TED2009
Robert Full: Learning from the gecko's tail
Filmed:
Readability: 3.7
722,710 views
Biologist Robert Full studies the amazing gecko, with its supersticky feet and tenacious climbing skill. But high-speed footage reveals that the gecko's tail harbors perhaps the most surprising talents of all.
Robert Full - Biologist
Robert Full studies cockroach legs and gecko feet. His research is helping build tomorrow's robots, based on evolution's ancient engineering. Full bio
Robert Full studies cockroach legs and gecko feet. His research is helping build tomorrow's robots, based on evolution's ancient engineering. Full bio
Double-click the English transcript below to play the video.
00:18
Let me share with you today an original discovery.
0
0
5000
00:23
But I want to tell it to you the way it really happened --
1
5000
3000
00:26
not the way I present it in a scientific meeting,
2
8000
2000
00:28
or the way you'd read it in a scientific paper.
3
10000
3000
00:31
It's a story about beyond biomimetics,
4
13000
3000
00:34
to something I'm calling biomutualism.
5
16000
3000
00:37
I define that as an association between biology and another discipline,
6
19000
3000
00:40
where each discipline reciprocally advances the other,
7
22000
4000
00:44
but where the collective discoveries that emerge are beyond any single field.
8
26000
4000
00:48
Now, in terms of biomimetics,
9
30000
2000
00:50
as human technologies take on more of the characteristics of nature,
10
32000
3000
00:53
nature becomes a much more useful teacher.
11
35000
3000
00:56
Engineering can be inspired by biology
12
38000
2000
00:58
by using its principles and analogies when they're advantageous,
13
40000
3000
01:01
but then integrating that with the best human engineering,
14
43000
3000
01:04
ultimately to make something actually better than nature.
15
46000
5000
01:09
Now, being a biologist, I was very curious about this.
16
51000
3000
01:12
These are gecko toes.
17
54000
2000
01:14
And we wondered how they use these bizarre toes
18
56000
2000
01:16
to climb up a wall so quickly.
19
58000
2000
01:18
We discovered it. And what we found was
20
60000
3000
01:21
that they have leaf-like structures on their toes,
21
63000
2000
01:23
with millions of tiny hairs that look like a rug,
22
65000
3000
01:26
and each of those hairs has the worst case of split-ends possible:
23
68000
3000
01:29
about 100 to 1000 split ends that are nano-size.
24
71000
4000
01:33
And the individual has 2 billion of these nano-size split ends.
25
75000
4000
01:37
They don't stick by Velcro or suction or glue.
26
79000
2000
01:39
They actually stick by intermolecular forces alone,
27
81000
3000
01:42
van der Waals forces.
28
84000
2000
01:44
And I'm really pleased to report to you today
29
86000
2000
01:46
that the first synthetic self-cleaning, dry adhesive has been made.
30
88000
5000
01:51
From the simplest version in nature, one branch,
31
93000
3000
01:54
my engineering collaborator, Ron Fearing, at Berkeley,
32
96000
3000
01:57
had made the first synthetic version.
33
99000
3000
02:00
And so has my other incredible collaborator,
34
102000
2000
02:02
Mark Cutkosky, at Stanford --
35
104000
2000
02:04
he made much larger hairs than the gecko,
36
106000
2000
02:06
but used the same general principles.
37
108000
3000
02:09
And here is its first test.
38
111000
2000
02:11
(Laughter)
39
113000
1000
02:12
That's Kellar Autumn, my former Ph.D. student,
40
114000
2000
02:14
professor now at Lewis and Clark,
41
116000
2000
02:16
literally giving his first-born child up for this test.
42
118000
4000
02:20
(Laughter)
43
122000
1000
02:21
More recently, this happened.
44
123000
2000
02:23
Man: This the first time someone has actually climbed with it.
45
125000
3000
02:26
Narrator: Lynn Verinsky, a professional climber,
46
128000
2000
02:28
who appeared to be brimming with confidence.
47
130000
2000
02:30
Lynn Verinsky: Honestly, it's going to be perfectly safe. It will be perfectly safe.
48
132000
3000
02:33
Man: How do you know?
49
135000
2000
02:35
Lynn Verinsky: Because of liability insurance. (Laughter)
50
137000
2000
02:37
Narrator: With a mattress below and attached to a safety rope,
51
139000
2000
02:39
Lynn began her 60-foot ascent.
52
141000
3000
02:42
Lynn made it to the top in a perfect pairing
53
144000
3000
02:45
of Hollywood and science.
54
147000
3000
02:48
Man: So you're the first human being to officially emulate a gecko.
