The astounding athletic power of quadcopters | Raffaello D'Andrea

10,754,297 views ใƒป 2013-06-12

TED


์•„๋ž˜ ์˜๋ฌธ์ž๋ง‰์„ ๋”๋ธ”ํด๋ฆญํ•˜์‹œ๋ฉด ์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค.

00:00
Translator: Joseph Geni Reviewer: Morton Bast
0
0
7000
๋ฒˆ์—ญ: Inho Jeong ๊ฒ€ํ† : Kwangmin Lee
00:23
So what does it mean for a machine to be athletic?
1
23544
2523
๊ธฐ๊ณ„์˜ ์šด๋™์„ฑ์ด๋ผ๋Š” ๊ฒƒ์€ ๋ฌด์—‡์ผ๊นŒ์š”?
์ง€๊ธˆ๋ถ€ํ„ฐ ๊ธฐ๊ณ„ ์šด๋™์„ฑ์˜ ๊ฐœ๋…๊ณผ
00:27
We will demonstrate the concept of machine athleticism
2
27059
2825
00:29
and the research to achieve it
3
29908
1696
์ด๋ฅผ ๋‹ค๋ฃฌ ์—ฐ๊ตฌ์— ๋Œ€ํ•ด
00:31
with the help of these flying machines called quadrocopters,
4
31628
2858
'์ฟผ๋“œ์ฝฅํ„ฐ', ํ˜น์€ ์งง๊ฒŒ '์ฟผ๋“œ'๋ผ๊ณ  ๋ถˆ๋ฆฌ๋Š” ์ด ๋น„ํ–‰ ๊ธฐ๊ตฌ๋ฅผ ํ†ตํ•ด
00:34
or quads, for short.
5
34510
1047
๋ณด์—ฌ ๋“œ๋ฆฌ๊ฒ ์Šต๋‹ˆ๋‹ค.
00:38
Quads have been around for a long time.
6
38291
1893
์ฟผ๋“œ๊ฐ€ ๋งŒ๋“ค์–ด์ง„ ์ง€๋Š” ๊ฝค ๋์Šต๋‹ˆ๋‹ค.
00:40
They're so popular these days because they're mechanically simple.
7
40501
3277
๊ทธ๋Ÿฐ๋ฐ ์š”์ƒˆ ์ฟผ๋“œ๊ฐ€ ์ด๋ ‡๊ฒŒ ์ธ๊ธฐ๋ฅผ ๋„๋Š” ์ด์œ ๋Š”
๊ธฐ๊ณ„์ ์œผ๋กœ ๋‹จ์ˆœํ•˜๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
00:44
By controlling the speeds of these four propellers,
8
44255
2389
์ด๋Ÿฐ 4๊ฐœ์˜ ํ”„๋กœํŽ ๋Ÿฌ์˜ ์†๋„๋ฅผ ์กฐ์ ˆํ•จ์œผ๋กœ์จ
00:46
these machines can roll, pitch, yaw,
9
46668
2745
๋Œ๊ธฐ, ๋˜์ง€๊ธฐ, ํ•œ์ชฝ์œผ๋กœ ๊ธฐ์šธ์ด๊ธฐ๋ฅผ ํ•  ์ˆ˜ ์žˆ๊ณ 
00:49
and accelerate along their common orientation.
10
49437
2171
์—ฌ๋Ÿฟ์ด ๊ฐ™์€ ๋ฐฉํ–ฅ์œผ๋กœ ๋™์‹œ์— ๋‚  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
00:52
On board are also a battery, a computer,
11
52179
3198
์—ฌ๊ธฐ์—๋Š” ๋ฐฐํ„ฐ๋ฆฌ, ์ปดํ“จํ„ฐ,
00:55
various sensors and wireless radios.
12
55401
1953
์—ฌ๋Ÿฌ ๊ฐ€์ง€ ์„ผ์„œ์™€ ๋ฌด์„  ํ†ต์‹  ์žฅ์น˜๊ฐ€ ์žฅ์ฐฉ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค.
00:59
Quads are extremely agile, but this agility comes at a cost.
13
59298
3309
์ฟผ๋“œ๋Š” ๊ต‰์žฅํžˆ ๋ฏผ์ฒฉํ•ฉ๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ ์ด๋ ‡๊ฒŒ ๋งŒ๋“œ๋Š” ๋ฐ ๋Œ“๊ฐ€๊ฐ€ ๋”ฐ๋ฅด์ฃ .
01:03
They are inherently unstable,
14
63762
2088
๋ฏผ์ฒฉํ•œ ๋Œ€์‹  ๋ณธ์งˆ์ ์œผ๋กœ ๋ถˆ์•ˆ์ •ํ•˜๋ฉฐ ๋‚ ์•„๋‹ค๋‹ˆ๊ธฐ ์œ„ํ•ด์„œ๋Š”
01:05
and they need some form of automatic feedback control
15
65874
2571
์ผ์ข…์˜ ์ž๋™ ํ”ผ๋“œ๋ฐฑ ์ œ์–ด๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค.
01:08
in order to be able to fly.
16
68469
1373
01:16
So, how did it just do that?
17
76259
2253
๋ฐฉ๊ธˆ ์ด๊ฑด ์–ด๋–ป๊ฒŒ ํ•œ ๊ฑธ๊นŒ์š”?
01:19
Cameras on the ceiling and a laptop
18
79497
1794
์ฒœ์žฅ์— ๋‹ฌ๋ฆฐ ์นด๋ฉ”๋ผ์™€ ๋žฉํ†ฑ ์ปดํ“จํ„ฐ๊ฐ€
01:21
serve as an indoor global positioning system.
19
81315
2430
์ „๋ฐ˜์ ์ธ ์‹ค๋‚ด ์œ„์น˜ ์ œ์–ด ์‹œ์Šคํ…œ์œผ๋กœ ์ž‘๋™ํ•ฉ๋‹ˆ๋‹ค.
๊ทธ ์‹œ์Šคํ…œ์€ ์„ค์น˜๋œ ๊ณต๊ฐ„์—์„œ
01:24
It's used to locate objects in the space
20
84301
2039
01:26
that have these reflective markers on them.
21
86364
2088
์ด๋Ÿฐ ๋ฐ˜์‚ฌ์ฒด๋กœ ํ‘œ์‹œ๋œ ๋ฌผ์ฒด์˜ ์œ„์น˜๋ฅผ ํŒŒ์•…ํ•˜๋Š” ๋ฐ์— ์‚ฌ์šฉ๋ฉ๋‹ˆ๋‹ค
01:28
This data is then sent to another laptop
22
88476
2038
ํŒŒ์•…๋œ ์œ„์น˜๋Š” ๋‹ค๋ฅธ ๋žฉํ†ฑ ์ปดํ“จํ„ฐ๋กœ ๋ณด๋‚ด์ง€๊ณ ,
01:30
that is running estimation and control algorithms,
23
90538
2358
๊ทธ ์ปดํ“จํ„ฐ๋Š” ์ถ”์ •๊ฐ’๊ณผ ์ œ์–ด ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๊ณ„์‚ฐํ•ด์„œ
01:32
which in turn sends commands to the quad,
24
92920
1954
์ฟผ๋“œ์— ์ฐจ๋ก€๋กœ ๋ช…๋ น์„ ๋ณด๋ƒ…๋‹ˆ๋‹ค.
01:34
which is also running estimation and control algorithms.
25
94898
2633
์ฟผ๋“œ ๋˜ํ•œ ์ถ”์ •๊ฐ’๊ณผ ์ œ์–ด ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๊ณ„์‚ฐํ•˜๊ณ  ์žˆ์ฃ 
01:41
The bulk of our research is algorithms.
26
101864
2242
์ €ํฌ ์—ฐ๊ตฌ์˜ ๋Œ€๋ถ€๋ถ„์€ ์•Œ๊ณ ๋ฆฌ์ฆ˜์ž…๋‹ˆ๋‹ค.
01:44
It's the magic that brings these machines to life.
27
104130
2576
์•Œ๊ณ ๋ฆฌ์ฆ˜์€ ์ด ๊ธฐ๊ณ„๋“ค์—๊ฒŒ ์ƒ๋ช…์„ ๋ถˆ์–ด๋„ฃ๋Š” ๋งˆ๋ฒ•์ด์ฃ .
01:48
So how does one design the algorithms that create a machine athlete?
28
108291
4033
๊ทธ๋ ‡๋‹ค๋ฉด ์ด๋ ‡๊ฒŒ ๊ธฐ๊ณ„๊ฐ€ ์šด๋™์„ฑ์„ ๊ฐ–๊ฒŒ ๋งŒ๋“œ๋Š”
์•Œ๊ณ ๋ฆฌ์ฆ˜์€ ์–ด๋–ป๊ฒŒ ๋งŒ๋“ค๊นŒ์š”?
