TED2012
Vijay Kumar: Robots that fly ... and cooperate
Vijay Kumar 談論：會飛行還懂得團隊合作的機器人
Filmed:
Readability: 4.5
5,188,706 views
Vijay Kumar 和他的團隊在賓州大學的實驗室中製作了使用四個螺旋槳飛行、小巧的機器人，這些機器人可以聚集在一起，感覺彼此的存在，並且形成任務編組 -- 進行建造、災區偵查以及其它事情。
Vijay Kumar - Roboticist
As the dean of the University of Pennsylvania's School of Engineering and Applied Science, Vijay Kumar studies the control and coordination of multi-robot formations. Full bio
As the dean of the University of Pennsylvania's School of Engineering and Applied Science, Vijay Kumar studies the control and coordination of multi-robot formations. Full bio
Double-click the English transcript below to play the video.
00:20
Good morning.
0
5000
2000
早安。
00:22
I'm here today to talk
1
7000
2000
今天我想要來談一談
00:24
about autonomous, flying beach balls.
2
9000
3000
會自動飛行的海灘球。
00:27
No, agile aerial robots like this one.
3
12000
4000
不是啦，是靈巧的飛行機器人，就像這一個。
00:31
I'd like to tell you a little bit about the challenges in building these
4
16000
3000
我想告訴大家製作這種東西的挑戰性
00:34
and some of the terrific opportunities
5
19000
2000
以及一些很棒的可能性
00:36
for applying this technology.
6
21000
2000
來運用這種技術。
00:38
So these robots
7
23000
2000
這些機器人
00:40
are related to unmanned aerial vehicles.
8
25000
3000
算是一種無人的飛行器。
00:43
However, the vehicles you see here are big.
9
28000
3000
不過，如你所見，它們的尺寸都比較大。
00:46
They weigh thousands of pounds,
10
31000
2000
它們都有幾千磅重，
00:48
are not by any means agile.
11
33000
2000
一點都不靈巧。
00:50
They're not even autonomous.
12
35000
2000
它們甚至並不是自動操作的。
00:52
In fact, many of these vehicles
13
37000
2000
事實上，大部分這些飛行器
00:54
are operated by flight crews
14
39000
2000
是由飛行小組所操作，
00:56
that can include multiple pilots,
15
41000
3000
可能有好幾個駕駛員
00:59
operators of sensors
16
44000
2000
同時在操控著感應器
01:01
and mission coordinators.
17
46000
2000
以及任務協調器。
01:03
What we're interested in is developing robots like this --
18
48000
2000
我們想要開發的機器人是像這個樣子 --
01:05
and here are two other pictures --
19
50000
2000
左邊這裡另外兩張照片--
01:07
of robots that you can buy off the shelf.
20
52000
3000
這些你都可以買到現成的。
01:10
So these are helicopters with four rotors
21
55000
3000
這些是一種具有四個螺旋槳的直昇機，
01:13
and they're roughly a meter or so in scale
22
58000
4000
它們大約是一公尺大小，
01:17
and weigh several pounds.
23
62000
2000
也有好幾磅重。
01:19
And so we retrofit these with sensors and processors,
24
64000
3000
於是我們將它們進行感應器與處理器的改良，
01:22
and these robots can fly indoors
25
67000
2000
讓這些機器人能夠在室內
01:24
without GPS.
26
69000
2000
不靠GPS飛行。
01:26
The robot I'm holding in my hand
27
71000
2000
我手中所拿的這個機器人
01:28
is this one,
28
73000
2000
就是這種飛行器，
01:30
and it's been created by two students,
29
75000
3000
這是由兩位學生所製作的，
01:33
Alex and Daniel.
30
78000
2000
Alex 以及 Daniel。
01:35
So this weighs a little more
31
80000
2000
它的重量大概是
01:37
than a tenth of a pound.
32
82000
2000
十分之一磅左右。
01:39
It consumes about 15 watts of power.
33
84000
2000
它消耗的能量大概是15瓦。
01:41
And as you can see,
34
86000
2000
如你所見，
01:43
it's about eight inches in diameter.
35
88000
2000
它的直徑大概是8英吋大。
01:45
So let me give you just a very quick tutorial
36
90000
3000
讓我替大家簡單介紹一下
01:48
on how these robots work.
37
93000
2000
這些機器人的原理。
01:50
So it has four rotors.
38
95000
2000
這裡有四個螺旋槳。
01:52
If you spin these rotors at the same speed,
39
97000
2000
當這四個螺旋槳速度相同時，
01:54
the robot hovers.
