Photo Source: Flickr
Computers continue their march of triumph over humans in increasingly diverse areas of mastery. From chess to GO to Jeopardy, machines driven by artificial intelligence have achieved stunning wins over human masters in many fields. The drive continues in the dangerous and complex world of aerial combat between fighter planes, otherwise known as a dogfight.
Over three days in mid-August, 2020 from the 18th-20th, the Defense Advanced Research Projects Agency (DARPA) virtually hosted the final competition in a year-long contest called the “AlphaDogfight Trials.” The competition began in the fall of 2019 with eight teams ranging from the big defense contractor, Lockheed Martin, The Georgia Tech Research Institute to smaller companies such as Heron Systems and Aurora Flight Sciences. The competition was designed to build trust in AI systems among the military. DARPA challenged the companies to develop artificial intelligence systems to control simulated F-16 fighter jets in aerial combat. Over the past year, the companies have engaged with each other in simulated combat. The first engagement, Trial 1, occurred in November 2019, and Trial 2 took place in January 2020. Later in the spring, the companies work to improve their simulated fighter pilots. Finally, Trial 3 in mid-August pitted the companies against each other the first two days. On day three, the winner then challenged a top human F-16 fighter pilot in a virtual reality flight simulator (darpa.mil).
After two days of competition, Heron Systems Inc. defeated the Lockheed Martin virtual pilot to battle the human pilot on day three. Founded in 1993, Heron Systems describes itself as “a woman-owned, small disadvantaged business with locations in California, MD and Alexandria, VA. Heron Systems builds practical, deployable autonomous agents and multi-agent systems powered by artificial intelligence.” (heronsystems.com/about). The AI from Herron Systems achieved high proficiency in dogfighting through a process called reinforcement learning. The system practices through thousands of simulations, allowing it to accelerate its learning far faster than any person could achieve. The final competition pitted Herron’s AI and an anonymous Air Force F-16 fighter pilot recently graduated from the Weapons School. The Herron Systems AI resoundingly won all five contests over the human pilot, notching another victory for AI.
Aerial combat, or dogfighting, emerged as a vital and increasingly complex type of warfare. Remarkably, only ten years after the Wright brothers achieved the first motor-powered flight in a heavier-than-air aircraft at Kitty Hawk, NC, on December 17, 1903, the first chronicled dog fight occurred during the Mexican Revolution between two American mercenaries. “American pilots Phil Rader and Dean Ivan Lamb, who were on opposite sides of the conflict, fired revolvers at each other while airborne. Neither one got hit.” (airspacemag.com). World War 1 proved to spur the rapid advancement of technology and aviation skills to pit aviators against each other in the skies over Europe. At this time, one of the most famous fighter pilots of all time, the German fighter pilot Baron Manfred von Richthofen, gained great notoriety as one to the most famous flying aces of all time. Generally known as the Red Baron, Richthofen shot down 80 planes and helped define the tactics of aerial combat.
Over the years, faster planes with more sophisticated technology and weapons pushed the skills and physical capabilities of human pilots. With supersonic speeds and aircrafts able to turn sharply, pilots must work at the physical abilities of their bodies. Certain maneuvers can render a pilot unconscious due to the intense amount of gravitational force acting on the pilot in the plane. Technology helps fighter pilots monitor many enemy crafts simultaneously as well as maintain control of today’s highly complex aircraft. It makes sense that in aerial combat technology that can react and make decisions faster than a human would have an advantage. Richard Lawler for Engadget notes that, “the AI’s ‘superhuman’ aiming ability produced an edge even as the simulated F-16s circled low to the ground at high speed and would’ve experienced extreme G forces.” (engadget.com) In other words, because the AI would not have had a physical reaction to the intense vertical gravitational force, it would have an edge over the human being with physical limitations.
Flying aces transformed air combat over the past hundred years through their innovation, bravery, and willingness to push the edge of human ability. It follows that as computers improve and exceed human reaction time, the spirit of the ace will continue through artificial intelligence. However, no longer needing a human inside a fighter jet may just be the next step in the long history of military aviation.
Dr. Smith’s career in scientific and information research spans the areas of bioinformatics, artificial intelligence, toxicology, and chemistry. He has published a number of peer-reviewed scientific papers. He has worked over the past seventeen years developing advanced analytics, machine learning, and knowledge management tools to enable research and support high level decision making. Tim completed his Ph.D. in Toxicology at Cornell University and a Bachelor of Science in chemistry from the University of Washington.
You can buy his book on Amazon in paperback here and in kindle format here.