General Atomics Aeronautical Systems Inc. (GA-ASI) said on Jan. 11 that a company-owned MQ-20 Avenger drone advanced the company’s internally funded work on Collaborative Combat Aircraft (CCA) in three autonomous missions last month.
On Dec. 14, the Avenger, paired with “digital twin” planes, conducted simulated combat missions from GA-ASI’s Desert Horizons flight operations facility in El Mirage, Calif., and the company’s Reinforcement Learning (RL) architecture, developed using industry-standard tools, including Docker and Kubernetes, allowed the artifical intelligence/machine learning (AI/ML)-enabled Avenger “to validate three deep learning RL algorithms in an operationally relevant environment,” GA-ASI said.
The company said that its CCA “ecosystem” will provide “a new and innovative tool for next-generation military platforms to make decisions under dynamic and uncertain real-world conditions.”
During the Dec. 14 Avenger missions, the “single agent RL model successfully navigated the live plane while dynamically avoiding threats to accomplish its mission,” GA-ASI said on Jan. 11. “Multi-agent RL models flew a live and virtual Avenger to collaboratively chase a target while avoiding threats. The hierarchical RL agent used sensor information to select courses of action based on its understanding of the world state…This live operational data describing AI pilot performance will be fed into GA-ASI’s rapid retaining process for analysis and used to refine future agent performance.”
The company said that the Dec. 14 tests used a government-furnished Collaborative Operations in Denied Environment (CODE) autonomy engine, government-standard Open Mission System messaging for linking the RL agents, and a General Dynamics
[GD] EMC2 open-architecture, multi-function, multi-level security processor–the so-called “Einstein box”–to speed testing for various missions.
GA-ASI has said that it conducted a manned-unmanned teaming test to sense airborne targets in the infrared spectrum last Nov. 17–a test that featured a company MQ-20 Avenger drone, a Lockheed Martin [LMT]-operated Sabreliner acting as a fighter, and two F-5 Advanced Tiger fighters from Nevada-based Tactical Air Support, Inc.
Last September, General Atomics said that a company-owned MQ-20A flew an AI-enabled fully autonomous test flight for nearly 30 minutes as part of a “cooperative live, virtual, and constructive UAS swarm” (Defense Daily, Sept. 22, 2022).
The company had pitched the MQ-20 Avenger, known as the Predator C, as an option for the Navy’s former Unmanned Carrier-launched Airborne Surveillance and Strike (UCLASS) program, dubbed Sea Avenger.
When the Navy shifted from UCLASS to an aircraft carrier-based tanker, GA participated in the MQ-25 Stingray competition, bidding with an aircraft similar to the MQ-20 Sea Avenger, but larger. The Navy ultimately chose Boeing’s [BA] design for the MQ-25A tanker.
Last year, Air Combat Command (ACC) said that increasing fighter capacity through such programs as CCA is a top priority for U.S. Air Force Gen. Mark Kelly, the commander of ACC (Defense Daily, Nov. 15, 2022). Kelly has said that he envisions a future fleet of multi-role CCAs and that maintainers would be able to switch out CCA electronics and weapons regularly, based on force needs.
“If I was doing a clean sheet design, I would look for something you could iterate,” Kelly said last September. “You could bolt on growth. You could get an order one day going, ‘Hey, these guys and gals at Nellis [AFB, Nev.] want an eight-ship, and they want them all to be X-Band radars.’ So, let’s shift to X-Band radars. Unlock a nose, bolt on another nose. The next day…they want all jammers. So really quickly, let’s take off the X-band radars and put on the jammers.”
“I would not lock myself into, ‘It’s a sensor and can’t do anything else, or it’s a jammer and can’t do anything else, or it’s not armed, and you’ll never arm it,’” he said.
Kelly said in September that he believes in the next two to three years the Air Force will conduct operational testing of several CCAs out of an Air Force location or locations that can launch drones–Creech AFB, Nev., Tonopah Test Range, Nev., Holloman AFB, N.M., and/or Tyndall AFB, Fla.
Air Force Secretary Frank Kendall has said that his top technology priority is “autonomous behaviors and artificial intelligence decision support,” and he has suggested that the Next Generation Air Dominance (NGAD) manned fighter could serve as a play caller for autonomous CCAs (Defense Daily, Aug. 11, 2022).
For CCA, the Air Force may glean lessons learned from manned-unmanned teaming efforts, such as Skyborg, and may take a step toward equipping fighter squadrons with some initial stage CCAs in the upcoming fiscal 2024 budget request.
The Air Force has been developing an acquisition strategy and timeline for CCA, up to five of which may be employed by each “quarterback,” a manned Next Generation Air Dominance fighter, which Kendall said has entered the engineering and manufacturing development stage. Air Force acquisition chief Andrew Hunter has said that he envisions more than one supplier of CCA, rather than a “winner take all strategy.”