By Jim Morris, Warrior Vice President, News
Lockheed Martin is teaming up with the Defense Advanced Research Projects Agency (DARPA) for the next stage of an artificial intelligence program involving drones.
Lockheed was awarded a $4.6 million contract to develop AI tools for dynamic airborne missions as part of the agency’s Artificial Intelligence Reinforcements (AIR) program. Over an 18-month period, the company will use AI and machine learning techniques to build surrogate models of aircraft, sensors, electronic warfare and weapons. The goal is to provide advanced modeling and simulation approaches for live, multi-ship beyond- visual-range missions.
AIR was unveiled in November, 2022 and is scheduled to last four years. Its mission is to build tools for achieving what DARPA calls “dominant tactical autonomy” – in other words, for US drones to have the upper hand in aerial combat. The plan is for tools to be developed and tested on “human-on-the-loop” F-16s and then eventually moved to drones.
A DARPA notice to developers said AIR would address two technical areas: creating fast and accurate models that capture uncertainty and automatically improve with more data, and developing AI-driven algorithmic approaches which enable real-time distributed autonomous tactical execution within uncertain, dynamic and complex operational environments.
Army Pioneers New Commercial Classified Computer Network
Lockheed will use its ARISE system for the DARPA project. The company describes ARISE as a family of integrated toolkits that are used to build a weapon simulation tool. Lockheed says it allows developers to build a fight simulation in two to three months, instead of two years as in the past.
“This will provide significant cost savings opportunities for the Department of Defense and serve as a foundation for future AI defense solutions,” said Gaylia Campbell, vice president of engineering and technology for Lockheed’s Missiles and Fire Control unit.
DARPA is engaged in a number of projects involving artificial intelligence – in fact, the agency says AI, machine learning and autonomy are being used in about 70 percent of its programs. One of the best known is the Air Combat Evolution (ACE) program, a joint project with the Air Force.
In April, the agency said that a test plane – dubbed the X-62A VISTA – had become the first autonomous fighter jet ever to successfully engage in aerial dogfighting maneuvers against another fighter – in this case, an F-16.
ACE is focusing on how to give an AI-controlled aircraft the capability to take part in within-visual range combat.
“The X-62A team demonstrated that cutting-edge machine learning-based autonomy could be safely used to fly dynamic combat maneuvers,” said Air Force Secretary Frank Kendall.
In May, Kendall took a hands-off flight in the X-62A, sitting in the front seat with a “safety pilot” in the back. The controls of the plane remained untouched during the entire flight.
In all of these programs, DARPA is trying to develop systems that warfighters can trust implicitly. Arms control experts have raised the issue of that AI might one day be able to drop bombs without any humans being in the loop. Kendall, among others, has said that humans will always be consulted whenever weapons are used.
