The Perfect Soldier: Special Operations, Commandos, and the Future of Us Warfare by James F. Dunnigan
More Books by James Dunnigan
The Bleeding Edge Of Combat Software
by James Dunnigan
October 26, 2014
The U.S. Air Force and Navy agree on one thing; AI (artificial intelligence) is going to be a major element in future combat aircraft avionics (aircraft electronics). This is in large part due to the fact that AI has been a growing factor in avionics for decades in the form of software controlled landing systems for commercial aircraft, system monitoring and a growing number of other uses. All this began with the software developed for instrument landings (where the pilot relied on instruments, weather preventing a good view of the ground). In the last few years the U.S. Navy has taken that sort of thing to a higher degree of complexity by perfecting software that could handle landing on a moving air field (an aircraft carrier). This is considerably more complex than the usual situation (landing on an airfield) and requires that the software make lots of decisions correctly and at high speed. Thus carrier landings require more powerful hardware and software aboard the aircraft. The navy expected glitches and bugs but sees itself catching up to the reliability of commercial landing software (which has been used very successfully on UAVs) within years rather than decades.
Meanwhile there are similar software systems used to quickly let the pilot know the best route through bad weather, enemy air defenses and equipment failures aboard his aircraft. Because all these AI systems are simply showing up as (according to the developers) “normal development” of avionics they are often simply adapted from systems already being used in commercial aircraft. The military aviation community has long been working on how they could use AI for purely military uses. Actually, there is already commercial software available for purely military applications and that is the stuff used in computer games to control AI run aircraft (enemy of friendly). This sort of AI has been in development for commercial air combat games for over two decades. But military leaders are reluctant to embrace this sort of software.
Meanwhile many military people, including pilots and software engineers, and even some generals in the air force, believe that its next generation fighter will not have a pilot on board. Many air force generals admit that the F-35 is probably the last manned fighter. But some believe that the F-35 will be facing stiff competition from pilotless fighters before F-35 production is scheduled to end in 2034.
Not surprisingly UAV (unmanned aerial vehicles) are not particularly popular with many U.S. Air Force and naval aviation leaders, but that is not the case in many other countries. Air force generals around the world see the unpiloted jet fighter as a way to break the monopoly the U.S. Air Force has had on air supremacy for the last sixty years. Most Americans don't even think of this long domination of the air, but potential enemies of the United States are well aware of it, and that domination has a profound effect on how those nations do their military planning. In effect, if you think about going to war with the United States in the immediate future, you take for granted that American aircraft will control the skies above. Robotic jet fighters could change that. And this is forcing American air force generals to confront a very unsavory prospect; a sixth generation fighter that is flown by software, not a pilot.
It's not just that most of those American aviation commanders began their careers as fighter pilots. No, the reason is more practical. American air superiority has largely been the result of superior pilots. The U.S. didn't always have the best aircraft, but they always had the most talented and resourceful pilots. And that's what gave the U.S. its edge. Will that translate to software piloted fighters? Research to date seems to indicate it will.
Meanwhile, simulations, using fighter flown by software, versus those flown by humans, have been used for over two decades. The "software pilots" have gotten better, and better. Moreover, a fighter without a pilot is more maneuverable (because some maneuvers are too stressful on the human body.) UAV fighters can be smaller, cheaper, stealthier and more expendable. But the key to software pilots is the development of superior tactics, and AI that is more capable than anything your opponent can come up with.
The U.S. Air Force, and several other air forces, have already created fighter pilot software, and now the United States, and Russia, are creating pilotless fighters. Many air force generals are convinced that the pilotless fighters will perform as well for real, as they have in the simulations. So convinced are U.S. Air Force generals that they are seriously considering a sixth generation fighter that will not carry a human pilot. Otherwise, enemy pilotless fighters would have an edge over the U.S. sixth generation aircraft.
The potential superiority of U.S. pilotless fighters is partly driven by the fact that most American fighter pilots are geeks. Many can create software, and have a deep understanding of the many computers, and their software, that modern aircraft contain. It's the fighter pilots who will play a key role in creating the best "software pilots." Thus the thinking is that American control of the air will be maintained by a new generation combat aircraft controlled by software, not someone in a cockpit.