{"id":131165,"date":"2025-09-12T01:34:25","date_gmt":"2025-09-12T01:34:25","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/131165\/"},"modified":"2025-09-12T01:34:25","modified_gmt":"2025-09-12T01:34:25","slug":"this-humanoid-robot-is-playing-ping-pong-better-than-most-amateur-players-ive-seen","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/131165\/","title":{"rendered":"This humanoid robot is playing ping pong better than most amateur players I’ve seen"},"content":{"rendered":"

HITTER: A HumanoId Table TEnnis Robot via Hierarchical Planning and Learning – YouTube
\n\"HITTER:<\/p>\n

Watch On <\/a><\/p>\n

I tend to approach robot videos with a mild air of defeat and scepticism, in no small part because a lot of hype has been generated over what has turned out to be just choreographed movements in the past. Not here, though. This little HITTER (HumanoId Table TEnnis Robot)<\/a> is hitting bona fide table tennis rallies against humans. And yes, that’s table tennis, not ping pong; I’m right, you’re wrong.<\/p>\n

Alright, alright, we’ll go with ‘ping pong’. HITTER, the ping pong-playing robot, comes from researchers at the FHL Vive Center for Enhanced Reality at UC Berkeley<\/a>. In a recent video<\/a> (via Techeblog<\/a>), it’s shown forehanding, backhanding, and smashing ping pong balls with impressive dexterity and, dare I say it, human-like movement.<\/p>\n

<\/p>\n

What’s so visually impressive about it, at least to my eyes, is how it keeps balance and makes quick decisions and very human-like arm and leg movements. In particular, as someone who used to play a lot of table tennis and who wasn’t awful at it, I was genuinely impressed by how HITTER returns smashes, and also with the return at 0:28 right off the edge of the table.<\/p>\n

Related articles<\/p>\n

The robot achieves all this thanks to its design which “employs a hierarchical decomposition that separates high-level trajectory planning from low-level whole-body control execution.” In other words, there’s one part of its ‘brain’ that takes camera input and predicts where the ball will land and how to return the shot, then there’s a second part that translates this higher-order analysis into actual ‘body’ movement.<\/p>\n

Reinforcement learning\u2014the fancy AI industry term for trial and error\u2014was used to train it, including looking at videos of “human motion references” to improve its movements. The result, as you can see, is a robot that can adjust to different shots on the fly and return the shots very naturally for some long rallies.<\/p>\n

\n

\"UC\n<\/p>\n

(Image credit: Hybrid Robotics @ YouTube, FHL Vive Center for Enhanced Reality, UC Berkeley)<\/p>\n

HITTER “achieved 106 consecutive shots in a rally against a human opponent, surpassing typical casual play duration and demonstrating consistent tracking, returning, and balance recovery capabilities.”<\/p>\n

Very cool stuff. Though I’d be remiss if I didn’t end with an obligatory ‘it’s a robot’ reminder. I’m saying this as much as a reminder to myself as to anyone else: We need to remember not to attribute our awe and praise to the robot, but instead to the engineers who have managed to create a piece of technology that can adapt to changing environments so naturally. That would be the case whether it was a humanoid robot or a giant slab of metal being swung around by a cable.<\/p>\n

Keep up to date with the most important stories and the best deals, as picked by the PC Gamer team.<\/p>\n

HITTER also serves as a reminder that robotics seems to be heading in a very humanoid, or at least animal-oid (did you see those Chinese robo-wolves<\/a>?) direction. It really is hard to beat nature’s\u2014or God’s\u2014designs in many areas.<\/p>\n

Plus, I doubt humans would feel as comfortable playing ping pong against a generic mass of fast-moving and distinctly non-humanoid metal as they would playing against little HITTER.<\/p>\n

\"Secretlab<\/p>\n

Best PC gaming kit 2025<\/p>\n

All our favorite gear<\/p>\n","protected":false},"excerpt":{"rendered":"HITTER: A HumanoId Table TEnnis Robot via Hierarchical Planning and Learning – YouTube Watch On I tend to…\n","protected":false},"author":2,"featured_media":131166,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-131165","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/131165","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=131165"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/131165\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/131166"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=131165"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=131165"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=131165"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}