Name: 埃隆-馬斯克在2022年特斯拉人工智能日上展示了一個人形機器人 (Elon Musk reveals a humanoid robot at Tesla AI Day 2022)
Uploaded: 2022-10-03T17:10:17.000Z
Duration: 11 min 35 s

I do want to set some expectations with respect to our optimist robot.

我確實想對我們的樂觀主義機器人設定一些期望。

程度副詞

Um, as as you know, last year, it was just a person in a robot suit.

嗯，如你所知，去年，它只是一個穿著機器人衣服的人。

Uh, but we've not we've come a long way and it's I think, you know, compared to that, it's going to be very impressive.

呃，但我們沒有，我們已經走了很長的路，而且我認為，你知道，與之相比，這將是非常令人印象深刻。

So should we, should we bring out the vote before we do that?

那麼，我們是否應該，是否應該在這樣做之前把投票權拿出來？

We have one one little bonus tip for the day.

我們今天有一個小小的獎勵提示。

This is actually the first time we try this robot without any backup support, cranes, mechanical mechanisms, no cables, nothing.

這實際上是我們第一次嘗試這種機器人，沒有任何備份支持、起重機、機械裝置、沒有電纜，什麼都沒有。

So this is essentially the same full self driving computer that runs in your Tesla cars by the way.

是以，這基本上是與你的特斯拉汽車中運行的完全自動駕駛計算機相同。

This is this is literally the first time the robot has operated without a tether was on stage tonight.

這是這是該機器人第一次在沒有繫繩的情況下運作，今晚在舞臺上。

So the robot can actually do a lot more than we just showed you.

是以，機器人實際上可以做的事情比我們剛才展示的多得多。

We just didn't want it to fall on its face.

Uh, so we'll show you some videos now of the robot doing a bunch of other things.

呃，所以我們現在會給你看一些機器人做其他事情的視頻。

Yeah, we wanted to show a little bit more what we've done over the past few months with about and just walking around and dancing on stage.

是的，我們想更多地展示我們在過去幾個月裡所做的事情，關於和只是在舞臺上走動和跳舞。

Um, just humble beginnings, but you can see the autopilot, neural networks running as is just retrained for the bud directly on that on that new platform.

嗯，只是簡陋的開始，但你可以看到自動駕駛儀，神經網絡的運行，因為是直接在那個新平臺上為花蕾重新訓練。

That's my watering can when you, when you see a rendered view.

這是我的澆水壺，當你，當你看到一個渲染的視圖。

That's that's the robot, that's the that's the world the robot sees.

那是那是機器人，那是機器人看到的世界。

So it's it's very clearly identifying objects like this is the object.

所以它是它是非常清楚地識別物體，比如說這就是物體。

We use the same process as we did for autopilot to collect data and train your networks that we didn't deploy on the robot.

我們使用與自動駕駛相同的程序來收集數據，並訓練你的網絡，我們沒有在機器人上部署。

That's an example that illustrates the upper body a little bit more.

這是一個例子，更多說明了上半身的情況。

That what what you saw was what we call bubble.

你所看到的就是我們所說的保麗龍。

See that's our uh sort of rough development robot using semi off the shelf actuators.

看，那是我們的呃排序粗糙的開發機器人，使用半現成的執行器。

Um But we actually have gone a step further than that already.

嗯，但實際上我們已經比這更進一步了。

這個團隊做了一個令人難以置信的工作。

Um And we actually have an optimist with uh fully Tesla designed and both actuators um battery pack control system everything.

而我們實際上有一個樂觀主義者，呃，完全是特斯拉設計的，兩個執行器呃，電池組控制系統一切。

Um It wasn't quite ready to walk but I think it will walk in a few weeks but we wanted to show you the robot the something that's actually fairly close to what will go into production and um and show you all all the things that can do.

它還沒有準備好行走，但我認為它將在幾周內行走，但我們想向你們展示機器人，它實際上相當接近將投入生產的東西，嗯，向你們展示它能做的所有事情。

是以，讓我們把它提出來，做到這一點。

With the degrees of freedom that we expect to have an optimist production unit one which is the ability to move all the fingers independently move the uh to have the thumb have two degrees of freedom.

我們期望有一個樂觀主義的生產組織、部門的自由度，這是獨立移動所有手指的能力，呃，有拇指有兩個自由度。

Uh So it has opposable thumbs and both left and right hand.

呃，所以它有對立的拇指和左、右手。

So it's able to operate tools and do useful things.

所以它能夠操作工具，做有用的事情。

Our goal is to make a a useful humanoid robot as quickly as possible.

我們的目標是儘快製造出一個有用的仿人機器人。

But optimist is designed to be an extremely capable robot, but made in very high volume, probably ultimately millions of units.

