TED2017
Marc Raibert: Meet Spot, the robot dog that can run, hop and open doors
Filmed:
Readability: 3.8
4,082,182 views
That science fiction future where robots can do what people and animals do may be closer than you think. Marc Raibert, founder of Boston Dynamics, is developing advanced robots that can gallop like a cheetah, negotiate 10 inches of snow, walk upright on two legs and even open doors and deliver packages. Join Raibert for a live demo of SpotMini, a nimble robot that maps the space around it, handles objects, climbs stairs -- and could soon be helping you out around the house.
Marc Raibert - Roboticist
Marc Raibert is the founder and CEO of robot maker Boston Dynamics. Full bio
Marc Raibert is the founder and CEO of robot maker Boston Dynamics. Full bio
Double-click the English transcript below to play the video.
00:19
(Laughter)
0
7265
1800
00:24
(Laughter)
1
12503
1150
00:36
That's SpotMini.
2
24890
1150
00:38
He'll be back in a little while.
3
26064
1618
00:39
I --
4
27706
1164
00:40
(Applause)
5
28894
3838
00:45
I love building robots.
6
33652
1737
00:48
And my long-term goal is to build robots
7
36599
2452
00:51
that can do what people and animals do.
8
39075
2082
00:54
And there's three things in particular
9
42144
3099
00:57
that we're interested in.
10
45721
1703
01:00
One is balance and dynamic mobility,
11
48188
3302
01:03
the second one is mobile manipulation,
12
51514
2630
01:06
and the third one is mobile perception.
13
54168
2365
01:09
So, dynamic mobility and balance --
14
57124
2911
01:12
I'm going to do a demo for you.
15
60059
1637
01:15
I'm standing here, balancing.
16
63279
1547
01:18
I can see you're not very impressed.
OK, how about now?
OK, how about now?
17
66301
2714
01:21
(Laughter)
18
69039
1160
01:22
How about now?
19
70223
1193
01:23
(Applause)
20
71440
2043
01:26
Those simple capabilities mean that people
can go almost anywhere on earth,
can go almost anywhere on earth,
21
74229
4373
01:30
on any kind of terrain.
22
78626
1652
01:32
We want to capture that for robots.
23
80302
2655
01:35
What about manipulation?
24
83952
1392
01:37
I'm holding this clicker in my hand;
25
85794
1716
01:39
I'm not even looking at it,
26
87534
1292
01:40
and I can manipulate it
without any problem.
without any problem.
27
88850
2730
01:43
But even more important,
28
91604
2640
01:46
I can move my body while I hold
the manipulator, the clicker,
the manipulator, the clicker,
29
94268
5160
01:52
and stabilize and coordinate my body,
30
100295
2174
01:54
and I can even walk around.
31
102493
1522
01:56
And that means
I can move around in the world
I can move around in the world
32
104547
2898
01:59
and expand the range
of my arms and my hands
of my arms and my hands
33
107469
3302
02:02
and really be able to handle
almost anything.
almost anything.
34
110795
2169
02:04
So that's mobile manipulation.
35
112988
1592
02:07
And all of you can do this.
36
115634
1775
02:09
Third is perception.
37
117433
1917
02:11
I'm looking at a room
with over 1,000 people in it,
with over 1,000 people in it,
38
119374
3265
02:14
and my amazing visual system
can see every one of you --
can see every one of you --
39
122663
4775
02:19
you're all stable in space,
40
127462
1961
02:21
even when I move my head,
41
129447
1319
02:22
even when I move around.
42
130790
1546
02:24
That kind of mobile perception
is really important for robots
is really important for robots
43
132360
3779
02:28
that are going to move and act
44
136163
1644
02:29
out in the world.
45
137831
1213
02:32
I'm going to give you
a little status report
a little status report
46
140045
2278
02:34
on where we are in developing robots
toward these ends.
toward these ends.
47
142347
3846
02:40
The first three robots are all
dynamically stabilized robots.
dynamically stabilized robots.
