Air Force must build trust to add drone wingmen, report says

WASHINGTON — The Air Force is all-in on the concept of drone wingmen teamed up with human piloted fighters as the future of air combat.

But how exactly will the human pilots inside F-35s or the Next Generation Air Dominance platform work with and guide collaborative combat aircraft? And will these drone wingmen be ready to fight under their human pilots’ instructions in a future war?

The Air Force needs must do more to lay the groundwork for human-machine interaction, Heather Penney, a former F-16 pilot and senior resident fellow for the Mitchell Institute for Aerospace Studies, wrote in a report.

And if those steps aren’t taken soon, she said, drone wingmen might not be fully trusted by the pilots flying alongside them and the concept of manned-unmanned teaming might not be as effective as it could be.

“Teaming must be built in from the very beginning,” Penney wrote in her paper “Five Imperatives for Developing Collaborative Combat Aircraft for Teaming Operations.” “The effectiveness of CCA in combat will be primarily driven by how well they team with humans, not capabilities such as the weapons and sensors they carry.”

This must include making sure the human pilots can handle the tasks required to direct the autonomous or semi-autonomous wingmen and not get overwhelmed, she said.

Operators also need to be brought into the process of developing drone wingmen early on, and given the tools they need to understand how they should perform in a fight, Penney wrote.

Pilots must be able to depend on the drones and trust their autonomous capability will work as intended, and make sure they can keep control over them even during rapidly shifting, chaotic battles, Penney said. And the teams of human pilots and drone wingmen need to be tailored based on the strengths of each, she wrote.

‘Great potential’

Air Force Secretary Frank Kendall has made creating a fleet of drone wingmen teamed up with the F-35 and Next Generation Air Dominance platform a top priority, and said that a competition to create them is likely to come in 2024.

In her report, Penney said that the CCA concept “has great potential,” and could even give the Air Force a decisive combat advantage in a war against an adversary such as China.

If the Air Force fields them in enough numbers to make a difference, she said, they could increase the service’s combat capability, create a force mix that is able to better weather losses in battle, and give theater commanders a reserve of aircraft that could be used to surge operations.

They could also give the Air Force the ability to stage more complex operations that could overwhelm an enemy’s defenses, Penney wrote. These drones could be used to conduct airstrikes, jam enemy signals, conduct reconnaissance, or serve as decoys to distract enemy forces or lure out their air defenses, and tailored with a mix of sensors, weapons or other equipment to perform different missions.

While the drones will need to have a mix of traditional software and machine-learning algorithms to guide its autonomous capabilities, she wrote, they will still need humans to direct them.

To the drawing board

For that reason, Penney wrote, it’s crucial that the Air Force focus on “human factors engineering,” or how humans interact with the machines they work with.

This means bringing warfighters into the process of developing CCAs as soon as possible, and working alongside software engineers, to shape how they will work in battle, she said.

If that doesn’t happen — and so far, Penney said, it’s largely not — the Air Force runs the risk of developing a CCA design that isn’t as effective in battle as it could be.

“Collaborative combat aircraft are meant to team with a human warfighter, yet, for the most part, warfighters are being left out of CCA development,” Penney wrote.

The Air Force can’t wait until drone wingmen are fielded to flesh out how they will behave because the dynamics of how they will interact with humans have to be part of their programming, algorithm development, and training processes.

It also remains to be seen exactly how pilots will direct their drone wingmen, Penney told reporters. So far, she said, industry is leaning toward voice commands instead of diverting the pilot’s attention by tapping instructions on a tablet or display.

Voice control has its own complications that need to be worked out, she said, such as how to control multiple drones, how well it would work in battlefield environments, and whether the system would clearly understand what the pilot is trying to say.

“Siri still doesn’t understand me, right?” Penney said. She said CCAs will probably need to have multiple methods of control for resiliency, so the pilot could switch to a different control system if necessary.

The human touch

Penney told reporters she doesn’t favor having a second operator, apart from the pilot of a manned fighter, directing or sharing control of that fighter’s drone wingman because that could lead to confusion and contradictory instructions.

“It’s really important that you don’t split leadership of that formation,” Penney told reporters.

Autonomous or semi-autonomous drones could also operate “untethered” to a particular manned aircraft, perhaps as part of a swarm, or as a network of electronic warfare aircraft controlled by an air battle manager. These air battle managers controlling the untethered CCAs could operate from a nearby E-7 Wedgetail or E-8 Joint Surveillance Target Attack Radar System, or JSTARS, she said.

CCAs aren’t going to take the place of human pilots, Penney said, since humans can improvise and adapt to get the job done.

“Humans will always have a cognitive advantage within the battlespace, in terms of making decision through uncertainty, relying on intuition, breaking rules and norms, and being able to use adaptive thinking from past life experience, and applying it in new and different ways,” Penney said. “Autonomy just simply is not going to be able to do that.”