ABOUT THE SPEAKER
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.

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.

More profile about the speaker
Robert Full | Speaker | TED.com
TED2002

Robert Full: Robots inspired by cockroach ingenuity

Robert Full谈生物进化启发工程设计

Filmed:
1,087,679 views

尽管动物和昆虫进化出了令人咂舌的生存技能,Robert Full认为它们适用面窄,太细致,被“过分设计”了:倘若要使这些“尖端科技”为人类造福,我们只需取其精髓。请看Robert Full展示工程仿生妙诀。
- 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

Double-click the English transcript below to play the video.

00:19
Welcome欢迎. If I could have the first slide滑动, please?
0
1000
5000
欢迎大家!能调到第一张幻灯片吗?
00:33
Contrary相反 to calculations计算 made制作 by some engineers工程师, bees蜜蜂 can fly,
1
15000
5000
与工程师们所计算的正好相反:蜜蜂能飞
00:38
dolphins海豚 can swim游泳, and geckos壁虎 can even climb
2
20000
7000
海豚能游泳,壁虎甚至可以
00:45
up the smoothest流畅的 surfaces. Now, what I want to do, in the short time I have,
3
27000
6000
在最光滑的表面垂直爬行。我现在利用很短的一点时间
00:51
is to try to allow允许 each of you to experience经验
4
33000
4000
想让你们都体验一下
00:55
the thrill颤抖 of revealing揭示 nature's大自然 design设计.
5
37000
6000
揭开大自然设计的奥秘有多么刺激
01:01
I get to do this all the time, and it's just incredible难以置信.
6
43000
2000
我每次都有机会做这个,真是难以置信
01:03
I want to try to share分享 just a little bit of that with you in this presentation介绍.
7
45000
6000
现在要和你们一同分享
01:09
The challenge挑战 of looking at nature's大自然 designs设计 --
8
51000
2000
窥探自然匠心的挑战
01:11
and I'll tell you the way that we perceive感知 it, and the way we've我们已经 used it.
9
53000
4000
我们怎么去理解它、应用它
01:15
The challenge挑战, of course课程, is to answer回答 this question:
10
57000
2000
其中的挑战自然是围绕着一个问题:
01:17
what permits许可证 this extraordinary非凡 performance性能 of animals动物
11
59000
3000
是什么使得动物拥有某些能力,让它们的分布
01:20
that allows允许 them basically基本上 to go anywhere随地?
12
62000
3000
广泛得如此出乎意料,几乎无处不在
01:23
And if we could figure数字 that out, how can we implement实行 those designs设计?
13
65000
7000
倘若我们能掌握这些设计的话会发生什么呢?
01:30
Well, many许多 biologists生物学家 will tell engineers工程师, and others其他,
14
72000
3000
好吧,许多生物学家都会跟工程师们说
01:33
organisms生物 have millions百万 of years年份 to get it right;
15
75000
3000
生物们花了上百万年才把这些东西弄对
01:36
they're spectacular壮观; they can do everything wonderfully奇妙 well.
16
78000
3000
让人叹为观止,这些家伙没什么做不了的
01:39
So, the answer回答 is bio-mimicry生物拟态: just copy复制 nature性质 directly.
17
81000
4000
所以它们的答案是:直接复制自然
01:43
We know from working加工 on animals动物 that the truth真相 is
18
85000
5000
通过研究动物我们知道
01:48
that's exactly究竟 what you don't want to do -- because evolution演化 works作品
19
90000
4000
其实这样根本不行。因为进化
01:52
on the just-good-enough刚刚足够好 principle原理, not on a perfecting完善 principle原理.
20
94000
3000
遵循的是“见好就收”的原则,而不是“精益求精”
01:55
And the constraints限制 in building建造 any organism生物, when you look at it,
21
97000
4000
造化在“设计”任何一种生物时面临的障碍
01:59
are really severe严重. Natural自然 technologies技术 have incredible难以置信 constraints限制.
22
101000
5000
都十分严峻。“自然科技”遇到的困难复杂得无可想象
02:04
Think about it. If you were an engineer工程师 and I told you
23
106000
3000
试想你是一个工程师,我告诉你
02:07
that you had to build建立 an automobile汽车, but it had to start开始 off to be this big,
24
109000
5000
你得造一辆汽车,一开始只能有这么点儿大
02:12
then it had to grow增长 to be full充分 size尺寸 and had to work every一切 step along沿 the way.
25
114000
4000
然后要渐渐长成,而且在每个阶段都能工作
02:16
Or think about the fact事实 that if you build建立 an automobile汽车, I'll tell you that you also -- inside it --
26
118000
4000
想象我要你造这么一辆汽车
02:20
have to put a factory that allows允许 you to make another另一个 automobile汽车.
27
122000
4000
符合要求的汽车里得藏着一个工厂来制造另一辆汽车
02:24
(Laughter笑声)
28
126000
2000
(笑)
02:26
And you can absolutely绝对 never, absolutely绝对 never, because of history历史
29
128000
4000
并且你绝对绝对没法从零开始,因为有进化史
02:30
and the inherited遗传 plan计划, start开始 with a clean清洁 slate石板.
30
132000
4000
和遗传编码的一大堆限制
02:34
So, organisms生物 have this important重要 history历史.
31
136000
3000
所以生物的历史很重要
02:37
Really evolution演化 works作品 more like a tinkerer工匠 than an engineer工程师.
32
139000
5000
进化的工作方式更像修理工而不是工程师,太过琐碎了
02:42
And this is really important重要 when you begin开始 to look at animals动物.
33
144000
3000
你看看动物体就知道这一点有多重要了
02:45
Instead代替, we believe you need to be inspired启发 by biology生物学.
34
147000
7000
我觉得应该是从生物中得到灵感
02:52
You need to discover发现 the general一般 principles原则 of nature性质,
35
154000
4000
你得寻找自然的一般原理
02:56
and then use these analogies类比 when they're advantageous有利.
