Skip to main content

How Disney Designed a Robotic Spider-Man

The stunts we love to watch in movies like 'Spider-Man: Homecoming' usually rely on a combination of green screen, stunt performers, and computer-generated models. Stunts like these can be dangerous, especially when performed live...over and over. At the new Avengers Campus at Disneyland Resorts, radical stunts are being performed everyday by a robotic acrobat. Disney Imagineers Tony Dohi and Morgan Pope sit down with WIRED to talk about how they designed this amazing stuntronic robot.

Released on 10/01/2021

Transcript

[Narrator 1] The stunts we love to watch

in movies like Spider-Man: Homecoming,

usually rely on a mix

of stunt performers,

green screen

and computer generated models.

[Narrator 2] Stunts like these,

where Spider-Man flies through the air

can be dangerous,

especially if they're being performed live

over, and over, and over again.

But at the new adventures campus

at Disneyland Resorts,

these stunts are being done everyday

by this guy, a robotic acrobat.

[Narrator 1] It's part of a Centronics technology system

developed by a team of roboticist and engineers.

Wired spoke with Imagineers, Tony Dohi and Morgan Pope

to discover what it took to design,

launch, and catch

the stuntronic acrobat.

[music swells then fades]

[Narrator 1] While this fine Spider-Man may look human,

it's actually a complex robotic system

covered by a 3D printed shell.

Ultimately, the two driving design things

for the Stuntronic robot

were this idea

of robustness

and grace.

It also had to communicate

the fluidity of a human performance.

It had to be believable and alive.

[Tony] Yeah. It had to look like Spider-Man.

So, Tony and I were working

on kind of parallel paths.

He had this idea for throwing a robot across the room,

and meanwhile, over on the research side,

I was working on how do you control something

as it's free-falling through space.

[Narrator 2] To tackle the challenge of creating

a totally controllable robotic system

that mimics the uncontrollability of flight,

they started their design with this,

The Brick.

[Morgan] The idea here was

that we could spin it to the air,

and it had these weights inside that could move.

[Tony] It didn't look like a character,

but it had all the intelligence in it.

It had sensors,

uh

it knew how high it was from the ground.

[Narrator 1] Their next prototype

didn't look much like a character either.

So, this right over here, is our prototype.

This actually has that bend that's a lot more.

This is how a human shifts inertia, right, kind of do

tuck and do a flip and lay out.

And so from that, we moved really quickly.

Maybe the next day we started building Stickman.

[Tony] It's actually a Z-shaped series of linkages.

[Morgan] It's like the engineering version

of a Lincoln logs or Legos.

And you get this kind of double pendulum,

which is like a classically chaotic system,

which means it's kind of a bear to control,

but it also means you can do really fun stuff.

If you squint at it starts to look like a human.

This is the first time

we could get asymmetric motion, right?

This guy can move this arm

and not this arm,

and do all sorts of like twists and stuff.

[Narrator 2] Then, they moved to half scale figures.

[Narrator 1] And finally their series

of full-sized stuntronic robots.

The weigh about 95 pounds.

Mhmm

And their height,

I would say, is

5'9.

[Morgan] We constructed them out of mainly

3D printed plastic and aluminum,

[Both] and a lot of screws.

Yeah, right, a lot of screws.

[transition music]

So, we started with a 40 foot high throw

Right.

Not even that. Yeah, it was like 40, 45,

which was twice what we were doing inside.

So we started to, to tune up power on the winch

and we kept throwing this higher and higher

and higher until we pretty much peaked out the winch.

Yeah, around 65 ft

Yeah throws about 65 feet high in the air.

And there was this kind of magical moment where

I think it was it's around 55 feet.

It feels like it shouldn't it have come down by now.

Like, you know, like it's, it feels like it's floating.

That was a cool moment.

Then internally,

the robot uh keeps track of its position

using the same basic sensors that are in your phone

uses an accelerometer and gyroscope.

So the same thing that tells you

if it's portrait or landscape.

The only external sensor I can say we have

is one that's really more tied into the show control system.

And it's the anemometers

because we really need to be very,

very aware of the wind speeds.

[Morgan] We took measurements of the weather

at Disneyland for like a year,

so that we can be really conservative

about how the robot flies through the air

and make sure that we always hit the net.

[Narrator 1] Creating the illusion of Peter Parker,

as Spider-Man flying through the air

was an ongoing design

and engineering challenge.

Robots are designed to be precise and they're not designed

to, to make mistakes and be clumsy or, or show panic.

[Morgan] Right.

So the question is now, how can we capture Spider-Man

out of control flying through the air?

[Morgan] Oh, that was a really fun challenge

[Tony] That took,

that just took so many iterations to get that dialed in.

It's a lot of trying.

When we were indoors, we had a whole motion capture system

set up, so that we could

validate that our sensors were actually giving us precision

to within a few centimeters.

Which is the kind of position we needed

to do what we wanted to do.

So I think as human beings,

we understand like

things spinning in one plane pretty well.

But when you start flipping in 3D space,

things get weird physics wise.

For instance,

if I'm doing a front flip like this,

and I have my arms up,

if I throw my arm down like this,

I'm going to tip a little bit.

But then I'm also out of nowhere,

I'm going to start like twisting around this axis.

Now you've gone from just a straight front flip to

a twisted, slightly tilted, tumbling.

And the, the, the physics you can ride on one line,

it's a very simple equation,

but then what pops out of it is so counterintuitive

and honestly so beautiful.

And I think that's kind of the fun part about how,

you know, moving from

the brick where it was very much in two dimensions

to this, to the more complicated

where you've got cross products of inertia.

You've got this more complicated human shaped object

that can do all these weird things in 3D dynamics.

And that suddenly made it feel really alive.

[music playing]

[Tony] The catchment system is very, very specific.

So not only do

we have to be able to catch the figure,

decelerate it very quickly,

but that net also has to be robust enough

to do this over and over and over again,

because we don't want to replace it every show.

The net area is only about 10 feet by 10 feet.

[Morgan] I think 14, which sounds big.

But then when you get on the roof and if you like look from

the position of the robot, 65 feet in the air,

it doesn't feel very big.

So another good reason we don't do this with people.

And it has a deceleration system.

That's, it's just this beautifully simplistic

in its design.

[Morgan] Yep.

[Tony] But it's also very quickly resettable.

The robot was designed

to actually have some breakaway linkages

Right.

So that if we

uh land in a funny way,

for instance,

uh we're only gonna break

a 3D printed part

that's designed to snap

so that we don't transfer that impact to the more

delicate or more expensive parts, like the servos.

That, that actually has been a cool thing.

We've done some pretty violent things to our robots

in the course of testing.

You design it so that if it breaks, it's not a big deal.

It goes all the way back to Stickman.

We're always thinking

of what the next thing might be.

[Morgan] I think ultimately we designed this system

with the hopes of it being flexible and adaptable,

and there are so many dynamic characters

in the Disney Pantheon,

and we hope that we can deliver more of them for real.

[Tony] Centronics isn't a robot,

it's a category of,

of stunt robots.

So, we are

really hoping that we just scratched

the surface with this one

and that we can keep taking it as far as we can.

Yeah. Yeah. Fingers crossed.

I think it would be really fun to see more.

[music fading]

Up Next