B1 中級 1 タグ追加 保存
Hello, everyone, and welcome to a another This time recorded episode of Robots Weekly.
This is robots weekly number three.
We have Raquel Villas, also known on the Internet as Rock Pot comes from us More out of the robotic side into the node world.
I made her big debut on the conference, seen at this year's jazz comp US 2013 down Amelia in Florida with a presentation called a i dot Js.
There's a some title to that.
I think it's something like learning to love the nuclear bomb or something.
I don't know.
It was absolutely amazing.
Great time.
The video is out.
We will have it in the show notes and, uh, with me, like always, is Matt Hamasaki Matt saying a little the nice people.
Hello, everyone.
Um, awesome to be here again.
It's great toe.
Have Raquel here.
Especially since I learned a lot from her at her Jess.
Calm presentation.
Especially about her history.
So Yeah, it was awesome.
So why don't we Why don't we get started?
Ah, right now, looking at you.
Gotta tell us a little bit about yourself.
Hi, everyone.
I'm rock back on Twitter for Callen person on, and I am a roboticist turned Notre.
I am have been a practicing roboticists for over a decade.
Basically means I've been building robots since high school, and, uh, and now I play around with node.
I work at sauce Labs in the mobile are indeed team.
And, uh, not much to say about that, except that, you know, we're doing awesome stuff.
Yeah, awesome.
That's that's really cool.
Ah, so in your presentation, you talked a little bit about some of your early days.
How did you get started in robotics?
Uh, so basically I So I started high school or yet started high school.
Want to be a filmmaker?
Believe it or not, I want to be the next Steven Spielberg.
And I was like, Yes, I'm totally gonna build movies in a move to California and win Oscars and things like that.
And then I saw presentation at a local university.
Uh, that basically featured a bunch of different women engineers, and one of them was a mechanical engineer, and she was like, these other engineers were really awesome and stuff, but I got to go this and she, like, whips out this robot and puts it on the table on this.
Like what?
It was like Whoa, You can totally do that.
I didn't even notice that.
I had no idea that you could build robots.
Like I thought that was just something that you were with that, um, And from that point on, it was literally like a light switch.
And I think I think maybe you later that day maybe the next day.
I can't imagine that.
I waited more than 24 hours.
I asked my parents.
It was in November.
It was like for Christmas.
I want Christmas to come early.
And I bought a Lego Mindstorms kit.
Put me out.
That's awesome.
Especially toe have.
Ah, a female role model is in there.
You know, it's not just guys playing with robots, but anyone can.
Yeah, Yeah, it was awesome.
And so I basically started my career with Lego Mindstorms was then eventually worked my way up to everything from robotic vehicles like automated cars to, uh, using industrial robots to do biomechanical research.
Thio, uh, just working on all sorts of things.
Those, uh, those projects did you just did on your own during college, or were they tied with.
I have.
So because of the nature of the robotics industry, Uh, most of those things I was not able to do by myself all the Lego Mindstorms stuff.
I totally did by myself.
Not a problem.
But, um, all the bigger robots had to be with some sort of either industry or college.
So I've done work at NASA, J P L.
I've done work at the University of Duisburg, Essen and Germany.
I've done work, um, at applied minds, which is this little shop in Southern California.
Um, and I've done some stuff that the University of Joe Genera in Italy.
Um, and then, uh, also did a little bit of stuff at M i t.
Lincoln Laboratory.
But that was more outreach than, like, big projects.
Ah, I'm sure there's something I'm just Yeah, sounds pretty varied.
So what kind of what kind of programming interface?
What kind of things did you do programing wise against those robots?
Eso in terms of programming.
So are you asking like, more like languages, or like like actual, like, algorithmic type stuff?
Well, kind of a mixture of both.
That is how did you approach it?
Um, so most The time, especially with something is large, is like a robotic vehicle.
You can't program the whole thing by yourself.
You have be on a team.
Um and so, uh, basically, I've been kind of lucky.
I guess you could call it that.
You have always, like, found a spot on the team and just had some say you own this ticket.
But most of the projects have been other C c++ or python as the main programming languages.
And, um, those make a lot of sense.
