Historical documents suggest that people have used dogs to fight crime since the Middle Ages, when bloodhounds first helped track down ne’er-do-wells. Centuries on, technology is slowly beginning to catch up with dogs’ sensitive noses in terms of detecting the tiniest trace amounts of target substances, but nothing can rival canines in terms of agility and versatility.
People who train and use dogs for security and defence applications are discovering that technology and dogs can be most effective when used together, and handlers need as much training as their canine sidekicks.
Training sniffer dogs and handlers
Chester-based Sniffer Dog Training offers international search dog and handler training to mostly police force customers around the world, and to contractors for the Transport Security Authority (TSA) and Department of Defence (DoD) in the US.
“We have two arms,” explains CEO John McHugh. “The main one is training explosive detection dogs, and our main areas are Africa, the United States, and the Middle East. We train the dogs from start to finish, we breed some as well, and we go through the whole process of selection, training, assessing and shipping the dogs out. We also train the handlers how to work the dogs and do searches for explosives. The second arm is deployments to festivals, music concerts and things like that.”
McHugh says that one of the main challenges is getting the dog’s focus. While anyone can play ball with a dog in the park, once it can no longer see the ball, it stops searching for it and switches off.
“That’s no good if we’re going to need a search dog,” he says. “It’s easy to overcome if it’s in the dog, but if it’s not interested and gets bored easily, that’s harder. We say we have to kiss a lot of frogs to get a prince. If we look at 10 dogs, we might come back with two. When we get to train those two, chances are they’re going to make it, but even then there’s about a 30% failure rate.”
Another key problem to overcome is handlers giving subconscious signals to the dog that may result in false positives, something the handler has to be trained to avoid.
“It’s what we call cueing in, when they cue the dog by a change in their body language or tone in of voice,” says McHugh. “A lot of handlers wear Velcro trousers and keep a ball in a pocket, and the sound of the Velcro opening can cue a dog in. So we need to avoid false indicators in the training of the dog but more so in the training of the handlers. We can train the dogs a lot easier than the people!”
McHugh says that by travelling and meeting up with international police forces he is always learning and picking up new ideas for training techniques. The US Office of Naval Research, in particular, has made much of its research and papers resulting from millions of dollars of research funding available to the detector dog training community. The FBI is also forthcoming with some of its information, but McHugh believes some of their research is inappropriate.
“They’re trying to train a dog in a week instead of six to eight weeks,” he says. “They’re also trying to work out if you can mix all the chemicals together whether the dog can detect it, but my opinion is the dog’s not consolidated enough on that.”
Sniffer Dog Training also undertakes its own research and is currently investigating how long it takes dog-detectable odours to permeate soil made of different composites and at different depths.
“We just buried some stuff at two metres and to find out how long it takes before the dog can find that odour,” says McHugh. “Different composites affect the timescale. When you bury a cache of arms or explosives, how long will it be before the dog can find it? It could be a day or two days depending on the composite – we call this the soak time. We’re trying to put all that together for future searches because if the police say we believe criminals hid it here this morning, we might have to tell them to secure the area because we can’t do anything until tomorrow.”
The technology improving dog detection
While nothing can beat a dog’s ability to detect concentrations of substance as low as parts per million, a variety of technology is available and a number of research projects are underway to improve detector dog training and operation for defence and security purposes.
The UK Defence Science and Technology Laboratory (Dstl) has made a solution to overcome cueing by handlers during training available to industry through its Easy IP service. A hidden marker and detector provide a double-blind animal training system that replicates a real scenario where the handler does not know the location of the target. Made from commercial off-the-shelf components including a detector and hidden RFID tag markers, the system can also track the dog’s movements through the search area, which can then be downloaded to computer software and displayed as a timeline.
Setting up a scenario to train dogs to detect explosives is prone to contamination, where traces of the target substance can be transferred to control samples. Dr Ta-Hsuan Ong of the Massachusetts Institute of Technology has developed technology that enables handlers to check whether a dog that seems to be mistakenly indicating against a control is, in fact, correct. Ong’s solution uses electrospray ionisation technology to ionise molecules in the air around a sample and direct it to a mass spectrometer to confirm a dog’s indication to a handler.
Taking a different approach, researchers from the US National Institute of Standards and Technology, Massachusetts Institute of Technology’s Lincoln Laboratory and the US Food and Drug Administration set out to discover whether biomimicry held the secret to making a machine as effective as a dog’s nose. The team found they could improve the ability of current technology to detect explosives and contraband by more than 10 times by using “active sniffing” to repeatedly suck streams of air containing the sample through a dog-nose shaped collector.