The Stack Archive

The rock-paper-scissors robot with a 100% win rate

Fri 18 Sep 2015

A team of robotics scientists from Tokyo has introduced a rock-paper-scissors robot that never loses, beating human competitors before they have even completed their move.

The robot, named Janken, has been created at the University of Tokyo’s Ishikawa Watanabe Laboratory. The machine relies on high-speed visual technology and tracking systems ‘1ms Auto Pan-Tilt’ and ‘Lumipen 2’ to predict and counter its competitor’s moves. Within one millisecond, Janken is able to record the initial movements of the human hand before it has fully formed the final shape.

This is the third and latest version of the Janken rock-paper-scissors robot, which the team claims provides evidence to promote future human-robot cooperation in real-time. In the first iteration, the robot completed its move in 20 milliseconds after the human had already performed their move. In the second version, the robot achieved a 100% win rate but finished its gesture at the same time as the human. Janken now manages a millisecond advantage over the human to win the game.


The team is also developing a range of baseball-playing robots, which can throw, track, bat, run and catch. In a similar fashion to Janken, the tasks are achieved by controlling high-speed robotic systems based on real-time visual feedback.

A throwing robot mirrors the human action, although generally achieves greater precision. The technology currently has a lower success rate than the rock-paper-scissors prototype, with 90% accuracy for throwing a ball into the strike zone.

Two catching robots use high-speed hand movements, closing their fingers 10 times per second. These robots are able to catch flying balls by controlling their manipulators in response to the recorded ball movements.

The researchers suggest that although Janken and the baseball robots are limited for the time being, the technology could help realise faster robotic processes used in manufacturing, improving productivity and reducing costs.


Asia Japan news robotics
Send us a correction about this article Send us a news tip