DARPA pop-up testbed takes on spectrum management
- By George Leopold
- Apr 07, 2017
As use of the electromagnetic spectrum grows, so does the noise from signals from wireless devices and other emitters, which can make it more difficult for warfighters to maintain critical communications links, interpret ambiguous radar returns, or defend against electronic warfare tactics.
One way to address the spectrum gap is through a signal analysis technique called modulation recognition. (The most common modulation schemes are AM and FM: amplitude and frequency modulation.) Researchers have been working on developing algorithms that can be used to automatically recognize communication signals. If signals can be parsed, operators can use that intelligence to make more efficient use of scarce frequencies.
With that in mind, the Defense Advanced Research Projects Agency sponsored a recent "Battle of the ModRecs," as in modulation recognition, to test new approaches to navigating the "thicket of waveforms" within the radio frequency spectrum.
Among the goals was demonstrating new modulation recognition techniques for identifying signal origins and types. The DARPA event, which used about 30 modulation schemes, also sought better techniques for sending and receiving information like real-time intelligence data over the least crowded spectral bands.
"Classically, we've described the spectrum strictly by occupancy: There are signals present or not. As the spectrum becomes increasingly filled, we now need more information," explained Paul Tilghman of DARPA's Microsystems Technology Office.
"Modulation recognition is that first step towards getting beyond just describing 'presence' or 'absence' [and] actually describing what is present," Tilghman added.
Another goal was pushing spectrum utilization approaches such as modulation recognition out the laboratory and demonstrating them in real-world scenarios, according to Tom Rondeau, the DARPA program manager.
The DARPA-sponsored "battle" pitted hand-coded expert systems against newer platforms based on emerging machine learning techniques. The hand-coded systems performed better in identifying signal characteristics, but Rondeau predicted the machine learning approach would soon close the gap.
"We now have a better understanding of the state of the art and which directions to explore as we pursue our goal of more effectively managing the spectrum," he added.
The agency argues that modulation recognition is a key step in the effort to achieve "wireless situational awareness" that could help squeeze more capacity out of congested electromagnetic spectrum. That awareness could then be used to predict spectrum use and ultimately boost throughput to move data faster via wireless airwaves.
This article was first posted to Defense Systems, a sister site to GCN.