但樂觀主義者被設計成一個能力極強的機器人，但製造量非常大，可能最終會有數百萬臺。

Um, and it is expected to cost much less than a car.

嗯，而且預計它的成本會比汽車低得多。

So, uh, I would say probably less than $20,000.

所以，呃，我想說可能不到2萬美元。

The potential for optimistic is I think appreciated by very few people.

我想，很少有人能體會到樂觀的潛力。

Hey as usual, Tesla demos are coming in hot.

嘿，像往常一樣，特斯拉的演示正在火熱進行中。

So that robot that came out and did the little routine for you guys.

是以，那個出來為你們做小動作的機器人。

We had that within six months built, working on software integration, hardware upgrades over the months since then.

我們在6個月內完成了這個項目，在此後的幾個月裡致力於軟件整合和硬件升級。

But in parallel, we've also been designing the next generation, this one over here.

但與此同時，我們也一直在設計下一代，即這邊的這個。

Obviously there's a lot that's changed since last year, but there's a few things that are still the same.

顯然，自去年以來有很多變化，但有幾件事仍然是一樣的。

You'll notice we still have this really detailed focus on the true human form.

你會注意到我們仍然有這個真正詳細的重點，即真正的人類形式。

So on the screen here, you'll see in orange are actuators, which we'll get to in a little bit and in blue, our electrical system.

是以，在這裡的螢幕上，你會看到橙色的是執行器，我們稍後會講到，藍色的是我們的電氣系統。

This is sized at 2.3 kilowatt hours, which is perfect for about a full day's worth of work.

這款產品的尺寸為2.3千瓦時，非常適合於大約一整天的工作。

What's really unique about this battery pack is it has all of the battery electronics integrated into a single PCB within the pack.

這款電池組真正獨特之處在於，它將所有的電池電子元件集成在電池組內的一塊PCB上。

So going on to sort of, our brain, it's not in the head, but it's pretty close.

是以，繼續排序，我們的大腦，它不在頭部，但它很接近。

Um, also in our torso we have our central computer.

嗯，在我們的軀幹裡也有我們的中央計算機。

So we still are gonna, it's gonna do everything that a human brain does, processing vision data, making split second decisions based on multiple sensory inputs and also communications.

是以，我們仍然要，它要做人腦所做的一切，處理視覺數據，根據多種感官輸入做出瞬間決定，還要進行通信。

So to support communications it's equipped with wireless connectivity as well as audio support.

是以，為了支持通信，它配備了無線連接以及音頻支持。

And then it also has hardware level security features which are important to protect both the robot and the people around the robot.

然後，它還具有硬件級別的安全功能，這對保護機器人和機器人周圍的人都很重要。

So now that we have are sort of core, we're gonna need some limbs on this guy.

是以，現在我們有了某種核心，我們需要在這個傢伙身上找到一些肢體。

Um and we'd love to show you a little bit about our actuators and are fully functional hands as well.

嗯，我們很想向你展示一下我們的執行器，也是功能齊全的手。

So there are many similarities between a car and the robot when it comes to power train design.

是以，在動力系統設計方面，汽車和機器人之間有許多相似之處。

The most important thing that matters here is energy mass and cost.

這裡最重要的是能源品質和成本。

In the particular case, you see a car with two drive units and the drive units are used in order to accelerate the car 0 to 60 MPH time or drive a city drive side while the robot that has 28 actuators.

在特殊情況下，你看到一輛汽車有兩個驅動單元，而驅動單元是為了加速汽車0到60英里/小時的時間或驅動一個城市的驅動側，而機器人即有28個執行器。

And it's not obvious what are the tasks that actuator level.

而且，執行器水準的任務是什麼並不明顯。

So we have tasked that are higher level like walking or climbing stairs or carrying a heavy object which need to be translated into joint into joint specs, the rotary actuator in particular has a mechanical class integrated on the high speed side, angular contact ball bearing and on the high speed side and on the low speed side, cross roller bearing and the year train is a strain wave year.

是以，我們有更高層次的任務，如行走或爬樓梯或攜帶重物，需要轉化為關節的規格，特別是旋轉推杆有一個機械類的集成，在高速側，角接觸球軸承，在高速側和低速側，交叉滾子軸承，年輪是應變波年。

Um there are three integrated sensors here and bespoke permanent magnet machine.

嗯，這裡有三個集成傳感器和定製的永磁機。

So our actuator is able to lift a half ton, 9-foot concert grand piano.

是以，我們的推杆能夠提升半噸重、9英尺高的音樂會大鋼琴。

Our fingers are driven by metallic tendons that are both flexible and strong.

我們的手指由金屬肌腱驅動，既靈活又堅固。

We have the ability to complete wide aperture power grasps while also being optimized for precision gripping of small, thin and delicate objects.

