Retired F-16 Fighting Falco?ns sit at the 309th Aerospace Maintenance and Regeneration Group’s Aircraft and Missile Storage and Maintenance Facility on Davis-Monthan AFB, Ariz, Aug. 2, 2017. USAF Photo by SSgt. Perry Aston.
An AFA Mitchell Institute recommendation that the Air Force reactivate some of its retired aircraft as unmanned drones will get a hearing from Air Force Research Laboratory chief Maj. Gen. William Cooley said Tuesday.
“I’m delighted they are thinking outside the box,” Cooley told reporters on Tuesday after a Mitchell paper on manned/unmanned teaming was released on Capitol Hill, “and we will be looking at” resurrecting boneyard fighters “as one of the alternatives” to building new aircraft to generate the numbers needed to counter growing numbers of Chinese and Russian combat aircraft.
Cooley said the preferred choice, however, will be to build inexpensive airframes that are “attritable,” which he described as not so cheap that they are disposable, but not so expensive that their loss would represent a major deficit in combat air strength. A strike against the idea of resurrecting retired aircraft is that the cost to restore them to flying condition and “sustain” them, even for limited or single-use missions in combat, might exceed that of building dedicated new aircraft with all the desired sensors and performance, Cooley said.
Douglas Birkey, executive director of Mitchell, one of the authors of the paper, said the Air Force has experimented with F-16s operated by artificial intelligence under a program called “Have Raider,” in which the aircraft performed a series of mission tasks cooperatively with manned F-16s.
The Air Force “used to have ‘attrition reserve’ aircraft,” Birkey said at the paper’s release, but tight budgets have squeezed those extra-margin platforms out of the force, and US adversaries “know it,” he said. Using retired aircraft and basing them in forward areas would require potential adversaries to count those aircraft “in their calculus” of the number of weapons necessary to fight a limited or protracted war with the US. It might be a useful deterrent, and the technology to make the idea work exists, Birkey argued.
It’s “a terrific idea,” Cooley said, “but we have to look closely at the business case.” He also said more experimentation and prototyping is necessary to explore the idea, and “I think we’re poised to start doing that.”
Cooley, in a brief speech, said the US is growing more confident in artificial intelligence and machine learning as it applies to the cockpit, noting that automated systems already do most mission route planning. He also pointed out that Automatic Ground Collision Avoidance Systems in the F-16 have saved eight lives. Auto-GCAS takes the controls away from the pilot and flies F-16s that are so equipped away from the ground if the pilot does not take action to avoid a crash.
He also noted there is ongoing work with the FAA on getting clearance for autonomous aircraft permission to operate in US airspace, and that such permission is frequently given out by States before it happens at the Federal level. Cooley said driverless cars are operating throughout the US on an experimental basis, but in direct proximity to ordinary traffic, and he sees no reason why that latitude can’t also be applied to aircraft.
Tim Grayson, director of the Strategic Technology Office at the Defense Advanced Research Projects Agency, said the Pentagon ought to be “agnostic” about the platforms it puts into the air to build a network of sensors and knowledge that the whole force can draw on. “It should be a heterogeneous mix … with no single type of platform” relied on more than others, he said. Machine learning and artificial intelligence will be crucial to studying the information collected, he said, and also at prioritizing which sensors should be given attention if the communications environment is “degraded.” It’s all part of pushing decision-making down to the lowest level and enhancing the speed of making choices, he said. He called this “mosaic” warfare.
Asked to distinguish between Artificial Intelligence and autonomy, Cooley said he’s come to use the terms interchangeably, and that the key concept is “machine learning,” such that computers can automatically assess data and help humans draw the right conclusions from it.
Cooley revealed that the first flight of the XQ-58—a Mach 0.8 armed, unmanned demonstrator—will take place this fall. The program is known as the Low Cost Attritable Strike UCAS demonstrator, and the contractor, Kratos, calls it the Valkyrie.
“Our job will be to drive the cost of these platforms as low as possible,” Cooley said of projects like the XQ-58, suggesting he believes it will be “lower than the cost of refurbishing” retired F-16s. Regardless of which platform or platforms move on to operational status. “We need something like this as soon as possible” to address the growing imbalance of numbers, Cooley said.