40
99000
2000
機器人就會懸浮在空中。
01:56
If you increase the speed of each of these rotors,
41
101000
3000
如果這些螺旋槳速度增加，
01:59
then the robot flies up, it accelerates up.
42
104000
3000
機器人就會飛起來，往上加速。
02:02
Of course, if the robot were tilted,
43
107000
2000
當然，如果機器人傾斜了，
02:04
inclined to the horizontal,
44
109000
2000
相對於水平線來說，
02:06
then it would accelerate in this direction.
45
111000
3000
它就會往這個方向前進。
02:09
So to get it to tilt, there's one of two ways of doing it.
46
114000
3000
想讓它傾斜的話，這裡有兩種方法可以辦到。
02:12
So in this picture
47
117000
2000
在這圖片中，
02:14
you see that rotor four is spinning faster
48
119000
2000
你可以看見4號螺旋槳轉速變快一點，
02:16
and rotor two is spinning slower.
49
121000
2000
而2號螺旋槳轉速變慢一點。
02:18
And when that happens
50
123000
2000
當這種情況發生時，
02:20
there's moment that causes this robot to roll.
51
125000
3000
就會讓機器人進行翻轉。
02:23
And the other way around,
52
128000
2000
另一種狀況是，
02:25
if you increase the speed of rotor three
53
130000
3000
當3號螺旋槳的速度上升，
02:28
and decrease the speed of rotor one,
54
133000
2000
1號螺旋槳的速度下降時，
02:30
then the robot pitches forward.
55
135000
3000
機器人就會往前傾斜。
02:33
And then finally,
56
138000
2000
而最後一種可能，
02:35
if you spin opposite pairs of rotors
57
140000
2000
當對角線的兩組螺旋槳
02:37
faster than the other pair,
58
142000
2000
轉得比另外一組快時，
02:39
then the robot yaws about the vertical axis.
59
144000
2000
機器人就會在垂直方向偏移。
02:41
So an on-board processor
60
146000
2000
有一個內置處理器
02:43
essentially looks at what motions need to be executed
61
148000
3000
一直在監控著該進行什麼動作，
02:46
and combines these motions
62
151000
2000
並且將這些動作進行組合，
02:48
and figures out what commands to send to the motors
63
153000
3000
然後以每秒600次的速度
02:51
600 times a second.
64
156000
2000
決定出該對這些螺旋槳下達什麼指令。
02:53
That's basically how this thing operates.
65
158000
2000
這就是它操作的基本概念。
02:55
So one of the advantages of this design
66
160000
2000
這種設計的其中一項優點是，
02:57
is, when you scale things down,
67
162000
2000
當你將它的尺寸縮小時，
02:59
the robot naturally becomes agile.
68
164000
3000
機器人自然就會變得很靈巧。
03:02
So here R
69
167000
2000
這邊的 R
03:04
is the characteristic length of the robot.
70
169000
2000
代表著機器人特性的長度。
03:06
It's actually half the diameter.
71
171000
3000
事實上是直徑的一半。
03:09
And there are lots of physical parameters that change
72
174000
3000
而當你將 R 縮減時，
03:12
as you reduce R.
73
177000
2000
許多物理係數就會跟著變動。
03:14
The one that's the most important
74
179000
2000
其中最重要的
03:16
is the inertia or the resistance to motion.
75
181000
2000
就是慣性或稱為阻止變動的抵抗力。
03:18
So it turns out,
76
183000
2000
結果，
03:20
the inertia, which governs angular motion,
77
185000
3000
控制了角運動的慣性，
03:23
scales as a fifth power of R.
78
188000
3000
大小約是 R 的 5 次方。
03:26
So the smaller you make R,
79
191000
2000
所以當 R 變小時，
03:28
the more dramatically the inertia reduces.
80
193000
3000
慣性會急遽的下降。
03:31
So as a result, the angular acceleration,
81
196000
3000
結果，角加速度，
03:34
denoted by Greek letter alpha here,
82
199000
2000
這裡用希臘字母的 α 表示，
03:36
goes as one over R.
83
201000
2000
變成了 1 / R 。
03:38
It's inversely proportional to R.
84
203000
2000
它和 R 成反比。
03:40
The smaller you make it the more quickly you can turn.
85
205000
3000
當尺寸越小時，它就越容易旋轉。
03:43
So this should be clear in these videos.
