them dropping CBU-24
cluster bombs, each fitted with 600 or so "bomblets."  These bomblets are
fragmentation munitions, fuzed to explode as they strike the ground, but
some of them had their fall softened by foliage or struck in an anomalous
position, and some 10% or so may have failed to explode promptly.  These
bomblets are mechanically fuzed, and remain deadly for many decades.

     Unlike landmines, however, the bomblets contain large amounts of
metal, and they are extremely readily observed on the crudest mine
detector.  So the problem of finding and disposing of bomblets in Laos is
quite different from that of finding and disposing of APL elsewhere.

     What is needed is a specialized "magnetometer" or metal detector that
can operate from a considerable distance, and that will assuredly detect a
bomblet and can be used in such a way as to  give an accurate indication of
its location.  Since the bomblets are identical, a "signature" can be
determined to help distinguish a bomblet from a random piece of metal in
the ground.

     Various approaches then present themselves to the elimination of these
bomblets.  For instance, a rapid hardening foam can be used to encapsulate
the bomblet and its surroundings, together with a pull cord, so that after
a few minutes the cord can be pulled remotely from behind a shield, and the
bomblet taken to a convenient location in the neighborhood for detonation.
Alternatively, even if the bomblet location is not well known, an explosive
foam (LEXFOAM) can be used, together with normal detonating fuze, so that
all of the bomblets that have been located in an area can be remotely
exploded by an electrical signal applied to the fuze system.

     The MIT workshop report (Tsipis, 1996) available on the Web gives an
example assuming that one hundred million bomblets were dropped in Laos and
that five million of them still remain, assumed to occupy 10% of the total
area of Laos.

     Current or near term techniques would allow an individual with a
hand-held enhanced metal detector walking 2 km per day and searching a 2-m
swath to mark 1 sq km per year. The job will take 25 years and one assumes
that another worker is involved in the clearance operations (marking,
neutralization) so that 2,000 workers for 25 years at $3,000 per year will
contribute a total labor cost of $150 million.

       The total 25-year cost of detectors is about $50 million so the
total cost of clearing Laos of bomblets would be about $200 million and
take 25 years.  However, improved metal detectors and clearance techniques
such as the air knife and smart probe should allow a factor five increase
in productivity so that the job could be done in five years and the overall
cost would be about $50 million.

THE POLICY CHALLENGES:  COORDINATION, COMMUNICATION,
AND LEARNING

     As noted above, since the landmine crisis came to public attention in
the early 1990s substantial effort by governments, private firms, and NGOs
has gone into finding better ways to detect and clear landmines.  An
ever-growing list of conferences explore the problems and discuss potential
solutions. (9)  There is some evidence of cross-fertilization and
learning from one conference to the next, but it is hard to avoid
the suspicion that a significant percentage of the effort is
expended on discussing the same issues in different venues with
limited cumulative benefit.

     Perhaps that is how it must and should be.  It might be argued that
the problem is sufficiently compelling and the potential rewards of success
sufficiently attractive that a free market of ideas will in time yield
answers.  Attempts to encourage or impose coordination (the latter could
only be done by governments or multilateral organizations on the research
they support, if that) might well end up excluding people with potentially
valuable contributions to make and perhaps impede the creative process.
But it appears worth considering how one might at least encourage
coordination and communication in four important areas, two in the research
and development phase and two during testing and evaluation.

     1)  Improving communication and coordination among those doing
demining R&D. The funds available for demining R&D are relatively limited
and it seems obviously desirable for different projects and programs to
learn from one another. (10)  There are two alternatives to promote this
communication, one traditional and one that would take advantage of the
advances in communication offered by the Internet and the WorldWide Web.
The more traditional option would be an annual conference explicitly
devoted to sharing and evaluating the previous year's R&D, in order to
promote more rapid progress.  (The annual international AIDS research
conferences come to mind as a potential model.)  It would have to include
the government officials charged with the R&D programs, the scientists and
engineers actually doing the research, as well as representatives from the
private sector, academia, and potential funding sources such as
foundations.  It would be very important to include reports and discussions
of field testing and experience, and to have representatives of demining
organizations actively participating so that researchers had a sense of the
genuine operational requirements for new technologies.

