0:00:00.929,0:00:05.929
As a roboticist,[br]I get asked a lot of questions.

0:00:05.929,0:00:08.909
"When we will they start[br]serving me breakfast?"

0:00:08.909,0:00:14.066
So I thought the future of robotics[br]would be looking more like us.

0:00:14.066,0:00:18.139
I thought they would look like me,

0:00:18.139,0:00:22.214
so that I built eyes[br]that would simulate my eyes,

0:00:22.214,0:00:28.882
I built fingers that are dextrous enough[br]to serve me baseballs.

0:00:32.075,0:00:34.968
Classical robots like this are built

0:00:34.968,0:00:40.979
and can function based on[br]the fixed number of joints and actuators,

0:00:40.979,0:00:44.157
and this means their[br]functionality and shape

0:00:44.157,0:00:47.376
are already fixed at the moment[br]of their conception.

0:00:47.697,0:00:50.587
So even though this arm[br]has really nice throw,

0:00:50.587,0:00:53.614
to even hit the tripod at the end,

0:00:54.182,0:00:57.646
it's not meant for cooking you[br]breakfast per se.

0:00:57.646,0:01:01.075
It's not really suited for scrambled eggs.

0:01:01.075,0:01:06.307
So this was when I was hit[br]by a new vision of future robotics:

0:01:06.307,0:01:09.355
the transformers.

0:01:09.355,0:01:11.720
They drive, they run, they fly,

0:01:11.720,0:01:17.435
all depending on the ever-changing[br]new environment and task at hand.

0:01:17.601,0:01:19.405
To make this a reality,

0:01:19.405,0:01:23.822
you really have to rethink[br]how robots are designed.

0:01:23.822,0:01:27.637
So imagine a robotic module[br]in a polygon shape

0:01:27.637,0:01:29.679
and using that simple polygon shape[br]to reconstruct multiple different forms

0:01:29.679,0:01:33.174
to create a new form of robot[br]for different tasks.

0:01:33.174,0:01:41.348
In CG, computer graphics,[br]it's not only news,

0:01:41.348,0:01:44.864
it's been done for a while, and that's how[br]most of the movies are made.

0:01:44.864,0:01:48.757
But if you're trying to make a robot[br]that's physically moving,

0:01:48.757,0:01:50.647
it's a completely new story.

0:01:50.898,0:01:54.118
It's a completely new paradigm.

0:01:54.118,0:01:57.071
But you've all done this.

0:01:57.071,0:02:01.569
Who hasn't made a paper airplane,[br]paper boat, paper crane.

0:02:01.569,0:02:07.937
Origami is a versatile[br]platform for designers.

0:02:07.937,0:02:12.142
From a single sheet of paper,[br]you can make multiple shapes,

0:02:12.142,0:02:16.114
and if you don't like it,[br]you unfold and fold back again.

0:02:16.114,0:02:22.194
Any 3D form can be made[br]from 2D surfaces by folding,

0:02:22.194,0:02:25.005
and this is proven mathematically.

0:02:26.762,0:02:31.027
And imagine if you were to have[br]an intelligent sheet

0:02:31.027,0:02:34.793
that can self-fold into any form it wants

0:02:34.793,0:02:36.323
anytime.

0:02:37.386,0:02:38.977
And that's what I've been working on,

0:02:38.977,0:02:41.926
and I call this robotic origami

0:02:41.926,0:02:43.909
"robogami."

0:02:45.617,0:02:49.019
This is our first robogami transformation

0:02:49.019,0:02:52.351
that was made by me about 10 years ago.

0:02:52.351,0:02:54.055
From a flat-sheeted robot,

0:02:54.055,0:02:57.200
it turns into a pyramid[br]and back into a flat sheet

0:02:57.200,0:03:00.392
and into a space shuttle.

0:03:00.614,0:03:02.726
Quite cute.

