ADM-201 dump PMP dumps pdf SSCP exam materials CBAP exam sample questions

2012最佳善于学习的飞行机器人 – 译学馆
未登陆,请登陆后再发表信息
最新评论 (0)
播放视频

2012最佳善于学习的飞行机器人

Robots that fly ... and cooperate | Vijay Kumar

早上好
Good morning.
今天我想谈谈自主飞行沙滩球
I’m here today to talk about autonomous flying beach balls.
(笑声)
(Laughter)
不 其实是灵巧的遥控机器人 就像这个一样
No, agile aerial robots like this one.
我想告诉你一个在这设计时的小小的挑战
I’d like to tell you a little bit about the challenges in building these,
和一些极好的应用这项技术的机会
and some of the terrific opportunities for applying this technology.
这些飞行器和无人机连接在一起
So these robots are related to unmanned aerial vehicles.
然而 你能看见这些飞行器是如此巨大
However, the vehicles you see here are big.
他们有上千磅重 也决不灵敏
They weigh thousands of pounds, are not by any means agile.
他们甚至不是自动化机器
They’re not even autonomous.
事实上 大量飞行器都由航空人员运行
In fact, many of these vehicles are operated by flight crews
包括各种各样的飞行员
that can include multiple pilots,
传感器的操纵者
operators of sensors,
和任务协调者
and mission coordinators.
我们感兴趣的是正在改进的飞行器
What we’re interested in is developing robots like this —
比如这里的两幅画
and here are two other pictures —
一副是你能在货架上买到的飞行器
of robots that you can buy off the shelf.
这架直升飞机有4个螺旋桨
So these are helicopters with four rotors,
每一个螺旋桨不超过一米
and they’re roughly a meter or so in scale,
只有几磅重
and weigh several pounds.
如果我们可以改造一下传感器和处理器
And so we retrofit these with sensors and processors,
这些飞行器就能在室内飞行
and these robots can fly indoors.
不需要GPS
Without GPS.
我手里拿着的这个飞行器
The robot I’m holding in my hand
就是这个
is this one,
这由两个学生创造
and it’s been created by two students,
Alex和Daniel
Alex and Daniel.
这个只比0.1磅重一点的飞行器
So this weighs a little more than a tenth of a pound.
只消耗大约15瓦特的能量
It consumes about 15 watts of power.
你能看见直径大概只有八英寸
And as you can see, it’s about eight inches in diameter.
让我给你一个简短的解说
So let me give you just a very quick tutorial
看这些飞行器如何工作
on how these robots work.
四个螺旋桨
So it has four rotors.
当四个螺旋桨转速相同
If you spin these rotors at the same speed,
飞行器就开始运转
the robot hovers.
如果增加每个螺旋桨的速度
If you increase the speed of each of these rotors,
飞行器飞得越高 速度就增加越快
then the robot flies up, it accelerates up.
当然如果飞行器倾斜了
Of course, if the robot were tilted,
就得减速 这样它会回到水平方向
inclined to the horizontal,
再从这个方向加速
then it would accelerate in this direction.
怎么能让它侧过来呢
So to get it to tilt,
有两个途径
there’s one of two ways of doing it.
在这照片里 你能看到四号螺旋桨加速
So in this picture, you see that rotor four is spinning faster
同时二号螺旋桨变慢
and rotor two is spinning slower.
如果真的是这样
And when that happens,
这时飞行器会摇摆不停
there’s a moment that causes this robot to roll.
反之亦然
And the other way around,
如果你增加三号螺旋桨的速度 再减少一号螺旋桨的速度
if you increase the speed of rotor three and decrease the speed of rotor one,
飞行器就会向前倒
then the robot pitches forward.
最后一种情况
And then finally,
如果旋转反方向的一对螺旋桨
if you spin opposite pairs of rotors
使他们的速度大于另外一对
faster than the other pair,
飞行器就会原地旋转
then the robot yaws about the vertical axis.
所以飞行器上的处理器
So an on-board processor
基本可以判断出动作的执行
essentially looks at what motions need to be executed
和动作的组合
and combines these motions,
并且决定下达发送出指令
and figures out what commands to send to the motors —
一秒六百次
600 times a second.
简而言之飞行器就是这么工作的
That’s basically how this thing operates.
这种设计的好处之一
So one of the advantages of this design
是当你把降低飞行器的尺寸
is when you scale things down,
其灵活性自然会提高
the robot naturally becomes agile.
在这里R代表飞行器的特征长度
So here, R is the characteristic length of the robot.
即直径的一半 半径
It’s actually half the diameter.
而且当你减小半径时有大量物理因素影响
And there are lots of physical parameters that change as you reduce R.