01:52
We use something broadly called model-based design.
29
112792
2587
์ €ํฌ๋Š” ํ”ํžˆ ๋ชจ๋ธ ๊ธฐ๋ฐ˜ ์„ค๊ณ„๋ผ๊ณ  ๋ถˆ๋ฆฌ๋Š” ๊ฒƒ์„ ์‚ฌ์šฉํ–ˆ์Šต๋‹ˆ๋‹ค.
01:55
We first capture the physics
30
115839
1366
์šฐ์„  ๊ธฐ๊ณ„๊ฐ€ ๋ฌผ๋ฆฌ์ ์œผ๋กœ ์–ด๋–ป๊ฒŒ ์›€์ง์ด๋Š”์ง€
01:57
with a mathematical model of how the machines behave.
31
117229
2983
์ˆ˜ํ•™ ๋ชจ๋ธ์„ ์ด์šฉํ•ด ๋‚˜ํƒ€๋ƒˆ์Šต๋‹ˆ๋‹ค.
02:00
We then use a branch of mathematics called control theory
32
120636
3950
๊ทธ ๋‹ค์Œ ์ œ์–ด ์ด๋ก ์ด๋ผ ๋ถˆ๋ฆฌ๋Š” ์ˆ˜ํ•™์˜ ํ•œ ๋ถ„์•ผ๋ฅผ ํ†ตํ•ด
๊ทธ ์ˆ˜ํ•™ ๋ชจ๋ธ์„ ๋ถ„์„ํ•˜๊ณ 
02:04
to analyze these models
33
124610
1657
๊ธฐ๊ณ„๋ฅผ ์ œ์–ดํ•  ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๊ตฌํ˜„ํ–ˆ์Šต๋‹ˆ๋‹ค.
02:06
and also to synthesize algorithms for controlling them.
34
126291
2770
02:09
For example, that's how we can make the quad hover.
35
129820
3024
์˜ˆ๋ฅผ ๋“ค์–ด, ์ฟผ๋“œ๋ฅผ ๊ณต์ค‘์— ๋– ์žˆ๊ฒŒ ํ•˜๋Š” ๊ฒƒ๋„ ๋งˆ์ฐฌ๊ฐ€์ง€์ž…๋‹ˆ๋‹ค
02:12
We first captured the dynamics with a set of differential equations.
36
132868
3290
๋จผ์ € ์ผ๋ จ์˜ ๋ฏธ๋ถ„ ๋ฐฉ์ ์‹์„ ํ†ตํ•ด
์—ญํ•™ ๊ด€๊ณ„๋ฅผ ๋‚˜ํƒ€๋‚ด๊ณ 
02:16
We then manipulate these equations with the help of control theory
37
136182
3452
์ œ์–ด ์ด๋ก ์˜ ๋„์›€์„ ๋ฐ›์•„ ๊ทธ ๋ฏธ๋ถ„ ๋ฐฉ์ •์‹์„ ์กฐ์ž‘ํ•˜์—ฌ
์ฟผ๋“œ๋ฅผ ์•ˆ์ •์ ์œผ๋กœ ๋– ์žˆ๊ฒŒ ํ•˜๋Š” ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๋งŒ๋“  ๊ฒ๋‹ˆ๋‹ค.
02:19
to create algorithms that stabilize the quad.
38
139658
2937
02:23
Let me demonstrate the strength of this approach.
39
143130
2293
์ด์ œ ์ด ์ ‘๊ทผ๋ฒ•์˜ ๊ฐ•์ ์ด ๋ญ”์ง€ ๋ณด์—ฌ ๋“œ๋ฆฌ์ฃ .
02:29
Suppose that we want this quad to not only hover
40
149198
2633
์ด ์ฟผ๋“œ๊ฐ€ ๊ณต์ค‘์— ๋– ์žˆ๋Š” ๊ฒƒ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ
02:31
but to also balance this pole.
41
151855
1659
์ด ๋ง‰๋Œ€๊ฐ€ ๋„˜์–ด์ง€์ง€ ์•Š๊ฒŒ ๊ท ํ˜•์„ ์žก๊ฒŒ ํ•˜๊ณ  ์‹ถ๋‹ค๊ณ  ์นฉ์‹œ๋‹ค.
02:34
With a little bit of practice,
42
154520
1774
์‚ฌ๋žŒ์€ ๋ˆ„๊ตฌ๋‚˜ ์กฐ๊ธˆ๋งŒ ์—ฐ์Šตํ•˜๋ฉด
02:36
it's pretty straightforward for a human being to do this,
43
156318
2683
์ด๋ ‡๊ฒŒ ๊ท ํ˜• ์žก๋Š” ๋ฒ•์„ ์‰ฝ๊ฒŒ ์•Œ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
๋ฌผ๋ก  ์šฐ๋ฆฌ์—๊ฒ ๋•…์„ ์งš์„ ์ˆ˜ ์žˆ๋Š”
02:39
although we do have the advantage of having two feet on the ground
44
159025
3191
๋‘ ๋ฐœ์ด ์žˆ๊ณ 
๋ฌด์Šจ ์ผ์ด๋“  ํ•  ์ˆ˜ ์žˆ๋Š” ์†์ด ์žˆ๊ธด ํ•˜์ง€๋งŒ์š”.
02:42
and the use of our very versatile hands.
45
162240
2059
02:44
It becomes a little bit more difficult
46
164804
2738
๊ทธ๋Ÿฐ๋ฐ ์†์„ ์“ฐ์ง€ ์•Š๊ณ 
02:47
when I only have one foot on the ground
47
167566
1979
ํ•œ ๋ฐœ๋งŒ ์งš๋Š”๋‹ค๋ฉด
02:49
and when I don't use my hands.
48
169569
1602
๊ฝค ์–ด๋ ค์›Œ์ง‘๋‹ˆ๋‹ค.
02:52
Notice how this pole has a reflective marker on top,
49
172733
2562
์ด ๋ง‰๋Œ€์˜ ์œ—๋ถ€๋ถ„์€ ๋ฐ˜์‚ฌ์ฒด๋กœ ํ‘œ์‹œ๋ผ ์žˆ์–ด์„œ
02:55
which means that it can be located in the space.
50
175319
2661
์ปดํ“จํ„ฐ๊ฐ€ ๊ทธ ์œ„์น˜๋ฅผ ํŒŒ์•…ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
03:04
(Audience) Oh!
51
184276
1067
(๋ฐ•์ˆ˜)
03:05
(Applause)
52
185367
1259
03:09
(Applause ends)
53
189695
1668
์ฟผ๋“œ๊ฐ€ ๋ฏธ์„ธํ•œ ์กฐ์ •์„ ํ†ตํ•ด ๋ง‰๋Œ€๊ฐ€ ๋„˜์–ด์ง€์ง€ ์•Š๊ฒŒ
03:11
You can notice that this quad is making fine adjustments
54
191387
2641
๊ท ํ˜•์„ ์žก๋Š” ๊ฒƒ์ด ๋ณด์ผ ๊ฒƒ์ž…๋‹ˆ๋‹ค.
03:14
to keep the pole balanced.
55
194052
1374
์ด ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์–ด๋–ป๊ฒŒ ์„ค๊ณ„ํ–ˆ์„๊นŒ์š”?
03:16
How did we design the algorithms to do this?
56
196417
2072
03:19
We added the mathematical model of the pole
57
199027
2081
๋ง‰๋Œ€์— ๋Œ€ํ•œ ์ˆ˜ํ•™ ๋ชจ๋ธ์„
03:21
to that of the quad.
58
201132
1447
์ฟผ๋“œ์— ์ž…๋ ฅํ•œ ๊ฒ๋‹ˆ๋‹ค.
03:22
Once we have a model of the combined quad-pole system,
59
202603
2978
์ฟผ๋“œ์™€ ๋ง‰๋Œ€๊ฐ€ ๊ฒฐํ•ฉ๋œ ๋ชจ๋ธ์ด ๋งŒ๋“ค์–ด์ง€๋ฉด
03:25
we can use control theory to create algorithms for controlling it.
60
205605
3319
์ œ์–ด ์ด๋ก ์„ ํ†ตํ•ด ๊ทธ๊ฒƒ์„ ์ œ์–ดํ•  ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ์ฃ .
03:30
Here, you see that it's stable,
61
210538
1564
์ง€๊ธˆ ์•ˆ์ •๋œ ์ƒํƒœ๊ฐ€ ๋ณด์ด์‹ค ๊ฒ๋‹ˆ๋‹ค.
