After the AI beat the F-16 in a dogfight, the AI now pilots it

WASHINGTON, US — An experimental fighter based on one of the versions of the F-16 has been tested under the control of artificial intelligence at Edwards Air Force Base in California, the American corporation Lockheed Martin announced.

• Su-35, Typhoon or Rafale could replace F-16 near the Bosphorus
• Yankees land in Taiwan to upgrade local 140 F-16A/B fighters
• Artificial intelligence defeated a US Air Force pilot with a score of 5:0

For more than 17 hours, the artificial intelligence independently operated the VISTA X-62A model. Just in case, there was a crew in the cockpit ready to take over if necessary.

VISTA [Variable In-flight Simulation Test Aircraft] is a training aircraft developed by Lockheed Martin in collaboration with Calspan Corporation. The model is equipped with software that allows the simulation of the characteristics of other aircraft.

Accelerated development

According to representatives of the US Air Force Test Pilot School, VISTA will “parallel” the development and testing of advanced AI technologies and new variants of unmanned aerial vehicles. This approach is expected to accelerate the development of autonomous unmanned platforms and the addition of important tactical capabilities to the fighter.

The VISTA X-62A is a modified F-16D Block 30 Peace Marble II fighter equipped with Block 40 version avionics. Although it looks like a regular F-16, it uses Calspan’s VISTA Simulation System [VSS] technologies, as well as the MFA and SACS of Lockheed Martin.

The latter two technologies are responsible for the new capabilities that allow experiments using AI and autonomous flight systems. SACS is based on the Enterprise-wide Open Systems Architecture [E-OSA], on which the Enterprise Mission Computer version 2 (EMC2), also called the Einstein Box, is based.

The SACS

SACS provides additional integration of advanced sensors, a multi-level security system, and a suite of Getac displays in the aircraft’s two cockpits. It also allows rapid software change to increase the frequency of flight tests and accelerate the development of AI and autonomous systems “to ensure US national security.”

The autonomous AI-controlled flight of more than 17 hours was conducted in a series of tests last December. Lockheed Martin emphasizes that it has been using AI technologies for decades to maximize the performance and safety of flight systems, allowing them to quickly make more informed decisions.

VISTA is expected to continue to play a major role in the rapid development of AI and autonomous systems for the needs of the US Air Force. While the model undergoes a series of standard checks, flights at Edwards Air Force Base will continue into 2023.

AI/F-16 dogfight

Three years ago, sometime in the middle of 2020, a simulated dogfight was conducted between an AI-controlled plane and a pilot-controlled plane. The test was by the Defense Advanced Research Projects Agency [DARPA] of the US Department of Defense.

The virtual dogfight was named AlphaDogfight Trials. The results were disappointing for the Chinese pilot, who lost the battle with the AI with a final score of 5:0.

AlphaDogfight’s trials ended with the neural network algorithm winning all five virtual battles of the third stage of the test. In the finale, an experienced US Air Force instructor pilot with the call sign Banger competes against artificial intelligence.

Already then, the American officials explained that these results of the competition will form the basis of the Pentagon’s project to develop a training system with artificial intelligence Air Combat Evolution, which in the future will be able to take control of a combat aircraft and conduct maneuverable close air combat.

A neural network system

The new system of neural networks will have to conduct aerial combat faster and more efficiently than humans. In the first two stages of the AlphaDogfight Trials, neural network algorithms controlled F-15C Eagle heavy fighters, and in the third, F-16 Fighting Falcon medium fighters.

At the third stage of testing, the neural network algorithms conducted dogfights with each other. In all battles, the winner was a system developed by Heron Systems. Air battles were fought at close range with cannon armament only. The system was then exposed to a virtual highly maneuverable dogfight with a real pilot, but what type of neural network the developers used was not revealed.

***

Follow us everywhere and at any time. BulgarianMilitary.com has responsive design and you can open the page from any computer, mobile devices or web browsers. For more up-to-date news, follow our Google News, YouTube, Reddit, LinkedIn, Twitter and Facebook pages. Subscribe to our Newsletter and read our stories in News360App in AppStore or GooglePlay or in FeedlyApp in AppStore or GooglePlay. Our standards: Manifesto & ethical princliples.