86
208000
2000
用這個影片說明會清楚一點。
03:45
At the bottom right you see a robot
87
210000
3000
在右下角，你可以看見一個機器人
03:48
performing a 360 degree flip
88
213000
2000
正在進行 360 度翻轉
03:50
in less than half a second.
89
215000
2000
在不到 1/2 秒的時間內。
03:52
Multiple flips, a little more time.
90
217000
3000
多次的翻轉，只要稍微長一點點的時間。
03:55
So here the processes on board
91
220000
2000
在這種狀況下，內置的處理器
03:57
are getting feedback from accelerometers
92
222000
2000
接收了加速器
03:59
and gyros on board
93
224000
2000
以及陀螺儀回傳的資訊，
04:01
and calculating, like I said before,
94
226000
2000
然後進行計算，如先前所說，
04:03
commands at 600 times a second
95
228000
2000
用每秒600次的速度發出指令，
04:05
to stabilize this robot.
96
230000
2000
讓機器人保持平衡。
04:07
So on the left, you see Daniel throwing this robot up into the air.
97
232000
3000
在左下角，Daniel 正將機器人拋向空中。
04:10
And it shows you how robust the control is.
98
235000
2000
這會讓你知道它的操控能力有多強大。
04:12
No matter how you throw it,
99
237000
2000
不論你怎麼丟，
04:14
the robot recovers and comes back to him.
100
239000
4000
機器人都能恢復平衡然後回到他的手中。
04:18
So why build robots like this?
101
243000
2000
為什麼要將機器人設計成這樣呢？
04:20
Well robots like this have many applications.
102
245000
3000
嗯，這種機器人有很多種運用方式。
04:23
You can send them inside buildings like this
103
248000
3000
你可以將它派遣到這種建築物裡，
04:26
as first responders to look for intruders,
104
251000
3000
擔任先遣部隊去找出侵入者，
04:29
maybe look for biochemical leaks,
105
254000
3000
或是去找尋生化物質洩漏，
04:32
gaseous leaks.
106
257000
2000
或是瓦斯洩漏等。
04:34
You can also use them
107
259000
2000
你也可以將它們運用在
04:36
for applications like construction.
108
261000
2000
例如建築上面。
04:38
So here are robots carrying beams, columns
109
263000
4000
這裡的機器人正運送著橫梁、柱子，
04:42
and assembling cube-like structures.
110
267000
3000
並且組合成立方體形狀的建築物。
04:45
I'll tell you a little bit more about this.
111
270000
3000
我再告訴大家詳細一點。
04:48
The robots can be used for transporting cargo.
112
273000
3000
這些機器人可以用來運送貨櫃。
04:51
So one of the problems with these small robots
113
276000
3000
但這些小機器人的困難在於
04:54
is their payload carrying capacity.
114
279000
2000
它們對於重物的負載能力有限。
04:56
So you might want to have multiple robots
115
281000
2000
所以如果你可能會希望能有多一點機器人
04:58
carry payloads.
116
283000
2000
一起來搬運這個重物。
05:00
This is a picture of a recent experiment we did --
117
285000
2000
這是我們近期實驗的照片 --
05:02
actually not so recent anymore --
118
287000
2000
事實上已經不算是近期了 --
05:04
in Sendai shortly after the earthquake.
119
289000
3000
在地震過後的仙台市(日本)。
05:07
So robots like this could be sent into collapsed buildings
120
292000
3000
這種機器人可以被派遣進入傾倒的建築物裡面
05:10
to assess the damage after natural disasters,
121
295000
2000
去評估天災造成的損害，
05:12
or sent into reactor buildings
122
297000
3000
或是派遣到反應爐裡
05:15
to map radiation levels.
123
300000
3000
去勘查輻射等級。
05:19
So one fundamental problem
124
304000
2000
如果這些機器人想有自主能力的話，
05:21
that the robots have to solve if they're to be autonomous
125
306000
3000
它們必須先解決這個問題，
05:24
is essentially figuring out
126
309000
2000
就是必須能夠判斷
05:26
how to get from point A to point B.
127
311000
2000
怎麼從 A 點到達 B 點。
05:28
So this gets a little challenging
128
313000
2000
這有一點難度，
05:30
because the dynamics of this robot are quite complicated.
129
315000
3000
因為這個機器人的動力學是相當複雜的。
05:33
In fact, they live in a 12-dimensional space.
130
318000
2000
事實上，它們活在 12 維空間裡。
05:35
So we use a little trick.