55
150000
3000
02:51
Lynn Verinsky: Ha! Wow. And what a privilege that has been.
56
153000
6000
02:57
Robert Full: That's what she did on rough surfaces.
57
159000
2000
02:59
But she actually used these on smooth surfaces --
58
161000
2000
03:01
two of them -- to climb up, and pull herself up.
59
163000
2000
03:03
And you can try this in the lobby,
60
165000
2000
03:05
and look at the gecko-inspired material.
61
167000
4000
03:09
Now the problem with the robots doing this
62
171000
2000
03:11
is that they can't get unstuck,
63
173000
2000
03:13
with the material.
64
175000
2000
03:15
This is the gecko's solution. They actually peel their toes away
65
177000
3000
03:18
from the surface, at high rates,
66
180000
2000
03:20
as they run up the wall.
67
182000
2000
03:22
Well I'm really excited today to show you
68
184000
3000
03:25
the newest version of a robot, Stickybot,
69
187000
3000
03:28
using a new hierarchical dry adhesive.
70
190000
3000
03:31
Here is the actual robot.
71
193000
3000
03:37
And here is what it does.
72
199000
3000
03:45
And if you look,
73
207000
2000
03:47
you can see that it uses
74
209000
3000
03:50
the toe peeling,
75
212000
2000
03:52
just like the gecko does.
76
214000
2000
03:56
If we can show some of the video, you can see it climbing up the wall.
77
218000
3000
03:59
(Applause)
78
221000
2000
04:01
There it is.
79
223000
2000
04:03
And now it can go on other surfaces because of the new adhesive
80
225000
3000
04:06
that the Stanford group was able to do
81
228000
3000
04:09
in designing this incredible robot.
82
231000
3000
04:12
(Applause)
83
234000
3000
04:15
Oh. One thing I want to point out is, look at Stickybot.
84
237000
3000
04:18
You see something on it. It's not just to look like a gecko.
85
240000
4000
04:22
It has a tail. And just when you think you've figured out nature,
86
244000
3000
04:25
this kind of thing happens.
87
247000
2000
04:27
The engineers told us, for the climbing robots,
88
249000
2000
04:29
that, if they don't have a tail,
89
251000
2000
04:31
they fall off the wall.
90
253000
2000
04:33
So what they did was they asked us
91
255000
2000
04:35
an important question.
92
257000
2000
04:37
They said, "Well, it kind of looks like a tail."
93
259000
4000
04:41
Even though we put a passive bar there.
94
263000
2000
04:43
"Do animals use their tails when they climb up walls?"
95
265000
3000
04:46
What they were doing was returning the favor,
96
268000
2000
04:48
by giving us a hypothesis to test,
97
270000
3000
04:51
in biology, that we wouldn't have thought of.
98
273000
3000
04:54
So of course, in reality, we were then panicked,
99
276000
3000
04:57
being the biologists, and we should know this already.
100
279000
2000
04:59
We said, "Well, what do tails do?"
101
281000
2000
05:01
Well we know that tails store fat, for example.
102
283000
3000
05:04
We know that you can grab onto things with them.
103
286000
3000
05:07
And perhaps it is most well known
104
289000
2000
05:09
that they provide static balance.
105
291000
3000
05:12
(Laughter)
106
294000
1000
05:13
It can also act as a counterbalance.
107
295000
3000
05:16
So watch this kangaroo.
108
298000
3000
05:19
See that tail? That's incredible!
109
301000
2000
05:21
Marc Raibert built a Uniroo hopping robot.
110
303000
4000
05:25
And it was unstable without its tail.
111
307000
3000
05:31
Now mostly tails limit maneuverability,
112
313000
2000
05:33
like this human inside this dinosaur suit.
113
315000
4000
05:37
(Laughter)
114
319000
1000
05:38
My colleagues actually went on to test this limitation,
115
320000
4000
05:42
by increasing the moment of inertia of a student, so they had a tail,
116
324000
4000
05:46
and running them through and obstacle course,
117
328000
2000
05:48
and found a decrement in performance,
118
330000
2000
05:50
like you'd predict.
119
332000
3000
05:53
(Laughter)
120
335000
1000
05:54
But of course, this is a passive tail.
121
336000
3000
05:57
And you can also have active tails.
122
339000
2000
05:59
And when I went back to research this, I realized
123
341000
2000
06:01
that one of the great TED moments in the past,
124
343000
2000
06:03
from Nathan,
125
345000
2000
06:05
we've talked about an active tail.
126
347000
2000
06:07
Video: Myhrvold thinks tail-cracking dinosaurs
127
349000
2000
06:09
were interested in love, not war.