我們有能力完成大口徑的動力抓取，同時也為精確抓取小的、薄的和精緻的物體進行了優化。

Some basic stats about her hands that has six actuators and 11° of freedom.

關於她的手的一些基本數據，它有六個執行器和11個自由度。

It has an in hand controller which drives the fingers and receive sensor feedback reported directly from autopilot to the bots situation.

它有一個手控器，可以驅動手指，並接收由自動駕駛儀直接報告給機器人的傳感器反饋情況。

It's exactly the same occupancy network that we're talking to a little bit more details later with the autopilot team that is now running on the But here in this video, the only thing that changed really is the training data that we have to recollect.

這正是我們稍後與自動駕駛團隊談論的佔用網絡，它現在正在運行。 但在這段視頻中，唯一真正改變的是我們必須回憶的訓練數據。

We're also trying to find ways to improve those occupancy networks um using work made on your radiance fields to get really great volumetric rendering of the butts environments for example, here.

我們也在努力尋找方法來改善這些佔用網絡嗯使用在你的輻射場上所做的工作來獲得真正偉大的體積渲染的屁股環境，例如，這裡。

So machinery that the but might have to interact with.

所以機械，但可能必須與之互動。

So we've been training more neural networks to identify high frequency features, key points within the body, camera streams and track them across frames over time as the boat navigates with its environment.

是以，我們一直在訓練更多的神經網絡來識別高頻特徵，身體內的關鍵點，相機流，並在船與環境的導航過程中跨幀追蹤它們。

And we're using those points to get a better estimate of the butts post and tragic trade within its environment as it's walking.

我們利用這些點來更好地估計屁股的位置和它行走時在環境中的悲慘交易。

And this is a video of the motion control code running in the opposite.

而這是一段運動控制代碼在對面運行的視頻。

Similar simulator showing the evolution of the robots work overtime.

類似的模擬器顯示了機器人工作加班的演變。

So as you can see, we studied quite slowly in april and start accelerating as we run a lot more joints and depart.

是以，正如你所看到的，我們在4月份研究得相當緩慢，隨著我們運行更多的關節和出發，開始加速。

More advanced techniques like arms balancing over the past few months.

在過去的幾個月裡，更多的高級技術，如手臂平衡。

We wanted to manipulate objects while looking as natural as possible um and also get there quickly.

我們希望在操縱物體的同時，看起來儘可能的自然嗯，也能快速達到目的。

So what we've done is we've broken this process down into two steps.

是以，我們所做的是將這一過程分解為兩個步驟。

First is generating a library of natural motion references um or we can call them demonstrations and then we've adapted these motion references online to the current real world situation.

首先是生成一個自然運動參考庫嗯，或者我們可以稱之為示範，然後我們在網上對這些運動參考庫進行了調整，以適應當前的現實世界情況。

So let's say we have a human demonstration of picking up an object.

是以，假設我們有一個人類拿起一個物體的示範。

We can get a motion capture of that demonstration which is visualized right here as a bunch of key frames representing the locations, the hands, the elbows, the torso.

我們可以得到該演示的動作捕捉，它在這裡被可視化為一堆關鍵幀，代表著位置、手、肘部和軀幹。

We can map that to the robot using inverse cinematics.

我們可以使用逆向電影學將其映射到機器人身上。

And if we collect a lot of these now we have a library that we can work with.

如果我們現在收集了很多這樣的東西，我們就有一個可以使用的圖書館。

But a single demonstration is not generalize a ble to the variation in the real world.

但是，一個單一的示範並不能把一個出血量概括為現實世界的變化。

For instance, this would only work for a box in a very particular local location.

例如，這隻對一個非常特殊的地方的盒子有效。

So what we've also done is run these reference trajectories through a trajectory optimization program which solves for where the hand should be, how the robot should balance during, uh, when it needs to adapt the motion to the real world.

是以，我們還做的是通過一個軌跡優化程序來運行這些參考軌跡，該程序解決了手應該在哪裡，機器人應該如何平衡，呃，當它需要適應現實世界的運動。

So for instance, if the box is in this location, then our optimizer we'll create this trajectory instead.

是以，例如，如果盒子在這個位置，那麼我們的優化器我們將創建這個軌跡來代替。

I think the first thing within the next few weeks is to get optimists at least that part with bumble see the other, but prototype you saw earlier and probably beyond.

我認為在未來幾周內的第一件事是讓樂觀主義者至少與大黃蜂的那部分看到其他，但原型你看到早些時候，可能超越。

Um, we're also going to start focusing on the real use case at one of our factories, um, and make this project a reality and change the entire economy.

嗯，我們也將開始關注我們的一個工廠的真實使用案例，嗯，並使這個項目成為現實，改變整個經濟。

All of this was done in barely six or eight months.

所有這些都是在短短的六或八個月內完成的。