Technology‎ > ‎

Rock, Paper, Scissors Robot Wins Hands Down Every Time

posted 8 Nov 2013, 08:31 by Mpelembe   [ updated 8 Nov 2013, 08:32 ]

There's a new champion in the world of "Rock, Paper, Scissors", and it's not a human. Researchers at the University of Tokyo have refined an earlier version of their "Rock, Paper, Scissors" robot and are showing it off to potential human challengers at the International Robot Exhibition in Tokyo.

TOKYOJAPAN (NOVEMBER 8, 2013) (REUTERS) - While you may think you're unbeatable when it comes to rock, paper, scissors, Professor Masatoshi Ishikawa of the Univesiry of Tokyo suggests you think again.

He says his new and improved rock, paper, scissors robot can match any challenger, albeit through a sleight of hand that's undetectable to humans.

The system utilizes a high-speed camera and sensors that enable it to recognize the emerging shape its opponent's hand, one millisecond before it plays its own winning hand.

The first version of the robot was able to recognize its opponent's choice of hand position as the fingers began to move, but took 20 milliseconds to respond with a position of its own. The new version begins to form its position slightly before the human can complete his or her own.

"Before it was the case that the human hand completely opened and then the other opened. But that happened so fast that human eyes couldn't see it. So from the human's point of view it seemed like a 100% win rate. Now however, we've made it so that, while not always, it opens before the human hand does," Professor Ishikawa said.

Some may see the millisecond detection advantage as a means of cheating, but Ishikawa says the purpose of the robot is not to win the game, but to demonstrate how such machines can be used in real world settings, like factories, where quick recognition of objects on the production line is essential.

"On a factory assembly line, there are places where humans worked that robots couldn't and I want to change it to this sort of machine," said Ishikawa.

The robot and its accompanying system was originally created in 2011. But Ishikawa says the new version beats the original, hands down.