Artificial intelligence developer Shield AI has notched another milestone in its campaign to develop autonomous flight, for the first time simultaneously controlling multiple aircraft.
The San Diego-based company on 21 August said it recently completed flight tests involving two unmanned Kratos MQM-178 Firejet target drones, which were each piloted by Shield AI’s “Hivemind” autonomy agent.
The two jets performed coordinated manoeuvres and demonstrated “operationally relevant collaborative behaviours”, according to Shield AI.
“This demonstration highlighted the capabilities of future unmanned jets and drones, with all computing executed on board and in-air agent communication seamlessly integrated,” the company says.
Two manned jets – an Aero Vodochody L-39 and L-29 – were used as observation and chase aircraft during the tests.
Shield AI is best known as the company whose algorithm controlled a specially modified Lockheed Martin F-16 fighter during recent dogfighting drills with the US Air Force (USAF) and Defense Advanced Research Projects Agency (DARPA) technology incubator.
Those tests saw Hivemind take control of the X-62 Variable In-Flight Stability Test Aircraft and engage conventionally piloted fighters during within-visual-range combat manoeuvres – better known as dogfighting.
While DARPA has declined to reveal many details about the man-against-machine dogfighting drills, the secretive agency has described the effort as a success, with Shield AI’s autonomy agent safely executing at least 21 sorties over 10 months.
The latest tests mark the second time Hivemind has taken control of a Firejet, building on flights involving a single MQM-178 in March. The Kratos type is the second jet-powered and sixth overall aircraft Hivemind has controlled.
While the technology achieved a new milestone in controlling two aircraft simultaneously, Shield AI chief executive Brandon Tseng says that number is only limited by test budgetary constraints.
“We limited the number of Firejets to two, but it could have been four eight, 16, 32 and so on,” he notes.
Notably, the Hivemind technology has proven capable of performing complex, dynamic tasks like within-visual-range fighter manoeuvres, not just simple pre-programmed flight profiles.
“It will improvise and come up with novel tactics,” Tseng told FlightGlobal in April at the annual Special Operations Forces Week conference in Tampa, Florida.
The former naval special warfare commando predicts such technologies will enable the USA and allies to deploy “millions of intelligent, resilient drones” within the next decade.
The Pentagon is in the midst of a major autonomy push, as it seeks to address recruiting challenges and China’s numerical advantage in ships and missiles in the Western Pacific.
Uncrewed aircraft developer Kratos is hoping to capitalise on that initiative, after being passed over for the initial round of contracts to develop a so-called collaborative combat aircraft for the air force.
The company’s XQ-58 Valkyrie was an early leader in the autonomous jet space, with multiple examples purchased by the USAF and US Marine Corps for testing and evaluation.
Shield AI has previously said it is working with Kratos to adapt Hivemind for use with the XQ-58.