     The first conference would necessarily repeat prior meetings if it
sought to assess the "state of the art" of demining, but subsequent
meetings could be devoted explicitly to the work done in the previous year.
Given the politics of the landmine issue (at least in the United States)
and the natural reluctance of demining organizations to adopt new
approaches to such an high-risk enterprise, the choice of a sponsor for
such a meeting would be important.  Funding should probably come from a
number of sources to enhance the effect of a "neutral tent."

     The alternative to a regular conference would be to utilize the
WorldWide Web to promote sharing of ideas and research results.  A number

of clearinghouses are emerging in the United States and Europe devoted to
collecting and sharing demining information, but none appears to be devoted
exclusively to R&D and the promotion of communication about it. (11)
There
are a number of relatively minor technical issues to be overcome to make a
WWW "conference" site useful, but it would offer significant advantages in
cost and accessibility to a wider range of researchers. (12)  It would be
particularly helpful if, in addition to the information about conferences
now available on a number of sites, digests of papers, and summaries or
full proceedings could be available.  The use of on-line conferencing could
also enable periodic discussions of the reported research to encourage
greater "learning" and to provide incentives for cross-fertilization.

     2)  Assuring communication between demining R&D and those doing R&D in
potentially relevant areas.  As mentioned above, the need for better
detection of explosives arises in many fields and one would expect at least
some of the R&D to be potentially relevant to demining.  The "artificial
dog nose" in which the U.S. Defense Advanced Research Projects Agency is
investing $25 million over a three year period is an obvious example
(Rouhi, p.21; see also JASON for an evaluation of the potential of this
technology). Some of the individuals doing demining research will naturally
look to these other fields for ideas.  But one could imagine periodic
meetings of R&D managers from the government and private sector that would
be explicitly devoted to analyzing how and whether research for other
purposes could serve demining.  This kind of coordination is already being
attempted among U.S. government-sponsored programs devoted to developing
sensors to detect chemical and biological weapons.  In spite of
proprietary, and perhaps classification considerations, this kind of
systematic effort could yield significant benefits.  This is another area
where, if the security considerations could be overcome, WWW facilities
could be a reasonable substitute for meeting face-to-face.

     3)  Achieving realistic testing of candidate technologies.  At least
in the United States, there is criticism that new and potential
technologies the Department of Defense is developing are not being tested
under the kinds of conditions they would encounter if used for actual
demining.  This is a frequent and persistent criticism of military R&D in
the United States, and demining research may simply be one more example.
But given the natural wariness of those doing demining to adopt new
technologies, the lack of "real world" tests adds additional -- and
unnecessary -- difficulty.

     For their part, deminers would need to accept that a fairly
substantial percentage of new methods might work only in certain terrain or
under specific conditions, or might be practical only after several
iterations of field trial and technology modification.  If each failure is
taken as evidence that only the tried-and-true methods are appropriate, the
vital communication between researchers and consumers will soon wither and
valuable options will be rejected.

     4) Improving evaluation of candidate technologies.  Perhaps the
biggest problem with introducing effective technology into the humanitarian
demining area is one of psychology and reward.  Many of those who are most
creative, and most of those who operate in the commercial area want to see
their own ideas or products succeed.  They have far less desire or
incentive (no matter how strong their humanitarian urges) to see another
concept succeed.

     Those responsible for purchasing or choosing appropriate technologies
that they did not invent are beleaguered by individuals, more or less well
intentioned, who believe that a concept will solve the problem, or by those
who are hired as proponents to push a proprietary technology.  So there is
a lot of "noise" and "clutter" in the system of technologies that might be
chosen for application.

     What seems to be missing is a group of talented, technologically
knowledgeable people, whose job it is, however, in a positive fashion to
test technologies as they are proposed and provided by the proponents, and
to score them promptly in a fashion that will lead to their improvement or
abandonment.

     It is surprisingly uncommon for such technological intermediates to
say about a candidate system "Yes, it has improved a lot, but it still
fails."  But that is exactly what is needed, and a lot of it is needed in
order that new and effective technologies can be brought to bear on this
problem.


