Lockheed Martin Skunk Works and the U.S. Air Force Test Pilot School have completed a series of AI-controlled air intercepts using the X-62 VISTA test aircraft at Edwards Air Force Base, California. In eight flights conducted in April, the system carried out 27 autonomous intercepts against a T-38 Talon target aircraft using live infrared sensor data.
The demonstration used the Legion Pod infrared search-and-track sensor to feed real-world target information to the autonomy system, which then manoeuvred the aircraft into intercept positions without direct human piloting. The campaign was designed to show that the X-62’s autonomy architecture can work from operational sensor inputs rather than simulated data, a step toward sensor-to-action autonomy in combat aircraft.
The result is relevant to future autonomous aircraft and collaborative combat aircraft programmes, where live mission-system testing is essential to validate behaviour in realistic operating conditions. Earlier work in this area relied more heavily on simulated sensor inputs, making the April flights a more demanding test of autonomy performance in flight.