AI Pilot Could One Day Steer Aircraft In Crowded Skies On Its Own

Researchers at Carnegie Mellon University have announced an exciting innovation in aviation: possibly the first artificial intelligence (AI) pilot that could enable autonomous aircraft to navigate a crowded airspace all on its own. 

According to the team, the AI program will allow the aircraft to safely avoid collisions, predict the intent of other aircraft in its way, track and coordinate with planes in the air, and even communicate over the radio with air traffic controllers back at the tower.

In fact, the scientists hope to be able to develop the AI even further, so that its behaviors become indistinguishable from that of a human pilot.

The team believes that the system could eventually pass the Turing Test, which examines a program’s ability to exhibit human-level intelligent behavior. 

The system makes use of a six-camera vision inputs and natural language to interact with other aircraft in the area, thereby enabling it to identify different planes and interpret messages from air traffic controllers to navigate the skies safely. 

To achieve this, the researchers trained the AI on data collected at the Allegheny County Airport and the Pittsburgh-Butler Regional Airport, including air traffic patterns, images of other aircraft, and radio transmissions. 

While autopilot controls have long existed on commercial airliners, the AI used isn’t equipped to handle lower-altitude traffic operating under visual flight rules (VFR), which the team has now cracked with its latest innovation. 

“This is the first AI pilot that works in the current airspace. I don’t see that airspace changing for UAVs (unmanned aerial vehicles). The UAVs will have to change for the airspace,” explained Sebastian Scherer, Associate Research Professor. 

Going forward, the scientists hope to trial the AI on an actual aircraft instead of just flight simulators, and could eventually implement the program on the autonomous delivery drones and air taxis of the future.

[via New Atlas and Carnegie Mellon University, cover image via Carnegie Mellon University]