Name: 人型ロボットはここまで來た | WIRED.jp (人型ロボットはここまで来た | WIRED.jp)
Uploaded: 2022-06-21T17:45:54.000Z
Duration: 7 min 36 s

情態副詞

It's clearly robotic and also, well very human like the uncanny valley when something artificial triggers a feeling which can go from super cute, super creepy happens in an instant.

這顯然是機器人的，也是非常人性化的，就像不可思議的谷底，當人工的東西觸發了一種感覺，可以從超級可愛，超級令人毛骨悚然的瞬間發生。

But there are actually two in candy valleys, one static and one for motion.

但實際上在糖果谷有兩個，一個是靜態的，一個是運動的。

And this team of designers and roboticists think they figured out how to avoid the fall.

而這個由設計師和機器人專家組成的團隊認為他們想出瞭如何避免墜落的辦法。

We have this really strong notion of biological motion.

我們有這種非常強烈的生物運動的概念。

So you very, very rarely mistake a machine for a living thing.

所以你很少會把機器誤認為生物。

So what is it about the way that a machine moves that will make you believe in it as if it's a live wire talk with Will Jackson founder and Ceo of engineered arts to understand how they design prototype and test their humanoid robots, cracking the code of dynamic movement has been a continuous journey of prototypes and designs for engineered arts.

那麼，是什麼讓機器的運動方式讓你相信它就像一條活生生的電線，與威爾-傑克遜交談，瞭解他們如何設計原型和測試他們的仿人機器人，破解動態運動的密碼一直是工程藝術的原型和設計的持續旅程。

Amica is their latest instead of being sort of photo realistic.

阿米卡是他們最新的，而不是那種照片式的現實。

We're more concerned about what is the essence of being human.

我們更關心什麼是做人的本質。

And it's not just visual, it's this dynamic data, it's this movement data.

而且它不僅僅是視覺上的，它是這種動態數據，它是這種運動數據。

如果你接近了，應該發生什麼。

Um, so we have uh changing the expression so the focus of the eyes comes in tighter.

嗯，所以我們有呃改變表情，使眼睛的焦點更緊。

I think we went through 10 different attempts of getting a smile and it was just trying to get the crease in the, in the skin when when the robot smiles.

我想我們經歷了10次不同的微笑嘗試，它只是試圖在機器人微笑時，在皮膚上獲得摺痕。

還有眉毛中的那些皺紋也是如此。

So we ended up having to add various fibers and parts that with tension the skin inside.

是以，我們最終不得不添加各種纖維和部件，使其與裡面的皮膚產生張力。

Part of what the programmers here are doing is they're studying people and looking at how people behave and how can we can mimic these kind of behaviors?

這裡的程序員所做的部分工作是他們在研究人，看人的行為方式，以及我們如何能夠模仿這種行為？

So if the robot behaves and the way you expect and it makes expressions that you understand, you can do a lot of powerful communication with that.

是以，如果機器人的行為和你期望的方式，以及它做出你能理解的表達，你就可以用它做很多強大的交流。

You know, you go to the cinema and you don't look at the screen and say, oh look some red pixels and some blue pixels you immerse into it and you become part of that story.

你知道，你去電影院，你不會看著螢幕說，哦，看一些紅色的像素和一些藍色的像素，你沉浸在其中，你成為故事的一部分。

What we're doing is taking that kind of willing suspension of disbelief and putting it into the physical world.

我們正在做的是採取這種自願懸置不信的做法，並將其放入物理世界。

America is a communication robot built to entertain and educate engineered arts, focuses on designing the gestures.

美國是一個為娛樂和教育工程藝術而建造的通信機器人，專注於設計手勢。

Look and movement of the robotic systems.

They leave the Ai and what's being said to the universities and companies that buy them each unit takes about 12 to 16 weeks to make and cost between 125 $100,000.

他們把艾和所說的東西留給購買它們的大學和公司，每個組織、部門大約需要12到16周的時間來製作，成本在125-10萬美元之間。

The first thing you learn is that you're never going to match human.

你學到的第一件事是，你永遠不可能與人類相匹配。

So the first thing you learn is how to deal with disappointment.

是以，你學習的第一件事是如何處理失望。

And so we were trying to mimic the way that human vertebrae worked and then these red cords here are like the muscles, there were six motors on this as well.

是以，我們試圖模仿人類脊椎的工作方式，然後這些紅色的繩索在這裡就像肌肉，這上面也有六個馬達。

So enormously big and bulky problem is that we have these incredibly effective actuators.

是以，巨大的龐大問題是，我們有這些令人難以置信的有效的執行器。

Human muscle has this huge power density and very, very fine force control.

人類的肌肉具有這種巨大的力量密度和非常、非常精細的力量控制。

This one in front of me is a version two.

So one paper this looked kind of cool and I drew it all out and I came back into work and I thought with a mechanical design team and said hey guys can you turn this into reality?

是以，有一份文件看起來有點酷，我把它全部畫了出來，我回到工作中，我和機械設計團隊一起思考，並說嘿，夥計們，你們能把這個變成現實嗎？

And then we realized there was so much friction in all of this complexity in all of this kind of cable drive in here that just all the power of the motors was lost in this drive.

