‘All about time’: Army turns to AI to reprogram RF waveforms at the edge

WASHINGTON —  As part of the Army’s ongoing transformation in contact effort, the Army Reprogramming Analysis Team (ARAT) is in the midst of creating an artificial intelligence-enabled tool that can generate code to alter radio frequency waveforms, ultimately designed to save the warfighter precious minutes when reprogramming devices in accordance with a threat.

More time means more survivability and more lethality, Nicholaus Saacks, the Army’s Communications Electronics Command’s (CECOM) deputy to the commanding general, told Breaking Defense in an exclusive interview. 

“A lot of what we do in this space is all about time. How do we give soldiers’ units more time to do their mission? If there’s time where they’re vulnerable because of an EW [electronic warfare] threat, that’s time they can’t be doing their mission,” said Saacks, whose office is working with ARAT to develop the AI tool.

Those building the AI-enabled code generator, which still doesn’t have an official name as it’s still in the “minor” prototyping stages, are combining CECOM’s A.I. Flow tool with ARAT’s Simulation Modeling Framework tool, Eric Bowes, the program officer of ARAT, told Breaking Defense. 

Related: Army’s CECOM levels up modernization with rapid software updates, new AI tools

A.I. Flow was initially created as an experimental program that uses Microsoft Azure Government’s open AI structure to assist in various tasks such as contract writing and sorting. The Simulation Modeling Framework facilitates threat simulation and creates code to generate a radio frequency waveform from devices such as software defined radios, Bowes explained. 

The goal behind changing a waveform is to either jam an adversary’s system or become unidentifiable by adversaries. With the A.I. Flow involved, soldiers should be able to talk or text their prompts into the device and it will generate the correct code needed to reprogram the waveforms on their device. 

“The intent behind this is to provide for a deployed soldier an enhanced capability so that they can easily provide an RF [radio frequency] waveform and provide some of that reprogramming at the edge — some of that quick turn,” Bowes said.

“One of the ways to get around and expedite those timelines is to have it done right at the edge. Rather than having them try to develop code while they’re warfighting, it would certainly be much, much easier and much faster to simply say, ‘I need to degrade system X, and I have system Y, provide me a waveform or technique that can counter that, or that can identify that.’”

The plan to create such a tool came after the Army conducted the second phase of its RF Data Pilot earlier this year where soldiers showed the Simulation Modeling Framework creating code for an RF waveform, successfully jamming a digital mobile radio, Bowes said. 

Bowes’s team, along with other EW offices in the service, brainstormed how they could go even further in speeding up the time it takes to alter an RF signal. Bowes said much of this need for speed derives from lessons learned in Russia’s ongoing war against Ukraine. 

“What we’re seeing is, traditionally reprogramming cadence [took] weeks and months, and now we’re looking at hours and days as a result of those lessons learned from Ukraine and Russia, and so it’s really kind of flipping everything on its head in terms of how we do things faster,” he said. 

As far as when soldiers will start seeing the new devices at the edge, Saacks and Bowes said it won’t be for a while as A.I. Flow is still in the process of obtaining an authority to operate — a formal certification that gives a system permission to operate in a sensitive environment, which often takes several months. 

However, the team remains hopeful that by working across multiple offices in the service and by harnessing the Army’s transformation in contact initiative, they can get the tool out as fast as possible, Austin Fox, spokesperson for CECOM told Breaking Defense. 

“It’s all sort of that transformation in contact sort of thing, so we’re perfecting as we progress. But they’re in the very early, early stages, and the intent is as they engage more with other parts of the EW community, to get that more refined and stronger timeline,” he said.