A team of researchers at Carnegie Mellon University believe they have developed the first AI-based pilot that will enable autonomous aircraft to navigate crowded airspace. According to the researchers, the new artificial intelligence system can track aircraft; predict the intent of other aircraft; coordinate with their actions; and communicate over the radio with pilots and air traffic controllers. In short, the actions of the AI system will be indistinguishable from those of a human pilot.
The objective is for the system to eventually pass the famous “Turing Test,” meaning that the AI is able to exhibit intelligent behavior indistinguishable from that of a human. To interact with other aircraft as a human pilot would, the AI uses both vision and natural language to communicate its intent with other aircraft. Such behavior leads to safe and socially compliant navigation.
Researchers have achieved this implicit coordination by training the AI on data collected at the Allegheny County Airport and the Pittsburgh-Butler Regional Airport which included 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 used by a human pilot. Its automatic speech recognition function uses natural language processing techniques to both understand incoming radio messages and communicate with pilots and air traffic controllers using speech.
This advancement in operating autonomous aircraft without a pilot behind the controls will broaden opportunities for drones, air taxis, helicopters and other aircraft in applications as diverse as: moving people and goods, inspecting infrastructure, treating fields to protect crops, and monitoring the environment. And that will require these aircraft to fly in an airspace already crowded with small airplanes, medical helicopters and more. While autopilot controls are already common in aircraft operating at higher altitudes under instrument flight rules, developing an AI to handle the often crowded lower-altitude traffic operating under visual flight rules has challenged the aerospace industry.
The Carnegie Mellon team’s AI is designed to seamlessly interact with aircraft in visual flight rules airspace. Ideally, delivery drones and air taxis will eliminate pilots to save weight and insulate them from a pilot shortage. In that scenario, the kind of AI-based system pioneered at CMU will enable autonomous aircraft to safely deliver packages and ferry passengers.