128
351000
5000
06:14
Robert Full: He talked about the tail being a whip for communication.
129
356000
3000
06:17
It can also be used in defense.
130
359000
4000
06:21
Pretty powerful.
131
363000
2000
06:23
So we then went back and looked at the animal.
132
365000
2000
06:25
And we ran it up a surface.
133
367000
2000
06:27
But this time what we did is we put a slippery patch
134
369000
2000
06:29
that you see in yellow there.
135
371000
2000
06:31
And watch on the right what the animal is doing with its tail
136
373000
4000
06:35
when it slips. This is slowed down 10 times.
137
377000
2000
06:37
So here is normal speed.
138
379000
2000
06:39
And watch it now slip,
139
381000
2000
06:41
and see what it does with its tail.
140
383000
2000
06:46
It has an active tail that functions as a fifth leg,
141
388000
2000
06:48
and it contributes to stability.
142
390000
2000
06:50
If you make it slip a huge amount, this is what we discovered.
143
392000
4000
06:57
This is incredible.
144
399000
2000
06:59
The engineers had a really good idea.
145
401000
3000
07:02
And then of course we wondered,
146
404000
2000
07:04
okay, they have an active tail, but let's picture them.
147
406000
2000
07:06
They're climbing up a wall, or a tree.
148
408000
3000
07:09
And they get to the top and let's say there's some leaves there.
149
411000
3000
07:12
And what would happen if they climbed on the underside of that leaf,
150
414000
3000
07:15
and there was some wind, or we shook it?
151
417000
3000
07:18
And we did that experiment, that you see here.
152
420000
3000
07:21
(Applause)
153
423000
1000
07:22
And this is what we discovered.
154
424000
2000
07:24
Now that's real time. You can't see anything.
155
426000
2000
07:26
But there it is slowed down.
156
428000
2000
07:30
What we discovered was the world's fastest air-righting response.
157
432000
3000
07:33
For those of you who remember your physics, that's a zero-angular-momentum
158
435000
2000
07:35
righting response. But it's like a cat.
159
437000
2000
07:37
You know, cats falling. Cats do this. They twist their bodies.
160
439000
3000
07:40
But geckos do it better.
161
442000
2000
07:42
And they do it with their tail.
162
444000
3000
07:45
So they do it with this active tail as they swing around.
163
447000
3000
07:48
And then they always land in the sort of superman skydiving posture.
164
450000
4000
07:54
Okay, now we wondered, if we were right,
165
456000
2000
07:56
we should be able to test this in a physical model, in a robot.
166
458000
3000
07:59
So for TED we actually built a robot,
167
461000
2000
08:01
over there, a prototype, with the tail.
168
463000
3000
08:04
And we're going to attempt the first air-righting response
169
466000
2000
08:06
in a tail, with a robot.
170
468000
2000
08:08
If we could have the lights on it.
171
470000
2000
08:10
Okay, there it goes.
172
472000
5000
08:15
And show the video.
173
477000
2000
08:20
There it is.
174
482000
2000
08:22
And it works just like it does in the animal.
175
484000
3000
08:25
So all you need is a swing of the tail to right yourself.
176
487000
4000
08:29
(Applause)
177
491000
2000
08:31
Now, of course, we were normally frightened
178
493000
2000
08:33
because the animal has no gliding adaptations,
179
495000
2000
08:35
so we thought, "Oh that's okay. We'll put it in a vertical wind tunnel.
180
497000
4000
08:39
We'll blow the air up, we'll give it a landing target, a tree trunk,
181
501000
3000
08:42
just outside the plexi-glass enclosure, and see what it does.
182
504000
4000
08:46
(Laughter)
183
508000
1000
08:47
So we did. And here is what it does.
184
509000
3000
08:50
So the wind is coming from the bottom. This is slowed down 10 times.
185
512000
5000
08:55
It does an equilibrium glide. Highly controlled.
186
517000
4000
08:59
This is sort of incredible. But actually it's quite beautiful,
187
521000
3000
09:02
when you take a picture of it.
188
524000
3000
09:05
And it's better than that, it -- just in the slide -- maneuvers in mid-air.
189
527000
5000
09:10
And the way it does it, is it takes its tail
190
532000
2000
09:12
and it swings it one way to yaw left, and it swings its other way to yaw right.
191
534000
4000
09:16
So we can maneuver this way.
192
538000
2000
09:18
And then -- we had to film this several times to believe this --
193
540000
3000
09:21
it also does this. Watch this.
194
543000
3000
09:24
It oscillates its tail up and down like a dolphin.
195
546000
2000
09:26
It can actually swim through the air.