It's nice, you know, low level enough room, otherwise that things are nice and fast you threatening while the lock.
Um, but for those projects, I've done things like, uh So I've done some computer vision stuff I've done, but mostly have done half planning.
So if you imagine, Uh eh, you know, you have some world and you want your robot to navigate within that world, uh, figuring out the decision making process to properly navigate from point A to point B.
Uh, that's being it's kind of most my my experience, Um, which is interesting because it really varies depending on the type of robot.
What kind of algorithms you using and stuff like that.
Very cool.
Yeah, it was kind of what I I I wanted to kind of cover both the languages and the algorithms because, yeah, it's it's all very, very complex.
Especially with the move.
It's stuff that I studied back in in my college days as well.
So very, very cool and so kind of two to flash forward.
How did all of this kind of tie in with bringing you to node?
Um uh, so So my my my way.
I got to know it is kind of a special little been Basically, I got tired of doing the academic thing.
Um, it's there's a lot of in terms of academia, it costs a lot of money.
And in order to move up in academia, you have to get more degrees right on.
You have to get research grants and you have to do all that stuff.
And that's what papers after paper after paper.
And I basically got to the point where I was just like, I'm tired of this.
He's gonna work on stuff that's fun.
Whenever I want to work on it and not have to deal with all of this.
Um, so I've got some opportunities to just kind of totally switch gears, do like a startup and just kind of play around with, like, managing people and stuff like that.
I thought maybe I would be happy with that.
Um right then I basically found myself a start brand of money, which happens on then I was just like, um okay, now what?
And I wanted to work.
I was living in Charlotte, North Carolina, the time, and, uh, all my cool friends were working this one company, so I said, Okay, how do I get a job at the company?
And so I talked to the CEO and he said, Learn node.
I said okay.
And so I did.
So I learned Node and then basically was doing this whole note thing for a while.
And then I heard about Chris and Rick.
Uh, Chris Williams.
Rick Waldron, who you probably both know from last week.
You were there.
You didn't see it, uh, and and they were trying to this thing with robots, and I was like, That's so adorable.
It's cute.
And like Rick, have this little, uh oh, a little robot that you presented it at note calm and it was cute.
Wasn't like, really basic object avoidance.
And I was like these air super adorable.
But there's so much more that they could do if we just apply.
It's an extra robotics principles.
And then I was that note cough summer Camp 2012 which is where I met Chris.
And, uh, I was talking to somebody else about like, Well, there's no good Matrix module in N.
P M.
And I was like, Well, I could build that like, really easily.
And I was trying to figure out how do I get into this whole note community because I was like, I know nothing about core and Streams kind of scare me and there's, like, a changing in the next row.
I'm sorry.
That's what I've been dealing with today, but our wheels.
But basically it was It was just like, uh, some was like, Yeah, we need a good matrix module And I was like, Okay, well, I can build that.
So that was my mind Vector and Pam module and then, uh, and then Chris de Ems me.
And it's like I hear you like robots and I was like, Yeah, who doesn't like robots?
And you just like, maybe you should think about speaking about robots like I want to speak.
I don't about robots.
And he was like, No, no, you should speak about robots.
So, uh, but I made a condition that it was a mean condition.
I made I ask you not to speak of any other events up until that point.
It's like a little as you phrased it.
Huge robotics.
Uh, yeah.
No, it was And it was a great presentation.
You came as the organizer of dance company was arguably one of the hardest ones.
Just for because we had, like, a whole town escape and hide from the audience.
And then all of a sudden, it shows up.
We gotta cover with a blanket or run around finding a blanket.
And, uh, you get up there and I'm like, Jimmy and I had to present it up on the screen.
You won this year's award that we don't give out for the most complex presentation, but it was super fun.
I was just like like when I got off the stage, I was just, like, completely, just like Oh, my God, This is amazing.
And it is all the time.
And then the next day, we had no bots day from the first note box event.
And that was just like you could more.
And I was like, All right, that's it.
I'm sold.
Good lots, community.
Let's do this.
Do you wanna give him a quick description of sort of what that was like and how that worked?
You were one of the four core organizer's that started it and sort of got it going.
I know that there we had Daniel Shaw on a bit ago to do the International note Bots Day.
