The Future of Flying Robots | Vijay Kumar | TED Talks

翻译人员: Zhiting Chen 校对人员: Yinchun Rui

In my lab, we build autonomous aerial robots

在我的实验室， 我们打造了自动飞行的机器人，

like the one you see flying here.

就是你们眼前的这种。

Unlike the commercially available drones that you can buy today,

不像当今市面上销售的那些，

this robot doesn't have any GPS on board.

这个机器人没有GPS定位系统。

So without GPS,

在没有GPS的情况下，

it's hard for robots like this to determine their position.

这样的机器人很难定位自己。

This robot uses onboard sensors, cameras and laser scanners,

这个机器人用机载传感器， 相机和激光扫描仪，

来扫描环境。

to scan the environment.

It detects features from the environment,

它能够探测到环境的特征，

and it determines where it is relative to those features,

并使用三角测量的方式来决定 不同的特征之间

using a method of triangulation.

有怎样的联系。

And then it can assemble all these features into a map,

然后它能够把所有这些信息 整合到一张地图上，

like you see behind me.

就是我背后的这张图。

And this map then allows the robot to understand where the obstacles are

这个地图能让机器人确定 障碍物的位置，

and navigate in a collision-free manner.

并巧妙地避开各种碰撞。

What I want to show you next

我接下来要展示的是

is a set of experiments we did inside our laboratory,

一组我们在户外进行的实验，

where this robot was able to go for longer distances.

证明机器人能够在户外 进行长距离飞行。

So here you'll see, on the top right, what the robot sees with the camera.

你们可以看到，在右上方， 是机器人通过照相机看到的影像。

And on the main screen --

在主屏幕上——

and of course this is sped up by a factor of four --

当然这是以4倍速度在播放展示——

on the main screen you'll see the map that it's building.

在主屏幕上，你们可以看到 地图的创建过程。

So this is a high-resolution map of the corridor around our laboratory.

这是一个高分辨率的地图， 展示了我们实验室周围走廊的样子。

And in a minute you'll see it enter our lab,

很快你们就能看到飞行机器人 进入了我们的实验室，

which is recognizable by the clutter that you see.

这么乱的地方，一看 就知道是我们的实验室。

(Laughter)

（笑声）

But the main point I want to convey to you

但我重点想说的是，

is that these robots are capable of building high-resolution maps

这些飞行器可以创建 高分辨率的地图，

at five centimeters resolution,

达到5厘米的分辨率，

allowing somebody who is outside the lab, or outside the building

可以使那些在实验室外， 或者是房屋外的人，

to deploy these without actually going inside,

无需进入房间即可观察到这些内容，

and trying to infer what happens inside the building.

并尝试了解房屋内发生的状况。

Now there's one problem with robots like this.

当然这样的机器人也有问题。

The first problem is it's pretty big.

首先，它有些大了。

Because it's big, it's heavy.

因为很大，所以很重。

And these robots consume about 100 watts per pound.

这些机器人每磅重量可以消耗 100瓦的电能。

And this makes for a very short mission life.

这使得它的运作时间很短。

The second problem

第二个问题就是，

is that these robots have onboard sensors that end up being very expensive --

这些机器人有机载传感器， 它们非常昂贵——

a laser scanner, a camera and the processors.

一个激光扫描仪，一个相机 以及处理器。

That drives up the cost of this robot.

这些让这个飞行机器人的 制造成本十分高昂。

So we asked ourselves a question:

因此我们问了自己一个问题：

what consumer product can you buy in an electronics store

你能在电子产品商店买到怎样的物品，

that is inexpensive, that's lightweight, that has sensing onboard and computation?

它既不昂贵，又很轻， 还有机载传感器和计算能力？

And we invented the flying phone.

于是我们发明了会飞的手机。

(Laughter)

（笑声）

So this robot uses a Samsung Galaxy smartphone that you can buy off the shelf,

这个机器人使用了你可以轻松购买到的 三星银河系列手机，

and all you need is an app that you can download from our app store.

你所需要的只是一个应用程序， 可以从我们的应用程序商店下载。

And you can see this robot reading the letters, "TED" in this case,

你们可以看到这个飞行机器人在读字， “TED”，

looking at the corners of the "T" and the "E"

观察字母“T”和“E”角落的位置，

and then triangulating off of that, flying autonomously.

再应用三角测量法，实现自主飞行。

That joystick is just there to make sure if the robot goes crazy,

那个手柄的作用仅限于 当飞行器失控时，

Giuseppe就能让它失去功能。

Giuseppe can kill it.

(Laughter)

（笑声）

In addition to building these small robots,

除了制造这些飞行机器人外，

we also experiment with aggressive behaviors, like you see here.