131
320000
2000
所以我們運用了一些技巧。
05:37
We take this curved 12-dimensional space
132
322000
3000
我們將這個 12 維空間的曲線
05:40
and transform it
133
325000
2000
轉換成為
05:42
into a flat four-dimensional space.
134
327000
2000
一個平面的四維空間。
05:44
And that four-dimensional space
135
329000
2000
在這個四維空間之中，
05:46
consists of X, Y, Z and then the yaw angle.
136
331000
3000
包含了 X, Y, Z 還有偏移的角度。
05:49
And so what the robot does
137
334000
2000
所以這個機器人所做的是，
05:51
is it plans what we call a minimum snap trajectory.
138
336000
4000
去找出我們所說的最小震盪軌跡。
05:55
So to remind you of physics,
139
340000
2000
複習一下物理參數，
05:57
you have position, derivative, velocity,
140
342000
2000
我們有位置，接著衍生出速度，
05:59
then acceleration,
141
344000
2000
以及加速度，
06:01
and then comes jerk
142
346000
2000
還有加加速度，
06:03
and then comes snap.
143
348000
2000
然後是震盪。
06:05
So this robot minimizes snap.
144
350000
3000
所以機器人將震盪進行最小化。
06:08
So what that effectively does
145
353000
2000
這實際上的結果就是
06:10
is produces a smooth and graceful motion.
146
355000
2000
產生出柔順且優美的動作。
06:12
And it does that avoiding obstacles.
147
357000
3000
它還可以用來避開障礙物。
06:15
So these minimum snap trajectories in this flat space
148
360000
3000
而這些最小震盪軌跡在這個平面空間中
06:18
are then transformed back
149
363000
2000
又會被轉換回
06:20
into this complicated 12-dimensional space,
150
365000
2000
這個複雜的 12 維空間，
06:22
which the robot must do
151
367000
2000
才能夠讓機器人去進行
06:24
for control and then execution.
152
369000
2000
控制以及執行任務。
06:26
So let me show you some examples
153
371000
2000
讓我給大家看一些例子
06:28
of what these minimum snap trajectories look like.
154
373000
2000
說明這些最小震盪軌跡是什麼樣子。
06:30
And in the first video,
155
375000
2000
在第一段影片中，
06:32
you'll see the robot going from point A to point B
156
377000
2000
你可以看見機器人經過中繼點
06:34
through an intermediate point.
157
379000
2000
由 A 點到達 B 點。
06:42
So the robot is obviously capable
158
387000
2000
所以機器人確實可以
06:44
of executing any curve trajectory.
159
389000
2000
去執行任何曲線軌跡。
06:46
So these are circular trajectories
160
391000
2000
這些是環狀軌跡，
06:48
where the robot pulls about two G's.
161
393000
3000
機器人牽引著大約 2 G 的重力。
06:52
Here you have overhead motion capture cameras on the top
162
397000
4000
在上面有個置頂動態影像攝影機，
06:56
that tell the robot where it is 100 times a second.
163
401000
3000
它會以每秒100次的速度告訴機器人自己在哪裡。
06:59
It also tells the robot where these obstacles are.
164
404000
3000
它也會告訴機器人這些障礙物的位置。
07:02
And the obstacles can be moving.
165
407000
2000
這些也可以是移動中的障礙物。
07:04
And here you'll see Daniel throw this hoop into the air,
166
409000
3000
你將會看見 Daniel 將這個鐵環丟向空中，
07:07
while the robot is calculating the position of the hoop
167
412000
2000
機器人會計算鐵環的位置，
07:09
and trying to figure out how to best go through the hoop.
168
414000
4000
然後試著去找出穿過鐵環的最佳方式。
07:13
So as an academic,
169
418000
2000
身為一個學術人員，
07:15
we're always trained to be able to jump through hoops to raise funding for our labs,
170
420000
3000
我們總是被訓練得能夠赴湯蹈火才能籌措研究經費，
07:18
and we get our robots to do that.
171
423000
3000
所以我們也要我們的機器人做到。
07:21
(Applause)
172
426000
6000
(掌聲)
07:27
So another thing the robot can do
173
432000
2000
這機器人還能做另一件事，
07:29
is it remembers pieces of trajectory
174
434000
3000
就是去記住軌跡的片段，
07:32
that it learns or is pre-programmed.
175
437000
2000
不論是它自行發現的或是事先輸入的。
07:34
So here you see the robot
176
439000
2000
所以你可以看見機器人會去
07:36
combining a motion
177
441000
2000
組合一項動作
07:38
that builds up momentum
178
443000
2000
讓它產生動量，
07:40
and then changes its orientation and then recovers.