其中最重要的是惯性
The one that’s most important is the inertia,
或者运动阻力
or the resistance to motion.
所以解决惯性导致的角向运动
So it turns out the inertia, which governs angular motion,
就要缩放到半径的五分之一
scales as a fifth power of R.
你制作的半径越小
So the smaller you make R,
惯性就更急剧地减少
the more dramatically the inertia reduces.
结果角向的加速运动
So as a result, the angular acceleration,
这里被表示为希腊字母α
denoted by the Greek letter alpha here,
等于1除以半径
goes as 1 over R.
它是和半径相反的比例
It’s inversely proportional to R.
你制作得越小 你就能越快地旋转
The smaller you make it, the more quickly you can turn.
这在这些影像资料里是很清楚的
So this should be clear in these videos.
在右下角 一个飞行器正在360°旋转飞行
On the bottom right, you see a robot performing a 360-degree flip
只需要不到半秒
in less than half a second.
连续翻转 可能需要更多一点的时间
Multiple flips, a little more time.
飞行器上的处理器
So here the processes on board
从加速器和陀螺仪得到反馈
are getting feedback from accelerometers and gyros on board,
然后计算 正如我之前所说
and calculating, like I said before,
一秒做出六百次指令
commands at 600 times a second,
来稳定控制这个飞行器
to stabilize this robot.
再看左边 Daniel把飞行器扔向空中
So on the left, you see Daniel throwing this robot up into the air,
展示出控制是很稳定的
and it shows you how robust the control is.
无论你怎样扔出去
No matter how you throw it,
飞行器都会恢复平衡并且又回到他手中
the robot recovers and comes back to him.
所以为什么这样制作飞行器呢?
So why build robots like this?
嗯 这样的飞行器有许多应用
Well, robots like this have many applications.
你可以送他们进入这样的建筑里
You can send them inside buildings like this,
作为报警器寻找入侵者
as first responders to look for intruders,
或许寻找生物化学渗漏处
maybe look for biochemical leaks,
气体渗漏
gaseous leaks.
你也可以在建筑中应用他们
You can also use them for applications like construction.
也有这样的飞行器搬运房梁、塔梁
So here are robots carrying beams, columns
组合成方形的结构
and assembling cube-like structures.
我要告诉你更多一些事情
I’ll tell you a little bit more about this.
这些机器人可被用于货物运输
The robots can be used for transporting cargo.
这些小飞行器只有一个问题
So one of the problems with these small robots
是他们的有效负载能力
is their payload-carrying capacity.
你或许需要很多飞行器去搬运货物
So you might want to have multiple robots carry payloads.
这张图是一个最近我们做的实验
This is a picture of a recent experiment we did —
实际上也不是很近了
actually not so recent anymore —
在日本仙台市 地震之后不久
in Sendai, shortly after the earthquake.
这种飞行器送进倒塌建筑
So robots like this could be sent into collapsed buildings,
去探测自然灾害的破坏程度
to assess the damage after natural disasters,
或者送进核反应堆
or sent into reactor buildings,
测量核辐射等级
to map radiation levels.
一个这种飞行器必须解决的基本问题就是
So one fundamental problem that the robots have to solve
如果他们是自主控制的
if they are to be autonomous,
是要自主解决如何从一个地点到另一个地点
is essentially figuring out how to get from point A to point B.
这是一个挑战
So this gets a little challenging,
因为飞行器的动力学相当的复杂
because the dynamics of this robot are quite complicated.
事实上他们存在于十二维度空间
In fact, they live in a 12-dimensional space.
所以我们用了一个技巧
So we use a little trick.
我们倒置了这个十二纬度空间
We take this curved 12-dimensional space,
使他转换成平面四维空间
and transform it into a flat, four-dimensional space.
这个四维空间只由X,Y,Z轴组成
And that four-dimensional space consists of X, Y, Z,
和旋转角
and then the yaw angle.
现在这些飞行器需要
And so what the robot does,
执行一个我们叫最小化加加加速度轨道的计划
is it plans what we call a minimum-snap trajectory.
提醒大家一个物理知识
So to remind you of physics:
有位置向量、导数、速度
You have position, derivative, velocity;
加速度
then acceleration;
加加速度
and then comes jerk,
加加加速度
and then comes snap.
飞行器把加加加速度最小化
So this robot minimizes snap.
基本上它的工作是
So what that effectively does,
创造一个平滑优美的运动曲线
is produce a smooth and graceful motion.
来避免障碍
And it does that avoiding obstacles.