03:32
and even if I give it little nudges,
62
212126
2669
๊ทธ๋ฆฌ๊ณ  ์ด๋ ‡๊ฒŒ ์กฐ๊ธˆ ๋ฐ€์–ด๋„
03:34
it goes back --
63
214819
1166
๋ฐ”๋กœ ๊ท ํ˜•์ด ์ž˜ ์žกํžŒ ์œ„์น˜๋กœ ๋˜๋Œ์•„๊ฐ‘๋‹ˆ๋‹ค
03:37
to the nice, balanced position.
64
217589
1769
03:40
We can also augment the model
65
220087
1390
๋ชจ๋ธ์„ ํ™•์žฅ์‹œ์ผœ
03:41
to include where we want the quad to be in space.
66
221501
2371
์šฐ๋ฆฌ๊ฐ€ ์›ํ•˜๋Š” ๊ณณ์œผ๋กœ ์ฟผ๋“œ๋ฅผ ์ด๋™์‹œํ‚ฌ ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
03:44
Using this pointer, made out of reflective markers,
67
224428
2943
๋ฐ˜์‚ฌ์ฒด๋กœ ํ‘œ์‹œ๋œ ์ด ํฌ์ธํ„ฐ๋ฅผ ์ด์šฉํ•ด์„œ
03:47
I can point to where I want the quad to be in space
68
227395
2429
์›ํ•˜๋Š” ๋ฐฉํ–ฅ์„ ๊ฐ€๋ฆฌํ‚ค๋ฉด ์ €์™€์˜ ๊ฑฐ๋ฆฌ๋ฅผ ๋ฐ”๊พธ๋ฉฐ
03:49
a fixed distance away from me.
69
229848
1597
์ฟผ๋“œ๊ฐ€ ์›€์ง์ด๊ฒŒ ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
03:59
(Laughter)
70
239785
1880
04:07
The key to these acrobatic maneuvers is algorithms,
71
247763
3126
์ด๋ ‡๊ฒŒ ๊ณก์˜ˆ์— ๊ฐ€๊นŒ์šด ๋™์ž‘์˜ ํ•ต์‹ฌ์€
04:10
designed with the help of mathematical models
72
250913
2293
์ˆ˜ํ•™ ๋ชจํ˜•๊ณผ ์ œ์–ด ์ด๋ก ์˜ ๋„์›€์œผ๋กœ ์„ค๊ณ„๋œ
04:13
and control theory.
73
253230
1185
์•Œ๊ณ ๋ฆฌ์ฆ˜์ž…๋‹ˆ๋‹ค.
04:15
Let's tell the quad to come back here
74
255246
1985
์ฟผ๋“œ๊ฐ€ ์—ฌ๊ธฐ๋กœ ๋Œ์•„์™€์„œ ๋ง‰๋Œ€๋ฅผ ๋–จ์–ดํŠธ๋ฆฌ๋„๋ก
04:17
and let the pole drop,
75
257255
1664
์กฐ์ž‘ํ•ด ๋ณด์ฃ .
04:18
and I will next demonstrate the importance
76
258943
2131
๋‹ค์Œ์—๋Š” ๋ฌผ๋ฆฌ์  ๋ชจํ˜•์„ ์ดํ•ดํ•˜๋Š” ๊ฒƒ์ด
์™œ ์ค‘์š”ํ•œ์ง€์™€
04:21
of understanding physical models
77
261098
1989
04:23
and the workings of the physical world.
78
263111
2045
๋ฌผ๋ฆฌ์ ์ธ ์ž‘๋™ ์›๋ฆฌ๋ฅผ ๋ณด์—ฌ๋“œ๋ฆฌ๊ฒ ์Šต๋‹ˆ๋‹ค.
04:37
Notice how the quad lost altitude when I put this glass of water on it.
79
277462
3828
์ œ๊ฐ€ ๋ฌผ์ปต์„ ๋‚ด๋ ค๋†“์ž
์ฟผ๋“œ๊ฐ€ ์•ฝ๊ฐ„ ๋‚ด๋ ค๊ฐ€๋Š” ๊ฒƒ์„ ๋ณด์…จ์„ ๊ฒ๋‹ˆ๋‹ค.
04:41
Unlike the balancing pole,
80
281314
1403
๋ง‰๋Œ€์™€๋Š” ๋‹ฌ๋ฆฌ ์ด ์ž”์— ๋Œ€ํ•œ ์ˆ˜ํ•™ ๋ชจ๋ธ์€
04:42
I did not include the mathematical model of the glass
81
282741
3015
์‹œ์Šคํ…œ์— ์ž…๋ ฅํ•˜์ง€ ์•Š์•˜์Šต๋‹ˆ๋‹ค.
04:45
in the system.
82
285780
1307
์‚ฌ์‹ค ์‹œ์Šคํ…œ์€ ๋ฌผ์ปต์ด ๊ฑฐ๊ธฐ ์žˆ๋Š”์ง€๋„ ์ธ์‹ํ•˜์ง€ ๋ชปํ•ฉ๋‹ˆ๋‹ค.
04:47
In fact, the system doesn't even know that the glass is there.
83
287111
3013
์•„๊นŒ์™€ ๋งˆ์ฐฌ๊ฐ€์ง€๋กœ ํฌ์ธํ„ฐ๋ฅผ ์ด์šฉํ•ด
04:50
Like before, I could use the pointer to tell the quad
84
290148
2738
04:52
where I want it to be in space.
85
292910
1571
์ฟผ๋“œ๋ฅผ ์ด๋™์‹œ์ผœ ๋ณด๊ฒ ์Šต๋‹ˆ๋‹ค.
(๋ฐ•์ˆ˜)
04:57
(Applause)
86
297684
1799
05:03
(Applause ends)
87
303531
1540
05:05
Okay, you should be asking yourself,
88
305515
1739
์—ฌ๋Ÿฌ๋ถ„์€ ๊ถ๊ธˆํ•˜์‹ค ๊ฒ๋‹ˆ๋‹ค,
05:07
why doesn't the water fall out of the glass?
89
307278
2062
"์–ด๋–ป๊ฒŒ ์ปต์—์„œ ๋ฌผ์ด ๋–จ์–ด์ง€์ง€ ์•Š๋Š” ๊ฑฐ์ง€?"๋ผ๊ณ ์š”.
05:09
Two facts.
90
309888
1030
๋‘ ๊ฐ€์ง€ ๋ฒ•์น™์ด ์ž‘์šฉํ•˜๋Š”๋ฐ, ์ฒซ์งธ๋Š” ์–ด๋–ค ๋ฌผ์ฒด๋“ ์ง€
05:10
The first is that gravity acts on all objects in the same way.
91
310942
3633
์ค‘๋ ฅ์˜ ์˜ํ–ฅ์„ ๋˜‘๊ฐ™์ด ๋ฐ›๋Š”๋‹ค๋Š” ๊ฒƒ์ด๊ณ ,
05:15
The second is that the propellers
92
315051
1779
๋‘˜์งธ๋Š” ํ”„๋กœํŽ ๋Ÿฌ ๋„ค ๊ฐœ๊ฐ€ ๋†“์—ฌ์žˆ๋Š” ๋ฐฉํ–ฅ์ด
05:16
are all pointing in the same direction of the glass, pointing up.
93
316854
3073
๋ฌผ์ปต๊ณผ ๋˜‘๊ฐ™์ด ์œ„๋กœ ํ–ฅํ•˜๊ณ  ์žˆ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.
05:20
You put these two things together,
94
320642
1647
์ด ๋‘ ๊ฐ€์ง€ ์‚ฌ์‹ค์„ ์ข…ํ•ฉํ•ด ๋ณด๋ฉด
05:22
the net result is that all side forces on the glass are small
95
322313
3555
๋ฌผ์ปต ์–‘์ชฝ์— ๊ฐ€ํ•ด์ง€๋Š” ํž˜์€ ์ž‘์€ ๋ฐ˜๋ฉด
05:25
and are mainly dominated by aerodynamic effects,
96
325892
2277
๋งŽ์€ ๊ณต๊ธฐ ์—ญํ•™์˜ ์˜ํ–ฅ์ด ์ปต์˜ ์ธก๋ฉด์— ์ž‘์šฉํ•˜๋Š” ํž˜์„ ์••๋„ํ•˜์—ฌ
05:28
which at these speeds are negligible.
97
328193
1961
์ธก๋ฉด์— ๊ฐ€ํ•ด์ง€๋Š” ํž˜์ด ๋ฌด์‹œ๋  ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฒ๋‹ˆ๋‹ค.
05:35
And that's why you don't need to model the glass.
