NASA Makes Strides toward Self-Aware Aircraft

What would it take for an aircraft to be fully autonomous, to the point that you could step into an aircraft—like a taxi—and tell it where you needed to go, then simply sit back, relax, and, in a reasonable time, arrive at your destination, safe, sound, and unruffled?
NASA, working with partners like Boeing, Honeywell, and General Atomics Aeronautical Systems, is making baby-steps toward that goal.
The key word in this process is “self-aware.” This concept may eventually allow unmanned aircraft (i.e., no pilot at the controls, even remotely) to routinely fly in the nation’s airspace—sharing it with piloted airlines and the rest of us who want to do our own flying.
To be “self-aware,” the aircraft must be able to autonomously “sense and respond.”

A basic flight element that all pilots are familiar with are aerodynamic stalls. From their first lessons, pilots are taught to recognize and avoid or recover from stalls. Most aircraft have some form of stall warning—a horn, angle-of-attack indicator, or stick shaker, etc. In a self-aware, autonomous aircraft, it would not only detect an impending stall, but react immediately to prevent the stall. (Actually, it would never let the aircraft get into a situation even approaching a stall.)

This technology has been demonstrated by NASA as part of their Unmanned Aircraft Systems Integration into the National Airspace system (UAS-NAS) project. In a series of tests, NASA and its partner, Boeing built and tested 5.5% scale model of what could be a Boeing 757—they call it the GTM-T2. The T2 flyable model was equipped with a digital system to prevent stalls and other un-flyable states. Using remote controls, pilots tried to make the aircraft stall or otherwise make the aircraft lose control. The system was consistently able augment the pilot’s control and maintain stable flight. Through a variety of sensors, the aircraft was able to autonomously determine that the aircraft was approaching an unsafe flight condition, analyze the condition, and autonomously react with the proper responses, including deflection of flight controls, to recover and maintain stability.
Irene Gregory, senior technologist for advanced control theory and application at the National Aeronautics and Space Administration, was quoted in the Wall Street Journal saying that eventually aircraft would be “smart enough that people will be able to get around in on-demand self-flying taxis.”
While that day is probably far into the future, she also suggested that a system like the one demonstrated on the GTM-T2 model could be approved for use in the next generation of all-new airliners as a safety backup; if approved by the FAA.

Another, critical capability of a self-aware aircraft is the ability to sense and avoid other aircraft. NASA has sponsored at least two different test programs in this area. For test aircraft, NASA used an unmanned Ikhana (an unarmed version of the Predator UAV) and a remotely piloted S-3B and. The S-3B will carry a safety pilot, but the pilot will not be flying the aircraft during the tests but could assume control if the situation required.

These aircraft were equipped with sense-and-avoid sensors and software. The aircraft are flown along a predetermined route. Manned aircraft were deliberately flown into the path of the test aircraft, simulating what might happen if two aircraft were converging with insufficient separation. After 40 flight tests, the Ikhana autonomous test aircraft demonstrated the ability to successfully and safely maneuver to avoid the conflicting aircraft traffic and safely return to course after every encounter. A total of 200 such encounters were planned for this phase of testing.

A follow-on test will use a T-34 plane equipped with a proof-of-concept control and communications system to test how well the digital systems control the aircraft, interact with air traffic controllers and remain well clear of other aircraft while executing a planned mission. With a safety pilot onboard, the aircraft will fly a typical mission complying with air traffic control and safely avoiding other air traffic.

These are only two elements for a fully autonomous “air taxi.” A complete flight will require the aircraft to plan the route, check for adverse weather, file a flight plan, communicate with and follow Air Traffic Control instructions, perform a safe flight, etc.—the same requirements a pilot would follow when conducting a flight. So, while NASA and the aerospace community are making significant progress in autonomous flight, there is still much work to be done.

For now, taxicabs will operate on four wheels with an occasional surly driver. A least an air taxi autonomous pilot will not expect a tip!