Disney Research: Acrobat and Stunt Double Robots

the.dreamfinder

Well-Known Member
Original Poster
From CNET
Disney's robot acrobat Stickman could flip its way into the circus
Cirque du Machine?
Mike SorrentinoMay 22, 2018 10:49 AM PDT
These acrobatic robots can do flips and untuck for a safe landing.
Disney Research Hub
Acrobats beware. Disney's limber new armless robots may out-flip you one day.
The Disney Research Hub published its work Tuesday on Stickman, an acrobatic robot that can swing through the air, do flips and then untuck its body for a safe landing.

The robots use a processor, gyroscopes and a lot of data to predict how to safely pull off their high-flying stunts, which include a single backflip, a double backflip and a graceful free-fall.
Here's Stickman as he flies through the air, seemingly with the greatest of ease -- until you think of all the research that went into it.
The robot can sense when it's the right time to tuck its body.
Disney Reserach Hub
And if you want to dive into that research, the team behind Stickman have published their findings for you to read here.

EDIT: Updated title to reflect the public reveal of the stunt double AAs and to dissuade duplicat threads.
 
Last edited:

GlacierGlacier

Well-Known Member
Looks great! I wonder though if it will find its way into the parks?

Anybody know if Disney Research projects generally move forward to actual application?
They were the ones behind the transparent LCD screens used in Na'vi, as well as a lot of tech prototyped for Pandora (interactive plants and floor stuff) before that was scrapped. They've published quite a bit on the software logic that is used for linear storytelling (i.e. star wars land). It's hard, because they make an absolute ton of stuff and disney hasn't really built a ton of stuff recently.
 

winstongator

Well-Known Member
You know they are doing all this research so that in say 20 years these can replace real people performing in the parks and save on wages and union disputes lol
Would they really be cheaper? At least in the short term, their cost is going to be high. Just say $1M. That buys a lot of CM time! How long will they operate on a charge? How many days can they work continuously? How difficult is maintenance?

Well before that, they'll have self-driving trams, buses and Minnie Vans. However, how would Disney be portrayed if they were using bleeding edge driving tech and a guest were injured by a tram/bus/van? Would be worse than if someone were injured on a ride due to reduced maintenance for cost control.

Concerns over technology obsoleting jobs is well placed, but I'd worry about the driving tech, inside and out of the parks, before worrying about humanoid robots replacing CMs.
 

Bairstow

Well-Known Member
This one's been stewing a while. Previous research in the area even had an evolutionary bent... the computer designed the robots, and then they were evaluated for fitness based on locomotor success.
I think I remember seeing one of those reports. Is that the one where the computer "invented" the little shuffling robot with a pivoting "walking stick" leg?
 

GlacierGlacier

Well-Known Member
I think I remember seeing one of those reports. Is that the one where the computer "invented" the little shuffling robot with a pivoting "walking stick" leg?
The computer did not invent the robot. However, the computer tried anywhere from dozens to hundreds of ways to get the robot from point A to point B, tweaking the most effective runs to get the best out of the robot. In the video, they showed that the computer managed to find a more efficient locomotive pattern in nearly every scenario compared the the pattern programmed by humans.
 

Bairstow

Well-Known Member
The computer did not invent the robot. However, the computer tried anywhere from dozens to hundreds of ways to get the robot from point A to point B, tweaking the most effective runs to get the best out of the robot. In the video, they showed that the computer managed to find a more efficient locomotive pattern in nearly every scenario compared the the pattern programmed by humans.

We're thinking of different studies, then.
The one I remember reading documented the AI coming up with the idea of a hinged, unpowered leverage limb, essentially on its own, to provide a point of leverage for its other articulated limbs. The study's designers had only anticipated the AI utilizing articulated, powered limbs. The AI's "invention" of a completely unanticipated solution was one of the more celebrated findings of the experiment.
 

Register on WDWMAGIC. This sidebar will go away, and you'll see fewer ads.

Back
Top Bottom