48
148518
4540
02:45
This one goes back
a little over 10 years ago --
a little over 10 years ago --
49
153082
2648
02:47
"BigDog."
50
155754
1168
02:48
It's got a gyroscope
that helps stabilize it.
that helps stabilize it.
51
156946
3310
02:52
It's got sensors and a control computer.
52
160280
3336
02:55
Here's a Cheetah robot
that's running with a galloping gait,
that's running with a galloping gait,
53
163640
3167
02:58
where it recycles its energy,
54
166831
1696
03:00
it bounces on the ground,
55
168551
1567
03:02
and it's computing all the time
56
170142
1554
03:03
in order to keep itself
stabilized and propelled.
stabilized and propelled.
57
171720
2887
03:08
And here's a bigger robot
58
176114
1828
03:09
that's got such good
locomotion using its legs,
locomotion using its legs,
59
177966
2705
03:12
that it can go in deep snow.
60
180695
1404
03:14
This is about 10 inches deep,
61
182123
2411
03:16
and it doesn't really have any trouble.
62
184558
2111
03:20
This is Spot, a new generation of robot --
63
188346
2629
03:22
just slightly older than the one
that came out onstage.
that came out onstage.
64
190999
2808
03:26
And we've been asking the question --
65
194438
1832
03:28
you've all heard about drone delivery:
66
196294
1974
03:30
Can we deliver packages
to your houses with drones?
to your houses with drones?
67
198292
2623
03:32
Well, what about plain old
legged-robot delivery?
legged-robot delivery?
68
200939
2787
03:35
(Laughter)
69
203750
1132
03:36
So we've been taking our robot
to our employees' homes
to our employees' homes
70
204906
3397
03:40
to see whether we could get in --
71
208327
1583
03:41
(Laughter)
72
209934
1024
03:42
the various access ways.
73
210982
1228
03:44
And believe me, in the Boston area,
74
212234
1884
03:46
there's every manner
of stairway twists and turns.
of stairway twists and turns.
75
214142
3174
03:49
So it's a real challenge.
76
217340
1237
03:50
But we're doing very well,
about 70 percent of the way.
about 70 percent of the way.
77
218601
2817
03:54
And here's mobile manipulation,
78
222659
1565
03:56
where we've put an arm on the robot,
79
224248
2591
03:58
and it's finding its way through the door.
80
226863
2334
04:01
Now, one of the important things
about making autonomous robots
about making autonomous robots
81
229922
3159
04:05
is to make them not do
just exactly what you say,
just exactly what you say,
82
233105
3220
04:08
but make them deal with the uncertainty
of what happens in the real world.
of what happens in the real world.
83
236349
5141
04:14
So we have Steve there,
one of the engineers,
one of the engineers,
84
242234
3011
04:17
giving the robot a hard time.
85
245269
1608
04:18
(Laughter)
86
246901
1032
04:19
And the fact that the programming
still tolerates all that disturbance --
still tolerates all that disturbance --
87
247957
4190
04:24
it does what it's supposed to.
88
252171
1497
04:25
Here's another example,
where Eric is tugging on the robot
where Eric is tugging on the robot
89
253692
2791
04:28
as it goes up the stairs.
90
256507
1298
04:29
And believe me,
91
257829
1152
04:31
getting it to do what it's supposed to do
in those circumstances
in those circumstances
92
259005
3272
04:34
is a real challenge,
93
262301
1341
04:35
but the result is something
that's going to generalize
that's going to generalize
94
263666
2740
04:38
and make robots much more autonomous
than they would be otherwise.
than they would be otherwise.
95
266430
3871
04:43
This is Atlas, a humanoid robot.
96
271497
2431
04:46
It's a third-generation humanoid
that we've been building.
that we've been building.
97
274607
3905
04:50
I'll tell you a little bit
about the hardware design later.
about the hardware design later.
98
278851
2818
04:53
And we've been saying:
99
281693
1176
04:54
How close to human levels
of performance and speed could we get
of performance and speed could we get
100
282893
4423
04:59
in an ordinary task,
101
287340
1575
05:00
like moving boxes around on a conveyor?
