*related=http://videosift.com/video/Timo-Boll-vs-KUKA-robot-Teaser

Timo Boll vs. KUKA robot - Teaser has been added as a related post - related requested by Zawash.

That was terribly disappointing Just a marketing gimmick. No real table tennis. I feel ripped off.

This one can actually play. Although obviously it has it's limits but would be fun for just goofing around. http://youtu.be/imVNg9j7rvU

Yeah - would be much more interesting to see some "real" table tennis - see how good the machine actually is.

Comment hidden because you are ignoring dag.(show it anyway)

I too am disappointed for the same reasons. Though I bet KUKA considers this a success - considering we probably didn't know what a KUKA was before this.

I would think a robot could be built that can simply serve balls so insanely fast that a human could barely hope to return them. This serving robot could beat the best human players even if it had no way to play an actual back and forth game of ping pong... provided it got to serve first, or course.

Actually.. wait... ping pong serves have to bounce on both sides of the net, right? That probably implies a max speed since if the ball is going too fast, you couldn't bounce on your side, clear the net, and also land on your opponent's side of the table without traveling too far. There may be room for an actual competition here.

It's not just a speed (of the robot) challenge to defeat a human, it's a CV/AI challenge. From this video it's not clear there was even a CV system; likely all these vollies were scripted and rehearsed. Even in this fictional match Boll learns how to defeat the robot by using moves it can't effectively counter.

The CV/AI (computer vision/artificial intelligence, I assume) challenge is certainly the more interesting part of this competition. The trouble I have with many man vs machine competitions is that it's often possible to design a robot that focuses on just one specific aspect of the game that it can do so well that it renders complex CV/AI work obsolete. Robots can often win in uninteresting ways that make the competition boring to watch. And humans can often exploit AI flaws or physical limitations of the robot to claim victory in equally non-spectator friendly ways.

In ping pong, if a robot can position its paddle in the right place to return a ball at all, it should be able to return it so fast that a human couldn't stop it. Strength and speed are the easy part. At the very least, the robot is pulling its punches (or more likely, it was not designed to punch as hard as it could have).

The competition just seems staged as a result. My guess is that both the human player and the robot are purposefully playing in such a way to create the most interesting looking match, rather than playing in the way most likely to win.

Im guessing the possibility to one day create a competitive table tennis robot is doable but doubt they are anywhere near that now technologically (just a guess). The info that has to be gathered before each shot and the ability for the computer/robot to counter with the best type of shot seems like a daunting task for the computer/robot to make.

From this video im guessing that it wouldn't take much of a table tennis player to destroy a robotic competitor. I see nothing here that makes me think this robot could play worth a shit competitively. Way to easy to exploit it's shortcomings is my guess.

Now a robotic pool player seems much more realistic.

How about computer controlled car in a formula one race?

In case you can't tell, the ball is actually CG, added later. I was disappointed as well, hoping that the bot would be able to actually throw down some sick loops and counterloops, but alas it's all faked.

This is some of the most overproduced nonsense I've ever seen.

There is some slight irony there, kuka is a Finnish word that literally means "who?"

A human-robot dance duet | Huang Yi & KUKA has been added as a related post - related requested by Eklek on that post.