179
445000
3000
接著改變自己的行進方向在回復過來。
07:43
So it has to do this because this gap in the window
180
448000
3000
它必須這麼做，因為這個窗戶的缺口大小
07:46
is only slightly larger than the width of the robot.
181
451000
4000
只比機器人的寬度稍微大一點。
07:50
So just like a diver stands on a springboard
182
455000
3000
就像是跳水選手站在跳板上，
07:53
and then jumps off it to gain momentum,
183
458000
2000
接著會跳起來用以產生動量，
07:55
and then does this pirouette, this two and a half somersault through
184
460000
3000
然後快速旋轉，翻轉兩周半進行穿越，
07:58
and then gracefully recovers,
185
463000
2000
最後優雅的回復，
08:00
this robot is basically doing that.
186
465000
2000
這就是機器人所做的事。
08:02
So it knows how to combine little bits and pieces of trajectories
187
467000
3000
它懂得如何去結合這些零碎的軌跡
08:05
to do these fairly difficult tasks.
188
470000
4000
以達成這些相當困難的任務。
08:09
So I want change gears.
189
474000
2000
我想換個話題。
08:11
So one of the disadvantages of these small robots is its size.
190
476000
3000
這些小機器人的缺點之一就是尺寸。
08:14
And I told you earlier
191
479000
2000
如同先前所提，
08:16
that we may want to employ lots and lots of robots
192
481000
2000
我們想使用大量的機器人
08:18
to overcome the limitations of size.
193
483000
3000
來解決尺寸上的限制。
08:21
So one difficulty
194
486000
2000
但有個困難點是
08:23
is how do you coordinate lots of these robots?
195
488000
3000
你要如何去協調這些機器人呢？
08:26
And so here we looked to nature.
196
491000
2000
這部份我們觀察了自然界。
08:28
So I want to show you a clip
197
493000
2000
我想讓大家看一段影片，
08:30
of Aphaenogaster desert ants
198
495000
2000
關於沙漠盤腹蟻
08:32
in Professor Stephen Pratt's lab carrying an object.
199
497000
3000
在 Stephen Pratt 教授的實驗室裡搬運東西。
08:35
So this is actually a piece of fig.
200
500000
2000
事實上這是一小塊無花果。
08:37
Actually you take any object coated with fig juice
201
502000
2000
事實上你可以把任何東西沾附一層無花果汁
08:39
and the ants will carry them back to the nest.
202
504000
3000
螞蟻們就會將它搬回巢穴裡。
08:42
So these ants don't have any central coordinator.
203
507000
3000
這些螞蟻並沒有中樞協調者。
08:45
They sense their neighbors.
204
510000
2000
它們能感覺到旁邊的鄰居們。
08:47
There's no explicit communication.
205
512000
2000
不用進行明確的溝通。
08:49
But because they sense the neighbors
206
514000
2000
但因為它們能感覺到鄰居，
08:51
and because they sense the object,
207
516000
2000
因為它們能感覺到東西，
08:53
they have implicit coordination across the group.
208
518000
3000
它們在團體間有著隱性協調能力。
08:56
So this is the kind of coordination
209
521000
2000
這種協調能力
08:58
we want our robots to have.
210
523000
3000
就是我們希望機器人能有的。
09:01
So when we have a robot
211
526000
2000
當我們的一個機器人
09:03
which is surrounded by neighbors --
212
528000
2000
被周圍的機器人包圍時 --
09:05
and let's look at robot I and robot J --
213
530000
2000
看看機器人 I 和機器人 J --
09:07
what we want the robots to do
214
532000
2000
我們希望機器人做的事情是
09:09
is to monitor the separation between them
215
534000
3000
當它們以特定隊形飛行時
09:12
as they fly in formation.
216
537000
2000
去偵測它們之間的距離。
09:14
And then you want to make sure
217
539000
2000
你期望能夠確保
09:16
that this separation is within acceptable levels.
218
541000
2000
這個距離是在可接受的範圍內。
09:18
So again the robots monitor this error
219
543000
3000
於是機器人們偵測著這個誤差值
09:21
and calculate the control commands
220
546000
2000
然後以每秒100次的速度
09:23
100 times a second,
221
548000
2000
去估算控制指令，
09:25
which then translates to the motor commands 600 times a second.