98
335336
2334
๊ทธ๋ž˜์„œ ๋ฌผ์ปต์— ๋Œ€ํ•œ ๋ชจ๋ธ์€ ํ•„์š”ํ•˜์ง€ ์•Š์€ ๊ฑฐ์ง€์š”.
05:37
It naturally doesn't spill, no matter what the quad does.
99
337694
2741
์ฟผ๋“œ๊ฐ€ ์–ด๋–ป๊ฒŒ ์›€์ง์ด๋“  ๋ฌผ์„ ํ˜๋ฆฌ์ง€ ์•Š์Šต๋‹ˆ๋‹ค.
05:44
(Audience) Oh!
100
344500
2563
05:50
(Applause)
101
350340
1858
(๋ฐ•์ˆ˜)
05:55
(Applause ends)
102
355159
1255
05:57
The lesson here
103
357940
1114
์—ฌ๊ธฐ์„œ ์•Œ ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์€ ๊ณ ๋„๋กœ ์ •๋ฐ€ํ•œ ์กฐ์ž‘ ์ค‘์—๋Š”
05:59
is that some high-performance tasks are easier than others,
104
359078
4013
๊ตฌํ˜„์ด ์‰ฌ์šด ๊ฒƒ๊ณผ ์–ด๋ ค์šด ๊ฒƒ์ด ์žˆ๊ณ 
06:03
and that understanding the physics of the problem
105
363115
2337
์–ด๋–ค ๊ฒŒ ์‰ฝ๊ณ  ์–ด๋ ค์šด์ง€ ํŒ๋ณ„ํ•˜๋ ค๋ฉด
06:05
tells you which ones are easy and which ones are hard.
106
365476
2694
๋ฌผ๋ฆฌํ•™์„ ์ดํ•ดํ•ด์•ผ ํ•œ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.
06:08
In this instance, carrying a glass of water is easy.
107
368194
2510
๋ฐฉ๊ธˆ ๋ณด์‹  ๋ฌผ์ปต ์˜ฎ๊ธฐ๊ธฐ๋Š” ์‰ฌ์šด ํŽธ์ž…๋‹ˆ๋‹ค.
06:10
Balancing a pole is hard.
108
370728
1630
๋ง‰๋Œ€ ๊ท ํ˜•์„ ์žก๋Š” ๊ฑด ์–ด๋ ต๊ณ ์š”.
06:14
We've all heard stories of athletes performing feats while physically injured.
109
374171
4271
์šด๋™ ์„ ์ˆ˜๋“ค์ด ๋ถ€์ƒ๋‹นํ•œ ๋’ค์—๋„
๋›ฐ์–ด๋‚œ ํ™œ์•ฝ์„ ๋ณด์ด๋Š” ๊ฒฝ์šฐ๋ฅผ ๋“ค์–ด๋ณด์…จ์„ ๊ฒ๋‹ˆ๋‹ค.
06:18
Can a machine also perform with extreme physical damage?
110
378466
3185
๊ทธ๋Ÿฌ๋ฉด ๊ธฐ๊ณ„ ๋˜ํ•œ ์ƒ๋‹นํ•œ ๋ฌผ๋ฆฌ์  ์†์ƒ์„ ์ž…์—ˆ์„ ๋•Œ
์ œ๋Œ€๋กœ ์ž‘๋™ํ•  ์ˆ˜ ์žˆ์„๊นŒ์š”?
06:22
Conventional wisdom says
111
382636
1485
์ƒ์‹์ ์œผ๋กœ ์ƒ๊ฐํ•ด๋ณด๋ฉด
06:24
that you need at least four fixed motor propeller pairs in order to fly,
112
384145
3883
๋‚ ๊ธฐ ์œ„ํ•ด์„œ๋Š” ์ ์–ด๋„ 4๊ฐœ์˜ ๊ณ ์ •๋œ ํ”„๋กœํŽ ๋Ÿฌ๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค.
06:28
because there are four degrees of freedom to control:
113
388052
2524
์ œ์–ด์—๋Š” 4๋‹จ๊ณ„์˜ ์ž์œ ๋„๊ฐ€ ์žˆ๊ธฐ ๋•Œ๋ฌธ์ด์ฃ :
06:30
roll, pitch, yaw and acceleration.
114
390600
1965
๋Œ๊ธฐ, ์†Ÿ์•„์˜ค๋ฅด๊ธฐ, ๊ธฐ์šธ๊ธฐ์™€ ๊ฐ€์†ํ™”์˜ 4 ๋‹จ๊ณ„ ๋ง์ž…๋‹ˆ๋‹ค.
06:33
Hexacopters and octocopters, with six and eight propellers,
115
393192
3221
ํ”„๋กœํŽ ๋Ÿฌ๊ฐ€ ๊ฐ๊ฐ 6๊ฐœ์ธ ํ—ฅ์‚ฌ์ฝฅํ„ฐ์™€ 8๊ฐœ์ธ ์˜ฅํ† ์ฝฅํ„ฐ๋Š”
06:36
can provide redundancy,
116
396437
1459
ํ”„๋กœํŽ ๋Ÿฌ๊ฐ€ ํ•„์š” ์ด์ƒ์œผ๋กœ ๋งŽ์Šต๋‹ˆ๋‹ค.
06:37
but quadrocopters are much more popular
117
397920
1965
์ฟผ๋“œ์ฝฅํ„ฐ๊ฐ€ ์ธ๊ธฐ์žˆ๋Š” ์ด์œ ๋Š”
06:39
because they have the minimum number of fixed motor propeller pairs: four.
118
399909
3686
4๊ฐœ๋ผ๋Š” ์ตœ์†Œํ•œ์˜ ๊ณ ์ •๋œ ๊ตฌ๋™ ํ”„๋กœํŽ ๋Ÿฌ๋กœ
๊ตฌ๋™์‹œํ‚ฌ ์ˆ˜ ์žˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
06:44
Or do they?
119
404168
1016
์‚ฌ์‹ค์ผ๊นŒ์š”?
06:52
(Audience) Oh!
120
412827
1000
06:54
(Laughter)
121
414407
1000
07:01
If we analyze the mathematical model of this machine
122
421334
2718
์ฟผ๋“œ์˜ ์ˆ˜ํ•™ ๋ชจ๋ธ์„ ๋‘ ๊ฐœ์˜ ํ”„๋กœํŽ ๋Ÿฌ๋งŒ ๋ฉ€์ฉกํ•  ๊ฒฝ์šฐ๋กœ
07:04
with only two working propellers,
123
424076
2152
๋ถ„์„ํ•˜๋ฉด
07:06
we discover that there's an unconventional way to fly it.
124
426252
2740
๊ธฐ๊ตฌ๋ฅผ ๋น„ํ–‰์‹œํ‚ค๋Š” ์˜ˆ์ƒ๋ฐ–์˜ ๋ฐฉ๋ฒ•์ด ์žˆ์Œ์„ ๋ฐœ๊ฒฌํ•˜๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.
07:19
We relinquish control of yaw,
125
439980
1682
๊ธฐ์šธ๊ธฐ ์ œ์–ด๋Š” ํฌ๊ธฐํ•˜๊ฒŒ ๋˜์ง€๋งŒ
07:21
but roll, pitch and acceleration can still be controlled
126
441686
3046
๋’ค์ง‘๊ธฐ, ์†Ÿ์•„์˜ค๋ฅด๊ธฐ, ๊ฐ€์†ํ™”๋Š” ์—ฌ์ „ํžˆ
07:24
with algorithms that exploit this new configuration.
127
444756
3346
์ด๋Ÿฌํ•œ ์ƒˆ๋กœ์šด ์„ค์ •์„ ์ฐพ์•„๋‚ด๋Š” ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์ด์šฉํ•ด ์ œ์–ดํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
07:33
Mathematical models tell us exactly when and why this is possible.
128
453668
4289
์ˆ˜ํ•™ ๋ชจ๋ธ์„ ์ดํ•ดํ•˜๋ฉด ์–ธ์ œ, ๊ทธ๋ฆฌ๊ณ  ์–ด๋–ป๊ฒŒ
์ด๋Ÿฐ ๊ฒƒ์ด ๊ฐ€๋Šฅํ•œ์ง€๋ฅผ ์ •ํ™•ํžˆ ์•Œ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
07:37
In this instance, this knowledge allows us to design
129
457981
2778
์ด๋Ÿฌํ•œ ์ง€์‹์ด ์ €ํฌ๋กœ ํ•˜์—ฌ๊ธˆ
07:40
novel machine architectures
130
460783
2473
์ฐธ์‹ ํ•œ ๊ธฐ๊ณ„ ์„ค๊ณ„๋‚˜
์ธ๊ฐ„ ์šด๋™ ์„ ์ˆ˜๋“ค์ฒ˜๋Ÿผ ๋ถ€์ƒ์„ ๋‹นํ•ด๋„ ๋ถ€๋“œ๋Ÿฝ๊ฒŒ ์›€์ง์ผ ์ˆ˜ ์žˆ๊ฒŒ ํ•ด์ฃผ๋Š”
07:43
or to design clever algorithms that gracefully handle damage,
131
463280
3668
07:46
just like human athletes do,
132
466972
1700
์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์„ค๊ณ„ํ•˜๋„๋ก ํ•ด ์ค๋‹ˆ๋‹ค.
