English-Video.net comment policy

The comment field is common to all languages

Let's write in your language and use "Google Translate" together

Please refer to informative community guidelines on TED.com

TEDxNijmegen

Peter van Manen: Better baby care -- thanks to Formula 1

彼得·范-梅南: 如何用F1方程式赛车帮助……婴儿?

Filmed
Views 794,639

在F1方程式比赛中,一辆赛车会传送数以千万的数据到它的车库(控制中心)进行实时的数据分析并接收反馈。那么是否可以把赛车使用的能够处理如此海量数据的系统应用到别的地方,比如……儿童医院?彼得·范梅南向我们展示了他们的成果。

- Electronic systems expert
Peter van Manen is the Managing Director of McLaren Electronics, which provides data systems to major motorsports series. Full bio

Motor发动机 racing赛跑 is a funny滑稽 old business商业.
赛车是一项充满乐趣的经典运动。
00:12
We make a new car汽车 every一切 year,
每年我们都造出新的赛车,
00:14
and then we spend the rest休息 of the season季节
然后在这一年剩下的时间中
00:16
trying to understand理解 what it is we've我们已经 built内置
分析和理解新赛车的特性,
00:19
to make it better, to make it faster更快.
改进它,让它更快。
00:21
And then the next下一个 year, we start开始 again.
然后第二年,重新开始这个过程。
00:25
Now, the car汽车 you see in front面前 of you is quite相当 complicated复杂.
你现在看到的这辆车是非常复杂的。
00:28
The chassis机壳 is made制作 up of about 11,000 components组件,
光底盘就用了超过一万个零件,
00:32
the engine发动机 another另一个 6,000,
引擎的零件数量超过六千个,
00:36
the electronics电子产品 about eight and a half thousand.
电控部分有大概八千五百个组件。
00:38
So there's about 25,000 things there that can go wrong错误.
总零件数超过2.5万个,每个零件保证不能出错。
00:41
So motor发动机 racing赛跑 is very much about attention注意 to detail详情.
所以赛车运动需要非常注重细节。
00:46
The other thing about Formula 1 in particular特定
F1方程式赛车的另一个特别的地方,
00:51
is we're always changing改变 the car汽车.
是我们一直在更新和改装赛车。
00:54
We're always trying to make it faster更快.
我们持续性的改进它使它更快。
00:56
So every一切 two weeks, we will be making制造
所以平均每两周时间,
00:58
about 5,000 new components组件 to fit适合 to the car汽车.
我们就会替换掉车内大约5000个零件。
01:01
Five to 10 percent百分 of the race种族 car汽车
赛车中大约5-10%的零部件
01:05
will be different不同 every一切 two weeks of the year.
在一年中每两周就会被更换一次。
01:08
So how do we do that?
那么我们是怎么做到的?
01:11
Well, we start开始 our life with the racing赛跑 car汽车.
我们的改进过程从比赛开始。
01:14
We have a lot of sensors传感器 on the car汽车 to measure测量 things.
我们用大量的传感器来记录车辆运行状态。
01:17
On the race种族 car汽车 in front面前 of you here
你们面前的这辆车在比赛时,
01:21
there are about 120 sensors传感器 when it goes into a race种族.
会携带大约120个不同的传感器。
01:23
It's measuring测量 all sorts排序 of things around the car汽车.
它们记录赛车运行时的所有数据。
01:26
That data数据 is logged记录. We're logging记录 about
数据被记录并保存。
01:30
500 different不同 parameters参数 within the data数据 systems系统,
我们的数据系统包含约500个内部参数,
01:32
about 13,000 health健康 parameters参数 and events事件
1.3万个车况信息及事件记录器,
01:36
to say when things are not working加工 the way they should do,
当系统的某些地方工作出现问题时,
01:39
and we're sending发出 that data数据 back to the garage车库
我们将这些数据回传到车库进行分析
01:44
using运用 telemetry遥测 at a rate of two to four megabits兆位 per second第二.
无线传送速率能达到2M到4M每秒
01:47
So during a two-hour两小时 race种族, each car汽车 will be sending发出
所以两个小时的比赛过程后,每辆车发送的数字
01:52
750 million百万 numbers数字.
数据量超过7.5亿。
01:55
That's twice两次 as many许多 numbers数字 as words that each of us
我们每个人一辈子说过的单词总数加起来,
01:57
speaks说话 in a lifetime一生.
还不到其中的一半。
02:00
It's a huge巨大 amount of data数据.