102
288939
2469
05:04
We're getting up to about two-thirds
of the speed that a human operates
of the speed that a human operates
103
292111
5306
05:09
on average.
104
297441
1162
05:10
And this robot is using both hands,
it's using its body,
it's using its body,
105
298938
2929
05:13
it's stepping,
106
301891
1169
05:15
so it's really an example
of dynamic stability,
of dynamic stability,
107
303084
2727
05:17
mobile manipulation
108
305835
1379
05:19
and mobile perception.
109
307238
1611
05:22
Here --
110
310372
1173
05:24
(Laughter)
111
312085
1907
05:26
We actually have two Atlases.
112
314569
1610
05:28
(Laughter)
113
316663
1185
05:30
Now, everything doesn't go exactly
the way it's supposed to.
the way it's supposed to.
114
318378
3396
05:33
(Laughter)
115
321798
1702
05:38
(Laughter)
116
326498
1528
05:40
(Laughter)
117
328716
1871
05:45
And here's our latest robot,
called "Handle."
called "Handle."
118
333472
2544
05:48
Handle is interesting,
because it's sort of half like an animal,
because it's sort of half like an animal,
119
336637
4118
05:52
and it's half something else
120
340779
2314
05:55
with these leg-like things and wheels.
121
343117
2756
05:58
It's got its arms on
in kind of a funny way,
in kind of a funny way,
122
346321
3026
06:01
but it really does some remarkable things.
123
349371
2096
06:03
It can carry 100 pounds.
124
351491
3331
06:06
It's probably going to lift
more than that,
more than that,
125
354846
2059
06:08
but so far we've done 100.
126
356929
1778
06:10
It's got some pretty good
rough-terrain capability,
rough-terrain capability,
127
358731
2436
06:13
even though it has wheels.
128
361191
1428
06:18
And Handle loves to put on a show.
129
366061
2366
06:20
(Laughter)
130
368866
1358
06:24
(Applause)
131
372824
5129
06:30
I'm going to give you
a little bit of robot religion.
a little bit of robot religion.
132
378920
2998
06:34
A lot of people think that a robot
is a machine where there's a computer
is a machine where there's a computer
133
382536
4329
06:38
that's telling it what to do,
134
386889
1634
06:41
and the computer is listening
through its sensors.
through its sensors.
135
389080
2816
06:44
But that's really only half of the story.
136
392529
2404
06:46
The real story is
that the computer is on one side,
that the computer is on one side,
137
394957
3109
06:50
making suggestions to the robot,
138
398090
1944
06:52
and on the other side
are the physics of the world.
are the physics of the world.
139
400058
2492
06:54
And that physics involves gravity,
friction, bouncing into things.
friction, bouncing into things.
140
402958
4922
07:00
In order to have a successful robot,
141
408418
1835
07:02
my religion is that you have to do
a holistic design,
a holistic design,
142
410277
4591
07:06
where you're designing the software,
the hardware and the behavior
the hardware and the behavior
143
414892
3553
07:10
all at one time,
144
418469
1236
07:11
and all these parts really intermesh
and cooperate with each other.
and cooperate with each other.
145
419729
3702
07:15
And when you get the perfect design,
you get a real harmony
you get a real harmony
146
423455
3107
07:18
between all those parts
interacting with each other.
interacting with each other.
147
426586
2879
07:22
So it's half software and half hardware,
148
430449
2260
07:24
plus the behavior.
149
432733
1291
07:26
We've done some work lately
on the hardware, where we tried to go --
on the hardware, where we tried to go --
150
434878
3542
07:30
the picture on the left
is a conventional design,
is a conventional design,
151
438444
2477
07:32
where you have parts
that are all bolted together,
that are all bolted together,
152
440945
2900
07:35
conductors, tubes, connectors.
153
443869
2740
07:38
And on the right
is a more integrated thing;
is a more integrated thing;
154
446633
2049
07:40
it's supposed to look like
an anatomy drawing.
an anatomy drawing.
