Autonomous Systems Took Center Stage At AFA

Walking into the Air and Space Forces Association’s Air, Space & Cyber this year, conference goers were immediately greeted with a large display from Google Cloud—a contender for the Pentagon’s major tactical cloud program and one of the conference’s major sponsors. On the floor, SAIC’s tower was right behind the Air Force Research Laboratory's booth, like a beacon of would-be connectivity. 

The exhibit hall seemed split in two—with one side dedicated to traditional military hardware like helicopters and plane engines, and the other focused on digital services. General Atomics showcased its model-sized MQ-9 drones, while companies like Blue Halo displayed their small unmanned systems and counter-UAS capabilities. Anduril, in a break from the crowd, brought its version of a mobile command and control post. 

And after three days of panels and speeches, there was one big tech takeaway: autonomous aircraft are coming en masse. Or at least that’s the plan. 

Air Force Secretary Frank Kendall has detailed what he calls operational imperatives: key areas the service must focus on. And while autonomy isn’t specifically listed, Kendall has made it clear that autonomy, along with other technologies, is needed to meet those imperatives. 

Humans will continue to decide when to use lethal force, but machines can fly the aircraft and execute other tasks, Lt. Gen. Clinton Hinote, the Air Force’s deputy chief of staff for strategy, integration, and requirements, said during a panel on unmanned systems. 

“Clearly autonomy in flight is going to revolutionize flight. That will not only be true for combat, it's likely to be true for a lot of different areas across our economy and across our country. And I'm really excited about the democratization” of it, Hinote said. “But clearly we do not need to have humans do the flying of these aircraft.”

For years, the Air Force has been working toward “Next Generation Air Dominance,” or NGAD, a system-of-systems concept that will include several aircraft, including a sixth-generation fighter jet. Part of the NGAD concept is a “Collaborative Combat Aircraft,” which has previously been referred to as “loyal wingman” and would involve manned and unmanned platforms working together. The Air Force wants to have a “meaningful” CCA platform fielded as soon as possible, Andrew Hunter, the Air Force’s top weapons buyer, said. 

The military and industry seem to agree on the way forward when it comes to autonomous aircraft. So why has it taken so long if the technology already exists? Why isn’t the Air Force closer to having the future-ready collaborative combat aircraft it says it needs so urgently?

Because it’s taken until 2022 for key elements to align, from the maturation of advanced manufacturing to dedicated funding priorities. 

“We understand in this next fight, in a highly contested engagement with our peer adversaries, China and Russia, we don't have enough mass, we don't have enough aircraft, we don't have enough, you know, widgets to put up in the sky,” said Andrew Van Timmeren, the vice president of government services at Blue Force Technologies. “And CCAs, with the confluence of these technologies, bring increased capabilities and [a] lower price. All these technologies have advanced to a point where here in 2022, we can finally put them together.”

Lockheed Martin’s Skunk Works announced its manned-unmanned teaming effort, Project Carrera—a clear answer to the Air Force’s CCA call—ahead of the conference. As the conference kicked off, General Atomics announced it will open two new centers: one secure facility and one additive manufacturing facility, dedicated to UAS. 

For the Air Force, the technology may have existed, but it wasn’t ready for what the service needed. Hunter told reporters that successful experiments have not translated into operational capabilities. 

“We have crude platforms that have been very successful going back decades. The challenge is that ability to team with manned aircraft. There have been some early successful experiments. We're doing that, but they're experiments. They're not operationally robust capabilities to date,” he said. 

Autonomy and CCA will likely be a “part of the family systems it needs to be able to assist the mission of the NGAD system,” which means that autonomy could be used with weapons platforms while working with piloted aircraft, Andrew Hunter, the Air Force’s top weapons buyer, told reporters. 

The challenge is getting the manned-unmanned teaming right, scaling, and making sure the platform can operate in denied airspace—all things CCA aims to address. 

The Air Force doesn’t have a timeline for the aircraft, but it should be delivered “soon,” Hunter said. “The first thing is to field something meaningful in the next several years to meet the threat,” working with industry to identify “the most effective mix of vehicles and mission systems.”

For industry, it’s a matter of funding prioritization and leadership buy-in.   

“Why isn't it already here, like, en masse? I don't know that we've really had the prioritization behind it,” Van Timmeren told Defense One.