222
550000
3000
接著以每秒600次的速度對螺旋槳進行動作指令。
09:28
So this also has to be done
223
553000
2000
這必須是在
09:30
in a decentralized way.
224
555000
2000
沒有中央控制的方式下進行。
09:32
Again, if you have lots and lots of robots,
225
557000
2000
當你有許許多多機器人的時候，
09:34
it's impossible to coordinate all this information centrally
226
559000
4000
想要以中央協調訊息的方式
09:38
fast enough in order for the robots to accomplish the task.
227
563000
3000
快速的讓所有機器人完成任務是不可能的。
09:41
Plus the robots have to base their actions
228
566000
2000
再加上機器人們必須依靠
09:43
only on local information,
229
568000
2000
它們自身去偵測到鄰近機器人
09:45
what they sense from their neighbors.
230
570000
2000
以獲得訊息來進行動作。
09:47
And then finally,
231
572000
2000
最後，
09:49
we insist that the robots be agnostic
232
574000
2000
我們堅持機器人必須無法預知
09:51
to who their neighbors are.
233
576000
2000
鄰近機器人會是誰。
09:53
So this is what we call anonymity.
234
578000
3000
也就是匿名的方式。
09:56
So what I want to show you next
235
581000
2000
接下來我將要給大家看
09:58
is a video
236
583000
2000
一段影片
10:00
of 20 of these little robots
237
585000
3000
關於20個這些小機器人
10:03
flying in formation.
238
588000
2000
以特定隊形進行飛行。
10:05
They're monitoring their neighbors' position.
239
590000
3000
它們正在偵測鄰近機器人的位置。
10:08
They're maintaining formation.
240
593000
2000
它們正在保持著這個隊形。
10:10
The formations can change.
241
595000
2000
這些隊形可以改變。
10:12
They can be planar formations,
242
597000
2000
可以是平面的隊形，
10:14
they can be three-dimensional formations.
243
599000
2000
也可以是三維空間的隊形。
10:16
As you can see here,
244
601000
2000
如你所見的，
10:18
they collapse from a three-dimensional formation into planar formation.
245
603000
3000
它們從三維空間的隊形變換成平面的隊形。
10:21
And to fly through obstacles
246
606000
2000
在穿越障礙物時，
10:23
they can adapt the formations on the fly.
247
608000
4000
它們可以在飛行中調整隊形。
10:27
So again, these robots come really close together.
248
612000
3000
這些機器人移動時真的靠得很近。
10:30
As you can see in this figure-eight flight,
249
615000
2000
在這個 8 字飛行隊形中，
10:32
they come within inches of each other.
250
617000
2000
它們的距離只有幾吋而已。
10:34
And despite the aerodynamic interactions
251
619000
3000
儘管在這些螺旋槳葉片之間
10:37
of these propeller blades,
252
622000
2000
有著空氣動力的交互影響，
10:39
they're able to maintain stable flight.
253
624000
2000
它們仍然能維持穩定的飛行。
10:41
(Applause)
254
626000
7000
(掌聲)
10:48
So once you know how to fly in formation,
255
633000
2000
一旦你知道要怎麼進行特定飛行隊形，
10:50
you can actually pick up objects cooperatively.
256
635000
2000
你就能準確的協力拿起物體。
10:52
So this just shows
257
637000
2000
而這是要告訴大家
10:54
that we can double, triple, quadruple
258
639000
3000
藉由將機器人組合成小組後，
10:57
the robot strength
259
642000
2000
我們可以將機器人們的力量
10:59
by just getting them to team with neighbors, as you can see here.
260
644000
2000
放大兩倍、三倍、四倍，就像是你將看到的這樣。
11:01
One of the disadvantages of doing that
261
646000
3000
但這樣做有一個缺點，
11:04
is, as you scale things up --
262
649000
2000
當你將尺寸放大以後 --
11:06
so if you have lots of robots carrying the same thing,
263
651000
2000
如果你有很多這些機器人載運同一個東西，
11:08
you're essentially effectively increasing the inertia,
264
653000
3000
你一定會有效地增加慣性，
11:11
and therefore you pay a price; they're not as agile.
265
656000
3000
於是你將會付出代價，它們會失去靈巧性。
11:14
But you do gain in terms of payload carrying capacity.
266
659000
3000
但你可以相對獲得載運負重能力。
11:17
Another application I want to show you --
267
662000
2000
另一項我想給大家看的運用 --
11:19
again, this is in our lab.
268
664000
2000
這也是在我們的實驗室裡進行的。
11:21
This is work done by Quentin Lindsey who's a graduate student.