36
158000
3000
然后在类比的广泛层面上让它们为我所用
03:02
This is a real真实 challenge挑战 to do this, because animals动物,
37
164000
3000
做到这点真不容易,因为
03:05
when you start开始 to really look inside them -- how they work --
38
167000
3000
当你真的去研究动物,它们的机理
03:08
appear出现 hopelessly绝望地 complex复杂. There's no detailed详细 history历史
39
170000
4000
看起来复杂得令人绝望。“设计方案”
03:12
of the design设计 plans计划, you can't go look it up anywhere随地.
40
174000
3000
没有详细的历史备案可以随时查阅
03:15
They have way too many许多 motions运动 for their joints关节, too many许多 muscles肌肉.
41
177000
4000
它们的关节能做太多种的运动,有太多块不同的肌肉
03:19
Even the simplest简单 animal动物 we think of, something like an insect昆虫,
42
181000
3000
哪怕是最简单的动物,像昆虫
03:22
and they have more neurons神经元 and connections连接 than you can imagine想像.
43
184000
3000
它们的神经元和其中的连接方式多到你想不到
03:25
How can you make sense of this? Well, we believed相信 --
44
187000
5000
你如何去理解这个?嗯..我们相信
03:30
and we hypothesized假设 -- that one way animals动物 could work simply只是,
45
192000
5000
也假设情况是这样:能让动物的机理简单些的办法
03:35
is if the control控制 of their movements运动
46
197000
3000
可能是将运动的控制
03:38
tended往往 to be built内置 into their bodies身体 themselves他们自己.
47
200000
6000
内置在身体的构造里
03:44
What we discovered发现 was that two-二-, four-四-, six-六- and eight-legged八只脚 animals动物
48
206000
7000
我们发现两条腿、四条腿、六条、八条腿的动物
03:51
all produce生产 the same相同 forces军队 on the ground地面 when they move移动.
49
213000
3000
动起来时都对地面产生相似的力
03:54
They all work like this kangaroo袋鼠, they bounce弹跳.
50
216000
4000
它们都跟这只袋鼠差不多,它们都在“弹跳”
03:58
And they can be modeled仿照 by a spring-mass弹簧质量 system系统 that we call the spring弹簧 mass system系统
51
220000
4000
这些都可以用一个弹簧-质量系统来模拟,我们这么叫它
04:02
because we're biomechanistsbiomechanists. It's actually其实 a pogo弹簧 stick.
52
224000
3000
因为我们是生物机械学家,其实也就是个弹簧跳跳棒
04:05
They all produce生产 the pattern模式 of a pogo弹簧 stick. How is that true真正?
53
227000
4000
它们都产生这一类的运动。为什么这样呢?
04:09
Well, a human人的, one of your legs works作品 like two legs of a trotting小跑 dog,
54
231000
6000
这个..一个人的一条腿腿走起来像一只疾跑的狗一侧的两条腿
04:15
or works作品 like three legs, together一起 as one, of a trotting小跑 insect昆虫,
55
237000
4000
或者像一只昆虫身上三只同时动作的腿
04:19
or four legs as one of a trotting小跑 crab螃蟹.
56
241000
2000
或是螃蟹体侧同时行动的四条腿
04:21
And then they alternate备用 in their propulsion动力,
57
243000
4000
它们提供助推的顺序是交替的
04:25
but the patterns模式 are all the same相同. Almost几乎 every一切 organism生物 we've我们已经 looked看着 at this way
58
247000
5000
但是规律是一样的。几乎所有的动物这样看来
04:30
-- you'll你会 see next下一个 week, I'll give you a hint暗示,
59
252000
2000
--下周你就能看到--暂且先给个提示,
04:32
there'll有会 be an article文章 coming未来 out that says that really big things
60
254000
3000
会有一篇文章登出来说
04:35
like T. rex雷克斯 probably大概 couldn't不能 do this, but you'll你会 see that next下一个 week.
61
257000
4000
像霸王龙那样的大家伙可能不会这么动,你下周等着瞧吧
04:39
Now, what's interesting有趣 is the animals动物, then -- we said -- bounce弹跳 along沿
62
261000
2000
有意思的是动物们就这么
04:41
the vertical垂直 plane平面 this way, and in our collaborations合作 with Pixar皮克斯,
63
263000
3000
在纵平面上“跳跃”,在我们和Pixar公司合作的
04:44
in "A Bug's错误的 Life," we discussed讨论 the
64
266000
2000
《一只虫子的生活》时,我们谈到
04:46
bipedal双足 nature性质 of the characters人物 of the ants蚂蚁.
65
268000
3000
那里面蚂蚁角色们用双脚行走这一点
04:49
And we told them, of course课程, they move移动 in another另一个 plane平面 as well.
66
271000
2000
我们告诉他们蚂蚁当然还在另外一个平面里动
04:51
And they asked us this question. They say, "Why model模型
67
273000
3000
他们反问了我们一个问题,他们说,“为什么
04:54
just in the sagittal矢状 plane平面 or the vertical垂直 plane平面,
68
276000
2000
只在径向平面或者纵向平面建模
04:56
when you're telling告诉 us these animals动物 are moving移动
69
278000
2000
你不是说这些动物
04:58
in the horizontal plane平面?" This is a good question.
70
280000
3000
在横向平面里也动吗?”这是个好问题
05:01
Nobody没有人 in biology生物学 ever modeled仿照 it this way.
71
283000
3000
生物学里还没有人这么建模过呢
05:04
We took their advice忠告 and we modeled仿照 the animals动物 moving移动
72
286000
4000
我们就接受了建议,在横向平面里
05:08
in the horizontal plane平面 as well. We took their three legs,
73
290000
3000
也给动物运动建模了。拿蚂蚁的三条腿
05:11
we collapsed倒塌 them down as one.
74
293000
1000
当做是一条
05:12
We got some of the best最好 mathematicians数学家 in the world世界
75
294000
3000
借来世界上最好的数学家们--
05:15
from Princeton普林斯顿 to work on this problem问题.