07:48
instead of building machines with redundancy.
133
468696
2264
๋ถˆํ•„์š”ํ•œ ํ”„๋กœํŽ ๋Ÿฌ๋ฅผ ๋” ๋‹ฌ์ง€ ์•Š์•„๋„ ๋ง์ด์ฃ .
07:52
We can't help but hold our breath
134
472688
1657
์šฐ๋ฆฌ๋Š” ๋‹ค์ด๋น™ ์„ ์ˆ˜๊ฐ€ ๋ฌผ์— ๋›ฐ์–ด๋“ค๋ฉฐ ๊ณต์ค‘์ œ๋น„๋ฅผ ๋„˜์„ ๋•Œ๋‚˜
07:54
when we watch a diver somersaulting into the water,
135
474369
2675
์žฅ๋Œ€ ๋†’์ด๋›ฐ๊ธฐ ์„ ์ˆ˜๊ฐ€ ๊ณต์ค‘์—์„œ ๋ชธ์„ ๋น„ํ‹€๋ฉฐ
07:57
or when a vaulter is twisting in the air,
136
477068
1953
๋ฐ‘์œผ๋กœ ๋น ๋ฅด๊ฒŒ ๋–จ์–ด์ง€๋Š” ๊ฒƒ์„ ๋ณผ ๋•Œ๋งˆ๋‹ค
์ˆจ์„ ์ฃฝ์ผ ์ˆ˜ ๋ฐ–์— ์—†์Šต๋‹ˆ๋‹ค.
07:59
the ground fast approaching.
137
479045
1507
08:00
Will the diver be able to pull off a rip entry?
138
480576
2544
"๋‹ค์ด๋น™ ์„ ์ˆ˜๊ฐ€ ๋ฆฝ ์ž…์ˆ˜๋ฅผ ํ•ด๋‚ผ ์ˆ˜ ์žˆ์„๊นŒ?"
08:03
Will the vaulter stick the landing?
139
483144
1807
"์žฅ๋Œ€ ๋†’์ด๋›ฐ๊ธฐ ์„ ์ˆ˜๊ฐ€ ๋™์ž‘์„ ๊น”๋”ํ•˜๊ฒŒ ๋งˆ๋ฌด๋ฆฌํ•  ์ˆ˜ ์žˆ์„๊นŒ?" ํ•˜๊ณ  ๋ง์ด์ฃ .
08:05
Suppose we want this quad here to perform a triple flip
140
485292
3180
์˜ˆ๋ฅผ ๋“ค์–ด ์ด ์ฟผ๋“œ๊ฐ€
์„ธ ๋ฒˆ ์—ฐ์†์œผ๋กœ ๋’ค์ง‘๊ธฐ๋ฅผ ํ•œ ๋‹ค์Œ
08:08
and finish off at the exact same spot that it started.
141
488496
2826
์ฒ˜์Œ์˜ ์œ„์น˜๋กœ ์ •ํ™•ํ•˜๊ฒŒ ๋˜๋Œ์•„๊ฐ€๊ฒŒ ํ•˜๊ณ  ์‹ถ๋‹ค๊ณ  ํ•ด๋ณด์ฃ .
08:11
This maneuver is going to happen so quickly
142
491736
2119
์ด ์กฐ์ž‘์€ ์•„์ฃผ ๋นจ๋ผ์„œ
08:13
that we can't use position feedback to correct the motion during execution.
143
493879
3845
์œ„์น˜์— ๋Œ€ํ•œ ํ”ผ๋“œ๋ฐฑ์œผ๋กœ ๋™์ž‘์„ ์‹œ์ •ํ•  ์ˆ˜ ์—†์Šต๋‹ˆ๋‹ค.
08:17
There simply isn't enough time.
144
497748
1597
๊ทธ๋Ÿด ์‹œ๊ฐ„์ด ์—†๊ฑฐ๋“ ์š”.
08:19
Instead, what the quad can do is perform the maneuver blindly,
145
499782
3670
๊ทธ ๋Œ€์‹ , ์ฟผ๋“œ๋Š” ์•„๋ฌด ๊ฒƒ๋„ ๋ชจ๋ฅธ ์ฑ„๋กœ ๋™์ž‘์„ ์ˆ˜ํ–‰ํ•˜๊ณ 
08:23
observe how it finishes the maneuver,
146
503476
2208
์ž์‹ ์ด ๊ทธ ๋™์ž‘์„ ์–ด๋–ป๊ฒŒ ๋๋‚ด๋Š”์ง€ ์ธก์ •ํ•˜์—ฌ
08:25
and then use that information to modify its behavior
147
505708
2477
์–ป์€ ์ •๋ณด๋ฅผ ๊ฐ€์ง€๊ณ  ์Šค์Šค๋กœ ์ž๊ธฐ ํ–‰๋™์„ ์ˆ˜์ •ํ•ฉ๋‹ˆ๋‹ค.
08:28
so that the next flip is better.
148
508209
1638
๊ทธ๋ ‡๊ฒŒ ํ•˜๋ฉด ๋‹ค์Œ ๋’ค์ง‘๊ธฐ๋Š” ํ–ฅ์ƒ๋  ํ…Œ๋‹ˆ๊นŒ์š”.
08:30
Similar to the diver and the vaulter,
149
510421
1788
๋‹ค์ด๋น™์ด๋‚˜ ์žฅ๋Œ€ ๋†’์ด๋›ฐ๊ธฐ ์„ ์ˆ˜์™€ ๋น„์Šทํ•ฉ๋‹ˆ๋‹ค.
08:32
it is only through repeated practice
150
512233
1895
์ˆ˜๋„ ์—†์ด ์—ฐ์Šต์„ ํ•ด์•ผ
08:34
that the maneuver can be learned and executed
151
514152
2143
๋™์ž‘์ด ํ•™์Šต๋˜๊ณ  ์ตœ๊ณ ์˜ ์ˆ˜์ค€๊นŒ์ง€
08:36
to the highest standard.
152
516319
1242
๋Œ์–ด์˜ฌ๋ฆด ์ˆ˜ ์žˆ๊ฑฐ๋“ ์š”.
08:45
(Laughter)
153
525100
1269
08:46
(Applause)
154
526393
3785
(๋ฐ•์ˆ˜)
08:51
Striking a moving ball is a necessary skill in many sports.
155
531164
3448
๋งŽ์€ ์Šคํฌ์ธ  ์ข…๋ชฉ์—์„œ๋Š” ์›€์ง์ด๋Š” ๊ณต์„ ์น˜๋Š” ๊ธฐ์ˆ ์ด ํ•„์ˆ˜์ ์ž…๋‹ˆ๋‹ค.
์–ด๋–ป๊ฒŒ ํ•˜๋ฉด ๊ธฐ๊ณ„๊ฐ€ ์šด๋™ ์„ ์ˆ˜๋“ค์ฒ˜๋Ÿผ
08:55
How do we make a machine do
156
535127
1492
08:56
what an athlete does seemingly without effort?
157
536643
2588
๊ฒ‰๋ณด๊ธฐ์—๋Š” ์‰ฝ๊ฒŒ ํ•˜๋Š” ์ผ์„ ํ•  ์ˆ˜ ์žˆ๋„๋ก ํ• ๊นŒ์š”?
09:09
(Laughter)
158
549547
1507
(๋ฐ•์ˆ˜)
09:16
(Applause)
159
556143
1857
09:20
(Applause ends)
160
560500
1298
09:22
This quad has a racket strapped onto its head
161
562542
2626
์ด ์ฟผ๋“œ๋Š” ์œ—๋ถ€๋ถ„์— ๋ผ์ผ“์„ ์žฅ์ฐฉํ–ˆ๋Š”๋ฐ,
09:25
with a sweet spot roughly the size of an apple, so not too large.
162
565192
3463
๊ทธ ์œ„์—๋Š” ์‚ฌ๊ณผ ์ •๋„ ํฌ๊ธฐ๋กœ ๋„ˆ๋ฌด ํฌ์ง€๋Š” ์•Š๊ฒŒ ๊ณต์„ ์น˜๊ธฐ ์ ํ•ฉํ•œ ์œ„์น˜๊ฐ€ ํ‘œ์‹œ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค.
