U.S. Air Force Koby Larsen, 16th Electronic Warfare Squadron B-1 Electronic Warfare specialist, analyzes radio frequency signals at the B-1 Lab at Eglin Air Force Base, Fla., Jan. 11, 2024. The 16th EWS provides EW support to U.S. Air Force Global Strike Command assets. (U.S. Air Force/Benjamin Aronson)

WASHINGTON — From Ukraine to Taiwan, the US is already waging an invisible battle of electrons, where nations nominally at peace constantly scan each others’ transmissions in search of threats and weak points. But modern radios and radars can switch frequencies as fast as new software can be written — and longstanding, cumbersome Pentagon processes can’t update systems fast enough to keep up.

So, last fall, the commander of the Air Force’s 350th Spectrum Warfare Wing — a rapidly growing unit less than three years old — made an ambitious promise to reporters at the annual Air Force Association conference: “We updated F-15s and B-52s on a quarterly schedule,” Col. Josh Koslov said. “That’s not good enough… Our goal, our moonshot, is three hours.”

Now, just nine months later, Koslov says his wing is actually making good on that 180-minute “moonshot.”

“In our wing, we just executed our first internal exercise — we called it Exercise Rapid Raven” to test the streamlined update process, Koslov told a C4ISRnet webinar on Wednesday. “An AFI [Air Force Instruction] said we need to be able to do this in three hours — that’s based off joint force commander requirements — and quite frankly, we’re very close to that, if not exceeding [it], in most of the systems that we cover.”

“More than half our systems are underneath or exceeding the deadline,” Koslov said. Doing the math, that means of some 70-odd systems across the Air Force that the 350th Wing supports, from short-range jamming pods to long-range radars, at least 35 can get updated within three hours of the 350th getting intelligence on a new threat.

Koslov didn’t specify which systems met which timelines, for obvious reasons of security. But it’s likely that the 350th has made the most progress with the most modern electronics, since they tend to have more spare computing power, more bandwidth to receive new data and software patches wirelessly (“over the air”), and above all more “open” architectures that can easily adapt to new software. Some longer-serving systems use decades-old electronics with bespoke software or data formats that require extensive custom work to update, and cramming all that effort into three hours would be much harder.

“It will always depend on the system,” Koslov acknowledged.

AI At The Tactical Edge

How can the 350th Spectrum Wing update electronic warfare systems so much faster than before? Even more important, how can other military branches replicate the Air Force unit’s success?

The solution is not just new technology, but also streamlined bureaucratic processes, Koslov and other EW experts said at the C4ISRnet event. Sometimes it’s possible to get a major speed boost simply by relaxing the rules on what data is classified at what level, making it easier to share with both human analysts and machine-learning systems.

“We spend an inordinate amount of time going through ‘cross-domain’ systems” that control the movement of information from one level of classification to another,” said Col. Gary Brock, the capability manager for electromagnetic warfare at Army Cyber Command. “What we don’t really have is data classification guidance that guides that data all the way from the tactical edge and all the way back.”

Only once the central hub has solid data, and lots of it, can it feed that data to AI. But “just add AI” is not an easy button to speed up EW updates.

For years, the high-tech holy grail has been a concept called “cognitive electronic warfare”: an AI compact enough to fit on a fighter jet and smart enough to detect unfamiliar signals, analyze them on the spot, and automatically update the plane’s EW systems to counter the new threat, near-instantly and on the spot — without having to wait for a central hub to process data and send an update back. Despite dramatic advances in AI, however, that grail remains elusive.

Training and retraining cutting-edge algorithms requires tremendous computing power, making it a matter for warehouses of server racks, rather than streamlined pods under an airplane’s wing. So while it is increasingly possible for limited “edge AI” to do some preliminary sorting of the raw data, meaning the frontline platform doesn’t have to send back as much to the central analytic hub, the hub remains essential.

What’s more, even with cognitive EW aboard, a single EW pod might not pick up enough instances of an unfamiliar signal to analyze. So the military would still need a way to share data from sensors all over the globe — what Koslov calls “crowdsourcing” EW data.

“Cognitive EW’s a real thing. But the idea that there’s boxes out on the edge that are going to be making decisions and executing is still a few years away,” Koslov said. For the near term, he went on, “the right answer is transporting that information to Eglin [Air Force Base] to have the biggest collection of folks working on it.”

“We have a very similar approach…out of our EW community down at [Fort] Eisenhower,” said Brig. Gen. Ed Barker, the Army’s Program Executive Officer for Intelligence, Electronic Warfare and Sensors (PEO-IEWS). “Our Army Reprogramming Enterprise [can] make adjustments at the edge. based on changes in the threat.” The Army’s has even gotten Pentagon funding for an “EW data pilot” studying how to modernize its “transport” layer carrying raw data from frontline sensors back to the central analytic hub.

The trick is running this cycle as fast as possible, from frontline forces detecting an unfamiliar signal, to the central hub analyzing it and creating an update, to the frontline receiving that update.

“The spectrum [is] an area of perpetual novelty,” Koslov said. “Our ability to detect something new and novel, and create new combat capability against that signal, is something that we really are focusing on and investing in quite a bit.”

Speed is of “the utmost importance,” agreed Barker. “We’re having to iterate every 24 hours based on changes in adversary TTPs [Tactics, Techniques, & Procedures].”

This approach requires high-bandwidth connections between the frontline forces and the central EW hubs. That’s hard enough to keep up in peacetime, with US units all over the world. It could be harder yet in combat, when a savvy adversary will attack communications systems with its own jammers, cyber warfare, or precision-guided explosives. The Chinese military calls such attacks “systems destruction warfare,” intended to defeat the US military not one ship, plane, or tank at a time, but by blinding its sensors and paralyzing its communications.

“If we lose in the spectrum… the joint force will lose and we’ll lose very quickly,” Koslov said. But, he said, “What I’m really encouraged by, across the DoD, is a resurgence of a focus, an organizational focus, on electronic warfare.”

That focus, and America’s decades of combat experience, gives Koslov confidence in the face of a rapidly growing and modernizing People’s Liberation Army. “The PLA talks a good game but they don’t have the luxury that we have of being a joint force that knows how to fight together,” he said. “We fight connected and integrated all of the time, even absent some of our links. China has never demonstrated that capability.”