CONCLUSIONS:  A ROLE FOR ACADEMIES OF SCIENCE
AND SCIENCE SOCIETIES

     All four of these areas where coordination and communication need
improvement appear to offer potential roles for academies of science,
perhaps in partnership with other organizations.  For example, European
academies, either individually or through their regional organization
(ALLEA, for All European Academies), might supplement or complement efforts
by NATO or the European Communities.  Their ties to academies in other
parts of the world, such as through the Inter-Academy Panel on
International Issues, might be used to facilitate the real world field
testing needed if new approaches are to prove themselves and be adopted.

                                   NOTES

     1.  Estimates of the number of mines and of mine victims have
substantial margins of error.  The State Department and ICRC reports cited
in the References respectively discuss the problems of estimating numbers
of mines and of casualties.  In addition, many of the estimates of
casualties, numbers of mines being cleared and laid, and costs and rates of
clearance are several years old.

     2.  The use of landmines is currently governed under international
humanitarian law, specifically a protocol of the UN Convention on
Prohibitions or Restrictions on the Use of Certain Conventional Weapons
Which May be Deemed to be Excessively Injurious or to Have Indiscriminate
Effects.  The Protocol, which was amended and strengthened in May 1996,
requires that the location of landmines should be recorded and encourages
that these records be made available to assist demining.

     3.  There are some reports that the pace of laying new mines has
slowed, perhaps in part as a result of the international campaign against
them.

     4.  This figure may be somewhat deceptive, since U.S. law prohibits
U.S. soldiers from clearing APL.  Instead, the United States has provided
training and other assistance for those doing demining; in announcing the
Demining 2010 Initiative, Secretary of Defense William Cohen stated that "a
full quarter of the demining efforts around the world are conducted by
experts trained by the U.S. military" (U.S. Department of State, 1997c).

     5.  It is now about 16 years since a relatively standard personal
computer (PC) was marketed, based on integrated-circuit microprocessor.
But now a system that cost $5,000 in 1981 can be bought for $1,000, except
that the processor raw speed is increased by a factor 20; there is another
factor 10 from sophistication of the processor; the semiconductor storage
is increased by a factor 500; and the mechanical (hard disk) increased in
capacity by a factor 30,000.  Most of this advance has come in the last few
years, in view of the exponential nature of progress in this industry, as
evidenced by Moore's law. So the goal for humanitarian demining should be
to obtain the same interaction with the user, and the same kind of
competition worldwide that has led to the enormous increase in capability
of the personal computer.

     6.  The  May  1996  amended Protocol on Prohibitions or Restrictions
on the Use of Mines, Booby-Traps, and Other Devices of the Convention on
Conventional Weapons, to which all the major nations are signatories, but
which has not yet been ratified by many or entered into force, mandates
that all anti-personnel landmines (APL) be readily detectable either
inherently or by a not easily removable attachment, providing a response
signal "equivalent to a signal from 8 grams or more of iron in a single
coherent mass."  Hand-emplaced mines, unless in marked areas, and all
remotely-delivered APL must self-destruct to at least 90% confidence within
30 days after emplacement and there must be a  back-up self-deactivation
feature so that no more than one in  1,000 activated mines will function as
a mine 120 days after emplacement.

     7.  At least in the United States, the battle over U.S. policy toward
an immediate and comprehensive ban on landmines has made that communication
more difficult with demining organizations with close ties to the
humanitarian community.  In general, there should not be a major barrier to
communication of technical knowledge.  The workers who remove the mines may
be technically unsophisticated, but many of the leaders of demining
organization have substantial practical technical knowledge and experience,
often acquired in the military. Extensive detail about the U.S. technology
research and development and test program is available at
http://www.demining.brtrc.com/ and an extensive summary of landmine
WorldWide Web pages can be found at
http://lenti.med.umn.edu/~mwd/landmines.html.

     8.  Some of the problem is whether to include in the total R&D the
traditional military countermine technologies that might be relevant to
humanitarian demining, a source of dispute particularly within the
humanitarian community.  Much of the rest is simply finding a way to
collect and collate information about what is being done in places like the
U.S. Department of Energy's national laboratories.

     9.  There is some dispute about whether the amount of funding is
sufficient or needs to be increased, which we do not feel qualified to
judge at this point.