76
297000
2000
普林斯顿的人,来弄这个问题
05:17
And we were able能够 to create创建 a model模型
77
299000
3000
我们最后造了一个模型
05:20
where animals动物 are not only bouncing蹦蹦 up and down,
78
302000
1000
不光模拟动物的上下“弹跳”
05:21
but they're also bouncing蹦蹦 side to side at the same相同 time.
79
303000
4000
同时还左右运动
05:25
And many许多 organisms生物 fit适合 this kind of pattern模式.
80
307000
2000
许多动物都遵循这个规律
05:27
Now, why is this important重要 to have this model模型?
81
309000
2000
为什么这个模型很重要呢?
05:29
Because it's very interesting有趣. When you take this model模型
82
311000
3000
因为很有意思,你拿这个模型
05:32
and you perturb干扰 it, you give it a push,
83
314000
3000
干扰它一下,给它一个推力
05:35
as it bumps颠簸 into something, it self-stabilizes自稳定, with no brain
84
317000
4000
比如说当它撞在什么上面,它自动会趋向稳定,不需要大脑思考
05:39
or no reflexes反射, just by the structure结构体 alone单独.
85
321000
4000
不用条件反射,依靠这个结构本身的性质就行
05:43
It's a beautiful美丽 model模型. Let's look at the mathematics数学.
86
325000
5000
很美妙的一个模型。我们来看看其中的数学原理
05:48
(Laughter笑声)
87
330000
2000
(笑)
05:50
That's enough足够!
88
332000
1000
够了。
05:51
(Laughter笑声)
89
333000
4000
(笑)
05:55
The animals动物, when you look at them running赛跑,
90
337000
2000
当你看动物们跑的时候
05:57
appear出现 to be self-stabilizing自稳定 like this,
91
339000
3000
他们看上去会这样进行稳定调整
06:00
using运用 basically基本上 springy弹性 legs. That is, the legs can do
92
342000
3000
用它们的“弹簧腿”。就是说那些腿
06:03
computations计算 on their own拥有; the control控制 algorithms算法, in a sense,
93
345000
3000
仿佛能自己做计算,控制这些东西的算法某种意义上
06:06
are embedded嵌入式 in the form形成 of the animal动物 itself本身.
94
348000
3000
被嵌入了动物的形态本身
06:09
Why haven't没有 we been more inspired启发 by nature性质 and these kinds of discoveries发现?
95
351000
7000
为什么我们没有能从这样的发现中得到更大的启示呢?
06:16
Well, I would argue争论 that human人的 technologies技术 are really different不同 from
96
358000
4000
我认为这是因为人类科技
06:20
natural自然 technologies技术, at least最小 they have been so far.
97
362000
3000
和“自然科技”差别很大,至少迄今一直如此
06:23
Think about the typical典型 kind of robot机器人 that you see.
98
365000
5000
想想你看到的典型的机器人
06:28
Human人的 technologies技术 have tended往往 to be large, flat平面,
99
370000
3000
人类科技趋向于又大又笨
06:31
with right angles, stiff僵硬, made制作 of metal金属. They have rolling压延 devices设备
100
373000
5000
有棱有角,僵硬,金属构造。它们装备的是滚动装置
06:36
and axles车轴. There are very few少数 motors马达, very few少数 sensors传感器.
101
378000
3000
和轴。一部机械里马达和传感器数量都很少
06:39
Whereas nature性质 tends趋向 to be small, and curved弯曲,
102
381000
5000
而自然造物趋向于小体积和曲线形
06:44
and it bends弯曲 and twists曲折, and has legs instead代替, and appendages附属物,
103
386000
3000
弯弯曲曲,带着腿和附加物
06:47
and has many许多 muscles肌肉 and many许多, many许多 sensors传感器.
104
389000
3000
许多的肌肉和一大堆的感应器
06:50
So it's a very different不同 design设计. However然而, what's changing改变,
105
392000
4000
所以设计思路很不同。然而,现在正在改变着的
06:54
what's really exciting扣人心弦 -- and I'll show显示 you some of that next下一个 --
106
396000
2000
真正刺激的东西--下面让你们看看这些--
06:56
is that as human人的 technology技术 takes on more of the characteristics特点
107
398000
3000
就是当人类科技越来越多地拥有
06:59
of nature性质, then nature性质 really can become成为 a much more useful有用 teacher老师.
108
401000
6000
自然物的特征,自然就能更好地教导、启发我们
07:05
And here's这里的 one example that's really exciting扣人心弦.
109
407000
2000
这儿有个真正带劲儿的例子
07:07
This is a collaboration合作 we have with Stanford斯坦福.
110
409000
2000
是我们和斯坦福大学合作的
07:09
And they developed发达 this new technique技术, called Shape形状 Deposition沉积 Manufacturing制造业.
111
411000
4000
他们发明了一种叫形状沉积制造的新技术
07:13
It's a technique技术 where they can mix混合 materials物料 together一起 and mold模子 any shape形状
112
415000
4000
利用这种技术他们可以把材料混合起来,铸成任意形状
07:17
that they like, and put in the material材料 properties性能.
113
419000
4000
并把材料的性质加进去
07:21
They can embed sensors传感器 and actuators执行器 right in the form形成 itself本身.
114
423000
3000
可以把传感器和促动器直接植入形态本身
07:24
For example, here's这里的 a leg: the clear明确 part部分 is stiff僵硬,
115
426000
5000
比方说,这儿有一条腿--透明的这一块是固定的
07:29
the white白色 part部分 is compliant兼容, and you don't need any axles车轴 there or anything.
116
431000
3000
白色的一块可以弯曲,你不需要装轴或是什么东西
07:32
It just bends弯曲 by itself本身 beautifully精美.
117
434000
3000
它自己会很漂亮地弯下来
07:35
So, you can put those properties性能 in. It inspired启发 them to show显示 off
118
437000
3000
所以你可以把那些性质加进来。这项技术给了他们可以卖弄的灵感
07:38
this design设计 by producing生产 a little robot机器人 they named命名 Sprawl蔓生.
