Janken (Rock-Paper-Scissors) Robot With 100% Winning Rate
The purpose of this study is to develop a janken (rock-paper-scissors) robot system with 100% winning rate as one example of human-machine cooperation systems.
Rock-paper-scissors should be a simple game. But it requires strategy, psychology, instinct. Does your opponent always pick rock, like Michael Bluth in Arrested Development or Bart in The Simpsons? If you put up paper and are met with scissors, do you pick the sheers yourself in the subsequent round, or stick with paper on the assumption that your rival will next pick rock.
If you assign your rock-paper-scissors to this robot surrogate, though, you’ll never fail. That’s because the machine’s high-speed vision analyzes the way its opponent is shaping his or her hand and is able to identify within one millisecond the weapon that is about to be thrown its way. Once it can tell whether it will face rock, scissors, or paper, the robot can figure out what it should do to win. “[I]t all happens so fast that it's more or less impossible to tell that the robot is waiting until you commit yourself before it makes its move,” says Evan Ackerman on IEEE Spectrum’s Automaton blog. The result: The bot wins 100 percent of the time.
Researchers at the University of Tokyo’s Ishikawa Oku Laboratory call this a “human-machine cooperation system.” Though it doesn’t feel like the robot is “cooperating” when it wins by taking advantage of our slower reflexes, its creators say the technology could make it easier for man and bot to work together—say, along a factory production line.