Robots are (sometimes) pretty good at playing catch, sometimes extraordinarily good, but many of the coolest demos involving robots interacting with fast-moving objects rely on extremely precise external vision systems. This reliance means that you aren't likely to see robots performing such tricks outside of a lab. DLR's Rollin' Justin, on the other hand, is using onboard cameras and processing, meaning that it could just as easily play catch with you on your front lawn:
Did I say play catch with you? I meant to say, play catch with you and someone else at the same time.
In case you missed those specs, the robot is able to position itself within two centimeters of where it needs to be in a time window of only five milliseconds, which yields an impressive catch rate of better than 80 percent. That's maybe not major league, but it could probably be little league, and undoubtedly there's still some optimizing to do. You know that robot buddy you wish you had when you were a kid, that would always be up for a game of catch? This could be it; they just need to teach it how to throw the ball back first.
[ IROS 2010 Paper and ICRA 2011 Paper Abstract ] via [ Hizook ]
Evan Ackerman is a senior editor at IEEE Spectrum. Since 2007, he has written over 6,000 articles on robotics and technology. He has a degree in Martian geology and is excellent at playing bagpipes.