An artificial intelligence software program that recently beat a US Air Force fighter pilot in a combat simulation requires so little computing power that it could run on a Raspberry Pi, according to its creators.
A team of computer scientists and aerospace experts at the University of Cincinnati developed the AI algorithm, which participated in a simulated dogfight with retired Air Force Colonel Gene Lee.
He told the university's magazine that the program, called "Alpha," is "the most aggressive, responsive, dynamic and credible AI I've seen to date."
Alpha was designed as a research tool to develop techniques for teams of fighter pilots flying together. In addition to besting Lee, it has also won dogfights against several Air Force simulation programs.
With most AI programs, "an experienced pilot can beat up on it [the AI] if you know what you're doing," Lee explained. "Sure, you might have gotten shot down once in awhile by an AI program when you, as a pilot, were trying something new, but, until now, an AI opponent simply could not keep up with anything like the real pressure and pace of combat-like scenarios."
Alpha's tests were performed while it was running on a $500 entry-level consumer PC, though according to the magazine, it could also run on a Raspberry Pi. The Raspberry Pi 3 has a 1.2GHz processor and 1GB of RAM.
Similar to other AI algorithms that can solve complex tasks, Alpha is built on language-based code that approximates the functioning of a human brain—it can make decisions on the fly using genetic algorithms. Alpha's designers hope it will be one day be used as a sort of virtual wingman for Air Force pilots.
"Alpha would be an extremely easy AI to cooperate with and have as a teammate," University of Cincinnati professor Kelly Cohen told the university's magazine. "Alpha could continuously determine the optimal ways to perform tasks commanded by its manned wingman, as well as provide tactical and situational advice to the rest of its flight."