- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
The robots were just a little too slick.
You must log in or register to comment.
A robot isn’t just software: the hardware for humanoid robots has proved quite difficult to build and if these robots have the hardware necessary to walk around and manipulate objects (in the real world, not in a lab where they get multiple “takes”) then they are remarkable even if their every action is directed by a human.
Also I would have guessed that Hasbro owned the IP for robots named “Optimus”. Maybe Tesla paid them, the way that Verizon paid Lucasfilm for the right to use the word “droid”.
While it is impressive, it’s impressive in a “college robotics classes were doing this six years ago” way. The control loop would be interesting with how the human gesticulation interacts with the locomotion algorithm they’re using, but that assumes that the robots can walk and chew bubblegum (or in this case, walk and interact with a crowd via teleoperation) at the same time. I wouldn’t be surprised to learn that the robots were entirely software controlled when walking. That seems like the case from what I’m seeing in these vids.
It has been a while since I went to college but I don’t remember college robotics classes building robots capable of bipedal motion at all.
Not trying to be contrary here: have you taken any robotics courses? We do walking algorithms for our sophomore classes now, as an intro to complex multivariate simulation. Desktop sized, but motion tracking for telemetry control of end effectors is a common enough student project that we’ve started requiring people to modify their project proposals so they’re not just implementing some online tutorial by rote. (SLAM on hexapods is another classic undergrad project, though one that doesn’t have nearly as much direct applicability to the topic at hand)
