Beating the Better Mousetrap, Part 1

In the world of landmines versus people, landmines are the mousetraps and we’re the mice.  Technologically, most landmines are pretty basic items.  A physical trigger sets off an explosive charge of varying size which propels shrapnel into the victim.  Some landmines have additional items, like ball bearings, in the casing to make them more destructive.  The casing itself may be made of plastic or metal, but most landmines have a minimal metal content to ensure they can be found using commercially-available metal detectors.  The design of landmines hasn’t changed much in decades, but as mice, we humans want alternative means of overcoming their threat.

In the last few weeks, I have seen a number of new technologies announced to find landmines in the field.  They all have some merit, however, I can’t help feeling that they are more gimmicks than anything else and as gimmicks, they get some headlines, but don’t actually get the job done.

The original gimmicky method of landmine detection is probably the only one that really works, rats, and they work precisely because the rats are used in the same way that dogs are.  For decades, dogs have been used to sniff out explosives in minefields (as well as food in customs halls, truffles in the woods, foxes in the fields, etc.).  Dogs are trainable and have the ready presence of their handlers to monitor and respond to behavior.  Norwegian Peoples Aid and the Marshall Legacy Institute use dogs for mine detection and recently, mine detection dogs have been introduced to Angola.  Coincidentally, APOPO’s Hero Rats have also recently been introduced to Angola to search minefields.  APOPO, a Belgian organization has been training and using Giant African Pouched Rats in Mozambique to find landmines with a strong record of success (hence the opportunity to expand into Angola).  APOPO’s founder, Bart Weetjens, has a very popular TED Talk in which he explains how he started the organization and realized that rats could be used to find landmines and detect tuberculosis.  Weetjens argues that rats are ideally suited to the work because they have exceptional senses of smell, they are trainable, they are very cheap to maintain, and, unlike dogs, do not want to please their handlers, so they are focused solely on the tangible reward of treats and not affection.  But again, the point I would make is the gimmick in this mine detection method is the use of rats over dogs, otherwise the protocols are very similar.  It is when the methods of detection and the tools to find them diverge that I have issues.

When reviewing a new technology and judging whether it is effective or a gimmick, ask yourself this: Would you feel comfortable playing soccer on a field declared clear using any of the following methods?  If yes, then it’s either reliable or you are very brave.  If no, then something in the set-up is wrong.  A few of the very gimmicky mine detection schemes are:

• Bees: The Defense Advanced Research Project Agency (DARPA) commissioned research on the “Stealth Insect Sensor Project” in 2004, but decided that “Bees are not reliable enough for military tactical use.”  However, the initial research suggested that bees could be used to locate explosive compounds at very low concentrations.  Using very small radio transmitters to track the bees location, one could identify likely sites of landmines by placing a hive within three kilometers of a minefield.  The thought is that bees would work like dogs without the handlers.  They would be cheap and easy to disperse.  The problem is that bees are not perfectly trainable.  There is no guarantee that the bees would not find something else within the three kilometer radius of the hive that is more appealing than a landmine, like a flower.  And there is still the case that another means would be needed to verify any possible landmine discovered by a bee and there is no way to ensure that the bees have absolutely surveyed the entire area within the radius of the hive (KSL.com).  In short, it’s a gimmick and not one that is reliable enough for me to test.
• MouSensors: I think a good sign that a mine detection system is a gimmick is when it comes with a cute name like “MouSensor.”  MouSensors are genetically modified mice that are designed to be “500 times more sensitive than normal [mice] to the smell of TNT.”  MouSensors “are so sensitive to TNT that encountering the molecule is likely to change their behavior involuntarily” so little or no special training would be necessary.  Basically, MouSensors freak out at the merest scent of landmines and by attaching a location chip to the MouSensors, handlers would just monitor the mice on a computer screen and record the location of the seizure.   Yes, scientists at Hunter College in New York City have built mice that go into fits when they landmines (io9.com).  If PETA were to find out about this, it would probably also go into seizures.  Like the bees, all locations identified by the MouSensors would need to be verified and cleared; unlike the bees, housing and storing the MouSensors is not addressed and how the demining teams would get the mice to return home predictably is an open question.  Very gimmicky, but no more so than bees.
• Nano Fiber Film: Staying within the academic community and just a short train ride from Hunter College, researchers at the University of Connecticut have developed a “fluorescent nanofiberous film” that can detect trace amounts of explosive vapors released by landmines and is activated by ultraviolet light.  The film can also be configured as a small paper test strip to spot check areas.  Awesome, it’s CSI Landmine.  But here is the hitch: to use the film to find landmines, “the nanofilm could be rolled out over a piece of land like a giant roll of paper towels.”  So to sample a field, one would need enough film to cover the entire field and then either wait for dark to expose it the UV light or use a specially-equipped camera.  Yes, it would work, but it’s a little complicated and the need to completely cover a field with the film to ensure it is clear of landmines feels a little off.  The process is undergoing a large-scale field test in Sweden and it may have some merit, but let’s see how the test goes (Discovery News).  Also, who would want to be the person who has to roll out the film over the minefield.  Less gimmicky than bees or MouSensors but still impractical.
• Drones: Quantum International Corporation won’t call them drones, but says it wants to develop “autonomous, robotic vehicles” that will use spectral analysis to find landmines.  In theory, the drones may also be able to clear the mines that they find, making this a distinct improvement over the other methods described above which only find the mines.  The drones could be programmed to search in a more systematic manner than bees or mice ever could, but there is a significant problem: cost.  Quantum International Corporation is getting into the robot demining business to “improve… our company’s shareholder value” and “estimate the market value for this technology to be in the billions” of dollars (Robotics Tomorrow).  Who is going to pay those billions of dollars?  Humanitarian demining is already under-funded and the additional competition from drones – which do not have the best reputation within the humanitarian disarmament community at the moment – will be unwelcome.  Demining isn’t an area one gets into to make money and so while this is much less gimmicky than the others, I’m very skeptical about how likely it is to move forward as proposed.
• Mine Kafon: Okay, I admit, this looks cool:

But will it work?  The Mine Kafon was designed by an Afghan-born designer and is wind-propelled.  Like the mice and the bees, it won’t work systematically; like the drones it has the capacity to destroy the mines as it goes; but unlike any of those, it can’t work in rough terrain or in wooded areas.  It’s based on toys the designer, Massoud Hassani, made as a child and probably should stay that way (CNET.com)

In 1996 the Massachusetts Institute of Technology (MIT) Program in Science and Technology for International Security “drew together a disparate group of participants, including a field worker from Laos with many years of demining experience; researchers with expertise in physics, chemistry, electrical engineering, material science, and anthropology; several people working on high-tech mine-detection schemes; and three experts on demining” to review the technologies and methodologies available for demining and concluded “the use of a metal detector, hand-held probe, and explosive charge is generally accepted as the most reliable demining method despite its laborious and perilous nature.”  The researchers predicted the development of chemical sensors like the nanofilm above and the spectral sensors like in the drones, but they supported further improvements on the slow and laborious work of manual demining through use of magnetometers to assist standard metal detection, an “air knife” that uses high-pressure air to expose buried landmines and Lexfoam which can be used to detonate found mines.  Their ultimate recommendation was for the “development of autonomous, mechanized demining systems to incorporate some of the more sophisticated detection technologies” available; essentially Quantum’s drones (MIT Technology Review).  Again though, I would point out that until the reliability of new technologies is certain, the slow and steady method of manual mine detection and clearance remains the only viable and humane means.

Michael P. Moore, October 31, 2012