But just so that first event, how that all went, Yeah.
So basically, what happened?
Waas, uh, we had a whole bunch of that.
We had a huge room full of depths.
Who I would say probably about.
I'd say 50% of them had never touched any hard work at all whatsoever on then.
Another 25 to 30% of them, uh, head on Li like, dabbled in it.
And so, like, most people didn't really know about hardware at all.
It certainly never used Java script for for any hardware type stuff.
And, uh uh So thanks to our amazing sponsors, Toby, we had, like, this huge lineup of just all the components you could possibly imagine from servos like regular servos to continuous servos Thio L E D's thio sauntering irons, Ted.
Oh, my I mean, like, I can't even describe all the pieces that we had, but they were just like it just looked like, um, like, eight of fruit exploded in this room and people just kind of came in.
And we're like, Okay, go build stuff.
And they're like, What way had a few bot box tohave a sumo stew about competition And that worked out really well because there are a few people were like, Brava I can build on this as opposed to having to start from scratch.
But there are some people who just had these crazy ideas and just went with them on, and we had Chris quickly.
Come on.
We have, like, six or seven hours for people that just kind of hack on stuff.
I mean, it depends on whether they took a lunch break and I got there on time.
I mean it.
If they got there on time they would have had from 9 a.m. till 6 p.m. You work pretty much free will at that point.
Um, yeah.
So which was really, really nice, as opposed Thio say, like a no calm where we did eight sessions of what essentially turned out to be 20 or 30 minutes each.
People just kind of hack on stuff, uh, which both situations were really valuable, but had a very, very different outcome.
One was just let me introduce you to know Box and Johnny five and had a kind of put together using a bread board to hook up in led in a button.
Whereas it jazz confit waas The sky's the limit.
Here are all the pieces you could possibly need.
Go build something awesome.
Let us know if you need help.
Um, and the things that came out of that, uh, that that day were just incredible.
Blew everybody's minds like we had a rubber band launcher that basically shot down Note copters.
Uh, we have this, like, automated, uh, box that new one.
Like the toilet paper and run out of, like, beautiful.
You need work full of paper.
It was it was like, typed out.
Yeah, I saw the typing one.
Yeah, it was awesome, huh?
So yeah.
Anyway, I could talk with it forever.
Yeah, it's definitely come a long way.
I can remember my first introduction to kind of the serial port of things at the Js Calm Fuck you.
Ah, where they were doing all of the mad attacks on with, uh, let's see.
It was It was you, Chris.
You're up there with, uh, with the, uh uh, look.
See boo, guys.
Yeah, Thanks.
Evil guys.
I I wasn't even up there.
I was up on the balcony with Toby Taylor on the other side who had done all the underpinning work?
Uh, yeah, Schneider.
And you know yet what?
You broke the week of the Wii controller.
Oh, yeah, that's right.
They didn't bring me out to do the Wii controller.
I almost forgot about that video whenever lies on that one.
Yes, that was That was a good little time.
And that was sort of the first time it touched the stage and go from there.
Go full day.
That's kind of crazy.
Then they did the mad rush in the next year and that she was the worst idea of thing.
Tasted horrible.
You don't know what the Russian is.
Don't look it up.
Just please don't go back to more important things as more pertinent to this conversation.
Were you mentioned?
Um, robotic principles.
Now, is that just trying to get these robots to obey the three laws or what do you mean by the robotic principles on?
Hey, um, okay.
When I hear about it principles, I mean, sure, we could talk with three laws.
Um, less about that.
More more about just what is possible in robotics, using basically using math to do more high level things.
Not just really basic if else luke type stuff, but, um, actually making interesting calculations, uh, that will eventually lead to decisions that will render a robot to do more than just like the really simple, like, picking place type stuff.
Uh, like I want to see robots Britain and javascript that do interesting things like, uh, deliver a soda from the refrigerator to you at your desk.
I mean, how great would that be like that?
If you just think about that, it's actually really complicated Problem, right?
You have that this robot, That has to a get your order.
Be, uh, know where the refrigerator is compared to where current its current position.
Get to the refrigerator, open the refrigerator, grab the soda.