飞行器在平面空间执行加加加速度轨道再转换
So these minimum-snap trajectories in this flat space are then transformed
回到复杂的十二维空间
back into this complicated 12-dimensional space,
这样就能获得控制和执行动作
which the robot must do for control and then execution.
我来举些关于
So let me show you some examples
像这样加加加速度轨道的例子
of what these minimum-snap trajectories look like.
在第一个视频里
And in the first video,
可以看见飞行器从一个地点到另一个地点
you’ll see the robot going from point A to point B,
通过了中间点
through an intermediate point.
(机翼旋转声)
(Whirring noise)
所以飞行器明显可以飞行任意曲线轨道
So the robot is obviously capable of executing any curve trajectory.
还有打圈儿的轨道
So these are circular trajectories,
这里飞行器对抗两倍重力
where the robot pulls about two G’s.
飞行器上方还有一个动作抓拍相机
Here you have overhead motion capture cameras on the top
一秒可以传送一百张画面
that tell the robot where it is 100 times a second.
也可以告诉飞行器哪里有障碍
It also tells the robot where these obstacles are.
障碍也可移动
And the obstacles can be moving.
这里可以看见Danie向空中扔这个铁环
And here, you’ll see Daniel throw this hoop into the air,
当飞行器计算铁环的位置时
while the robot is calculating the position of the hoop,
并且尝试解决如何最好地钻过铁环
and trying to figure out how to best go through the hoop.
作为一个学者
So as an academic,
我们总在试图钻出重重圈套,拿到更多经费
we’re always trained to be able to jump through hoops
来支持我们的实验室
to raise funding for our labs,
我们能用飞行器做到这点
and we get our robots to do that.
(掌声)
(Applause)
另一件飞行器可以做到的事是
So another thing the robot can do
他能记忆
is it remembers pieces of trajectory
学习过的或者预编了的轨道
that it learns or is pre-programmed.
这里 你可以看见飞行器正加入一个动作来积累动量
So here, you see the robot combining a motion that builds up momentum,
改变它的定向 再回到预设轨迹上来
and then changes its orientation and then recovers.
它必须这样是因为窗台上的缝隙
So it has to do this because this gap in the window
只比它宽度大一点点
is only slightly larger than the width of the robot.
就像一个站在跳板上的跳水员
So just like a diver stands on a springboard
向下跳来积累动量
and then jumps off it to gain momentum,
做这样的翻转 两圈半
and then does this pirouette, this two and a half somersault through
优雅地恢复平衡
and then gracefully recovers,
飞行器是自主做到的
this robot is basically doing that.
它知道怎么把小段的轨迹组合起来
So it knows how to combine little bits and pieces of trajectories
来做这些很难的动作
to do these fairly difficult tasks.
所以我想改一下发动装置
So I want to change gears.
这种小型飞行器是一个缺点是它的尺寸
So one of the disadvantages of these small robots is its size.
正如我已经说过的
And I told you earlier
我们需要制造大量的飞行器
that we may want to employ lots and lots of robots
来克服大小的限制
to overcome the limitations of size.
这又有一个困难是
So one difficulty is:
你如何定位这些飞行器呢?
How do you coordinate lots of these robots?
所以看这里 我们从自然寻找答案
And so here, we looked to nature.
我想展示一个关于Aphaenogaster沙漠蚁的视频
So I want to show you a clip of Aphaenogaster desert ants,
Stephen Pratt的实验室里 这些沙漠蚁在搬运物品
in Professor Stephen Pratt’s lab, carrying an object.
这是无花果汁的一部分
So this is actually a piece of fig.
事实上无论你在任何物品上抹上无花果汁
Actually you take any object coated with fig juice,
这些蚂蚁都会搬它回巢
and the ants will carry it back to the nest.
这些蚂蚁没有任何中央调控
So these ants don’t have any central coordinator.
他们感应旁边的蚂蚁
They sense their neighbors.
也没有任何明确的交流
There’s no explicit communication.
但因为他们能感应旁边的蚂蚁
But because they sense the neighbors
因为他们能感应物品
and because they sense the object,
他们在整个团队了有默契
they have implicit coordination across the group.
这种默契是我们想我们的飞行器有的
So this is the kind of coordination we want our robots to have.
当一个飞行器被其他飞行器围住
So when we have a robot which is surrounded by neighbors —
看飞行器I和飞行器J
and let’s look at robot I and robot J —
我希望飞行器做到
what we want the robots to do,