0:03:02.980,0:03:09.029
Ten years later, with my group[br]of ninja origami robotic researchers,

0:03:09.029,0:03:12.626
and there about 22 of them right now,

0:03:12.626,0:03:15.770
we have a new generation of robogamis,

0:03:16.057,0:03:19.368
and they're a little more effective[br]and we do more than that.

0:03:20.320,0:03:23.756
So the new generation of robogamis[br]actually serve a purpose.

0:03:23.756,0:03:28.826
For example, this one actually navigates[br]through different terrains autonomously.

0:03:28.826,0:03:31.143
So when it's a dry[br]and flat land, it crawls.

0:03:31.143,0:03:37.511
And if it meets some rough terrain,[br]it starts rolling.

0:03:38.201,0:03:40.573
It does this, it's the same robot,

0:03:40.573,0:03:43.619
but depending on which terrain it meets,

0:03:43.619,0:03:49.621
it activates a different sequence[br]of actuators that's on board.

0:03:50.706,0:03:54.434
And once it meets and obstacle,[br]it jumps over it.

0:03:54.434,0:03:58.748
It does this by storing energy[br]in each of its legs

0:03:58.748,0:04:03.003
and releasing it and catapulting[br]like a slingshot.

0:04:03.003,0:04:05.752
And it even does gymastics.

0:04:05.752,0:04:07.041
Yay.

0:04:08.753,0:04:13.222
So I just showed you[br]what a single robogami can do.

0:04:13.222,0:04:15.793
Imagine what they can do as a group.

0:04:15.793,0:04:20.047
They can join forces to tackle[br]more complex tasks.

0:04:20.047,0:04:23.186
Each module either active or passive,

0:04:23.186,0:04:26.476
we can assemble them[br]to create different shapes.

0:04:26.476,0:04:30.643
Not only that, by controlling[br]the folding joints, we are able

0:04:30.643,0:04:34.189
to create and attack different tasks.

0:04:34.189,0:04:37.156
The form is making new task space.

0:04:37.880,0:04:42.247
And this time, what is most[br]important is the assembly.

0:04:42.438,0:04:46.559
They need to autonomously[br]find each other in a different space,

0:04:46.559,0:04:50.712
attach and detach depending on[br]the environment and task.

0:04:51.950,0:04:55.174
And we can do this now.

0:04:55.174,0:04:56.624
So what's next?

0:04:56.624,0:04:57.958
Our imagination.

0:04:57.958,0:05:01.889
So this is a simulation of what[br]you can achieve with this type of module.

0:05:01.889,0:05:05.413
So we decided that we were going[br]to have a four-legged crawler

0:05:05.413,0:05:10.471
turning into a little dog[br]and making small gaits.

0:05:10.471,0:05:14.210
With the same module, we can actually[br]make it do something else:

0:05:14.210,0:05:17.452
a manipulator, a typical[br]classic robot task.

0:05:17.651,0:05:20.320
So with a manipulator,[br]it can pick an object.

0:05:20.320,0:05:24.164
Of course, you can add more modules[br]to make the manipulator legs longer,

0:05:24.164,0:05:27.133
to attack or pick up[br]objects that are bigger or smaller,

0:05:27.133,0:05:30.065
or even have a third arm.

0:05:31.752,0:05:35.312
For robogamis, there's no[br]one fix shape nor task.

0:05:35.312,0:05:40.492
They can transform into anything[br]anywhere anytime.

0:05:42.615,0:05:45.188
So how do you make them?

0:05:45.188,0:05:50.576
The biggest technical challenge[br]of robogami is keeping them super-thin,

0:05:50.576,0:05:53.585
flexible, but still remaining functional.

0:05:54.737,0:05:58.517
They're composed of multiple layers[br]of circuits, motors,

0:05:58.517,0:06:01.532
micro-controllers and sensors,[br]all in the single body,

0:06:01.532,0:06:06.323
and when you control[br]individual folding joints,

0:06:06.323,0:06:09.887
you'll be able to achieve[br]soft motions like that

0:06:10.043,0:06:12.158
upon your command.