我们还试着让它们做更激烈的动作， 比如这样。

So this robot is now traveling at two to three meters per second,

这个飞行机器人目前的速度是 每秒钟2到3米，

pitching and rolling aggressively as it changes direction.

在转变方向的时候 激烈地起降和翻转。

The main point is we can have smaller robots that can go faster

重点是，我们还能制造更小，更快的 飞行机器人，

and then travel in these very unstructured environments.

它们可以在非常复杂的环境中飞行。

And in this next video,

下一个影片中，

just like you see this bird, an eagle, gracefully coordinating its wings,

你们可以看到这只鹰， 优雅地协调它的翅膀，

its eyes and feet to grab prey out of the water,

眼睛和爪子之间的配合， 把猎物抓出水面，

our robot can go fishing, too.

我们的机器人也能去捕鱼。

(Laughter)

（笑声）

In this case, this is a Philly cheesesteak hoagie that it's grabbing out of thin air.

在这个试验中，它在空中抓起了 这个特大号菲力芝士牛排三明治。

(Laughter)

（笑声）

So you can see this robot going at about three meters per second,

你们可以看到这个机器人 以每秒3米左右的速度飞行，

比步行的速度快一些， 同时还能协调它的手臂和爪子，

which is faster than walking speed, coordinating its arms, its claws

and its flight with split-second timing to achieve this maneuver.

以极快的速度完成整套动作。

In another experiment,

在另一个试验中，

I want to show you how the robot adapts its flight

我想要展示， 机器人是如何依据悬浮载重

to control its suspended payload,

来调整飞行模式的，

whose length is actually larger than the width of the window.

它的总长度大于窗子的高度。

So in order to accomplish this,

为了完成任务，

it actually has to pitch and adjust the altitude

它需要向下倾斜，调整高度，

and swing the payload through.

然后把重物摆动过去。

But of course we want to make these even smaller,

但是，我们希望让机器人变得更小，

and we're inspired in particular by honeybees.

而蜜蜂启发了我们。

So if you look at honeybees, and this is a slowed down video,

如果你们观察一下蜜蜂， 在这个慢速的视频里，

they're so small, the inertia is so lightweight --

它们是这么小， 惯性是这么轻 ——

(Laughter)

（笑声）

that they don't care -- they bounce off my hand, for example.

它们并不在意—— 它们会撞击我的手，比如说。

This is a little robot that mimics the honeybee behavior.

这是个小机器人， 可以模拟蜜蜂的行为。

And smaller is better,

越小越好，

because along with the small size you get lower inertia.

因为尺寸小，惯性就会小。

Along with lower inertia --

惯性小——

(Robot buzzing, laughter)

（机器人嗡嗡声，笑声）

along with lower inertia, you're resistant to collisions.

惯性小，发生撞击的 可能性就低一些。

And that makes you more robust.

这使得机器人更耐用。

So just like these honeybees, we build small robots.

模仿这些蜜蜂， 我们制作了小型机器人。

And this particular one is only 25 grams in weight.

这个特殊的机器人只有25克重。

It consumes only six watts of power.

它的耗电量仅为6瓦。

And it can travel up to six meters per second.

它的飞行速度可达每秒6米。

因此，如果我把它的尺寸按比例放大，

So if I normalize that to its size,

it's like a Boeing 787 traveling ten times the speed of sound.

就好比一架波音787飞机 以10倍于音速的速度飞行。

(Laughter)

（笑声）

And I want to show you an example.

我还要给你们展示一个例子。

This is probably the first planned mid-air collision, at one-twentieth normal speed.

这可能是第一个计划中的 以20分之1正常速度进行的空中相撞。

These are going at a relative speed of two meters per second,

它们的相对速度为每秒2米，

这里展示了基本的原理。

and this illustrates the basic principle.

The two-gram carbon fiber cage around it prevents the propellers from entangling,

两克重的碳纤维笼子包围着它 使螺旋桨不会受损，

but essentially the collision is absorbed and the robot responds to the collisions.

但最关键的是，撞击被吸收了， 机器人能够对撞击做出反应。

And so small also means safe.

并且，小也意味着安全。

In my lab, as we developed these robots,

在我的实验室， 最初开发这些机器人的时候，

we start off with these big robots

我们从大的机器人开始

and then now we're down to these small robots.

现在着手做小的机器人。

And if you plot a histogram of the number of Band-Aids we've ordered

如果你们看一下我们购买 创可贴的数量统计图，

in the past, that sort of tailed off now.

就知道现在我们已经几乎不需要买了。

因为这些机器人非常安全。

Because these robots are really safe.

The small size has some disadvantages,

小尺寸也有一些缺点，

and nature has found a number of ways to compensate for these disadvantages.

