Invincible robot in the game of rock paper scissors

With the ability to identify and analyze an opponent's finger movement, the robot of Ishikawa Oku Laboratory, Tokyo University will know that the player will prepare scissors, sacks or hammers.

>>>Video: Invincible robot in the game rock paper scissors

We often think that scientists cannot replicate the complexity of the human brain on machines. But, no matter how strong our neurons and nervous systems are, we still can't win against the robot of Ishikawa Oku's lab. At least in the game rock paper scissors.

What makes Ishikawa Oku's robot can overcome creativity, intelligence and human prediction? The answer is a high-speed image recognition system.

In fact, we can say that this robot has cheated, because instead of making its decisions at the same time as its opponents like genuine "rock-paper-knights" , Oku's robot will analyze. Hand movement of the player in a very short period of time and then make the appropriate decision to pull / cover / hammer. The process of analysis and decision making of robots is so fast that you feel that this robot doesn't cheat.

According to Oku lab's website, the process of identifying human hands only takes place in 1 millisecond (1/1000 second) and the robot will recognize both the shape and position of the human hand.

Fortunately, according to the scientists at Ishikawa Oku, the robot's rock-paper-scissors victory today is not synonymous with human destruction at the hands of machines in the future. Instead, they argue that their invention will help robots and humans work better.

"This technology is an example of how the ability to control coordination between humans and machines in a few milliseconds is possible. And this technology can be applied to support your movements. people, coordinating work between people and robots . v . without time lag. "

The robot is shown as the generation of janken robots (janken: the name of the second Japanese rock paper scissors game developed by Oku). The first generation has an identity delay of up to . 20 milliseconds.

However, according to Sethu Vijayakuma, professor of robotics at Edinburgh University, the 1 millisecond speed of 2nd generation janken is still not enough.

"These robots respond very quickly, but there are 1-millisecond delays that are unacceptable, such as accident prevention or virtual stock market monitoring."

"Under these conditions, we need to combine high-speed response and high-speed predictability, using game rules and behavioral modeling theory."