155
448706
2401
07:43
Using the miracle of 3-D printing,
156
451510
2510
07:46
we're starting to build parts of robots
157
454044
2637
07:48
that look a lot more
like the anatomy of an animal.
like the anatomy of an animal.
158
456705
2897
07:51
So that's an upper-leg part
that has hydraulic pathways --
that has hydraulic pathways --
159
459626
3344
07:54
actuators, filters --
160
462994
1930
07:56
all embedded, all printed as one piece,
161
464948
2380
07:59
and the whole structure is developed
162
467352
3202
08:02
with a knowledge of what the loads
and behavior are going to be,
and behavior are going to be,
163
470578
3036
08:05
which is available from data
recorded from robots
recorded from robots
164
473638
2903
08:08
and simulations and things like that.
165
476565
1809
08:10
So it's a data-driven hardware design.
166
478398
2937
08:13
And using processes like that,
167
481727
1726
08:15
not only the upper leg
but some other things,
but some other things,
168
483477
2244
08:17
we've gotten our robots to go from big,
behemoth, bulky, slow, bad robots --
behemoth, bulky, slow, bad robots --
169
485745
5140
08:22
that one on the right,
weighing almost 400 pounds --
weighing almost 400 pounds --
170
490909
3619
08:26
down to the one in the middle
which was just in the video,
which was just in the video,
171
494552
3109
08:29
weighs about 190 pounds,
172
497685
1563
08:31
just a little bit more than me,
173
499272
1697
08:32
and we have a new one,
174
500993
1486
08:34
which is working but I'm not
going to show it to you yet,
going to show it to you yet,
175
502503
2742
08:37
on the left,
176
505269
1162
08:38
which weighs just 165 pounds,
177
506455
1644
08:40
with all the same
strength and capabilities.
strength and capabilities.
178
508123
2266
08:42
So these things are really getting
better very quickly.
better very quickly.
179
510413
2730
08:46
So it's time for Spot to come back out,
180
514460
3262
08:49
and we're going to demonstrate
a little bit of mobility,
a little bit of mobility,
181
517746
3748
08:53
dexterity and perception.
182
521518
1600
08:55
This is Seth Davis,
who's my robot wrangler today,
who's my robot wrangler today,
183
523861
3764
08:59
and he's giving Spot
some general direction
some general direction
184
527649
3040
09:02
by steering it around,
185
530713
1826
09:04
but all the coordination
of the legs and the sensors
of the legs and the sensors
186
532563
3071
09:07
is done by the robot's computers on board.
187
535658
2509
09:10
The robot can walk
with a number of different gaits;
with a number of different gaits;
188
538723
3446
09:14
it's got a gyro,
189
542193
2159
09:16
or a solid-state gyro,
190
544376
1337
09:17
an IMU on board.
191
545737
1420
09:19
Obviously, it's got a battery,
and things like that.
and things like that.
192
547622
3027
09:23
One of the cool things
about a legged robot is,
about a legged robot is,
193
551532
2506
09:26
it's omnidirectional.
194
554062
1449
09:27
In addition to going forward,
it can go sideways,
it can go sideways,
195
555535
2773
09:31
it can turn in place.
196
559593
1440
09:36
And this robot
is a little bit of a show-off.
is a little bit of a show-off.
197
564662
2268
09:39
It loves to use its dynamic gaits,
198
567855
1960
09:41
like running --
199
569839
1158
09:43
(Laughter)
200
571021
1044
09:44
And it's got one more.
201
572089
1516
09:47
(Laughter)
202
575284
1830
09:50
Now if it were really a show-off,
it would be hopping on one foot,
it would be hopping on one foot,
203
578038
3236
09:53
but, you know.
204
581298
1168
09:54
Now, Spot has a set of cameras
here, stereo cameras,
here, stereo cameras,
205
582490
4043
09:58
and we have a feed up in the center.
