One day the inventorying of electronic equipment in the workplace could be accomplished through radio frequency (RF) noise detection rather than labelling and tagging. The concept is based on the fact that all electronics always emit distinct radio noise when they’re running.
Those unique RF prints could be used instead of serial numbers or expensive, attached RFID identifying tags and could quickly ID the gear. Even gadgets of exactly the same model type appear unique when analyzed, say researchers (PDF).
“Electromagnetic emissions are highly structured and [are] a direct manifestation of the circuits that generate them,” says Chouchang Yang of Disney Research, one of the scientists working on the project, in a press release distributed on EruekAlert’s website.
A pile of laptops or a wall of racked modems or servers could conceivably be inventoried, for example. They all look the same in the racks, but in fact all have a unique electromagnetic emission print. Ultimately, that print could be logged on original purchase, then the item could be physically tracked throughout its life.
It would be done not by peering at the enclosure’s inconveniently located serial number or other labelling appendage, but through the integral, scanned RF noise print.
And because barcoded stickers, or enclosure-adhered RFID tags, often relate to just that—the enclosure only—this new process might be more accurate. The innards, such as a drive, could have been swapped out earlier, leaving the enclosure-mounted tag referring to the original, and wrong, drive.
It’s “like a fingerprint” the release says of the technique, under development, called EM-ID.
I’ve written about other wave-oriented identification before, including how sound-fingerprinting could be used to spot utility grid attackers. That system identifies the unique audible noises that emanate from cyber-physical systems. They’re systems that use digital controls to perform mechanical tasks, such as opening a water valve. Spoofing the opening of the valve won’t sound right, in that case.
Soundwaves are also being used to predict mechanical failure. In that case, an algorithm identifies the noise a machine should make and compares it to sounds it shouldn’t make—such as vibrations when a drive belt is failing.
In both of those cases, the detection uses audible sound waves, which are different from the electromagnetic waves used in the Disney work, but the idea is similar in that wave signatures, along with algorithms, are used to detect changes or anomalies.
The Disney scientists reckon their system of radio noise identification is 95 percent accurate. They obtained “72 percent accuracy for the iPhone 6 to 100 percent accuracy for toy light sabers,” they say.
“It's not foolproof,” says Alanson P. Sample of Disney Research's Wireless Systems group in the release. Because the signature isn’t “designed to be a unique ID, it is possible that the EM spectrums may overlap.” That would make it harder to pick out some equipment variants, Sample says.
However, we can “identify devices right out of the box,” says Jessica Hodgins, vice president at Disney Research, in the release. And in any case, their algorithm can predict whether the RF noise is unique enough to obtain an identification.
“It can alert the user whether the device's EM-ID is unique enough to be read or if an alternative strategy is needed,” Sample said.
This article is published as part of the IDG Contributor Network. Want to Join?