It’s hard to find a good idea that hasn’t been done [better] before.
For this post we’re going way back to the Rock-Paper-Scissors playing robot that was one of my first write-ups.
While reading up on some of the cutting-edge experiments going on in robotics, I came across this 2012 video from Ishikawa Watanabe Laboratory at the University of Tokyo. As you’ll see, the idea of having a robot playing cheating at rock paper scissors against a real human has been done quite well before.
Their lab focuses on finding applications for extremely high speed image recognition. Using some powerful computers and a 600fps camera, they were able to tune the entire operation of seeing/recognizing the opponent’s move and throwing the corresponding wining move down to a speed of 1ms—completely unnoticeable to the human eye.
In principle, however, it’s the exact same as my group’s “Handroid” bot. I’m confident that, given unlimited time and resources, we would have been able to optimize our build down to something closer to a millisecond.
GH
Bits, Bots, and Miscellania 