这是很多的数据。
02:02
But it's not enough足够 just to have data数据 and measure测量 it.
但是仅仅测量和记录这些数字还不够。
02:05
You need to be able能够 to do something with it.
你需要利用这些数据做出点什么。
02:07
So we've我们已经 spent花费 a lot of time and effort功夫
所以我们花了大量的时间和精力,
02:09
in turning车削 the data数据 into stories故事
赋予这些数据以意义,
02:12
to be able能够 to tell, what's the state of the engine发动机,
使我们能够知道,引擎的状态如何,
02:14
how are the tires轮胎 degrading降解,
轮胎磨损的程度如何,
02:17
what's the situation情况 with fuel汽油 consumption消费?
油耗的情势如何?
02:19
So all of this is taking服用 data数据
我们所做的这一切,
02:23
and turning车削 it into knowledge知识 that we can act法案 upon.
就是将数据转换成能够知道我们工作的知识。
02:26
Okay, so let's have a look at a little bit of data数据.
现在,让我们看看一眼原始数据。
02:29
Let's pick a bit of data数据 from
再看看从一个三个月大的
02:32
another另一个 three-month-old三个月大的 patient患者.
病人身上采集到的一些数据。
02:34
This is a child儿童, and what you're seeing眼看 here is real真实 data数据,
这是个孩子,而你现在看到的都是真实数据,
02:37
and on the far right-hand右手 side,
注意最右边的曲线,
02:41
where everything starts启动 getting得到 a little bit catastrophic灾难性的,
情况开始变得有点糟糕,
02:43
that is the patient患者 going into cardiac心脏的 arrest逮捕.
患者开始出现心跳骤停的症状。
02:46
It was deemed认为 to be an unpredictable不可预料的 event事件.
这被认为是一件无法预先判定的事件。
02:49
This was a heart attack攻击 that no one could see coming未来.
没有人预见到这次心脏病发作。
02:53
But when we look at the information信息 there,
但是当我们仔细看这些曲线,
02:56
we can see that things are starting开始 to become成为
我们能够看到在心跳骤停的5分钟左右,
02:59
a little fuzzy模糊 about five minutes分钟 or so before the cardiac心脏的 arrest逮捕.
记录仪的数据出现了一些征兆。
03:01
We can see small changes变化
我们能够看到心率等数据中
03:05
in things like the heart rate moving移动.
出现的一些细小的改变。
03:07
These were all undetected未被发现 by normal正常 thresholds阈值
这些改变的幅度很小,常规的检测值
03:10
which哪一个 would be applied应用的 to data数据.
无法甄别出这些改变。
03:12
So the question is, why couldn't不能 we see it?
那么问题变成了,为什么我们注意不到这个?
03:15
Was this a predictable可预测 event事件?
这个事件可以用来预测么?
03:18
Can we look more at the patterns模式 in the data数据
是不是我们对于数据中模式的分析越详尽
03:20
to be able能够 to do things better?
就可以预测的越准?
03:23
So this is a child儿童,
这个孩子
03:27
about the same相同 age年龄 as the racing赛跑 car汽车 on stage阶段,
跟台上的这台赛车一样大,
03:29
three months个月 old.
三个月了。
03:33
It's a patient患者 with a heart problem问题.
这个孩子患上了心脏病。
03:34
Now, when you look at some of the data数据 on the screen屏幕 above以上,
现在,看着屏幕上的这些数据,
03:37
things like heart rate, pulse脉冲, oxygen, respiration呼吸 rates利率,
有心率、脉搏、血氧量、呼吸频率,
03:40
they're all unusual异常 for a normal正常 child儿童,
它们跟正常的小孩子相比都存在差异,
03:45
but they're quite相当 normal正常 for the child儿童 there,
但是对于病房的小朋友们来说很正常,
03:48
and so one of the challenges挑战 you have in health健康 care关心 is,
所以医疗诊断的挑战之一,
03:51
how can I look at the patient患者 in front面前 of me,
就是我如何通过观察眼前的病人,
03:55
have something which哪一个 is specific具体 for her,
通过从她身上获取的特有的一些数据,
03:58
and be able能够 to detect检测 when things start开始 to change更改,
能够在情况开始恶化之前,
04:01
when things start开始 to deteriorate恶化?
检测到苗头?
04:04
Because like a racing赛跑 car汽车, any patient患者,
这点上病人与赛车是相似的,
04:06
when things start开始 to go bad, you have a short time
当事情恶化时,你只有很短的时间,
04:09
to make a difference区别.
来避免事态扩大。
04:12
So what we did is we took a data数据 system系统
