The Defense Advanced Research Projects Agency (DARPA) is halfway through the first phase of a program that conducts human-machine collaborative dogfighting to advance the development of a scalable, trusted artificial intelligence-driven autonomy for air combat.
DARPA said Thursday that the Air Combat Evolution program has made several achievements including live flights of an L-29 jet trainer to measure a pilot’s trust in AI and virtual AI dogfights within and beyond visual range multi-aircraft scenarios.
“Our biggest focus at the end of Phase 1 is on the simulation-to-real transition of the AI algorithms as we prepare for live-fly sub-scale aircraft scenarios in late 2021,” said Col. Dan “Animal” Javorsek, program manager at DARPA’s Strategic Technology Office. “Managing this transition to the real world is a critical test for most AI algorithms.”
In Feb. 2020, the program’s algorithm development teams introduced a missile and a gun for longer- and shorter-range targets into the first AI scrimmage to demonstrate simulated 2-v-1 aerial dogfights at Johns Hopkins University Applied Physics Laboratory.
In August 2020, APL conducted the AlphaDogfight Trials in which an AI ‘pilot’ on an F-16 Viper jet came out undefeated against a human Air Force fighter pilot in five rounds of simulated air combat.
“Adding more weapon options and multiple aircraft introduces a lot of the dynamics that we were unable to push and explore in the AlphaDogfight Trials,” Javorsek said. “These new engagements represent an important step in building trust in the algorithms since they allow us to assess how the AI agents handle clear avenue of fire restrictions set up to prevent fratricide.”