119
440000
6000
就创造出了这个叫Sprawl(“匍匐”)的机器人
07:44
Our work has also inspired启发 another另一个 robot机器人, a biologically生物 inspired启发 bouncing蹦蹦 robot机器人,
120
446000
4000
我们的研究还启发了另一个机器人,一个从生物学得到启发的弹跳机器人
07:48
from the University大学 of Michigan密歇根州 and McGill麦吉尔
121
450000
2000
出自密歇根大学和麦克吉尔大学
07:50
named命名 RHexRHex, for robot机器人 hexapod昆虫, and this one's那些 autonomous自主性.
122
452000
8000
名叫RHex,“六足机器人”(robot hexapod) 的缩写,它是自动的
07:58
Let's go to the video视频, and let me show显示 you some of these animals动物 moving移动
123
460000
3000
让我们看一段录像,看看一些动物运动
08:01
and then some of the simple简单 robots机器人
124
463000
2000
还有些简单的机器人
08:03
that have been inspired启发 by our discoveries发现.
125
465000
3000
他们都是从对动物的研究发现中受启发的
08:06
Here's这里的 what some of you did this morning早上, although虽然 you did it outside,
126
468000
4000
你们有些人早上会去干这个,只不过是在户外
08:10
not on a treadmill跑步机.
127
472000
2000
不是在跑步机上
08:12
Here's这里的 what we do.
128
474000
3000
而这个是我们做的
08:15
(Laughter笑声)
129
477000
2000
(笑)
08:17
This is a death's死亡的 head cockroach蟑螂. This is an American美国 cockroach蟑螂
130
479000
5000
这是一只骷髅蟑螂--美国品种
08:22
you think you don't have in your kitchen厨房.
131
484000
1000
不过估计你厨房里没有
08:23
This is an eight-legged八只脚 scorpion, six-legged六足 ant蚂蚁, forty-four-legged44条腿 centipede.
132
485000
7000
一只八条腿的蝎子,六条腿的蚂蚁,四十四条腿的蜈蚣
08:30
Now, I said all these animals动物 are sort分类 of working加工 like pogo弹簧 sticks --
133
492000
3000
我现在告诉你这些动物走起来都像弹簧跳跳棒
08:33
they're bouncing蹦蹦 along沿 as they move移动. And you can see that
134
495000
4000
你看他们朝一个方向移动时也上下跳动
08:37
in this ghost crab螃蟹, from the beaches海滩 of Panama巴拿马 and North Carolina卡罗来纳州.
135
499000
3000
比如这只来自巴拿马或北卡罗来纳海岸的幽灵蟹
08:40
It goes up to four meters per second第二 when it runs运行.
136
502000
3000
跑起来速度能高达四米每秒
08:43
It actually其实 leaps飞跃 into the air空气, and has aerial天线 phases
137
505000
3000
其实它经常有脚不沾地的滞空时间
08:46
when it does it, like a horse, and you'll你会 see it's bouncing蹦蹦 here.
138
508000
4000
而当它这样做的时候--就像匹马似的--你能看见他这里的跳动
08:50
What we discovered发现 is whether是否 you look at the leg of a human人的
139
512000
3000
我们发现无论你观察人类的腿
08:53
like Richard理查德, or a cockroach蟑螂, or a crab螃蟹, or a kangaroo袋鼠,
140
515000
6000
比如说理查的,或者一只蟑螂、螃蟹或是袋鼠的
08:59
the relative相对的 leg stiffness刚性 of that spring弹簧 is the same相同 for everything we've我们已经 seen看到 so far.
141
521000
5000
它们腿中那个“弹簧”的相对劲度都是相同的
09:04
Now, what good are springy弹性 legs then? What can they do?
142
526000
2000
那么弹簧腿有什么好处呢?能做些什么呢?
09:06
Well, we wanted to see if they allowed允许 the animals动物
143
528000
2000
我们想知道它们能否给动物的运动
09:08
to have greater更大 stability稳定性 and maneuverability机动性.
144
530000
3000
更大的稳定性和可操控性
09:11
So, we built内置 a terrain地形 that had obstacles障碍 three times the hip臀部 height高度
145
533000
4000
于是我们构建了一个
09:15
of the animals动物 that we're looking at.
146
537000
1000
拥有三倍我们所研究的动物髋高度障碍物的地形
09:16
And we were certain某些 they couldn't不能 do this. And here's这里的 what they did.
147
538000
4000
我们很确定它们翻不过去。可它们是如此反应的:
09:20
The animal动物 ran over it and it didn't even slow down!
148
542000
3000
动物飞奔而过,甚至都没有减速
09:23
It didn't decrease减少 its preferred首选 speed速度 at all.
149
545000
2000
它一点都不乐意牺牲自己习惯的速度
09:25
We couldn't不能 believe that it could do this. It said to us
150
547000
3000
我们一开始根本不相信的。事实告诉我们
09:28
that if you could build建立 a robot机器人 with very simple简单, springy弹性 legs,
151
550000
5000
如果我们能用简易的弹簧腿造一个机器人
09:33
you could make it as maneuverable机动性 as any that's ever been built内置.
152
555000
6000
就可以使它的可控性出类拔萃
09:39
Here's这里的 the first example of that. This is the Stanford斯坦福
153
561000
2000
这儿是第一个例子,斯坦福大学的
09:41
Shape形状 Deposition沉积 Manufactured制成的 robot机器人, named命名 Sprawl蔓生.
154
563000
3000
形状沉积技术制造的机器人“匍匐”
09:44
It has six legs -- there are the tuned调整, springy弹性 legs.
155
566000
6000
有着六条腿--协调好了的弹簧腿
09:50
It moves移动 in a gait步态 that an insect昆虫 uses使用, and here it is
156
572000
3000
它采用昆虫的步调行走,这个就是它
09:53
going on the treadmill跑步机. Now, what's important重要 about this robot机器人,
157
575000
7000
在跑步机上飞奔呢。与其他机器人相比
10:00
compared相比 to other robots机器人, is that it can't see anything,
158
582000
3000
它的关键就是它什么也看不见
10:03
it can't feel anything, it doesn't have a brain, yet然而 it can maneuver演习
159
585000
6000
也感觉不到,根本就没头脑,然而却能
10:09
over these obstacles障碍 without any difficulty困难 whatsoever任何.