์ด๊ฒƒ์— ํ•„์š”ํ•œ ๊ณ„์‚ฐ์€ ๋งค 20๋ฐ€๋ฆฌ์ดˆ๋งˆ๋‹ค ์ด๋ฃจ์–ด์ง€๋Š”๋ฐ
09:29
The following calculations are made every 20 milliseconds,
163
569017
2805
09:31
or 50 times per second.
164
571846
1317
1์ดˆ์—๋Š” 50๋ฒˆ ํ•˜๋Š” ์…ˆ์ž…๋‹ˆ๋‹ค.
09:33
We first figure out where the ball is going.
165
573602
2275
์šฐ์„  ๊ณต์ด ์–ด๋””๋กœ ๊ฐ€๋Š”์ง€ ์•Œ์•„๋‚ด์•ผ ํ•˜๊ณ 
09:36
We then next calculate how the quad should hit the ball
166
576414
2621
๊ทธ ๋‹ค์Œ์—๋Š” ์ฟผ๋“œ๊ฐ€ ์–ด๋–ป๊ฒŒ ๊ณต์„ ์น  ์ง€ ๊ณ„์‚ฐํ•ด์„œ
๊ณต์ด ๋‚ ์•„๊ฐ€๋Š” ๊ณณ์œผ๋กœ ๋‚ ์•„๊ฐ€๊ฒŒ ๋˜๋Š” ๊ฑฐ์ฃ .
09:39
so that it flies to where it was thrown from.
167
579059
2485
09:41
Third, a trajectory is planned that carries the quad
168
581950
4432
์…‹์งธ๋กœ, ์ฟผ๋“œ๊ฐ€ ํ˜„ ์ƒํƒœ์—์„œ ๊ณต์„ ์น˜๊ธฐ๊นŒ์ง€
09:46
from its current state to the impact point with the ball.
169
586406
3000
์ด๋Œ๋„๋ก ํ•˜๋Š” ๊ถค๋„๊ฐ€ ๋งŒ๋“ค์–ด์ง‘๋‹ˆ๋‹ค.
09:49
Fourth, we only execute 20 milliseconds' worth of that strategy.
170
589430
3579
๋„ท์งธ๋กœ ์ด๋Ÿฐ ์ „๋žต์˜ ์‹คํ–‰์€ 20๋ฐ€๋ฆฌ์ดˆ๊นŒ์ง€๋งŒ ์‹คํ–‰๋˜๋Š”๋ฐ
09:53
Twenty milliseconds later, the whole process is repeated
171
593366
2793
20๋ฐ€๋ฆฌ์ดˆ ํ›„์—๋Š” ์ฟผ๋“œ๊ฐ€ ๊ณต์„ ๋‹ค์‹œ ์น  ๋•Œ๊นŒ์ง€
09:56
until the quad strikes the ball.
172
596183
1767
์ „ ๊ณผ์ •์ด ๋˜ํ’€์ด๋˜๋Š” ๊ฒƒ์ด์ฃ .
10:07
(Applause)
173
607290
2404
(๋ฐ•์ˆ˜)
10:10
Machines can not only perform dynamic maneuvers on their own,
174
610131
3423
๊ธฐ๊ณ„๋“ค์€ ์ด๋Ÿฌํ•œ ๋™์ž‘์„ ํ˜ผ์ž๋ฟ๋งŒ ์•„๋‹ˆ๋ผ
10:13
they can do it collectively.
175
613578
1839
์—ฌ๋Ÿฟ์ด ๋™์‹œ์—๋„ ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
10:15
These three quads are cooperatively carrying a sky net.
176
615441
3587
์ด ์„ธ ์ฟผ๋“œ๋“ค์€ ํ•จ๊ป˜ ๊ทธ๋ฌผ ํ•˜๋‚˜๋ฅผ ๋“ค๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
10:27
(Applause)
177
627589
1679
(๋ฐ•์ˆ˜)
10:32
(Applause ends)
178
632372
1589
10:33
They perform an extremely dynamic and collective maneuver
179
633985
4341
์•„์ฃผ ๋™์ ์ด๊ณ  ์ง‘๋‹จ์ ์ธ
๋™์ž‘์„ ํ•ด ๋‚ด์ฃ .
10:38
to launch the ball back to me.
180
638350
1668
๊ณต์„ ์ €์—๊ฒŒ๋กœ ๋‹ค์‹œ ๋˜์งˆ ๋•Œ
10:40
Notice that, at full extension, these quads are vertical.
181
640042
3146
์ž˜ ๋ณด์„ธ์š”, ๊ทธ๋ฌผ์„ ํŽผ์ณค์„ ๋•Œ ์ฟผ๋“œ๋“ค์€ ์ˆ˜์ง์œผ๋กœ ์„œ๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.
10:48
(Applause)
182
648486
1748
(๋ฐ•์ˆ˜)
10:50
In fact, when fully extended,
183
650258
2143
์‚ฌ์‹ค ๊ทธ๋ฌผ์ด ์™„์ „ํžˆ ํŽผ์ณ์กŒ์„ ๋•Œ๋Š”
10:52
this is roughly five times greater than what a bungee jumper feels
184
652425
3263
๋ฒˆ์ง€ ์ ํ”„์˜ ๋งˆ์ง€๋ง‰์— ๋Š๋ผ๋Š” ๋‹น๊ธฐ๋Š” ํž˜์˜
10:55
at the end of their launch.
185
655712
1970
์•ฝ ๋‹ค์„ฏ ๋ฐฐ ์ •๋„์ž…๋‹ˆ๋‹ค.
11:03
The algorithms to do this are very similar
186
663203
2459
์ด๋Ÿฐ ๊ฒƒ์„ ํ•˜๊ธฐ ์œ„ํ•œ ์•Œ๊ณ ๋ฆฌ์ฆ˜์€
11:05
to what the single quad used to hit the ball back to me.
187
665686
2915
์ฟผ๋“œ ํ•˜๋‚˜๋กœ ๊ณต์„ ์น  ๋•Œ์™€ ์•„์ฃผ ๋น„์Šทํ•ฉ๋‹ˆ๋‹ค.
11:09
Mathematical models are used to continuously re-plan
188
669070
2840
์ˆ˜ํ•™ ๋ชจ๋ธ์ด ์‚ฌ์šฉ๋˜์–ด 1์ดˆ์— 50๋ฒˆ์”ฉ
11:11
a cooperative strategy 50 times per second.
189
671934
2831
๋Š์ž„์—†์ด ํ˜‘๋ ฅ ์ „๋žต์„ ๊ณ„์‚ฐํ•˜์ฃ .
11:16
Everything we have seen so far has been about the machines and their capabilities.
190
676506
4259
์ง€๊ธˆ๊นŒ์ง€ ๋ณด์‹  ๊ฒƒ๋“ค์€
๊ธฐ๊ณ„์™€ ๊ธฐ๊ณ„๊ฐ€ ํ•  ์ˆ˜ ์žˆ๋Š” ๊ฒƒ๋“ค์ด์—ˆ์Šต๋‹ˆ๋‹ค.
11:20
What happens when we couple this machine athleticism
191
680789
2992
์ด๋Ÿฐ ๊ธฐ๊ณ„์˜ ์šด๋™์„ฑ๊ณผ ์ธ๊ฐ„์˜ ์šด๋™์„ฑ์„
11:23
with that of a human being?
192
683805
1483
๊ฒฐํ•ฉํ•˜๋ฉด ์–ด๋–ป๊ฒŒ ๋ ๊นŒ์š”?
11:25
What I have in front of me is a commercial gesture sensor
193
685765
3596
์ œ ์•ž์— ์žˆ๋Š” ๊ฒƒ์€ ์ฃผ๋กœ ๊ฒŒ์ž„์— ์‚ฌ์šฉ๋˜๋Š”
11:29
mainly used in gaming.
194
689385
1285
์‹œํŒ์šฉ ๋ชธ์ง“ ์„ผ์„œ์ž…๋‹ˆ๋‹ค.
11:30
It can recognize what my various body parts
195
690694
2048
์ œ ๋ชธ์˜ ์—ฌ๋Ÿฌ ๋ถ€๋ถ„์ด ์–ด๋–ป๊ฒŒ ์›€์ง์ด๋Š”์ง€
11:32
are doing in real time.
196
692766
1317
์‹ค์‹œ๊ฐ„์œผ๋กœ ์ธ์‹ํ•˜์ฃ .