     10.  For example, in the References for this paper, see FOA, JASON,
Joint Research Centre, Swiss Federal Institute of Technology, and Tsipis.
Information on conferences can also be found at the websites cited in the
next footnote.

     11.  For example, the U.S. Department of Defense has funded the
Humanitarian Demining Information Center at James Madison University in
Virginia (http://www.hdic.jmu.edu/hdic).  In Europe, the Demining
Technology Center (http://diwww.epfl.ch/lami/detec/detec.html) of the Ecole
polytechnique federale de Lausanne (EPFL) in Switzerland may emerge as a
major source of information, but it is also directly engaged in research
and therefore may not entirely satisfy the need for a neutral site.

     12.  In view of the greater difficulty of getting pictures and
equations onto the Web, a significant advance at minor cost would be
achieved if a central non-profit organization were funded to receive
manuscripts in a number of formats, together with illustrations that could
be scanned, and in turn to post the transformed article onto the Inter-Mine
homepage (or whatever name was chosen).  There it could be reviewed by the
submitting organization, and any problems worked out before it is posted in
its final form there or anywhere else.  As is usual with "real publication"
each of the posted articles could have a line "Received by Inter-Mine
10/14/97" and "Received by Inter-Mine in revised form 10/29/97" or
whatever.  This would save money, accelerate the pace of interaction, and
the like.  It is inexpensive to do if one is handling only the mechanics
and not editorial aspects.



                                REFERENCES

The Arms Project, A Division of Human Rights Watch, and Physicians for
Human Rights, Landmines:  A Deadly Legacy (New York:  Human Rights Watch,
1993).

Associated Press, "U.S. Launches Demining Initiative," October 31, 1997.

Craib, J.A., "Survey of Mine Clearance Technology," Conducted for the
United Nations University and the United Nations Department of Humanitarian
Affairs, 1994.

FOA -- National Defense Research Establishment, "International Workshop of
Technical Experts on Ordnance Recovery and Disposal in the Framework of
International Demining Operations," Stockholm, Sweden, 8-10 June 1994.

International Committee of the Red Cross, Assistance for Victims of
Anti-Personnel Mines: Needs, Constraints, and Strategy (Geneva:  ICRC,
1996[?]).

JASON, New Technological Approaches to Humanitarian Demining (McLean, VA:
The Mitre Corporation, 1996).

Joint Research Centre, European Commission, "International Workshop and
Study on the State of Knowledge for the Localisation and Identification of
Anti-Personnel Mines," (Luxembourg:  Office for Official Publications of
the European Communities, 1995).

Morrison, Philip and Kosta Tsipis, "New Hope in the Minefields," Technology
Review, October 1997.

Rouhi, A. Maureen, "Landmines:  Horrors Begging for Solutions," Chemical
and Engineering News, March 10, 1997.

Salter, S.H. and C.N.G. Gibson, "Electronic Navigation Systems for the
Dervish and Other Mine-Detonating and Detecting Vehicles," prepared for the
International Workshop on Sustainable Humanitarian Demining, Zagreb, 29
September 1997.

Swiss Federal Institute of Technology (EPFL), "Proceedings:  Workshop on
Anti-personnel Mine Detection and Removal (WAPM'95)," Lausanne,
Switzerland, 30 June and 1 July, 1995.

Tsipis, Kosta, "Report on the Landmine Brainstorming Workshop of August
25-30, 1996," Program in Science & Technology for International Security,
Massachusetts Institute of Technology, 1996.

U.S. Department of State, Hidden Killers:  The Global Landmine Crisis
(Washington, DC: U.S. Government Printing Office, 1994).

U.S. Department of State, "Fact Sheet:  The Demining 2010 Initiative,"
October 31, 1997 (a).

U.S. Department of State, "Fact Sheet:  U.S. Government Humanitarian
Demining Program," October 31, 1997 (b).

U.S. Department of State, "Transcript:  Albright, Cohen Announce U.S.
Demining Initiative," October 31, 1997 (c).

Wolff, Jonathan, "The Landmine Acoustic Characterization System," April 18,
1997, available from Professor Paul Horowitz, Harvard University.


RLG:rlg:W304PHD:103197.PHD