A team of researchers at Carnegie Mellon University in Pittsburgh, Pennsylvania, believe they have developed the first AI pilot that enables autonomous aircraft to navigate a crowded airspace.
The artificial intelligence can safely avoid collisions, predict the intent of other aircraft, track aircraft and coordinate with their actions, and communicate over the radio with pilots and air traffic controllers, according to the researchers.
The researchers say their aim is to develop the AI pilot so its behaviors will be indistinguishable from a human pilot.
“We believe we could eventually pass the Turing Test,” said Jean Oh, an associate research professor at CMU’s Robotics Institute. The Turing Test refers to an AI’s ability to exhibit intelligent behavior equivalent to a human.
To interact with other aircraft as a human pilot would, the AI uses both vision and language to communicate its intent with other aircraft, both piloted and unmanned. Researchers achieved this by training the AI on data collected at Allegheny County Airport (KAGC) and Pittsburgh-Butler Regional Airport (KBTP), including air traffic patterns, images of aircraft, and radio transmissions.
The AI uses six cameras and a computer vision system to detect nearby aircraft in a manner similar to that of a human pilot, according to the researchers. Its automatic speech recognition function uses natural language processing techniques to understand incoming radio messages and communicate with pilots and air traffic controllers using speech, they explained.
“This is the first AI pilot that works in the current airspace,” said Sebastian Scherer, an associate research professor. “I don’t see that airspace changing for unmanned aerial vehicles (UAV). The UAVs will have to change for the airspace.”
More and more drones and autonomous aircraft are expected to fly in the National Airspace System, sharing the skies with piloted aircraft. The FAA and NASA have proposed dividing the airspace into lanes or corridors with restrictions on when, what kind, and how many aircraft can use them. This would significantly alter the current use and standard practices in the National Airspace System and could create air traffic jams, the researchers noted.
The research team has yet to test the AI pilot on actual aircraft, but it has performed well on flight simulators, they report.
To test the AI, the researchers set up two flight simulators. One is controlled by the AI, the other by a human. Both operate in the same airspace.
The AI can safely navigate around the piloted aircraft, even if the person behind the controls is not an experienced pilot, the researchers discovered.
“We need more pilots, and AI can help,” said Jay Patrikar, a Ph.D. student who worked on the project.
The research was supported by the U.S. Army Research Office and the Army Futures Command’s Artificial Intelligence Integration Center (AI2C).