11:34
Similar to the pointer that I used earlier,
197
694710
2079
์•„๊นŒ ์ œ๊ฐ€ ์ผ๋˜ ํฌ์ธํ„ฐ์™€ ๋น„์Šทํ•˜๊ฒŒ
11:36
we can use this as inputs to the system.
198
696813
2222
์ด๊ฒƒ์„ ์ด์šฉํ•ด ์‹œ์Šคํ…œ์— ๋ช…๋ น์„ ์ž…๋ ฅํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
11:39
We now have a natural way of interacting
199
699454
2372
์ด๋ ‡๊ฒŒ ํ•˜๋ฉด ์ œ ๋ชธ์ง“์„ ํ†ตํ•ด ๊ธฐ๊ณ„์˜ ๋‹ค๋“ฌ์–ด์ง€์ง€ ์•Š์€ ์šด๋™์„ฑ๊ณผ
11:41
with the raw athleticism of these quads with my gestures.
200
701850
3261
์ž์—ฐ์Šค๋Ÿฝ๊ฒŒ ์ƒํ˜ธ์ž‘์šฉ ํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.
12:22
(Applause)
201
742715
3857
(๋ฐ•์ˆ˜)
๊ฐ€์ƒ์ด ์•„๋‹ˆ๋ผ ์ง์ ‘์ ์œผ๋กœ ์ƒํ˜ธ์ž‘์šฉ ํ•  ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
12:36
Interaction doesn't have to be virtual.
202
756354
2016
12:38
It can be physical.
203
758394
1190
12:39
Take this quad, for example.
204
759989
1515
์ด ์ฟผ๋“œ๋ฅผ ์˜ˆ๋กœ ๋ณด์—ฌ ๋“œ๋ฆฌ๋ฉด
12:41
It's trying to stay at a fixed point in space.
205
761893
2223
ํ•œ ๊ณต๊ฐ„์— ๊ณ ์ •๋˜์–ด ๋จธ๋ฌด๋ฅด๋ ค๊ณ  ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
12:44
If I try to move it out of the way, it fights me,
206
764933
3689
์–ต์ง€๋กœ ์˜ฎ๊ธฐ๋ ค๊ณ  ํ•˜๋ฉด, ์ €ํ•ญํ•˜๋ฉฐ
12:48
and moves back to where it wants to be.
207
768646
1874
์žˆ์–ด์•ผ ํ•  ๊ณณ์œผ๋กœ ๋˜๋Œ์•„๊ฐ‘๋‹ˆ๋‹ค.
12:52
We can change this behavior, however.
208
772254
2087
ํ•˜์ง€๋งŒ ์ด๋Ÿฐ ํ–‰๋™์„ ์ˆ˜์ •ํ•  ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
12:55
We can use mathematical models
209
775373
1675
์ˆ˜ํ•™ ๋ชจ๋ธ์„ ์ด์šฉํ•ด์„œ
12:57
to estimate the force that I'm applying to the quad.
210
777072
2664
์ œ๊ฐ€ ์ฟผ๋“œ์— ๊ฐ€ํ•˜๋Š” ํž˜์„ ์ธก์ •ํ•˜๊ฒŒ ํ•ด์„œ
13:00
Once we know this force, we can also change the laws of physics,
211
780205
3214
์–ด๋Š ์ •๋„์˜ ํž˜์ธ์ง€ ์•Œ์•„๋‚ด๊ธฐ๋งŒ ํ•˜๋ฉด ๋ฌผ๋ฆฌ ๋ฒ•์น™๊นŒ์ง€ ๋ฐ”๊ฟ€ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
13:03
as far as the quad is concerned, of course.
212
783443
2087
๋ฌผ๋ก  ์ฟผ๋“œ์— ํ•œํ•ด์„œ ๋ง์ด์ง€๋งŒ์š”.
13:07
Here, the quad is behaving as if it were in a viscous fluid.
213
787942
3309
์ด ์ฟผ๋“œ๋Š” ๋งˆ์น˜ ์ ์•ก์งˆ ์†์— ์žˆ๋Š” ๊ฒƒ์ฒ˜๋Ÿผ
์›€์ง์ด๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
13:14
We now have an intimate way of interacting with a machine.
214
794570
3719
์ด์ œ ๊ธฐ๊ณ„์™€ ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š”
์นœ๋ฐ€ํ•œ ๋ฐฉ๋ฒ•์ด ์ƒ๊ธด ์…ˆ์ด์ฃ .
13:18
I will use this new capability to position
215
798593
2482
์ด ์ƒˆ๋กœ์šด ๋Šฅ๋ ฅ์„ ์‚ฌ์šฉํ•˜์—ฌ
13:21
this camera-carrying quad to the appropriate location
216
801099
2553
์ด ์นด๋ฉ”๋ผ ๋‹ฌ๋ฆฐ ์ฟผ๋“œ๊ฐ€ ์ ์ ˆํ•œ ์œ„์น˜์— ์žˆ๊ฒŒ ๋งŒ๋“ค์–ด
13:23
for filming the remainder of this demonstration.
217
803676
2428
์ด ๊ฐ•์˜์˜ ๋‚˜๋จธ์ง€๋ฅผ ์ดฌ์˜ํ•ด ๋ณด๊ฒ ์Šต๋‹ˆ๋‹ค.
13:36
So we can physically interact with these quads
218
816583
2383
์ด ์ฟผ๋“œ๋“ค๊ณผ ์ง์ ‘์ ์œผ๋กœ ์ƒํ˜ธ์ž‘์šฉ ํ•จ์œผ๋กœ์จ
13:38
and we can change the laws of physics.
219
818990
1960
๋ฌผ๋ฆฌ ๋ฒ•์น™์„ ๋ฐ”๊ฟ€ ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
13:41
Let's have a little bit of fun with this.
220
821393
1977
ํ•œ๋ฒˆ ์žฌ๋ฐŒ๋Š” ๊ฑธ ํ•ด ๋ณผ๊นŒ์š”?
13:43
For what you will see next,
221
823539
1369
์ด์ œ๋Š” ์„ธ ๊ฐœ์˜ ์ฟผ๋“œ๋“ค์ด
13:44
these quads will initially behave as if they were on Pluto.
222
824932
3294
๋ช…์™•์„ฑ์— ์žˆ๋Š” ๊ฒƒ์ฒ˜๋Ÿผ ์›€์ง์ผ ๊ฒ๋‹ˆ๋‹ค.
13:48
As time goes on, gravity will be increased
223
828906
2288
์‹œ๊ฐ„์ด ์ง€๋‚จ์— ๋”ฐ๋ผ ์ค‘๋ ฅ์ด ์ฆ๊ฐ€ํ•ด์„œ
13:51
until we're all back on planet Earth,
224
831218
2025
์ง€๊ตฌ์™€ ๋˜‘๊ฐ™์ด ๋˜๊ฒ ์ง€๋งŒ
13:53
but I assure you we won't get there.
225
833267
1766
๊ฑฐ๊ธฐ๊นŒ์ง€ ๊ฐ€์ง€๋Š” ์•Š๋„๋ก ํ•˜๊ฒ ์Šต๋‹ˆ๋‹ค.
13:55
Okay, here goes.
226
835462
1024
ํ•œ๋ฒˆ ํ•ด ๋ณผ๊นŒ์š”?
14:04
(Laughter)
227
844897
1551
(์›ƒ์Œ)
14:34
(Laughter)
228
874778
2960
(์›ƒ์Œ)
14:37
(Applause)
229
877858
3761
(๋ฐ•์ˆ˜)
14:41
Whew!
230
881760
1066
ํœด!
์ด์ œ ์ƒ๊ฐํ•˜์‹œ๊ฒ ์ฃ .
14:47
You're all thinking now,
231
887045
1227
14:48
these guys are having way too much fun,
232
888296
1951
๋„ˆ๋ฌด ์žฌ๋ฐŒ๊ธฐ๋งŒ ํ•œ ๊ฑด ์•„๋‹ˆ๋ƒ๊ณ ์š”.
14:50
and you're probably also asking yourself,
233
890271
2055
๊ทธ๋ฆฌ๊ณ  ๋˜. ์˜๋ฌธ์ด ์žˆ์œผ์‹ค ๊ฒ๋‹ˆ๋‹ค.
14:52
why exactly are they building machine athletes?
234
892350
2801
"๋Œ€์ฒด ์™œ ์ด๋Ÿฐ ์šด๋™์„ฑ ๊ธฐ๊ณ„๋ฅผ ๋งŒ๋“œ๋Š” ๊ฒƒ์ผ๊นŒ?"