But advances in autonomy capabilities and advanced manufacturing, alongside “government architectures, which enable swapping of these capabilities—all that has finally come together” in tandem with endorsements from the Air Force secretary and chief of staff on down, he said. 

“They've seen the briefs, and they've gotten the updates, and they see all of these disparate parts that finally arrived at a maturity level, where we can finally bring them all together to an integrated system,” he said. 

Competition with Russia and China have also highlighted that the U.S. doesn’t have enough unmanned systems. 

“We understand in this next fight, in a highly contested engagement with our peer adversaries, China and Russia, we don't have enough mass…we don't have enough aircraft, we don't have enough, you know, widgets to put up in the sky. And CCAs with the confluence of these technologies, bring increased capabilities and lower price,” Van Timmeren said. But it’s taken until 2022 to start putting it all together to make meaningful advancements.

“We're never going to have enough F-35s, we're never gonna have enough F-22s. We're not building anymore. And those are extremely exquisite, and expensive, expensive platforms. So we can enable survivability and we can enable combat capability if we use these advanced unmanned systems, while keeping the price point low.”

There’s also a complicated cultural problem, Mike Benitez, Shield AI’s product director, told Defense One. 

The process gets choked in the budgeting cycle, but there also hasn’t been a focus on autonomy research, he said. 

Of the Pentagon’s hundreds of AI programs, only a handful are doing advanced autonomy, and their budgets reflect it: “The cost of an engine for a fighter jet is probably more than all those programs put together,” Benitez said. 

“We're past a proof of concept. The technology exists, we've demonstrated it can be done. We just need the government to make the next move to actually get serious about it. If you want to get serious about a program, you’ve got to fund it seriously.” 

But making collaborative combat aircraft a reality will take more than money and experimentation. Human operators will also need to learn and trust the technology.

Maj. Gen. Evan Dertien, commander of the Air Force Test Center, told reporters that performing autonomy testing has a different set of needs. 

“We need different telemetry systems, we need different airspaces, we need different ways of testing to what speed, and we need ways of how we're going to actually develop the test plans for non-deterministic events,” he said, noting that the Air Force will soon host a week-long summit dedicated to autonomy to outline the different challenges. 

The service may also need a dedicated training pipeline for CCA pilots as well, Benitez said. 

“So it's not just how to use it, you got to know how it's engineered to an extent so you understand what the limitations are of your weapon systems. And we do that with sensors. We do that with flight controls, with engines,” Benitez said. “And so that same type of knowledge base, that general knowledge has to be implemented across the end user base, and right now there's just a chasm in between the two.”

Shrinking that gap means finding places to train with autonomous systems, which have federal airspace restrictions, and could mean taking things virtual. 

“They're going to need a place to go train,” Benitez said. “That's going to be required to actually make some of these real. Just like an MQ-9, which isn't autonomous, it's remotely piloted, but even then it has so many restrictions on where you can and can't use them in the United States. They have very, very narrow approved [Federal Aviation Administration] corridors to go back and forth from.”

But there could be an opportunity to introduce autonomous architecture into simulations so pilots can train on the exact systems they’d use in the air. Benitez said Shield AI is working with the Air Force on that concept to plug in some of the essential autonomy into simulators to speed up the training process. 

“Like, let's skip the aircraft. Let's put them in the sim. Let's make the humans smarter,” he said, while “developing trust and providing value to the end user before we even have one of these CCAs even in the air.” 

However, the need for aircraft is still there, and an increased focus on CCAs from the military and industry could signal exponential growth in unmanned capabilities as the technology matures in the next few years. 

“I think we're gonna see over the next two to five years, an explosion of unmanned aircraft in the Air Force, which is going to make everybody better,” Van Timmeren said. 

“So I imagine at some point here in the spring, we'll see the fiscal year [2024] president's budget, we'll probably see prioritization of CCAs in that. It is going to earnestly drive the conversation forward” as companies make their own investments in CCA development to meet the Air Force’s call. 

On the technical side, one thing to watch will be using autonomous platforms adversarially with the Air Force Research Laboratory’s Bandit program. Blue Force Technologies landed that contract and is building unmanned aircraft with open architecture.

“What's important to know is that if we're building an unmanned fighter for adversary air purposes, that same vehicle has unmanned fighter characteristics for blue air purposes,” Van Timmeren said. 

“The confluence of technologies, you know, advanced manufacturing, autonomy, and miniaturization of mission systems, it's all coming together. And we just gotta put it together into an actual aircraft.”