The US military recently undertook multidomain testing of its Scarlet Dragon platform, testing the platform’s machine learning capabilities in target identification.
The US XVIII Airborne Corp recently undertook multidomain testing of its Scarlet Dragon target identification platform, alongside the US Navy, Air Force and Marine Corps.
As part of the exercise, the Army used the platform’s machine learning capabilities to search 7,200 square kilometres of satellite imagery across US eastern seaboard to identify simulated targets.
The exercise then tested the interoperability and command-and-control integrations between the domains and F-35s providing air support once the targets were identified.
According to reports from the US-based Army Times, some targets were the “size of a 10-square-foot box”.
The platform utilises machine learning from Project Maven, which hit headlines in 2019 when Google departed the project after a series of concerns.
Writing for the Modern War Institute, Richard Schultz and General Richard Clarke explain the importance of Project Maven for target identification.
“Designated Project Maven, this effort’s initial objective is to automate the processing, exploitation and dissemination of massive amounts of full-motion video collected by intelligence, surveillance and reconnaissance (ISR) assets in operational areas around the globe,” the pair noted.
The exercise was the fourth iteration of testing, with the first Scarlet Dragon test fire being conducted in 2020.
Scarlet Dragon was conducted as NATO finalised Exercise Dynamic Monarch in September, the 11th iteration of NATO-sponsored live Submarine Escape and Rescue (SMER) exercises.
[Related: Defence targets enhanced cyber capability]