Portland, Ore. One of the United States' largest electromagnetic-emission "shielding" labs has switched roles to help armed-services personnel detect bombs. The University of Missouri at Rolla's Electromagnetic Compatibility Laboratory is trying to sense, rather than shield, emissions from the triggering mechanisms in improvised explosive devices (IEDs), which could alert troops to unexploded munitions.
EE professor Todd Hubing and assistant professor Daryl Beetner said they saw the benefit of reorienting their work when visiting nearby Fort Leonard Wood, Mo., where they viewed IED training tapes. They saw that the emissions from many command-initiated IEDs, which are normally placed at the side of a road, were sufficient to be sensed from a distance. Hubing and Beetner immediately began thinking that it would be possible to detect the signature of electronically triggered IEDs, even when they are not transmitting, but merely "waiting" beside a road in Iraq to receive a detonation signal.
"We started looking at unintended electromagnetic emission from improvised explosive devices and found that if we shifted their frequency to the audio range, almost anyone could recognize their distinctive 'signature' sound," Beetner said. "And if you can hear the difference, then there has got to be a way to build a device for soldiers to use that detects radio-controlled IEDs from a distance."
That's the opposite of the lab's typical practice. "At our lab we try to shield devices, so we know just how very difficult it is to make an electronic device not emit electromagnetic energy even when it's just a receiver," Beetner said.
According to the engineers, sophisticated electronic devices like cell phones cannot be completely turned off there is always active circuitry inside, if only to monitor whether the "on" button has been pushed. What's more, the receiver section of a typical IED triggering device, like a remote-controlled toy, emits even more electromagnetic radiation than a cell phone, making them even easier to "hear." Not only did the Electromagnetic Compatibility Laboratory know what to listen for, it had the equipment and anechoic chambers to hear it.
"From our experience at the lab, we knew intuitively that all these kinds of triggering devices have a unique kind of signature to their electromagnetic emission that we can identify," Beetner said.
To prove their case, the EEs have frequency-shifted to the audio range the unintentional sounds created by typical IED triggering devices. These unintentional sounds are made by the receivers of the various triggering devices while they are waiting for a remote transmitter to trigger them. Each receiver has a distinctive signature that can easily be memorized, even after only casually listening. That's true of all receivers. For instance, the receiver in a toy truck has a repetitive "chirp." A car alarm receiver has a higher-pitched chirp. A wireless phone makes a distinctive clicking sound and a cell phone receiver a high-pitched squeal. Even digital cameras and video cameras have different, distinct sound signatures. Hubing and Beetner have recorded these audio signatures and put them online at web.umr.edu/~daryl/IEDs.
"Two things we think will really work are, firstly, to train people to listen for these signatures with a prototype device we use to shift the frequencies, and secondly, to train a neural network to automatically detect them," Beetner said.
In remote locations, where the IED is not within an urban environment filled with radio receivers and emitters, it may be relatively easy just to listen for the signature of any receiver up ahead.
"There has been a lot of trouble with bombs sitting in the middle of the desert on the side of the road but if you were going down the road and you could tell there was a receiver in that pile of rocks up ahead, then you could deal with it from a safe distance," said Beetner.
In urban environments, on the other hand, filtering out the cacophony of noise in the environment may be the biggest engineering hurdle, according to Beetner.
"One of the things we need to do is to show we can filter out all the other ambient signals in the air there are a lot of emitters of electromagnetic radiation," he said. "There is still a lot of engineering work to do, but I think there is a good chance we can make this work. We would like to have a prototype within the next year."
Other possible related devices the lab is considering building, after proving the concept, are jammers to override an identified receiver or, alternatively, to send a preemptive signal to detonate the device from a safe distance. The EEs plan to train personnel in how to listen for IEDs, much like "sonar" operators on submarines. Once trained, an artificial neural network could learn to take over the job or at least provide an "autopilot" mode.
One problem is representing the signal, because it takes about 10 seconds to play back 100 ms of frequency-shifted sound, so algorithms will also have to be found to determine which parts of sounds are "ordinary." In that scheme, a neural network might act as an intelligent filter, only indicating to the operator out-of-the-ordinary sounds.