Northrop Grumman is developing software it says can simplify the high-stakes process of discovering, classifying and monitoring missile launches across the globe by leaning on pattern-recognition capabilities.
The defense company is in the process of refining what it calls False Track Reduction Using Machine Learning for the U.S. Space Force, with eyes on delivery in early 2025. It is anticipated for use in the Space-Based Infrared System program, or SBIRS, and has potential application in other overhead persistent infrared assignments.
Space Force personnel track thousands of potential missile incidents each month and must contend with false alarms. Increasingly delicate spying technologies, proliferating satellites, ever-evolving weapons and military flare-ups overseas can aggravate the already-complicated process.
Northrop’s offering is designed to ease the information avalanche analysts face by parsing what may not be an actual launch or outbound projectile while, at the same time, ensuring no “real event or real missile” is improperly sorted, according to John Stengel, the director of the company’s mission exploitation enterprise.
False Track Reduction Using Machine Learning is trained on real-world data and can be amended as foreign militaries advance their respective arsenals. The system uses what Stengel called profiles, or proven characteristics such as speed, shape and altitude, to detect and earmark objects for further inspection by users.
The Department of Defense has for years considered artificial intelligence and machine learning critical to the speedy sorting of battlefield information. Its implementation is gaining speed and spread.
