U.S. military develops IED detection training with Xbox Kinect and ‘virtual’ sniffer dogs
Tue 28 Oct 2014
The Naval Research Laboratory, the scientific research branch of the U.S. Navy which created TOR in the 1990s, is developing a virtualised training environment for training military dog handlers who must use canine personnel to detect Improvised Explosive Devices (IEDs) in battlefield environments.
NRL scientist Adam Moses has used the third version of the U.S. military’s Virtual Battlespace VR simulation programme to create ROVER, a simulation program featuring a virtual Labrador sniffer dog tasked with detecting an IED plume in a challenging environment.
The system involves the use of an Xbox Kinect motion sensor, which allows the participant to engage interactively with the virtual sniffer dog via a skeleton tracker system. “Gestures are important,” says Moses. “the whistle commands are important, even the voice commands are important.”
Even well-concealed IEDs generate a plume which a trained sniffer dog can engage with. Moses says: “It’s actually leaking all this gas subtly, at levels no human-built sensor could read because the particles are so small. Per billion is the level these dogs can sniff.”
The ROVER program, which originated during a period of combat operations in Afghanistan, models a scenario where the plume trace is strongest in close proximity to the virtual IED, and then becomes elusive due to atmospheric conditions.
The program’s main challenge is not so much to simulate plumes or other battlefield variables, but to realistically model dog psychology. “Not only should the dog be taking cues from the handler,” explains Moses. “the handler should be taking cues from the dog…If a dog’s going down a street, every now and then you’d see him glance to the left or right when he passed an alleyway. And sometimes he would glance longer or stop, and that’s one of those cues that’s really important; that’s when you have to read something from the dog.”
Since the dog is motivated by trying to please its handler, detection incidents can be completely lost if the trainer keeps the dog moving when it was attempting to show its handler something of interest. In this way, Moses says, “An inexperienced handler can un-train a dog by accident.”
ROVER employs Moses’ previous work on the NRL’s CT-Analyst program, which was used at the 2009 and 2013 presidential inaugurations and at a superbowl event.
Moses explains: “CT-Analyst is unique because it simulates worst-case scenario plumes based on minimal information.”
Since the U.S. Marine Corps is closing its IED dog detection facilities, Moses, who studied computer science at Virginia Tech and joined the NRL in 2003, believes that the program could be a potential repository for knowledge which will not otherwise maintain in circulation. He also sees potential for ROVER to be adapted for use in Law Enforcement, simulating difficult environments such as crowded airports and border crossings, in new Virtual Battlespace environments with multi-player capability, with real participants portraying both law enforcement and its antagonists.
But Moses believes that the scenarios need to become more sophisticated, with the ability to choose from a wider range of canine personality types. “This one’s super obedient,” he explains. “this one is distracted a lot, this one is more aggressive. [The handlers] don’t know which dog they’ll end up with, so if they train against 20 different kinds they’ll be better in the long haul.”