206
586557
1830
10:00
It's kind of dark out in the audience,
207
588771
1824
10:02
but it's going to use those cameras
in order to look at the terrain
in order to look at the terrain
208
590619
3220
10:05
right in front of it,
209
593863
1173
10:07
while it goes over
these obstacles back here.
these obstacles back here.
210
595060
2761
10:09
For this demo, Seth is steering,
211
597845
3313
10:13
but the robot's doing
all its own terrain planning.
all its own terrain planning.
212
601182
2505
10:15
This is a terrain map,
213
603711
1502
10:17
where the data from the cameras
is being developed in real time,
is being developed in real time,
214
605237
5170
10:22
showing the red spots,
which are where it doesn't want to step,
which are where it doesn't want to step,
215
610431
3237
10:25
and the green spots are the good places.
216
613692
1991
10:27
And here it's treating
them like stepping-stones.
them like stepping-stones.
217
615707
2449
10:30
So it's trying to stay up on the blocks,
218
618180
2755
10:32
and it adjusts its stride,
219
620959
1273
10:34
and there's a ton of planning
220
622256
1461
10:35
that has to go into
an operation like that,
an operation like that,
221
623741
2241
10:38
and it does all
that planning in real time,
that planning in real time,
222
626006
2262
10:40
where it adjusts the steps
a little bit longer
a little bit longer
223
628292
2430
10:42
or a little bit shorter.
224
630746
1276
10:45
Now we're going to change it
into a different mode,
into a different mode,
225
633396
2430
10:47
where it's just going to treat
the blocks like terrain
the blocks like terrain
226
635850
3423
10:51
and decide whether to step up or down
227
639297
3719
10:55
as it goes.
228
643040
1285
10:57
So this is using dynamic balance
229
645265
2969
11:00
and mobile perception,
230
648258
1858
11:02
because it has to coordinate what it sees
along with how it's moving.
along with how it's moving.
231
650140
5387
11:09
The other thing Spot has is a robot arm.
232
657168
4318
11:14
Some of you may see that
as a head and a neck,
as a head and a neck,
233
662517
2490
11:17
but believe me, it's an arm.
234
665031
1585
11:18
Seth is driving it around.
235
666640
1765
11:20
He's actually driving the hand
and the body is following.
and the body is following.
236
668429
3768
11:24
So the two are coordinated
in the way I was talking about before --
in the way I was talking about before --
237
672221
4133
11:28
in the way people can do that.
238
676378
1864
11:30
In fact, one of the cool things
Spot can do we call, "chicken-head mode,"
Spot can do we call, "chicken-head mode,"
239
678266
4643
11:34
and it keeps its head
in one place in space,
in one place in space,
240
682933
3184
11:38
and it moves its body all around.
241
686141
1845
11:40
There's a variation of this
that's called "twerking" --
that's called "twerking" --
242
688733
2602
11:43
(Laughter)
243
691359
1016
11:44
but we're not going to use that today.
244
692399
1823
11:46
(Laughter)
245
694246
1071
11:47
So, Spot: I'm feeling a little thirsty.
Could you get me a soda?
Could you get me a soda?
246
695341
3586
11:51
For this demo,
Seth is not doing any driving.
Seth is not doing any driving.
247
699473
3773
11:55
We have a LIDAR on the back of the robot,
248
703270
2113
11:57
and it's using these props
we've put on the stage
we've put on the stage
249
705407
2537
11:59
to localize itself.
250
707968
1393
12:01
It's gone over to that location.
251
709385
2281
12:03
Now it's using a camera that's in its hand
252
711690
2882
12:06
to find the cup,
253
714596
1786
12:09
picks it up --
254
717033
1255
12:10
and again, Seth's not driving.
255
718312
1951
12:13
We've planned out a path for it to go --
256
721552
3241
12:16
it looked like it was
going off the path --
going off the path --
257
724817
2107
12:18
and now Seth's going
to take over control again,
to take over control again,
258
726948
2358
12:21
because I'm a little bit chicken
about having it do this by itself.
about having it do this by itself.