我们的方法是使用一个数据系统,
04:14
which哪一个 we run every一切 two weeks of the year in Formula 1
F1赛车每两周运行一次的数据系统,
04:17
and we installed安装 it on the hospital醫院 computers电脑
我们把该系统安装在伯明翰儿童医院的
04:20
at Birmingham伯明翰 Children's儿童 Hospital醫院.
内部电脑上。
04:23
We streamed data数据 from the bedside床头 instruments仪器
我们将医院儿童重症监护室中病床周围的设备
04:25
in their pediatric小儿科的 intensive集约 care关心
接入到我们的系统中,
04:27
so that we could both look at the data数据 in real真实 time
这样我们就可以实时的查看到这些数据,
04:30
and, more importantly重要的, to store商店 the data数据
并且更重要的,我们将这些数据长期保存,
04:33
so that we could start开始 to learn学习 from it.
使我们能够从中寻找规律。
04:36
And then, we applied应用的 an application应用 on top最佳
然后,我们在系统上使用了一款软件,
04:39
which哪一个 would allow允许 us to tease out the patterns模式 in the data数据
能够帮助我们实时的将数据中包含的模式
04:44
in real真实 time so we could see what was happening事件,
梳理出来,让我们看到事情发展的过程,
04:47
so we could determine确定 when things started开始 to change更改.
让我们能够确定情势何时开始发生了变化。
04:50
Now, in motor发动机 racing赛跑, we're all a little bit ambitious有雄心,
在赛车比赛中,我们都怀有野心,
04:54
audacious胆大, a little bit arrogant傲慢 sometimes有时,
无所畏惧,有时候还会有些鲁莽,
04:58
so we decided决定 we would also look at the children孩子
所以我们觉得我们应该在孩子被送往医院的路上,
05:00
as they were being存在 transported to intensive集约 care关心.
就开始对他们进行数据采集和分析。
05:04
Why should we wait until直到 they arrived到达 in the hospital醫院
为什么要等到他们到了医院才开始
05:06
before we started开始 to look?
数据采集和分析呢?
05:09
And so we installed安装 a real-time即时的 link链接
于是我们在救护车和医院之间
05:11
between之间 the ambulance救护车 and the hospital醫院,
搭建了一个实时数据传送的链接,
05:14
just using运用 normal正常 3G telephony电话 to send发送 that data数据
采用普通的3G通信网络传送数据,
05:16
so that the ambulance救护车 became成为 an extra额外 bed
现在急救车也变成了移动版的
05:20
in intensive集约 care关心.
重症监护室。
05:23
And then we started开始 looking at the data数据.
然后我们开始分析这些数据。
05:26
So the wiggly蠕动 lines线 at the top最佳, all the colors颜色,
上面的这些弯弯绕绕、五颜六色的线条,
05:30
this is the normal正常 sort分类 of data数据 you would see on a monitor监控 --
跟你在病床监视器上看到过的数据是一样的——
05:32
heart rate, pulse脉冲, oxygen within the blood血液,
心率、脉搏、血氧含量,
05:36
and respiration呼吸.
以及呼吸速率。
05:39
The lines线 on the bottom底部, the blue蓝色 and the red,
现在,底下的两条红色和蓝色的曲线,
05:42
these are the interesting有趣 ones那些.
是我们关心的。
05:45
The red line线 is showing展示 an automated自动化 version
红线表示的是伯明翰儿童医院
05:46
of the early warning警告 score得分了
正在使用的早期预警阈值
05:49
that Birmingham伯明翰 Children's儿童 Hospital醫院 were already已经 running赛跑.
这是(通过数据分析)自动生成的。
05:51
They'd他们会 been running赛跑 that since以来 2008,
从2008年就开始运行了,
05:53
and already已经 have stopped停止 cardiac心脏的 arrests逮捕
已经在医院内成功的阻止了多次
05:56
and distress苦难 within the hospital醫院.
心搏骤停引发的悲剧。
05:58
The blue蓝色 line线 is an indication迹象
蓝色的曲线向我们指示出
06:01
of when patterns模式 start开始 to change更改,
病人的情势何时开始发生变化,
06:03
and immediately立即, before we even started开始
让我们甚至不需要进行临床诊断
06:06
putting in clinical临床 interpretation解释,
就能立即而直观的,
06:08
we can see that the data数据 is speaking请讲 to us.
看到数据自己向我们传达的信息。
06:10
It's telling告诉 us that something is going wrong错误.
数据告诉我们出问题了。
06:13
The plot情节 with the red and the green绿色 blobs斑点,
这些红色和绿色的小球体,
06:16
this is plotting绘制 different不同 components组件
表示的是来自不同样本群体的
06:20