分别监控他们的距离
is to monitor the separation between them,
正当他们编队飞行时
as they fly in formation.
我也希望
And then you want to make sure
这些距离在一个可接受的范围内
that this separation is within acceptable levels.
再次强调 飞行器监控这些错误
So again, the robots monitor this error
一秒计算一百次控制指令
and calculate the control commands 100 times a second,
这个指令传送到马达
which then translates into motor commands,
一米六百次
600 times a second.
所以这个程序是分散化执行的
So this also has to be done in a decentralized way.
再有 如果你有很多很多飞行器
Again, if you have lots and lots of robots,
要完成集体飞行任务
it’s impossible to coordinate all this information centrally
能足够快地集中协调所有这些信息是不可能的
fast enough in order for the robots to accomplish the task.
加上这些飞行器只能依靠局部的信息来决定做什么动作
Plus, the robots have to base their actions only on local information —
也就是从他们旁边的飞行器中感应
what they sense from their neighbors.
最后
And then finally,
我们坚持认为这些飞行器不知道旁边的飞行器是什么
we insist that the robots be agnostic to who their neighbors are.
也就是匿名飞行
So this is what we call anonymity.
下一个我想给大家展示的是这二十个飞行器的视频
So what I want to show you next is a video of 20 of these little robots,
他们编队成群飞行
flying in formation.
他们监控旁边飞行器位置
They’re monitoring their neighbors’ positions.
他们维持队形
They’re maintaining formation.
队形可以改变
The formations can change.
他们可以在平面上飞
They can be planar formations,
也可以在三维空间飞行
they can be three-dimensional formations.
如你所见
As you can see here,
他们从三维空间到平面空间时会改变队形
they collapse from a three-dimensional formation into planar formation.
飞过障碍物
And to fly through obstacles,
他们再飞行中能适应队形变化
they can adapt the formations on the fly.
我想强调 这些飞行器距离都很近
So again, these robots come really close together.
比如这个群队 八架飞行器
As you can see in this figure-eight flight,
相互距离不过几英寸
they come within inches of each other.
尽管在空气动力学上 这些螺旋桨相互干扰
And despite the aerodynamic interactions with these propeller blades,
它们还是能够维持平稳飞行
they’re able to maintain stable flight.
(掌声)
(Applause)
现在它们会成群飞了
So once you know how to fly in formation,
能让他们合作搬运物品
you can actually pick up objects cooperatively.
这只是展示我们能两倍、三倍、四倍地增加
So this just shows that we can double, triple, quadruple
飞行器的力量
the robots’ strength,
仅仅通过让他们与相邻的飞行器合作 如你所见
by just getting them to team with neighbors, as you can see here.
一个这样做的缺点是当你增加数量
One of the disadvantages of doing that is, as you scale things up —
比如你有大量飞行器搬同一个东西时
so if you have lots of robots carrying the same thing,
你实际上增加了惯性
you’re essentially increasing the inertia,
因此你得付出代价:他们不再灵活了
and therefore you pay a price; they’re not as agile.
但是你可以增加载荷承载量
But you do gain in terms of payload-carrying capacity.
我想告诉你另一个应用,同样在我们实验室里
Another application I want to show you — again, this is in our lab.
由刚刚毕业的Quentin Lindsey所研究的课题
This is work done by Quentin Lindsey, who’s a graduate student.
他的算法基本上告诉这些飞行器
So his algorithm essentially tells these robots
如何自主建立方形结构
how to autonomously build cubic structures
通过使用构架材料
from truss-like elements.
他的算法程序告诉飞行器哪一块捡起
So his algorithm tells the robot what part to pick up,
在什么时候 什么地方放下
when, and where to place it.
你可以看见这个视频里
So in this video you see —
是加速了10,14倍播放速度的
and it’s sped up 10, 14 times —
你可以看见三个不同的框架被这些飞行器建构
you see three different structures being built by these robots.
再次强调 一切都是自动的
And again, everything is autonomous,
Quentin所做的
and all Quentin has to do
只有给了他们他想建筑的设计蓝图
is to give them a blueprint of the design that he wants to build.
所有这里展示的实验
So all these experiments you’ve seen thus far,
所有的演习
all these demonstrations,
都是靠着它们自己的动态检测系统完成
have been done with the help of motion-capture systems.
那么当你离开实验室又会发生什么
