One of the big buzzwords in Washington today is the third offset strategy, championed by Deputy Secretary of Defense Bob Work. Work has described this effort as, in its simplest terms, leveraging advancements in machine learning and artificial intelligence to engender greater man-machine teaming.

The Loyal Wingman effort by the Air Force Research Laboratory (AFRL) fits neatly within this construct. Officially termed autonomous reliable teammate technologies (ART2), this effort seeks to "at the highest levels … develop and demonstrate the technologies that would allow an unmanned vehicle to fly as a teammate with a manned vehicle" with the unmanned vehicle knowing "how to balance the use of its resources to perform the mission that it has been given and be able to respond to things in the environment that weren't planned" such as changes in weather, Kris Kearns, AFRL autonomy portfolio lead and senior adviser for autonomy research, told C4ISRNET in an interview.

Spurred by the AFRL commander's identification three years ago of three game-changing technology areas — hypersonics, directed energy and autonomy — and direction to develop initiatives to advance them to realize their benefits within autonomy, Kearns said AFRL looked at where there could be opportunities to develop a capability for the Air Force that would build on technologies that were being worked on but not necessarily doing a thorough enough job of integration given the different technology focus areas across different organizations.

"As we look through that we kind of identified this idea of taking an unmanned vehicle and giving it the capability to fly, and fly a mission that it's been prescribed to fly. Doing in conjunction safely with a manned aircraft was a nice area to integrate our technologies and demonstrate capability of what autonomy and autonomy technologies could do," Kearns said.

Previous concepts for the Loyal Wingman program have been described by many, including Work, as taking an unmanned fourth-generation fighter aircraft such as an F-16 and pairing it with a fifth-generation fighter, such as an F-22 or F-35, to become, according to Work, a battle network node, operating together.

One of the concepts surrounding this effort is that the unmanned system can serve as a weapons truck for the manned aircraft given these new planes have limited magazine capacities.

While AFRL is merely developing the underlying technology to enable this, leaving the concepts to the Air Force's planning and doctrine departments, the weapons truck model is being used for demonstration purposes.

"We are currently going to target it as a weapons truck because in order to demonstrate the behaviors, we have to give it a scenario that it needs to fly and evolve from that what are the behaviors does it need to have," Kearns said. "In the end, what we want to do is show the Air Force how this system could behave, how we can test it, how we can teach it to fly and then allow the Air Force to decide: Do you want to use a current fourth-gen platform and unman it? Do you want to build a new platform, as well as what really do you want that mission to be?"

Kearns described that the program is really trying to develop autonomy algorithms to "build up behaviors in an air platform."

"What we envision that we'll build up is … what we call a line replacement unit or an LRU. The reasons that we want to do that is because we think that these algorithms, these behaviors, we should be able to do them and they are independent of the air platform," she said. "When I said we're going to build this LRU, that's where we will program in the behaviors. That LRU [is] a box that we should be able to plug into any aircraft. We need to teach it how to respond to the flight controls and how to interact with the flight controls and the flight safety systems, but for the most part, it should be able to — if we take it out of one system and teach it how to do a weapons delivery mission — it should be able to reason about how to do that for another vehicle."

Regarding funding, Kearns said the effort is supported through the AFRL budget, with the first integration program being planned out right now. Referencing the aforementioned three game-changing technology areas — hypersonics, directed energy and autonomy — Maj. Gen. Robert McMurry said at the Air Force Association's annual conference in September that the technology area is "not fully funded."

"That we are committed to getting fully funded this year is a demonstration on autonomy," McMurry said.

McMurry told C4ISRNET at the conference that the Loyal Wingman program is one of these autonomy programs of the future. But he noted they need buy-ins from the combatant commands and major component commands before they can go forward.

Kearns expressed a bit of skepticism regarding the ability to  transfer this to any aircraft, especially existing unmanned platforms such as the MQ-9 Reaper. Reapers are significantly slower and less maneuverable than F-22s, F-35s and F-16s.

MQ-9s and F-16s "are incredibly different aircraft," she said. "I don't know that it would be that transferable but we could think about it being useful on any class of aircraft that flies about or has a flight profile of something like a fighter airplane."

"The decision timeline that [an aircraft] has to figure out how to respond to an unplanned event would be very different based on the speed of the airplanes flying," she continued. "So it's more about driving the timelines that the system has to respond and do things as much as anything else."

In terms of actual capabilities, Kearns described demonstrations that took place utilizing the Air Force's test pilot school. Roughly two years ago, using an unmanned F-16 with a pilot in the cockpit, the aircraft using algorithms took various formations behind the lead F-16, broke off from formation to conduct a separate mission and returned to the piloted F-16.

Kearns pushed back against the notion that the pilots of manned aircraft would be able to control unmanned counterparts — or their unmanned wingman.

"I typically think about it as the manned aircraft would be able to command its other airplanes. I try to avoid calling it control it because we want the pilot in the manned vehicle to fly his plane and command the rest of the aircraft in the space with him but not have to fly his plane and fly another plane too," she said.

Ultimately, Kearns said this program seeks to lessen the workload of the pilot in the cockpit. "In the end, what we’re really trying to do is we’re trying to provide that pilot and his aircraft more capability and extended range [allowing] that pilot being able to command, being able to have more capabilities than what he has on his own platform, and being able to use those to better prosecute the mission that he’s been given is really how we see" how this technology will go into operations, she said.

AFRL is looking to conduct a full demonstration in the 2022 time frame. This technology, Kearns conceded, probably will not be ready for fielding for the next five to 10 years after the 2022 demonstrations "because we’re going to program in these behaviors, there’s still going to be a lot of work to develop additional responses and additional behaviors in that system even after we’re done."