160
591000
6000
娴熟地跨越障碍,毫不费力
10:15
It's this technique技术 of building建造 the properties性能 into the form形成.
161
597000
4000
正是这个把性质嵌入形态的技术造就了这一切。
10:19
This is a graduate毕业 student学生. This is what he's doing to his thesis论文 project项目 --
162
601000
3000
这是一名研究生的毕业设计内容
10:22
very robust强大的, if a graduate毕业 student学生
163
604000
2000
一个研究生的
10:24
does that to his thesis论文 project项目.
164
606000
2000
毕业设计里能干这个可真是了不起
10:26
(Laughter笑声)
165
608000
1000
(笑)
10:27
This is from McGill麦吉尔 and University大学 of Michigan密歇根州. This is the RHexRHex,
166
609000
4000
这是来自麦克吉尔和密歇根大学的RHex
10:31
making制造 its first outing游览 in a demo演示.
167
613000
3000
演示它首次出去放风
10:34
(Laughter笑声)
168
616000
4000
(笑)
10:38
Same相同 principle原理: it only has six moving移动 parts部分,
169
620000
5000
一样的原理。它只有六处能动的部分
10:43
six motors马达, but it has springy弹性, tuned调整 legs. It moves移动 in the gait步态 of the insect昆虫.
170
625000
6000
六个电机,可是它具有协调了的弹簧腿,昆虫的步法
10:49
It has the middle中间 leg moving移动 in synchrony同步 with the front面前,
171
631000
4000
中间的腿和前腿同步运动
10:53
and the hind leg on the other side. Sort分类 of an alternating交替 tripod,
172
635000
4000
后退则在另一边,交替地形成三角架形
10:57
and they can negotiate谈判 obstacles障碍 just like the animal动物.
173
639000
4000
它们能够像动物一样避开障碍物
11:01
(Laughter笑声)
174
643000
6000
(笑)
11:07
(Voice语音: Oh my God.)
175
649000
1000
我的天哪
11:08
(Applause掌声)
176
650000
5000
(掌声)
11:13
Robert罗伯特 Full充分: It'll它会 go on different不同 surfaces -- here's这里的 sand --
177
655000
2000
它现在要在不同的表面上走,这是沙子
11:15
although虽然 we haven't没有 perfected完善 the feet yet然而, but I'll talk about that later后来.
178
657000
5000
尽管我们尚未把腿再完善一些,我等一会儿会谈到这个
11:20
Here's这里的 RHexRHex entering进入 the woods树木.
179
662000
3000
RHex正在进入丛林
11:23
(Laughter笑声)
180
665000
2000
(笑)
11:38
Again, this robot机器人 can't see anything, it can't feel anything,
181
680000
4000
跟刚才一样,这个机器人也没有视力,没有知觉
11:42
it has no brain. It's just working加工 with a tuned调整 mechanical机械 system系统,
182
684000
6000
没有头脑。只是凭借着一个协调的机械系统
11:48
with very simple简单 parts部分, but inspired启发 from the fundamental基本的 dynamics动力学 of the animal动物.
183
690000
10000
和简单的各个部分。然而灵感都来源于动物的基本运动原理
11:58
(Voice语音: Ah, I love him, Bob短发.) RFRF: Here's这里的 it going down a pathway.
184
700000
8000
啊,我喜欢它--鲍勃,他正在下坡
12:06
I presented呈现 this to the jet喷射 propulsion动力 lab实验室 at NASANASA, and they said
185
708000
3000
我把这个展示给NASA的喷气动力实验室看,他们说
12:09
that they had no ability能力 to go down craters陨石坑 to look for ice,
186
711000
4000
他们缺一个能走下撞击坑寻找冰
12:13
and life, ultimately最终, on Mars火星. And he said --
187
715000
4000
最终寻找生命的技术,是用于火星的。他说--
12:17
especially特别 with legged-robots腿机器人, because they're way too complicated复杂.
188
719000
2000
他强调有腿的机器人不能胜任,因为造出来实在太复杂
12:19
Nothing can do that. And I talk next下一个. I showed显示 them this video视频
189
721000
5000
根本不现实。我接着发言。我给他们看了这个短片
12:24
with the simple简单 design设计 of RHexRHex here. And just to convince说服 them
190
726000
3000
和RHex的简单构造,为了使他们相信
12:27
we should go to Mars火星 in 2011, I tinted有色 the video视频 orange橙子
191
729000
4000
我们2011年得上一趟火星,我把短片背景设成橘黄色
12:31
just to give them the sense of being存在 on Mars火星.
192
733000
3000
让他们感觉仿佛是在火星上
12:34
(Laughter笑声)
193
736000
1000
(笑)
12:35
(Applause掌声)
194
737000
6000
(掌声)
12:43
Another另一个 reason原因 why animals动物 have extraordinary非凡 performance性能,
195
745000
3000
动物们生存能力极强
12:46
and can go anywhere随地, is because they have an effective有效 interaction相互作用
196
748000
3000
分布极广的另一个原因,就是他们能够与环境
12:49
with the environment环境. The animal动物 I'm going to show显示 you,
197
751000
3000
进行有效交流。我下面给你们看的
12:52
that we studied研究 to look at this, is the gecko壁虎.
198
754000
4000
我们所研究的动物是壁虎
12:56
We have one here and notice注意 its position位置. It's holding保持 on.
199
758000
7000
这儿就有一只,注意看它的姿势,它紧抓着不放
13:03
Now I'm going to challenge挑战 you. I'm going show显示 you a video视频.
