The Republic of Korea Air Force has accepted an AI-powered tactical development and flight simulation system — the first in the country to leverage reinforcement learning (RL) technology.
Reinforcement learning lets the system teach itself through trial and error, earning “rewards” for good results while gradually finding the smartest approach to meet the user’s objectives.
Developed by Korea Aerospace Industries (KAI), the trainer uses a virtual aircraft fed with real-world tactical data, allowing it to develop strategies without human intervention.
Training for KF-16, FA-50 Pilots
The 35.5 billion South Korean won ($25.6 million) equipment runs with two of the country’s supersonic airframes: the KF-16, a local variant of the American-made F-16 Fighting Falcon, and the FA-50, the combat version of KAI’s T-50 Golden Eagle.
Combat scenarios are played out by the platform through virtual and mixed reality, filling gaps in real-world training while avoiding airspace and noise restrictions.
The simulator also projects digital copies of friendly and enemy jets, generates detailed environment design, and even responds to voice commands.
KAI said its AI/RL flight trainer is the first in the world to combine all these capabilities in a single system.
In a statement, the Sacheon-si-based company added, “We will lead the development of core technologies such as next-generation modeling and simulation systems and live virtual constructive platforms to spearhead the military’s future power build-up.”
To date, the Republic of Korea Air Force operates approximately 118 KF-16 multirole fighters and 60 FA-50 light combat aircraft, according to publicly available sources.