196
548000
3000
09:29
But watch its front legs. Can you see what they are doing?
197
551000
5000
09:34
What does that mean for the origin of flapping flight?
198
556000
3000
09:37
Maybe it's evolved from coming down from trees,
199
559000
3000
09:40
and trying to control a glide.
200
562000
2000
09:42
Stay tuned for that.
201
564000
2000
09:44
(Laughter)
202
566000
2000
09:46
So then we wondered, "Can they actually maneuver with this?"
203
568000
3000
09:49
So there is the landing target. Could they steer towards it
204
571000
3000
09:52
with these capabilities? Here it is in the wind tunnel.
205
574000
2000
09:54
And it certainly looks like it.
206
576000
2000
09:56
You can see it even better from down on top.
207
578000
3000
09:59
Watch the animal.
208
581000
2000
10:02
Definitely moving towards the landing target.
209
584000
2000
10:04
Watch the whip of its tail as it does it. Look at that.
210
586000
4000
10:08
It's unbelievable.
211
590000
2000
10:10
So now we were really confused,
212
592000
2000
10:12
because there are no reports of it gliding.
213
594000
2000
10:14
So we went, "Oh my god, we have to go to the field,
214
596000
2000
10:16
and see if it actually does this."
215
598000
2000
10:18
Completely opposite of the way you'd see it on a nature film, of course.
216
600000
3000
10:21
We wondered, "Do they actually glide in nature?"
217
603000
3000
10:24
Well we went to the forests of Singapore and Southeast Asia.
218
606000
2000
10:26
And the next video you see is the first time we've showed this.
219
608000
2000
10:28
This is the actual video -- not staged, a real research video --
220
610000
3000
10:31
of animal gliding down. There is a red trajectory line.
221
613000
3000
10:34
Look at the end to see the animal.
222
616000
2000
10:36
But then as it gets closer to the tree,
223
618000
2000
10:38
look at the close-up. And see if you can see it land.
224
620000
3000
10:42
So there it comes down. There is a gecko at the end of that trajectory line.
225
624000
3000
10:45
You see it there? There? Watch it come down.
226
627000
2000
10:47
Now watch up there and you can see the landing. Did you see it hit?
227
629000
3000
10:50
It actually uses its tail too,
228
632000
2000
10:52
just like we saw in the lab.
229
634000
3000
10:55
So now we can continue this mutualism
230
637000
4000
10:59
by suggesting that they can make an active tail.
231
641000
3000
11:02
And here is the first active tail, in the robot,
232
644000
5000
11:07
made by Boston Dynamics.
233
649000
3000
11:10
So to conclude, I think we need to build biomutualisms, like I showed,
234
652000
4000
11:14
that will increase the pace of basic discovery in their application.
235
656000
3000
11:17
To do this though, we need to redesign education in a major way,
236
659000
3000
11:20
to balance depth with interdisciplinary communication,
237
662000
3000
11:23
and explicitly train people how to contribute to, and benefit from other disciplines.
238
665000
5000
11:28
And of course you need the organisms and the environment to do it.
239
670000
4000
11:32
That is, whether you care about security, search and rescue or health,
240
674000
3000
11:35
we must preserve nature's designs,
241
677000
2000
11:37
otherwise these secrets will be lost forever.
242
679000
3000
11:40
And from what I heard from our new president,
243
682000
4000
11:44
I'm very optimistic. Thank you.
244
686000
2000
11:46
(Applause)
245
688000
2000
ABOUT THE SPEAKER
Robert Full - BiologistRobert Full studies cockroach legs and gecko feet. His research is helping build tomorrow's robots, based on evolution's ancient engineering.
Why you should listen
UC Berkeley biologist Robert Full is fascinated by the motion of creatures like cockroaches, crabs and geckos having many legs, unusual feet or talented tails. He has led an effort to demonstrate the value of learning from Nature by the creating interdisciplinary collaborations of biologists, engineers, mathematicians and computer scientists from academia and industry. He founded CiBER, the Center for interdisciplinary Bio-inspiration in Education and Research, and the Poly-PEDAL Laboratory, which studies the Performance, Energetics and Dynamics of Animal Locomotion (PEDAL) in many-footed creatures (Poly).
His research shows how studying a diversity of animals leads to the discovery of general principles which inspire the design of novel circuits, artificial muscles, exoskeletons, versatile scampering legged search-and-rescue robots and synthetic self-cleaning dry adhesives based on gecko feet. He is passionate about discovery-based education leading to innovation -- and he even helped Pixar’s insect animations in the film A Bug's Life.
Robert Full | Speaker | TED.com