Robot Videos: Baseball Bots, UBTECH Walker, and More

Video Friday is your weekly number of superior robotics movies, collected by your pals at IEEE Spectrum robotics. We additionally put up a weekly calendar of upcoming robotics occasions for the following few months. Please send us your events for inclusion.

SOSV Robotics Matchup: 1–5 December 2025, ONLINE

ICRA 2026: 1–5 June 2026, VIENNA

Get pleasure from in the present day’s movies!

Researchers on the RAI Institute have constructed a low-impedance platform to review dynamic robot manipulation. On this demo, robots play a sport of catch and take part in batting observe, each with one another and with expert people. The robots are able to throwing 70mph [112 kph], approaching the velocity of a robust highschool pitcher. The robots can catch and bat at brief distances (23 ft [7 m]) requiring fast response occasions to catch balls thrown at as much as 41 mph [66kph] and hit balls pitched at as much as 30 mph [48kph].

That’s a pleasant contact with the customized “RAI” baseball gloves, however what I actually wish to know is how lengthy a pair of robots can preserve themselves entertained.

[ RAI Institute ]

This week’s finest bacronym winner is GIRAF: Drastically Elevated Attain AnyMAL Perform. And if that arm appears like magic, that’s as a result of it’s, though with some cautious pausing of the video you’ll be capable of see the way it works.

[ Stanford BDML ]

DARPA concluded the second 12 months of the DARPA Triage Challenge on October 4, awarding high marks to DART and MSAI in Programs and Information competitions, respectively. The three-year prize competitors goals to revolutionize medical triage in mass casualty incidents the place medical sources are restricted.

[ DARPA ]

We suggest a robotic agnostic reward operate that balances the achievement of a desired finish pose with impression minimization and the safety of crucial robotic components throughout reinforcement learning. To make the coverage sturdy to a broad vary of preliminary falling circumstances and to allow the specification of an arbitrary and unseen finish pose at inference time, we introduce a simulation-based sampling technique of preliminary and finish poses. By means of simulated and real-world experiments, our work demonstrates that even bipedal robots can carry out managed, gentle falls.

[ Moritz Baecher ]

Oh look, extra humanoid acrobatics.

My prediction: as soon as humanoid firms run out of mocapped dance strikes, we’ll begin seeing some freaky stuff that leverages the levels of freedom that robots have and people don’t. You heard it right here first, people.

[ MagicLab ]

I problem the following firm that makes a “lights-out” video to only lower to only a completely black display screen with a bit of “Profitable Picks” counter within the nook that simply goes up and up and up.

[ Brightpick ]

Thanks, Gilmarie!

The terrain stuff is cool and all however can we simply discuss concerning the trailer as a substitute?

[ LimX Dynamics ]

Presumably very choosy German birblets are getting customized nesting packing containers manufactured with excessively excessive precision by robots.

[ TUM ]

All these UBTECH Walker S2 robots weren’t pretend, it seems.

[ UBTECH ]

That is extra automation than what we’d actually be pondering of as robotics at this level, however I may nonetheless watch all of it day.

[ Motoman ]

Brad Porter (Cobot) and Alfred Lin (Sequoia Capital) talk about the way forward for robotics, AI, and automation on the Human[X] Convention, moderated by CNBC’s Kate Rooney. They discover why collaborative robots are accelerating now, how AI is remodeling bodily techniques, the position of humanoids, labor market shifts, and the funding traits shaping the following decade of robotics.

[ Cobot ]

Humanoid robots have lengthy captured our creativeness. Curiosity has skyrocketed together with the notion that robots are getting nearer to taking up a variety of labor-intensive duties. On this dialogue, we mirror on what we’ve discovered by observing manufacturing unit flooring, and why we’ve grown satisfied that chasing generalization in manipulation—each in {hardware} and conduct—isn’t simply fascinating, however needed. We’ll talk about AI analysis threads we’re exploring at Boston Dynamics to push this mission ahead, and spotlight alternatives our discipline ought to collectively make investments extra in to show the humanoid imaginative and prescient, and the reinvention of producing, right into a sensible, economically viable product.