0:06:14.347,0:06:19.056
Instead of being a single robot that is[br]specifically made for a single task,

0:06:19.056,0:06:23.620
robogamis are optimized[br]to do multi-tasks,

0:06:23.620,0:06:26.669
and this is quite important

0:06:26.669,0:06:29.351
for the difficult and unique[br]environments on the Earth

0:06:29.351,0:06:32.883
as well as in space.

0:06:33.782,0:06:37.745
Space is a perfect[br]environment for robogamis.

0:06:37.745,0:06:42.369
You cannot afford to have[br]one robot for one task.

0:06:43.347,0:06:47.045
Who knows how many tasks[br]you will encounter in space?

0:06:47.045,0:06:54.175
What you want is a single robotic platform[br]that can transform to do multi-tasks.

0:06:55.371,0:07:00.599
What we want is a deck[br]of thin robogami modules

0:07:00.599,0:07:06.017
that can transform to do multiples[br]of performing tasks.

0:07:06.552,0:07:10.418
So here, and don't take my word for it,

0:07:10.418,0:07:13.418
because it's the European Space Agency[br]and Swiss Space Center

0:07:13.418,0:07:15.821
are sponsoring this exact concept.

0:07:16.007,0:07:20.766
So here you see a couple of images[br]of reconfiguration of robogamis,

0:07:20.766,0:07:24.634
exploring the foreign land[br]aboverground, on the surface,

0:07:24.634,0:07:27.281
as well as digging into the surface.

0:07:27.483,0:07:29.478
It's not just exploration.

0:07:29.478,0:07:32.089
For astronauts, they need additional help,

0:07:32.089,0:07:35.077
because you cannot afford[br]to bring interns up there either.

0:07:36.589,0:07:39.589
They have to do every tedious task.

0:07:39.589,0:07:43.048
It may be simple, but super-interactive.

0:07:43.048,0:07:46.483
So you need robots[br]to facilitate their experiments,

0:07:46.483,0:07:49.196
assisting them with the communications,

0:07:49.196,0:07:54.749
and just docking onto surfaces to be[br]their third arm holding different tools.

0:07:55.006,0:07:59.588
But how will they be able[br]to control robogamis, for example,

0:07:59.588,0:08:00.698
outside the space station?

0:08:00.698,0:08:04.224
In this case, I show a robogami[br]that is holding a space debris.

0:08:04.224,0:08:07.779
You can work with your vision[br]so that you can control them,

0:08:07.779,0:08:12.382
but what would be better[br]is having the sensation of touch

0:08:12.382,0:08:16.392
directly transported onto[br]the hands of the astronauts.

0:08:16.392,0:08:19.152
And what you need is a haptic device,

0:08:19.152,0:08:22.860
a haptic interface that recreates[br]the sensation of touch.

0:08:23.234,0:08:26.267
And using robogamis, we can do this.

0:08:27.515,0:08:32.546
This is the world's[br]smallest haptic interface

0:08:32.546,0:08:38.391
that can recreate a sensation of touch[br]just underneath your fingertip.

0:08:38.391,0:08:41.115
We do this by moving the robogami

0:08:41.115,0:08:44.200
By microscopic and macroscopic[br]movements at the stage,

0:08:44.200,0:08:49.951
and by having this, not only[br]will you be able to feel

0:08:49.951,0:08:51.795
how big the object is,

0:08:51.795,0:08:54.465
the roundness and the lines,

0:08:54.465,0:08:56.987
but also the stiffness and the texture.

0:08:56.987,0:09:03.293
Alex has this interface[br]just underneath his thumb,

0:09:03.293,0:09:08.043
and if he were to use this[br]with a VR goggles and hand controllers,

0:09:08.043,0:09:11.590
now the virtual reality[br]is no longer virtual.

0:09:11.812,0:09:14.369
It becomes a tangible reality.