但是自然界有很多方法 来弥补这些缺陷。

The basic idea is they aggregate to form large groups, or swarms.

最基本的想法是， 它们可以聚集在一起组成大型的群落。

So, similarly, in our lab, we try to create artificial robot swarms.

因此，在实验室里， 我们也试着去组建机器人群组。

And this is quite challenging

这是非常有难度的工作，

because now you have to think about networks of robots.

因为我们要考虑飞行机器人网络。

And within each robot,

在每一个飞行器中，

you have to think about the interplay of sensing, communication, computation --

我们都需要考虑传感，沟通和计算， 这些互相影响的因素——

and this network then becomes quite difficult to control and manage.

这样的机器人网络 不易控制和管理。

So from nature we take away three organizing principles

从大自然中，我们汲取了 三个组织原则，

that essentially allow us to develop our algorithms.

最终帮助我们完成了算法的发展。

The first idea is that robots need to be aware of their neighbors.

第一个是，机器人 需要注意到它的邻居们。

They need to be able to sense and communicate with their neighbors.

它们要有能力去感知相邻的机器人 并与它们交流。

So this video illustrates the basic idea.

这个视频展示了最基本的想法。

You have four robots --

有4个机器人——

其中的一个被人手动控制了。

one of the robots has actually been hijacked by a human operator, literally.

But because the robots interact with each other,

但是因为机器人会相互通讯，

they sense their neighbors,

它们能够感知到旁边机器人的行动，

they essentially follow.

也会跟着它移动。

And here there's a single person able to lead this network of followers.

这个人能够领导整个 飞行机器人群体的行动。

So again, it's not because all the robots know where they're supposed to go.

再强调一次，这并不是因为 所有的机器人都知道它们要去哪里。

It's because they're just reacting to the positions of their neighbors.

而是它们能根据相邻机器人位置变化 做出相应的反应。

(Laughter)

（笑声）

So the next experiment illustrates the second organizing principle.

下一个实验展示了第二个组织原则。

And this principle has to do with the principle of anonymity.

这个原则与匿名原则有关。

Here the key idea is that

其中的关键在于，

the robots are agnostic to the identities of their neighbors.

机器人是不知道与它们相邻的 机器人的身份的。

They're asked to form a circular shape,

它们被要求形成一个圆圈，

and no matter how many robots you introduce into the formation,

不管你往这个阵列中放多少个机器人，

or how many robots you pull out,

或者是你拿出来多少个机器人，

每个机器人都会很简单地 根据它相邻机器人的行为做出反应。

each robot is simply reacting to its neighbor.

It's aware of the fact that it needs to form the circular shape,

它清楚地知道需要形成圆圈，

but collaborating with its neighbors

与它的相邻机器人配合行动，

it forms the shape without central coordination.

这个过程不需要中枢系统协调。

Now if you put these ideas together,

综合这两个想法之后，

the third idea is that we essentially give these robots

第三个想法就是我们会

mathematical descriptions of the shape they need to execute.

用数学方式来描述 它们需要组成的阵列形状。

And these shapes can be varying as a function of time,

这些形状会根据时间发生变化，

and you'll see these robots start from a circular formation,

你们看到这些飞行器 从一个圆形开始，

change into a rectangular formation, stretch into a straight line,

接着变为长方形， 然后变成一条直线，

back into an ellipse.

又变回椭圆形。

And they do this with the same kind of split-second coordination

它们通过瞬间的协调 来完成这些动作，

that you see in natural swarms, in nature.

就像自然界里的蜂群一样。

So why work with swarms?

为什么要模仿蜂群呢？

Let me tell you about two applications that we are very interested in.

我们对这项技术的两种应用 非常感兴趣。

The first one has to do with agriculture,

第一个有关于农业，

which is probably the biggest problem that we're facing worldwide.

农业应该是全球面临最严峻的问题。

As you well know,

大家都知道，

地球上每7个人中就有1个营养不良。

one in every seven persons in this earth is malnourished.

Most of the land that we can cultivate has already been cultivated.

我们耕种了绝大部分可耕的土地。

And the efficiency of most systems in the world is improving,

世界上很多系统的效率都在提高，

but our production system efficiency is actually declining.

但是我们生产系统的 效率事实上却是在下降。

And that's mostly because of water shortage, crop diseases, climate change

很大程度上源于水源短缺， 作物病害，气候变化

and a couple of other things.

和许多其他的原因。

So what can robots do?

那么，这些机器人能做什么呢？

Well, we adopt an approach that's called Precision Farming in the community.

我们称之为社区里的精密耕作。

And the basic idea is that we fly aerial robots through orchards,

我们基本的想法是 让这些飞行机器人飞越果园，

and then we build precision models of individual plants.