14:56
Some conjecture that the role of play in the animal kingdom
235
896286
2810
๋™๋ฌผ ์„ธ๊ณ„์—์„œ์˜ "๋†€์ด"์— ๋Œ€ํ•ด ์–ด๋–ค ๋ถ„๋“ค์€
๊ธฐ์ˆ ์„ ์—ฐ๋งˆํ•˜๊ณ  ๋Šฅ๋ ฅ์„ ๊ณ„๋ฐœํ•˜๋Š” ๋ฐ ์ด์šฉ๋œ๋‹ค๊ณ  ์ถ”์ธกํ•˜๊ณ 
14:59
is to hone skills and develop capabilities.
236
899120
2137
15:01
Others think that it has more of a social role,
237
901690
2215
๋‹ค๋ฅธ ๋ถ„๋“ค์€ ์ง‘๋‹จ์„ ํ•˜๋‚˜๋กœ ๋ฌถ๊ธฐ ์œ„ํ•œ
15:03
that it's used to bind the group.
238
903929
1650
์‚ฌํšŒ์  ๊ธฐ๋Šฅ์„ ํ•œ๋‹ค๊ณ  ์ถ”์ธกํ•ฉ๋‹ˆ๋‹ค.
15:05
Similarly, we use the analogy of sports and athleticism
239
905603
3336
๊ทธ๊ฒƒ๊ณผ ๋น„์Šทํ•˜๊ฒŒ ์ €ํฌ๋Š” ์šด๋™ ๊ฒฝ๊ธฐ์™€ ์Šคํฌ์ธ  ์ •์‹ ์— ์ด๊ฒƒ์„ ๋น„์œ ํ•จ์œผ๋กœ์จ
15:08
to create new algorithms for machines
240
908963
2055
๊ธฐ๊ณ„๋“ค์˜ ์„ฑ๋Šฅ์„ ํ•œ๊ณ„๊นŒ์ง€ ๋Œ์–ด์˜ฌ๋ฆฌ๊ธฐ ์œ„ํ•œ
15:11
to push them to their limits.
241
911042
1443
์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๊ฐœ๋ฐœํ•ฉ๋‹ˆ๋‹ค.
15:13
What impact will the speed of machines have on our way of life?
242
913350
3048
๊ทธ๋ ‡๋‹ค๋ฉด ์ด๋“ค ๊ธฐ๊ณ„์˜ ์†๋„๊ฐ€ ์šฐ๋ฆฌ ์ƒํ™œ์—๋Š” ์–ด๋–ค ์˜ํ–ฅ์„ ๋ฏธ์น ๊นŒ์š”?
15:16
Like all our past creations and innovations,
243
916866
2365
๊ณผ๊ฑฐ์˜ ๋ฐœ๋ช…ํ’ˆ์ด๋‚˜ ํ˜์‹ ๊ณผ ๋งˆ์ฐฌ๊ฐ€์ง€๋กœ
15:19
they may be used to improve the human condition
244
919255
2803
์ด ๊ธฐ๊ณ„๋“ค์€ ์ธ๊ฐ„์˜ ์ƒํ™œ ํ™˜๊ฒฝ์„ ๊ฐœ์„ ํ•  ์ˆ˜๋„,
15:22
or they may be misused and abused.
245
922082
1974
์•„๋‹ˆ๋ฉด ์˜ค์šฉ๋˜๊ฑฐ๋‚˜ ๋‚จ์šฉ๋  ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
15:24
This is not a technical choice we are faced with;
246
924595
2302
์šฐ๋ฆฌ๊ฐ€ ์ง๋ฉดํ•œ ๊ฒƒ์€ ๊ธฐ์ˆ ์  ๋ฌธ์ œ๊ฐ€ ์•„๋‹ˆ๋ผ
15:26
it's a social one.
247
926921
1015
์‚ฌํšŒ์  ๋ฌธ์ œ์ธ ๊ฑฐ์ฃ .
15:28
Let's make the right choice,
248
928332
1351
์˜ณ์€ ๊ฒฐ์ •์„ ๋‚ด๋ ธ์œผ๋ฉด ์ข‹๊ฒ ์Šต๋‹ˆ๋‹ค.
15:29
the choice that brings out the best in the future of machines,
249
929707
2912
๊ธฐ๊ณ„๋ฅผ ์‚ฌ์šฉํ•ด์„œ ์ตœ๊ณ ์˜ ๋ฏธ๋ž˜๋ฅผ ์ด๋Œ์–ด๋‚ผ ์ˆ˜ ์žˆ๋Š” ๊ฒฐ์ • ๋ง์ž…๋‹ˆ๋‹ค.
15:32
just like athleticism in sports can bring out the best in us.
250
932643
2865
์šด๋™ ๊ฒฝ๊ธฐ์—์„œ์˜ ์Šคํฌ์ธ  ์ •์‹ ์ด
์šฐ๋ฆฌ๋ฅผ ์ตœ์„ ์œผ๋กœ ์ด๋Œ๋“ฏ์ด ๋ง์ž…๋‹ˆ๋‹ค.
15:36
Let me introduce you to the wizards behind the green curtain.
251
936380
3075
๊ทธ๋Ÿผ ์ด์ œ ์ด๋Ÿฐ ์‹ ๊ธฐํ•œ ๊ฒƒ๋“ค์„ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•œ ๋งˆ๋ฒ•์‚ฌ๋“ค์„ ์†Œ๊ฐœํ•˜๊ฒ ์Šต๋‹ˆ๋‹ค.
15:39
They're the current members of the Flying Machine Arena research team.
252
939479
3334
"ํ”Œ๋ผ์ž‰ ๋จธ์‹  ์•„๋ ˆ๋‚˜"์˜ ์—ฐ๊ตฌํŒ€์ž…๋‹ˆ๋‹ค.
15:42
(Applause)
253
942837
4217
(๋ฐ•์ˆ˜)
15:47
Federico Augugliaro, Dario Brescianini,
254
947267
2278
ํŽ˜๋ฐ๋ฆฌ์ฝ” ์šฐ๊ตด๋ฆฌ์•„๋กœ, ๋‹ค๋ฆฌ์˜ค ๋ธŒ๋ ˆ์ƒค๋‹ˆ๋‹ˆ, ๋งˆ์ปค์Šค ํ—จ,
15:49
Markus Hehn, Sergei Lupashin, Mark Muller and Robin Ritz.
255
949569
3425
์„ธ๋ฅด๊ฒŒ์ด ๋ฃจํผ์‰ฐ, ๋งˆํฌ ๋ฌผ๋Ÿฌ, ๊ทธ๋ฆฌ๊ณ  ๋กœ๋นˆ ๋ฆฌ์ธ ์ž…๋‹ˆ๋‹ค.
15:53
Look out for them. They're destined for great things.
256
953018
2524
์ด ์นœ๊ตฌ๋“ค์„ ์ž˜ ์ง€์ผœ๋ณด์‹œ๊ธธ ๋ฐ”๋ž๋‹ˆ๋‹ค: ์•ž์œผ๋กœ ๋Œ€๋‹จํ•œ ์ผ๋“ค์„ ํ•ด๋‚ผ ๊ฒ๋‹ˆ๋‹ค.
๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
15:55
Thank you.
257
955566
1054
15:56
(Applause)
258
956644
3456
(๋ฐ•์ˆ˜)
์ด ์›น์‚ฌ์ดํŠธ ์ •๋ณด

์ด ์‚ฌ์ดํŠธ๋Š” ์˜์–ด ํ•™์Šต์— ์œ ์šฉํ•œ YouTube ๋™์˜์ƒ์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค. ์ „ ์„ธ๊ณ„ ์ตœ๊ณ ์˜ ์„ ์ƒ๋‹˜๋“ค์ด ๊ฐ€๋ฅด์น˜๋Š” ์˜์–ด ์ˆ˜์—…์„ ๋ณด๊ฒŒ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๊ฐ ๋™์˜์ƒ ํŽ˜์ด์ง€์— ํ‘œ์‹œ๋˜๋Š” ์˜์–ด ์ž๋ง‰์„ ๋”๋ธ” ํด๋ฆญํ•˜๋ฉด ๊ทธ๊ณณ์—์„œ ๋™์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค. ๋น„๋””์˜ค ์žฌ์ƒ์— ๋งž์ถฐ ์ž๋ง‰์ด ์Šคํฌ๋กค๋ฉ๋‹ˆ๋‹ค. ์˜๊ฒฌ์ด๋‚˜ ์š”์ฒญ์ด ์žˆ๋Š” ๊ฒฝ์šฐ ์ด ๋ฌธ์˜ ์–‘์‹์„ ์‚ฌ์šฉํ•˜์—ฌ ๋ฌธ์˜ํ•˜์‹ญ์‹œ์˜ค.

https://forms.gle/WvT1wiN1qDtmnspy7