200
765000
3000
我现在要给你们出个题目,给你们看一段短片
13:06
One of the animals动物 is going to be running赛跑 on the level水平,
201
768000
2000
其中一只动物将在平地上跑
13:08
and the other one's那些 going to be running赛跑 up a wall. Which哪一个 one's那些 which哪一个?
202
770000
4000
另一只在爬上一堵墙。哪一只对应哪一只呢?
13:12
They're going at a meter仪表 a second第二. How many许多 think the one on the left
203
774000
5000
它们的速度都是一米每秒。有多少人认为左边的这只
13:17
is running赛跑 up the wall?
204
779000
2000
正在爬墙?
13:19
(Applause掌声)
205
781000
4000
(掌声)
13:23
Okay. The point is it's really hard to tell, isn't it? It's incredible难以置信,
206
785000
5000
好的。看来是真的很难说,不是吗?难以置信啊
13:28
we looked看着 at students学生们 do this and they couldn't不能 tell.
207
790000
2000
我们的学生们也很难分辨出来
13:30
They can run up a wall at a meter仪表 a second第二, 15 steps脚步 per second第二,
208
792000
3000
壁虎以每秒一米的速度爬墙,每秒迈出15步
13:33
and they look like they're running赛跑 on the level水平. How do they do this?
209
795000
4000
看起来就像在平地上跑。怎么做到的呢?
13:37
It's just phenomenal非凡的. The one on the right was going up the hill爬坡道.
210
799000
6000
实在是了不起。右边的那只正在上墙
13:43
How do they do this? They have bizarre奇异的 toes脚趾. They have toes脚趾
211
805000
4000
是如何做的呢--它们的脚趾十分奇特--
13:47
that uncurl舒展 like party派对 favors好处 when you blow打击 them out,
212
809000
4000
它们像被吹过的聚会小礼物一样翻开
13:51
and then peel off the surface表面, like tape胶带.
213
813000
3000
如同带子一样从表面剥张开来
13:54
Like if we had a piece of tape胶带 now, we'd星期三 peel it this way.
214
816000
2000
比方说像我们这么剥开一条带子
13:56
They do this with their toes脚趾. It's bizarre奇异的! This peeling去皮 inspired启发
215
818000
7000
它们的脚趾能做这个。真是出了奇了。这个翻开的动作启发了
14:03
iRobot我是机器人 -- that we work with -- to build建立 Mecho-GeckosMecho,壁虎.
216
825000
3000
与我们共事的iRobot,于是制造出了Mecho-Geckos (机械壁虎)
14:06
Here's这里的 a legged version and a tractor拖拉机 version, or a bulldozer推土机 version.
217
828000
7000
这里有带腿型的、拖拉机型的和推土机型的
14:13
Let's see some of the geckos壁虎 move移动 with some video视频,
218
835000
2000
我们先从短片里看看壁虎的运动
14:15
and then I'll show显示 you a little bit of a clip of the robots机器人.
219
837000
3000
然后我再给你们看机器人
14:18
Here's这里的 the gecko壁虎 running赛跑 up a vertical垂直 surface表面. There it goes,
220
840000
3000
壁虎正在冲上一个垂直表面,一蹴而就
14:21
in real真实 time. There it goes again. Obviously明显, we have to slow this down a little bit.
221
843000
7000
这是实时播放,再来一遍。显然我们得慢放一点点
14:28
You can't use regular定期 cameras相机.
222
850000
2000
不能用普通镜头
14:30
You have to take 1,000 pictures图片 per second第二 to see this.
223
852000
3000
你得每秒捕捉1000张图像才能看到这个
14:33
And here's这里的 some video视频 at 1,000 frames per second第二.
224
855000
3000
这个短片是每秒1000帧的
14:36
Now, I want you to look at the animal's动物 back.
225
858000
2000
现在我要你们注意看这只动物的背部
14:38
Do you see how much it's bending弯曲 like that? We can't figure数字 that out --
226
860000
3000
你能看出它弯曲了多少吗?我们搞不清楚--
14:41
that's an unsolved未解 mystery神秘. We don't know how it works作品.
227
863000
3000
这还是个未解之谜。不知道它的工作原理是什么
14:44
If you have a son儿子 or a daughter女儿 that wants to come to Berkeley伯克利,
228
866000
3000
如果你的儿子或者女儿想来伯克利念书
14:47
come to my lab实验室 and we'll figure数字 this out. Okay, send发送 them to Berkeley伯克利
229
869000
4000
就来我的实验室我们把它弄清楚。是啊,送他们到伯克利来
14:51
because that's the next下一个 thing I want to do. Here's这里的 the gecko壁虎 mill.
230
873000
3000
因为我下面就想研究那个问题。这是壁虎跑步机
14:54
(Laughter笑声)
231
876000
1000
(笑)
14:55
It's a see-through透视 treadmill跑步机 with a see-through透视 treadmill跑步机 belt,
232
877000
3000
是个透明的跑步机,透明的跑步带
14:58
so we can watch the animal's动物 feet, and videotape录影带 them
233
880000
3000
我们能观察动物们的脚,给它们录像
15:01
through通过 the treadmill跑步机 belt, to see how they move移动.
234
883000
3000
透过这个跑步带,看它们如何运动
15:04
Here's这里的 the animal动物 that we have here, running赛跑 on a vertical垂直 surface表面.
235
886000
4000
这是我们的动物正在垂直面上跑
15:08
Pick a foot脚丫子 and try to watch a toe脚趾, and see if you can see what the animal's动物 doing.
236
890000
6000
选一只脚,试着观察一根脚趾,看你能否察觉动物在干嘛
15:14
See it uncurl舒展 and then peel these toes脚趾.
237
896000
2000
看看它翻开又剥张这些脚趾
15:16
It can do this in 14 milliseconds毫秒. It's unbelievable难以置信的.
238
898000
7000
14毫秒之内就完成了,实在是难以置信
15:23
Here are the robots机器人 that they inspire启发, the Mecho-GeckosMecho,壁虎 from iRobot我是机器人.
239
905000
4000
这就是受它们启发的机器人,iRobot的机械壁虎
15:27
First we'll see the animals动物 toes脚趾 peeling去皮 -- look at that.