259
729330
3449
12:24
Thank you, Spot.
260
732803
1376
12:28
(Applause)
261
736411
5283
12:35
So, Spot:
262
743069
1590
12:36
How do you feel about having just finished
your TED performance?
your TED performance?
263
744683
3567
12:41
(Laughter)
264
749194
2689
12:44
Me, too!
265
752248
1152
12:45
(Laughter)
266
753424
1032
12:46
Thank you all,
267
754480
1816
12:48
and thanks to the team at Boston Dynamics,
268
756320
2395
12:50
who did all the hard work behind this.
269
758739
2114
12:52
(Applause)
270
760877
2123
13:03
Helen Walters: Marc,
come back in the middle.
come back in the middle.
271
771239
2163
13:05
Thank you so much.
272
773426
1154
13:06
Come over here, I have questions.
273
774604
2095
13:08
So, you mentioned the UPS
and the package delivery.
and the package delivery.
274
776723
3415
13:12
What are the other applications
that you see for your robots?
that you see for your robots?
275
780162
3937
13:16
Marc Raibert: You know,
I think that robots
I think that robots
276
784123
2027
13:18
that have the capabilities
I've been talking about
I've been talking about
277
786174
2406
13:20
are going to be incredibly useful.
278
788604
1660
13:22
About a year ago, I went to Fukushima
279
790288
2497
13:24
to see what the situation was there,
280
792809
2031
13:26
and there's just a huge need
281
794864
1877
13:28
for machines that can go
into some of the dirty places
into some of the dirty places
282
796765
3338
13:32
and help remediate that.
283
800127
1968
13:34
I think it won't be too long until
we have robots like this in our homes,
we have robots like this in our homes,
284
802708
4599
13:39
and one of the big needs
is to take care of the aging
is to take care of the aging
285
807331
5321
13:44
and invalids.
286
812676
1366
13:46
I think that it won't be too long
till we're using robots
till we're using robots
287
814066
3989
13:50
to help take care of our parents,
288
818079
2462
13:52
or probably more likely,
have our children help take care of us.
have our children help take care of us.
289
820565
4415
13:57
And there's a bunch of other things.
290
825901
1750
13:59
I think the sky's the limit.
291
827675
1366
14:01
Many of the ideas
we haven't thought of yet,
we haven't thought of yet,
292
829065
2258
14:03
and people like you will help us
think of new applications.
think of new applications.
293
831347
3595
14:06
HW: So what about the dark side?
294
834966
1587
14:08
What about the military?
295
836577
1833
14:10
Are they interested?
296
838434
1503
14:12
MR: Sure, the military has been
a big funder of robotics.
a big funder of robotics.
297
840621
3634
14:16
I don't think the military
is the dark side myself,
is the dark side myself,
298
844279
4202
14:20
but I think, as with all
advanced technology,
advanced technology,
299
848505
3871
14:24
it can be used for all kinds of things.
300
852400
2047
14:26
HW: Awesome. Thank you so much.
301
854471
1641
14:28
MR: OK, you're welcome.
302
856136
1405
14:29
Thank you.
303
857565
1152
14:30
(Applause)
304
858741
1586
ABOUT THE SPEAKER
Marc Raibert - RoboticistMarc Raibert is the founder and CEO of robot maker Boston Dynamics.
Why you should listen
Working with his team at Boston Dynamics, Marc Raibert builds some of the world's most advanced robots, such as BigDog, Atlas, Spot and Handle. These robots are inspired by the remarkable ability of animals to move with agility, dexterity, perception and intelligence. A key ingredient of these robots is their dynamic behavior, which contributes to their lifelike qualities and their effectiveness in the real world.
Raibert founded Boston Dynamics as a spinoff from MIT, where he ran the Leg Laboratory, which helped establish the scientific basis for highly dynamic robots. He was a professor of EE&CS at MIT and before that associate professor of CS & Robotics at Carnegie Mellon University. Raibert is a member of the National Academy of Engineering.
Marc Raibert | Speaker | TED.com