Like, actually figure out which soda is up.
Uh, take it out.
Pick the right one.
Take it out of the refrigerator.
Closer refrigerator.
Go and find you.
Getting to you Making sure he doesn't bump into the wall or the other people.
Especially if you moved.
And if you're moving up, right, like you know what, you want it to be able to just come find you.
And I don't know how complex you want You're about to be.
But should it open your soda for you?
Should it just hand it to you?
Should it, like, throw it to you like that?
Relatively simple operation actually requires a lot of work, huh?
And the testing, too, because if it's throwing it while opening it, it's just it's just a fuzzy time.
Yeah, because if it's throwing at Young Mrs Yeah, I see what you did there.
So those principles arm or assertions on conditional logics in a more appropriate and multifaceted situation.
Um, any.
Like, I know that one of the things that I always battle with from setting up robots Coffin also doing the note boxes.
You have an established group that you know, has that same engendered thought as you had initially.
Oh, that's cute.
They're trying to make robots, Um, which is missing sort of what I see.
The best part about it is you get you get a new set of eyes new group that's coming in from the you know, we build cloudy servers and Web applications, and we we've done all this that the other and sure JAVASCRIPT sucks.
But it's a least one of those languages that has gone on past the sea world and views the world in a bigger, broader spread stroke versus the point of manipulation of the sea world and focused on threats collapsing in on each other and say faulting that sort of different or change perspective is something I find interesting to come back in.
One of things that I think a lot of people who might be watching now or or watch later they come into it mainly from the higher level language they have that that glossy perspective of everything is magical.
I could do it, or I'm gonna erupt in flames from a nine volt battery wth e.
One of the things that I've seen go on is that the entrenched or established group are They're doing robots.
They've already built a lot of things.
And then just like what?
Why Why are you rebuilding the wheel?
Why are you going back and doing led way solved?
Led these for all led solving things like roos robotic operating system.
Thank that nature.
Do you wanna you sort of sit in both worlds have played in the world.
Any any thoughts on that sort of domain?
Yeah, absolutely.
So So one thing that you kind of hinted at that I don't know if you were meant to hint at it.
Or maybe I just picked up or whatever.
Um, is this notion of Well, you know, that's cute.
Why bother?
Um and, uh, like, I think I get a really big high.
When people were brand new hardware white and led for the first time on purpose, Right?
Like if you put in, led into the yard win and you hook up yard, we know and like use Johnny five for the first time.
It just kind of starts blinking on its own because it's a firmer thing that doesn't count.
But when, like you actually purposefully turn that led on or use a button Anderson like that, that's such an amazing moment.
And people start to realize, Wait, this isn't a scary as I originally thought it would be, Um, And then you know, what else can I do?
And that's a really fantastic moment.
So that moving from there, uh, why Java script?
Why a Job's Robotics?
There's one thing that's really interesting and especially within node compared to say, hi founder C++.
I talk about multi threading a little bit.
Uh, node is and JavaScript are single threated, which I think presents a really interesting problem.
Uh, there's this.
There's this notion of, uh, everybody's heard.
The If you have a hammer, everything's in nail, right, Um, or if you have a nail, then normally you think I should use a hammer.
But what if you used like the bottom of a shoe?
It could also work.
It might actually even work better, depending what kind of nail it is.
And what you're trying to nail it into.
But most people they see a nail and they think, Well, I should be using a hammer because that's what everyone else always uses.
But I'm using a shoe.
What does that mean?
Is there a Is there a more interesting way of solving this problem using tools that other people use for other things and solve this problem differently but actually come to a solution?
So there are a lot of interesting problems in robotics that air still unsolved or they're really difficult, right?
So we talk about things like sensor fusion, which, in layman's terms, is just the idea of taking inputs for multiple sensors and combining them all into one juristic amount of data are interesting amount of data.
Uh, that tells you the story of where you are, what you're doing, etcetera so you can think about it in human terms, like eyes and ears and nose and mouth like Where am I?
Am I doing your balance?
Touch block being was to combine all that information in tow.
One, uh, piece of information like toe, for one like ST estimation, is really harm, and people use more multiple streams to do it right, So they do it with multi threading.