240
909000
5000
首先我们看到动物的脚趾剥开--看看这个
15:32
And here's这里的 the peeling去皮 action行动 of the Mecho-GeckoMecho,壁虎.
241
914000
4000
这就是机械壁虎的那个剥开的动作
15:36
It uses使用 a pressure-sensitive压敏 adhesive胶粘剂 to do it.
242
918000
3000
它用一个对压强敏感的粘合物来完成这个动作
15:39
Peeling去皮 in the animal动物. Peeling去皮 in the Mecho-GeckoMecho,壁虎 --
243
921000
3000
这是动物的剥开动作,这是机械壁虎的
15:42
that allows允许 them climb autonomously自主. Can go on the flat平面 surface表面,
244
924000
3000
都使他们能够自己攀爬,先走在平面上
15:45
transition过渡 to a wall, and then go onto a ceiling天花板.
245
927000
3000
移到墙上,再上天花板
15:48
There's the bulldozer推土机 version. Now, it doesn't use pressure-sensitive压敏 glue.
246
930000
6000
这是推土机版本的。这个用的可不是压强敏感性的粘胶
15:54
The animal动物 does not use that.
247
936000
2000
动物可不用那个
15:56
But that's what we're limited有限 to, at the moment时刻.
248
938000
2000
但是我们现在只有这个可用
15:58
What does the animal动物 do? The animal动物 has weird奇怪的 toes脚趾.
249
940000
5000
动物怎么做呢?它有怪异的脚趾
16:03
And if you look at the toes脚趾, they have these little leaves树叶 there,
250
945000
4000
如果你看仔细点它们有这些细小的叶片
16:07
and if you blow打击 them up and zoom放大 in, you'll你会 see
251
949000
2000
如果你把它们放大你会看见
16:09
that's there's little striations条纹 in these leaves树叶.
252
951000
3000
这些叶片里面有小条纹
16:12
And if you zoom放大 in 270 times, you'll你会 see it looks容貌 like a rug地毯.
253
954000
7000
你要是放大270倍,你会看见它其实长得像条毯子
16:19
And if you blow打击 that up, and zoom放大 in 900 times,
254
961000
3000
如果你再放大900倍
16:22
you see there are hairs there, tiny hairs. And if you look carefully小心,
255
964000
5000
你看到有毛,细小的毛,再看仔细些
16:27
those tiny hairs have striations条纹. And if you zoom放大 in on those 30,000 times,
256
969000
6000
这些毛上又有条纹。如果你再把它们都放大三万倍
16:33
you'll你会 see each hair头发 has split分裂 ends结束.
257
975000
3000
你看到每一根细毛头上都有分叉
16:36
And if you blow打击 those up, they have these little structures结构 on the end结束.
258
978000
5000
你要是再放大就会看到它们的头上都有一些小结构
16:41
The smallest最少 branch of the hairs looks容貌 like spatulae匙突,
259
983000
2000
毛最小的分枝看起来像小铲
16:43
and an animal动物 like that has one billion十亿 of these nano-size纳米尺寸 split分裂 ends结束,
260
985000
7000
这样一个小动物能用10亿根纳米级的分叉毛
16:50
to get very close to the surface表面. In fact事实, there's the diameter直径 of your hair头发 --
261
992000
5000
贴近物体的表面。这是你头发的直径
16:55
a gecko壁虎 has two million百万 of these, and each hair头发 has 100 to 1,000 split分裂 ends结束.
262
997000
6000
一只壁虎有200万个这样的,每根毛有100到1000个分叉
17:01
Think of the contact联系 of that that's possible可能.
263
1003000
3000
想象一下它们能形成的接触
17:04
We were fortunate幸运 to work with another另一个 group
264
1006000
2000
我们有幸和另外一个斯坦福的小组合作
17:06
at Stanford斯坦福 that built内置 us a special特别 manned载人 sensor传感器,
265
1008000
2000
造出了一个特殊控制的感应器
17:08
that we were able能够 to measure测量 the force of an individual个人 hair头发.
266
1010000
3000
能够帮我们测量一根单个毛发受到的力
17:11
Here's这里的 an individual个人 hair头发 with a little split分裂 end结束 there.
267
1013000
5000
这是一根带了点分叉的毛
17:16
When we measured测量 the forces军队, they were enormous巨大.
268
1018000
2000
当我们测量受力时发现大得惊人
17:18
They were so large that a patch补丁 of hairs about this size尺寸 --
269
1020000
3000
大到像一片这样尺寸的毛
17:21
the gecko's壁虎 foot脚丫子 could support支持 the weight重量 of a small child儿童,
270
1023000
4000
就壁虎脚这么大--能轻易撑起一个小孩的重量
17:25
about 40 pounds英镑, easily容易. Now, how do they do it?
271
1027000
4000
大约40磅。它们怎么做到的呢?
17:29
We've我们已经 recently最近 discovered发现 this. Do they do it by friction摩擦?
272
1031000
4000
我们最近发现了。是摩擦力吗?
17:33
No, force is too low. Do they do it by electrostatics静电学?
273
1035000
3000
不是,摩擦力太小了。是静电吗?
17:36
No, you can change更改 the charge收费 -- they still hold保持 on.
274
1038000
2000
不是,你改变一下电荷它们还是抓得牢牢地
17:38
Do they do it by interlocking联锁? That's kind of a like a Velcro-like魔术贴样 thing.
275
1040000
3000
是通过形状的互相咬合吗?是像维可劳(尼龙剌粘搭链)那样的吗?
17:41
No, you can put them on molecular分子 smooth光滑 surfaces -- they don't do it.
276
1043000
3000
不是,你把它们放到分子级的光滑平面上--就不起作用了
17:44
How about suction吸力? They stick on in a vacuum真空.
277
1046000
4000
空气吸力吗?它们在真空里一样能贴着
17:48
How about wet湿 adhesion粘着? Or capillary毛细管 adhesion粘着?