为每一株植物搭建精密的模型。

So just like personalized medicine,

就像个性化的机器，

while you might imagine wanting to treat every patient individually,

想象每个病人都有一对一服务，

what we'd like to do is build models of individual plants

我们要做的就是 为每株植物单独建立模型，

and then tell the farmer what kind of inputs every plant needs --

然后告诉农民 每株植物各需要些什么——

the inputs in this case being water, fertilizer and pesticide.

需要的因素在这里指的是水， 肥料和杀虫剂。

Here you'll see robots traveling through an apple orchard,

你们现在看到的是 机器人飞过一片苹果园，

and in a minute you'll see two of its companions

很快你们就会看到另外两个机器人

doing the same thing on the left side.

在左侧做着同样的事。

And what they're doing is essentially building a map of the orchard.

它们在制作果园的地图。

Within the map is a map of every plant in this orchard.

这个地图里有果园里每株植物的位置。

(Robot buzzing)

（机器人嗡嗡声）

我们来看一下 这些地图是什么样子的。

Let's see what those maps look like.

In the next video, you'll see the cameras that are being used on this robot.

在下一个视频中，你们会看到 机器人上安装的各种摄像头。

On the top-left is essentially a standard color camera.

左上角的是一个高清彩色摄像机。

On the left-center is an infrared camera.

左边中间的是一个红外线摄像机。

And on the bottom-left is a thermal camera.

左下方的是一个热成像摄像机。

And on the main panel, you're seeing a three-dimensional reconstruction

在主面板上，你们会看到 每一株果树的三维图像。

of every tree in the orchard as the sensors fly right past the trees.

这是当机器人飞过果树时 传感器采集到的数据。

Armed with information like this, we can do several things.

有了这些信息， 我们就可以做很多事情。

The first and possibly the most important thing we can do is very simple:

我们能做的第一件也可能是最重要的事 非常简单：

count the number of fruits on every tree.

清点每株果树上果实的数量。

By doing this, you tell the farmer how many fruits she has in every tree

这样，你就能告诉农民 每株果树上有多少果实，

and allow her to estimate the yield in the orchard,

农民就能估算出果园的产量，

optimizing the production chain downstream.

并调整下游的生产链。

The second thing we can do

我们能做的第二件事是

is take models of plants, construct three-dimensional reconstructions,

给植物建构三维立体图，

and from that estimate the canopy size,

从而推算出树冠的尺寸，

and then correlate the canopy size to the amount of leaf area on every plant.

然后通过树冠尺寸推算出 每株植物的树叶面积。

And this is called the leaf area index.

这被称为树叶面积指数。

So if you know this leaf area index,

那么如果我们知道一棵树的 树叶面积指数，

you essentially have a measure of how much photosynthesis is possible in every plant,

就能大概测算出这株果树 在进行多少光合作用，

which again tells you how healthy each plant is.

就能知道这株果树的健康状况。

By combining visual and infrared information,

将视觉信息和红外线信息合起来，

we can also compute indices such as NDVI.

我们还能计算出一些指数， 例如常态化差值植生指数。

And in this particular case, you can essentially see

在这个例子中， 你们可以明显看到

there are some crops that are not doing as well as other crops.

有一些植物并不像其他植物那样健康。

This is easily discernible from imagery,

在图像中很容易识别，

不仅是在视觉图像中，而是结合了

not just visual imagery but combining

both visual imagery and infrared imagery.

视觉和红外线的图像中。

And then lastly,

最后，

one thing we're interested in doing is detecting the early onset of chlorosis --

我们感兴趣的另一件事 是检测早期的植被萎黄病——

and this is an orange tree --

这是一棵橘子树——

你们可以看到泛黄的树叶。

which is essentially seen by yellowing of leaves.

But robots flying overhead can easily spot this autonomously

当机器人飞过这棵树的顶部 就能很快自主识别出来，

and then report to the farmer that he or she has a problem

然后报告农民 他/她遇到问题了，

in this section of the orchard.

就在果园的这个区域。

Systems like this can really help,

这样的系统真的非常实用，

and we're projecting yields that can improve by about ten percent

我们估计这能帮助农业产量提升10%，

and, more importantly, decrease the amount of inputs such as water

更重要的是，通过使用飞行器蜂群，

by 25 percent by using aerial robot swarms.

水资源的用量能够降低25%。

Lastly, I want you to applaud the people who actually create the future,

最后，我希望大家为创造未来的人们 致以热烈的掌声，

Yash Mulgaonkar, Sikang Liu and Giuseppe Loianno,

Yash Mulgaonkar, Sikang Liu 和Giuseppe Loianno,

who are responsible for the three demonstrations that you saw.

你们看到的三次演示都是他们完成的。

Thank you.

谢谢。

(Applause)

（掌声）