So what happens when you leave your lab,
当你去一个真实的世界的时候
and you go outside into the real world?
如果没有GPS怎么办?
And what if there’s no GPS?
飞行器都被装上摄像机
So this robot is actually equipped with a camera,
和一个激光测距仪 一个激光扫描仪
and a laser rangefinder, laser scanner.
它使用这些传感器绘制一张周围环境的地图
And it uses these sensors to build a map of the environment.
这副地图包括很多东西
What that map consists of are features —
比如正门、窗户、人、家具
like doorways, windows, people, furniture —
然后它还会计算自己的位置
and it then figures out where its position is,
相对于这些物体
with respect to the features.
所以这儿没有统一的坐标系统
So there is no global coordinate system.
坐标系统由飞行器而定
The coordinate system is defined based on the robot,
它在哪儿 它看见什么
where it is and what it’s looking at.
它利用周围的物体来导航
And it navigates with respect to those features.
我想给你展示一个小视频
So I want to show you a clip
关于Frank Shen和Nathan Michael教授的算法程序形成过程
of algorithms developed by Frank Shen and Professor Nathan Michael,
展示了飞行器第一次进入一栋建筑
that shows this robot entering a building for the very first time,
在飞行中绘制地图
and creating this map on the fly.
飞行器发现了有什么物体
So the robot then figures out what the features are,
它就加入地图中
it builds the map,
它计算出自己相对于那些物体的距离
it figures out where it is with respect to the features,
一秒一百次估计自己的位置
and then estimates its position 100 times a second,
允许我们使用我之前向你描述过的算法控制程序
allowing us to use the control algorithms that I described to you earlier.
飞行器实际上是被Frank远程控制的
So this robot is actually being commanded remotely by Frank,
但飞行器能计算它自己如何前进
but the robot can also figure out where to go on its own.
所以当我送它进入建筑时
So suppose I were to send this into a building,
我不知道建筑内部是什么样子的
and I had no idea what this building looked like.
我要求飞行器进入
I can ask this robot to go in,
绘制地图
create a map,
然后回来 再告诉我建筑内部的样子
and then come back and tell me what the building looks like.
这里 飞行器不仅仅解决了
So here, the robot is not only solving the problem
如何在地图上从一个地点到另一个地点的问题
of how to go from point A to point B in this map,
而且随时可以计算出到目的地最好的方法
but it’s figuring out what the best point B is at every time.
基本上它知道该去搜索哪里
So essentially it knows where to go
去寻找必要的信息
to look for places that have the least information,
来填满地图
and that’s how it populates this map.
这里我想展示给大家最后一个用途
So I want to leave you with one last application.
当然这个技术有很多很多用途
And there are many applications of this technology.
我是个教授 我们热衷于教育
I’m a professor, and we’re passionate about education.
这样的飞行器其实可以改变我们的小学和中学教育
Robots like this can really change the way we do K-12 education.
我们在南加州
But we’re in Southern California,
离洛杉矶很近
close to Los Angeles,
所以我不得不放点娱乐元素进去
so I have to conclude with something focused on entertainment.
我想给大家看一个音乐视频
I want to conclude with a music video.
我想向你们介绍负责剪辑视频的Alex和Daniel
I want to introduce the creators, Alex and Daniel, who created this video.
(掌声)
(Applause)
在我放视频前
So before I play this video,
我想告诉你 他们为制作这个视频花了三天时间
I want to tell you that they created it in the last three days,
因为主持人Chris给我打了个电话
after getting a call from Chris.
在这个视频中表演的飞行器全是完全自动的
And the robots that play in the video are completely autonomous.
你能看到九个机器人,演奏六种不同乐器
You will see nine robots play six different instruments.
当然了 这是为了今年的TED2012特别制作的
And of course, it’s made exclusively for TED 2012.
请欣赏
Let’s watch.
(空气从阀门逸出的声音)
(Sound of air escaping from valve)
(音乐)
(Music)
(机翼声)
(Whirring sound)
(音乐)
(Music)
(掌声)(欢呼声)
(Applause) (Cheers)

发表评论

译制信息
视频概述

手掌大小的飞行器带来什么奇妙能力?维杰·库马告诉我们小型飞行器的种种应用。从如今的腕带式飞行器,到未来潜水飞行器,上天入地,无所不能!

听录译者

收集自网络

翻译译者

Ljimnn

审核员

赖皮

视频来源

https://www.youtube.com/watch?v=4ErEBkj_3PY

相关推荐