然後我們意識到，在所有這些複雜的電纜驅動中，有如此多的摩擦，以至於在這個驅動中失去了電機的所有動力。

So I spent maybe three or four months just working on this and then you just at the end of it, it's just like, wow doesn't work at all.

所以我可能花了三四個月的時間來做這件事，然後你在結束的時候，它就像，哇，一點都不工作。

I think we actually went through seven different versions by the time we got to this when we dropped the idea of having seven vertebrae because we realized you couldn't actually see it.

我想我們實際上經歷了七個不同的版本，到了這個時候，我們放棄了有七個椎體的想法，因為我們意識到你實際上無法看到它。

But human movement isn't just affected by how our bones, joints and muscles move together, but also by the movement of our skin.

但是，人類的運動不僅受到我們的骨骼、關節和肌肉如何共同運動的影響，而且還受到我們皮膚運動的影響。

You have 45 I think it is different facial action units that you can do.

你有45個我認為是不同的面部動作單元，你可以做。

So it's everything from a little tiny muscle pool in here to your inner browse, middle brows out of browse mouth corners, how do they transition from into a smile and then back into a neutral expression.

是以，從這裡的一個小的微小的肌肉池到你的內部瀏覽，中間的眉毛出瀏覽嘴角，他們如何從過渡到微笑，然後回到中性的表情，這就是一切。

So we did a lot of work actually studying human faces to these faces are made of silicone.

是以，我們做了大量的工作，實際上是研究人臉，以這些臉是由硅膠製成的。

So we end up having to put all kinds of fibers and fabric and things inside that gave it some more strength.

是以，我們最終不得不在裡面放上各種纖維和織物之類的東西，給它增加一些強度。

The head is actually used up with the motors for driving the face and then we try to make the skin as Then as we can so it goes down to about one thick.

頭部實際上已經用完了驅動面部的馬達，然後我們儘量使皮膚變得更厚，所以它可以下降到大約一個厚度。

Their models are all made in house allowing ideas to go quickly from design to production.

他們的模型都是在公司內部製造的，使想法能夠快速從設計到生產。

They categorize and record every part they make if something doesn't work now it may work later.

他們對自己製作的每一個零件進行分類和記錄，如果有些東西現在不能用，以後可能會用。

America is built on the years of learning and design from earlier models.

美國是建立在多年來對早期模型的學習和設計之上的。

So robo thespian was our first kind of robot that was a hybrid pneumatic electric robot which is quite an unusual design.

是以，robo thespian是我們的第一種機器人，它是一個混合氣動電動機器人，這是一個相當不尋常的設計。

We were really interested how can we make a compliant robot?

我們真的很感興趣，我們怎樣才能做出一個符合要求的機器人？

How can we make a robot that's safe around people?

我們如何才能製造一個在人周圍安全的機器人？

We did that with the early ones by using pneumatic, so air is compressible whereas highly geared motors, you can't really back drive them.

我們在早期的產品中使用了氣動，所以空氣是可壓縮的，而高度齒輪電機，你不能真正地反驅動它們。

You'll notice some of the motion is quite jerky, it's a little bit clunky.

你會發現有些動作相當生硬，有點笨拙。

There is actually a very small compressor inside America which is used for the arm balance mechanism.

實際上，美國內部有一個非常小的壓縮機，用於手臂平衡機制。

So the air is used more like a spring within the system.

是以，空氣的使用更像是系統內的一個彈簧。

This is definitely something we discovered with robo thespian, you know, 15 year old robot that using this hybrid pneumatic electric approach could be really efficient.

這絕對是我們在機器人劇團發現的東西，你知道，15歲的機器人，使用這種混合氣動電動方法可能真的很有效。

He's a Mesmer type robot which means he's got natural skin actually touching his face.

他是一個Mesmer類型的機器人，這意味著他有天然的皮膚實際接觸到他的臉。

Every little piece of hair on his face is inserted manually.

他臉上的每一小塊頭髮都是手動插入的。

Even his wobbly little he's gonna punch me.

甚至他那搖搖欲墜的小他要給我一拳。

It's a very very different aesthetic to America.

這是一種與美國非常非常不同的審美。

So amica we wanted to make robot as robot.

是以，阿米卡我們想把機器人作為機器人。

No pretense that it's human and we were more concerned about the movement with the measurement type robots they made to have a more human look.

沒有假裝它是人類，我們更關心的是他們製作的測量型機器人的運動，以使其具有更多的人類外觀。

So we've got this soft skin and you see all these wrinkles on the hands.

是以，我們有這種柔軟的皮膚，你看到手上的所有這些皺紋。

Mm ca於2000年12月首次亮相，21日。

The team is still fine tuning its design and functionality with plans to perfect its hands dexterity.

該團隊仍在對其設計和功能進行微調，計劃完善其雙手的靈活性。

 So all these subtle things come and kind of mean that thing doesn't work. 

是以，所有這些微妙的事情都來了，有點意味著那件事不起作用了。