269
666000
2000
這是由 Quentin Lindsey 完成的，他是一位研究生。
11:23
So his algorithm essentially tells these robots
270
668000
3000
他的演算法告訴這些機器人們
11:26
how to autonomously build
271
671000
2000
如何能夠自主性的
11:28
cubic structures
272
673000
2000
將綑狀的材料
11:30
from truss-like elements.
273
675000
3000
建造成立體建築。
11:33
So his algorithm tells the robot
274
678000
2000
他的演算法告訴機器人
11:35
what part to pick up,
275
680000
2000
該拿起哪一個部份，
11:37
when and where to place it.
276
682000
2000
以及什麼時候該把它放在哪裡。
11:39
So in this video you see --
277
684000
2000
你可以在這短片中看到 --
11:41
and it's sped up 10, 14 times --
278
686000
2000
這是以 10 倍、14 倍速播放 --
11:43
you see three different structures being built by these robots.
279
688000
3000
你可以看見這些機器人們建造了三種不同建築。
11:46
And again, everything is autonomous,
280
691000
2000
再次提醒，一切都是自主性進行的，
11:48
and all Quentin has to do
281
693000
2000
而 Quentin 所做的是
11:50
is to get them a blueprint
282
695000
2000
給這些機器人一張藍圖
11:52
of the design that he wants to build.
283
697000
4000
記載著他想要的建築設計。
11:56
So all these experiments you've seen thus far,
284
701000
3000
你所看見的這些實驗，
11:59
all these demonstrations,
285
704000
2000
這些展示，
12:01
have been done with the help of motion capture systems.
286
706000
3000
都使用了動作擷取系統。
12:04
So what happens when you leave your lab
287
709000
2000
如果離開了實驗室，
12:06
and you go outside into the real world?
288
711000
3000
走進真實世界會變成怎麼樣呢？
12:09
And what if there's no GPS?
289
714000
3000
如果沒有 GPS 會怎樣呢？
12:12
So this robot
290
717000
2000
這個機器人
12:14
is actually equipped with a camera
291
719000
2000
裝置了一具攝影機，
12:16
and a laser rangefinder, laser scanner.
292
721000
3000
一具雷射H搜尋器，雷射掃描器。
12:19
And it uses these sensors
293
724000
2000
它使用這些感應器
12:21
to build a map of the environment.
294
726000
2000
來製作一張周圍的地圖。
12:23
What that map consists of are features --
295
728000
3000
這地圖然有著一些環境特徵 --
12:26
like doorways, windows,
296
731000
2000
例如大門、窗戶、
12:28
people, furniture --
297
733000
2000
人、家具 --
12:30
and it then figures out where its position is
298
735000
2000
接著它會辨識出相對於這些環境特徵
12:32
with respect to the features.
299
737000
2000
它所處的位置。
12:34
So there is no global coordinate system.
300
739000
2000
這裡並沒有整體座標系統。
12:36
The coordinate system is defined based on the robot,
301
741000
3000
座標系統是機器人自身定義出來的，
12:39
where it is and what it's looking at.
302
744000
3000
藉由它所在的位置以及它所看到的東西。
12:42
And it navigates with respect to those features.
303
747000
3000
接著它對這些環境特徵進行探索。
12:45
So I want to show you a clip
304
750000
2000
我想給大家看一段影片，
12:47
of algorithms developed by Frank Shen
305
752000
2000
關於 Frank Shen 以及 Nathan Michael 教授
12:49
and Professor Nathan Michael
306
754000
2000
所開發出來的演算法，
12:51
that shows this robot entering a building for the very first time
307
756000
4000
這個機器人第一次進入一個建築物，
12:55
and creating this map on the fly.
308
760000
3000
然後在飛行中製作了這個地圖。
12:58
So the robot then figures out what the features are.
309
763000
3000
於是機器人知道環境特徵是什麼東西。
13:01
It builds the map.
310
766000
2000
它製作出地圖。
13:03
It figures out where it is with respect to the features
311
768000
2000
它知道自己相對於環境特徵的位置，
13:05
and then estimates its position
312
770000
2000
然後每秒100次的速度
13:07
100 times a second
313
772000
2000
估算出自己的位置，
13:09
allowing us to use the control algorithms
314
774000
2000
讓我們可以利用
13:11
that I described to you earlier.
315
776000
2000
剛剛說過的控制演算法。
13:13
So this robot is actually being commanded
316
778000
2000
事實上這個機器人正被
13:15
remotely by Frank.
