Intelligent flight controls react to systems failures — and 'heal' them — sometimes before the flight crew knows what happened

BY WILLIAM COLE

What happens when a military aircraft is severely damaged, losing half its wing in air combat maneuvering?

In one case, swift reactions and smart work by an F-15 combat pilot involved in a midair collision brought him back alive — and saved the aircraft. He was able to compensate for the damaged wing by configuring the remaining flight controls in unusual ways. Many Vietnam-era combat pilots were not so lucky.

But flying a combat-damaged or otherwise crippled aircraft may be easier on the pilot in the future — thanks in part to Boeing research.

An airplane of tomorrow likely will be able to instantly "think" its way out of concurrent systems failures that cause a catastrophic loss of control. And neural network flight control laws will allow control systems to quickly "learn" to compensate for aberrant flight conditions that sometimes cannot be duplicated in wind tunnel tests, saving time and money.

As a result of "intelligent" flight-control solutions developed by Boeing in partnership with the U.S. Air Force, the U.S. Navy and NASA, modern military and commercial airplanes will have the ability to learn how to deal with a multitude of scenarios and "heal" themselves in-flight after experiencing damage or failed systems. And intelligent flight controls will have great applications for unmanned air vehicles and reusable space vehicles.

"Breakthrough programs using self-learning neural network-based control systems offer increased robustness in battle damage or failure conditions," said James Urnes Sr., manager of engineering, and an aerodynamics and flight control expert for Phantom Works.

Three consecutive programs, said Urnes, have been using different approaches to apply intelligent flight control technology to flight control systems.

In a self-repairing flight control test, an F-15 Eagle with its right stabilator completely disabled was able to fly safely by using a reconfiguration software process especially designed for major malfunctions.

Flight accidents were duplicated with the loss of control surfaces in separate propulsion-controlled aircraft tests, which were designed to improve and recover flight qualities of aircraft with control systems damaged to the level of complete inoperability. Tested first on an F-15 and then on an MD-11, this system uses motion sensor feedback and pilot commands to use the engine power for flight path steering. In the emergency situation of loss of all conventional control surfaces, the pilot could command the disabled aircraft to pitch and bank flight path changes through use of collective and differential thrust changes. Using this system, the F-15 achieved a perfect landing at Edwards Air Force Base in California with only the thrust changes of its two engines steering the plane. The MD-11 subsequently achieved a similar perfect landing with all of its conventional controls disabled.

Later, a prototype intelligent flight control system using neural network-based control was successfully flown on a NASA fly-by-wire flight-test F-15 with impressive results. A self-learning capability is being added to the F-15 controls that can adapt to many kinds of control surface damage.

The NASA program is now in its second generation with a concept called RESTORE, and has been test flown by NASA in its quarter-scale uninhabited Boeing-built X-36 tailless aircraft. And NASA is working with Boeing on a program to evaluate the intelligent control system on the C-17. If successful, the technology could one day be used on commercial aircraft.