278
1050000
3000
湿粘合呢?或者毛细管作用呢?
17:51
They don't have any glue, and they even stick under water just fine.
279
1053000
3000
它们可没有用胶,而且在水下抓得一样地紧
17:54
If you put their foot脚丫子 under water, they grab on.
280
1056000
2000
把脚放在水里也能抓住
17:56
How do they do it then? Believe it or not, they grab on
281
1058000
4000
怎么回事呢?信不信由你
18:00
by intermolecular forces军队, by Van面包车 derDER Waals瓦尔斯 forces军队.
282
1062000
4000
它们用了分子间作用力,范德华力
18:04
You know, you probably大概 had this a long time ago in chemistry化学,
283
1066000
2000
你可能还是很久以前在化学里学的这个
18:06
where you had these two atoms原子, they're close together一起,
284
1068000
2000
两个原子,靠得挺近
18:08
and the electrons电子 are moving移动 around. That tiny force is sufficient足够
285
1070000
3000
电子绕着它们动。这个小小的力足够
18:11
to allow允许 them to do that because it's added添加 up so many许多 times
286
1073000
3000
完成这个因为它被加成了许多许多次
18:14
with these small structures结构.
287
1076000
3000
通过这些小小的结构
18:17
What we're doing is, we're taking服用 that inspiration灵感 of the hairs,
288
1079000
5000
我们正在做的就是利用这些毛带来的灵感
18:22
and with another另一个 colleague同事 at Berkeley伯克利, we're manufacturing制造业 them.
289
1084000
5000
和我在伯克利的另一位同事一起制造它们
18:27
And just recently最近 we've我们已经 made制作 a breakthrough突破, where we now believe
290
1089000
3000
就是在最近我们有了一项突破,现在我们相信
18:30
we're going to be able能够 to create创建 the first synthetic合成的, self-cleaning自洁,
291
1092000
5000
我们就要制出第一种合成的,自我清洁
18:35
dry adhesive胶粘剂. Many许多 companies公司 are interested有兴趣 in this.
292
1097000
5000
干性的粘合物。许多公司都很感兴趣
18:40
(Laughter笑声)
293
1102000
3000
(笑)
18:43
We also presented呈现 to Nike耐克 even.
294
1105000
2000
我们甚至给耐克做了展示
18:45
(Laughter笑声)
295
1107000
3000
(笑)
18:48
(Applause掌声)
296
1110000
6000
(掌声)
18:54
We'll see where this goes. We were so excited兴奋 about this
297
1116000
3000
我们得看看它的前景了。真是让人兴奋呢
18:57
that we realized实现 that that small-size小尺寸 scale规模 --
298
1119000
3000
发现在极小的尺寸下
19:00
and where everything gets得到 sticky, and gravity重力 doesn't matter anymore --
299
1122000
3000
所有的东西都粘在一块儿,重力再也不重要了
19:03
we needed需要 to look at ants蚂蚁 and their feet, because
300
1125000
3000
我们需要看看蚂蚁和它们的脚,因为
19:06
one of my other colleagues同事 at Berkeley伯克利 has built内置 a six-millimeter六毫米 silicone
301
1128000
5000
伯克利的另一位同事,造了一个六毫米的
19:11
robot机器人 with legs. But it gets得到 stuck卡住. It doesn't move移动 very well.
302
1133000
3000
带腿的硅制机器人。但是它老卡住,动得不灵活
19:14
But the ants蚂蚁 do, and we'll figure数字 out why, so that ultimately最终
303
1136000
3000
可是蚂蚁相当灵活,我们就要知道为什么,就可以
19:17
we'll make this move移动. And imagine想像: you're going to be able能够
304
1139000
3000
相应作出改变。想象一下,你很快就可以
19:20
to have swarms成群 of these six-millimeter六毫米 robots机器人 available可得到 to run around.
305
1142000
5000
让一大群的六毫米大小的机器人到处乱跑
19:25
Where's哪里 this going? I think you can see it already已经.
306
1147000
3000
这有什么前景呢?我觉得你现在已经想到了
19:28
Clearly明确地, the Internet互联网 is already已经 having eyes眼睛 and ears耳朵,
307
1150000
4000
显然互联网已经拥有了眼睛和耳朵
19:32
you have web卷筒纸 cams凸轮 and so forth向前. But it's going to also have legs and hands.
308
1154000
4000
有网络摄像头等等东西,但是以后它还会有腿与手
19:36
You're going to be able能够 to do programmable可编程的
309
1158000
2000
你就可以用这些机器人
19:38
work through通过 these kinds of robots机器人, so that you can run,
310
1160000
4000
完成可编程的工作,可以到处跑啊
19:42
fly and swim游泳 anywhere随地. We saw David大卫 Kelly黄绿色 is at the beginning开始 of that with his fish.
311
1164000
9000
飞啊游啊。我们开始的时候看到了David Kelly和他的鱼
19:51
So, in conclusion结论, I think the message信息 is clear明确.
312
1173000
2000
所以综上所述,我觉得这个信息很清楚
19:53
If you need a message信息, if nature's大自然 not enough足够, if you care关心 about
313
1175000
4000
如果你需要一个信息,如果光是自然造物还不够,如果你关注
19:57
search搜索 and rescue拯救, or mine clearance净空, or medicine医学,
314
1179000
2000
搜救工作、矿井清理、医药
19:59
or the various各个 things we're working加工 on, we must必须 preserve保留
315
1181000
4000
还有我们在做的各种各样的事情,我们必须好好保管
20:03
nature's大自然 designs设计, otherwise除此以外 these secrets秘密 will be lost丢失 forever永远.
316
1185000
4000
大自然的精妙设计,不然这些秘密会永远消失
20:07
Thank you.
317
1189000
1000
谢谢你们
20:08
(Applause掌声)
318
1190000
9000
(掌声)
Translated by Miao Li
Reviewed by HUI YU

▲Back to top

ABOUT THE SPEAKER
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.

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.

More profile about the speaker
Robert Full | Speaker | TED.com