0:09:16.704,0:09:20.877
The blue ball, red ball,[br]and black ball that he's looking

0:09:20.877,0:09:23.175
is no longer differentiated by colors.

0:09:23.175,0:09:27.714
Now it is a rubber blue ball,[br]sponge red ball, and billiard black ball.

0:09:28.944,0:09:31.044
This is now possible.

0:09:31.581,0:09:33.339
Let me show you.

0:09:34.381,0:09:38.294
This is really the first time[br]this is shown live

0:09:38.294,0:09:42.317
in front of a public grand audience,

0:09:42.317,0:09:44.169
so hopefully this works.

0:09:44.169,0:09:48.343
So what you see here[br]is an atlas of anatomy,

0:09:48.343,0:09:51.033
and the robogami haptic interface.

0:09:51.033,0:09:54.032
So, like all the other[br]reconfigurable robots,

0:09:54.032,0:09:54.999
it multitasks.

0:09:54.999,0:09:57.305
Not only is it going to serve as a mouse,

0:09:57.305,0:09:59.540
but also a haptic interface.

0:09:59.540,0:10:05.336
So for example, we have a white background[br]where there is no object.

0:10:05.336,0:10:06.632
That means there is nothing to feel,

0:10:06.632,0:10:08.828
so we can have a very,[br]very flexible interface.

0:10:09.622,0:10:12.797
Now I use this as a mouse[br]to approach a skin,

0:10:12.797,0:10:16.343
a muscular arm,[br]so now let's feel his biceps,

0:10:16.343,0:10:17.672
or shoulders.

0:10:17.854,0:10:20.672
So now you see how stiffer it becomes.

0:10:20.672,0:10:22.066
Let's explore even more.

0:10:22.066,0:10:24.730
Let's approach the ribcage,

0:10:24.730,0:10:27.531
and as soon as I move[br]on top of the ribcage

0:10:27.531,0:10:29.741
and between the ?? muscles,

0:10:29.741,0:10:33.711
which is softer and harder, I can feel[br]the difference of the stiffness.

0:10:33.711,0:10:34.998
Take my word for it.

0:10:34.998,0:10:38.725
So now you see it's much stiffer[br]in terms of the force

0:10:39.233,0:10:41.559
it's giving back to my fingertip.

0:10:41.997,0:10:45.968
So I showed you the surfaces[br]that's not moving.

0:10:45.968,0:10:49.401
How about if I were to approach[br]something that moves,

0:10:49.401,0:10:51.911
for example like a beating heart?

0:10:51.911,0:10:53.666
What would I feel?

0:10:59.573,0:11:02.957
(Applause)

0:11:07.381,0:11:10.217
This can be your beating heart.

0:11:10.217,0:11:13.748
This can actually be inside your pocket

0:11:13.748,0:11:16.504
while you're shopping online.

0:11:16.504,0:11:20.316
Now you'll be able to feel the difference[br]of the sweater that you're buying,

0:11:20.316,0:11:21.647
how soft it is,

0:11:21.647,0:11:24.181
if it's actually cashmere or not,

0:11:24.181,0:11:26.550
or the bagel that you're trying to buy,

0:11:26.550,0:11:28.162
how hard it is or how crispy it is.

0:11:28.162,0:11:32.106
This is now possible.

0:11:34.996,0:11:39.802
The robotics technology is advancing[br]to be more personalized and adaptive,

0:11:39.802,0:11:44.516
to adapt to our everyday needs.

0:11:44.728,0:11:48.499
This unique specie[br]of reconfigurable robotics

0:11:48.499,0:11:54.276
is actually the platform to provide[br]this invisible, intuitive interface

0:11:54.276,0:11:57.277
to meet our exact needs.

0:11:58.455,0:12:03.129
These robots will no longer look like[br]the characters from the movies.

0:12:03.129,0:12:07.150
Instead, they will be whatever[br]you want them to be.

0:12:07.493,0:12:09.398
Thank you.

0:12:09.398,0:12:12.996
(Applause)