of the data数据 against反对 each other.
同一种类型的数据。
06:23
The green绿色 is us learning学习 what is normal正常 for that child儿童.
我们通过绿色的小球表示的数据来学习“正常状态”
06:25
We call it the cloud of normality常态.
我们称之为“常态云”。
06:29
And when things start开始 to change更改,
然后当情势开始变化,
06:32
when conditions条件 start开始 to deteriorate恶化,
当病情开始恶化,
06:34
we move移动 into the red line线.
数据就跳到了红色的范围。
06:37
There's no rocket火箭 science科学 here.
这并不复杂。
06:39
It is displaying显示 data数据 that exists存在 already已经 in a different不同 way,
我们只是将已经存在的数据用另一种方式呈现,
06:41
to amplify放大 it, to provide提供 cues线索 to the doctors医生,
放大数据的差异,供医护人员诊断,
06:45
to the nurses护士, so they can see what's happening事件.
让他们更容易看出情势的变化。
06:48
In the same相同 way that a good racing赛跑 driver司机
这跟一个优秀的赛车手,
06:51
relies依赖 on cues线索 to decide决定 when to apply应用 the brakes刹车,
依赖于各种线索来决定何时刹车,
06:54
when to turn into a corner,
何时转向是一个道理,
06:58
we need to help our physicians医师 and our nurses护士
我们需要帮助我们的医生和护士
06:59
to see when things are starting开始 to go wrong错误.
在病情恶化的开始提前发现和处理。
07:02
So we have a very ambitious有雄心 program程序.
我们有一个很雄伟的计划。
07:06
We think that the race种族 is on to do something differently不同.
我们在这场比赛中需要打破常规。
07:09
We are thinking思维 big. It's the right thing to do.
我们目标远大,我们在做正确的事。
07:14
We have an approach途径 which哪一个, if it's successful成功,
如果我们的方法是可行的,那么没有任何理由,
07:17
there's no reason原因 why it should stay within a hospital醫院.
将这种方法的应用范围局限在医院内。
07:20
It can go beyond the walls墙壁.
它可以有更广的使用范围。
07:22
With wireless无线 connectivity连接 these days,
无线网络连接在今天已经无处不在,
07:24
there is no reason原因 why patients耐心, doctors医生 and nurses护士
没有理由还要求病人、医生和护士
07:26
always have to be in the same相同 place地点
一定要在同一个时间,出现在
07:30
at the same相同 time.
同一个地点。
07:32
And meanwhile与此同时, we'll take our little three-month-old三个月大的 baby宝宝,
与此同时,我们将继续改进我们的赛车,
07:34
keep taking服用 it to the track跟踪, keeping保持 it safe安全,
确保其车况良好,确保其安全,
07:38
and making制造 it faster更快 and better.
并且使之更快、更好。
07:42
Thank you very much.
谢谢大家。
07:44
(Applause掌声)
(掌声)
07:45
Translated by Psycho Decoder
Reviewed by Maggie Zhu

▲Back to top

About the speaker:

Peter van Manen - Electronic systems expert
Peter van Manen is the Managing Director of McLaren Electronics, which provides data systems to major motorsports series.

Why you should listen

To say that Peter van Manen has a high-speed job would be an understatement. As Managing Director of McLaren Electronics, which provides electronics and data collection software to motorsports events, he and his team work in real time during a race to improve cars on about 500 different parameters. That's about 750 million data points in two hours.

But recently van Manen and his team have been wondering: Why can't the extremely precise and subtle data-collection and analysis systems used in motorsports be applied elsewhere, for the benefit of all? They have applied their systems to ICU units at Birmingham Children's Hospital with real-time analysis that allows them to proactively prevent cardiac arrests. The unit has seen a 25 percent decrease in life-threatening events. And it's just the beginning.

More profile about the speaker
Peter van Manen | Speaker | TED.com