317
780000
2000
Frank 以遠端遙控的方式下指令。
13:17
But the robot can also figure out
318
782000
2000
但這個機器人也能自行判斷
13:19
where to go on its own.
319
784000
2000
它應該往哪裡走。
13:21
So suppose I were to send this into a building
320
786000
2000
假設我把它送進一個建築物，
13:23
and I had no idea what this building looked like,
321
788000
2000
而我完全不知道這個建築物的樣子，
13:25
I can ask this robot to go in,
322
790000
2000
我可以命令機器人進入，
13:27
create a map
323
792000
2000
製作出一張地圖，
13:29
and then come back and tell me what the building looks like.
324
794000
3000
然後回來告訴我建築物的樣子。
13:32
So here, the robot is not only solving the problem,
325
797000
3000
所以機器人並不只是解決
13:35
how to go from point A to point B in this map,
326
800000
3000
如何從地圖上的A點到B點這個問題，
13:38
but it's figuring out
327
803000
2000
它甚至知道
13:40
what the best point B is at every time.
328
805000
2000
每一次的最佳B點是哪個位置。
13:42
So essentially it knows where to go
329
807000
3000
於是它知道該往哪裡去
13:45
to look for places that have the least information.
330
810000
2000
以找出還沒有訊息的位置。
13:47
And that's how it populates this map.
331
812000
3000
這就是它如何把地圖裝滿的方法。
13:50
So I want to leave you
332
815000
2000
最後，
13:52
with one last application.
333
817000
2000
我想再給大家看一樣應用。
13:54
And there are many applications of this technology.
334
819000
3000
這個技術有許多運用方式。
13:57
I'm a professor, and we're passionate about education.
335
822000
2000
我是一個教授，我們對教育充滿熱情。
13:59
Robots like this can really change the way
336
824000
2000
這種機器人可以改變
14:01
we do K through 12 education.
337
826000
2000
我們進行12年國教的方式。
14:03
But we're in Southern California,
338
828000
2000
我們身在南加州，
14:05
close to Los Angeles,
339
830000
2000
很靠近洛杉磯，
14:07
so I have to conclude
340
832000
2000
所以我想用關於娛樂的例子
14:09
with something focused on entertainment.
341
834000
2000
來作為最後的結尾。
14:11
I want to conclude with a music video.
342
836000
2000
我想用一段音樂影片來作為結尾。
14:13
I want to introduce the creators, Alex and Daniel,
343
838000
3000
我要為大家介紹
14:16
who created this video.
344
841000
2000
這個影片的作者，Alex 和 Daniel。
14:18
(Applause)
345
843000
7000
(掌聲)
14:25
So before I play this video,
346
850000
2000
在我播放影片之前，
14:27
I want to tell you that they created it in the last three days
347
852000
3000
我想告訴大家他們在接到 Chris 電話後的三天內
14:30
after getting a call from Chris.
348
855000
2000
就將這段影片製作完了。
14:32
And the robots that play the video
349
857000
2000
影片中演奏的機器人
14:34
are completely autonomous.
350
859000
2000
都是完全自主性的進行。
14:36
You will see nine robots play six different instruments.
351
861000
3000
你可以看見 9 個機器人們演奏著 6 種不同的樂器。
14:39
And of course, it's made exclusively for TED 2012.
352
864000
4000
當然，這是為了 TED 2012 特別製作的。
14:43
Let's watch.
353
868000
3000
讓我們一起來欣賞。
15:19
(Music)
354
904000
10000
(音樂聲)
16:23
(Applause)
355
968000
17000
(掌聲)
ABOUT THE SPEAKER
Vijay Kumar - RoboticistAs the dean of the University of Pennsylvania's School of Engineering and Applied Science, Vijay Kumar studies the control and coordination of multi-robot formations.
Why you should listen
At the General Robotics, Automation, Sensing and Perception (GRASP) Lab at the University of Pennsylvania, flying quadrotor robots move together in eerie formation, tightening themselves into perfect battalions, even filling in the gap when one of their own drops out. You might have seen viral videos of the quads zipping around the netting-draped GRASP Lab (they juggle! they fly through a hula hoop!). Vijay Kumar headed this lab from 1998-2004. He's now the dean of the School of Engineering and Applied Science at the University of Pennsylvania in Philadelphia, where he continues his work in robotics, blending computer science and mechanical engineering to create the next generation of robotic wonders.
Vijay Kumar | Speaker | TED.com