German Ping-Pong Robot Can Analyze Players, Will Eventually Surpass Its Human Masters
In 1996, when supercomputer Deep Blue beat the world’s greatest chess player, Gary Kasparov, we opened the door for machines to not only take the place human beings, but to surpass us with cold logic and massive processing power. In 2012, humanity is on the road to losing another game to the superior programming of a machine. That’s right, there’s a machine that’s better than you at ping-pong. Robotics experts at the Technical University of Darmstadt in Germany have created a robot arm that learns to adapt its game as it plays. Like the Highlander, this arm will inevitably absorb the skills of the world’s greatest ping-pong players, becoming an unstoppable force to be worshipped by anyone holding a paddle.
Katharina Muelling and her team attached the robot arm to the ceiling and attached a camera to watch and analyze the game. Slowly but surely, the team taught the arm to play ping pong by feeding it increasingly difficult shots. Eventually, the arm was able to use the information from the motions it had been taught to generate its own shots. By taking everything it’s learned, it has been able to create idealized answers to scenarios it had never been trained to respond to.
Though other robots have been successfully programmed to play ping-pong, this one is the first to “learn” the game. From it sound of things, it also sounds like the best player yet. After only an hour of practice, the robot arm was successfully returning 88 percent of shots. I’ve been playing the game for years and I don’t hit that many shots. (Then again, I’m pretty bad at ping-pong.)