So you have one threat percent, sir.
Well in node if we don't have one thread percent, sir, haven't we saw this problem, right?
Could we solve this problem in a more interesting and the more fascinating and more in a simpler way than what other people have always been doing.
Right now it's the nail is sensor fusion, and the hammer is multi threaded.
You know, operations.
Well, we don't have a hammer.
We have a shoe.
How do we do that?
So I think at least from a more academic standpoint, I think javascript could be really interesting for that.
At a bare minimum.
That's exactly it's interesting.
They may be completely and totally bunk, but at least it's interesting.
Yesterday, evaluated cools it exactly like like, for example, when some of my research that I've done with within within Microsoft, especially when we were doing some connect development as we're finding that streams based programming was kind of the the right way to go, not node streams, because node streams can't have multiple subscribers.
But if you're talking about having stream compose herbal streams Or you can say this and this until this allows for a lot more flexible kind of operations.
I keep moving.
Keep moving until you find this particular thing is is having a language that or or or library, Which kind of helps you do that?
So, uh, we just have a little bit of time left.
One of things I wanted to make sure to ask you work.
It was, uh, wear next.
We asked this question to Rick last time, and we asked it to Daniel Shaw all the other time.
I think it's a good sort of anything.
We're just We're doing this all by the seat of our pants.
So there's that, uh, Where do you see this evolving to next?
Where would you like it to see?
As as a first sort of painting A picture.
Where would you like to see it go?
And then where do you actually see it going?
Cause those two sometimes disjointed, but, um, be interesting.
Get your take on.
Where do you say God?
Um, so I think, uh, I definitely see more adoption across just the developer culture.
I think more people are gonna start dabbling in hardware.
Uh, mostly because they think we're becoming We're coming full circle in terms of like, I think e way run in cycles like software, the cloud, etcetera.
It's time is kind of nothing at all.
The problems are solved, not saying that's done.
But I think that software and has come to a point where now we can go back to hard road, right?
Because for a long time, carr hardware was our like, limiting factor.
We couldn't go.
We couldn't do interesting things with software.
And so we had the hardware, uh, ready.
And the hunter is there in terms of my cloud computing, et cetera, et cetera.
Uh, so now the software has to eat has to be some place, and I think we're at the point in software.
We're now we can start doing things with hardware again.
It'll be very interesting to see what happens, whether that happens in job switch or something else.
I'm not entirely sure, but I do think that war software developers are gonna start playing with hardware more.
Um, and and I think that will be very interesting.
Uh, I am noticing that there are a lot of people who want to get started who don't know how to have started.
Uh, fortunately, there's a group of us in the note in the note bots community who are about to do something about that.
So stay tuned.
Um, way have a whole batch of Google affects your home.
It may take a perfect, uh, she's laughing because in the know lots planning sessions, I went ballistic and found all of these crazy effects in here so you can do it again.
T me up, huh?
Oh, I know how to say, like we in the note bots community.
We're doing something about it.
Uh, okay, you're timing and your instructions, but eso And thanks to things like robots come off and international note Bots Day and all of these things, all of which I am super, super pumped to be a part of, um, those are all helping developers get to that level of bringing more harder into their lives and playing around with it on things like that.
Uh, that's what I see for the future.
I personally will continue adding more levels of interesting robotic stuff as examples so that people can kind of play around with it.
I'll be a telex dress in about six weeks, I guess, uh, talking about computer vision and, uh, and identifying things.
Um, you know, target for that open CV.
Because I know that in the note copper community, That's generally when someone says computer vision, the first thing the marine is open CV is that sort of the basis of it, or you go down a different path?
Right now, I'm actually attempting to do it without open CV.
Um, I have been trying to get ahold of Peter Brayden, the guy behind the note version or the node wrapper for open CV.
But he hasn't returned my calls yet.
Listen, yeah, um, so right now I'm completely avoiding open CB and seeing what I could do without it.
Uh, there's an interesting library that I'm playing around with, Uh, and all of it has to be done on Google chrome canary and using, uh, yeah, using interesting, perhaps stuff.
So, uh, super exciting.
But, uh, yeah, I can't wait for Alex.
J asked to kind of show up, but I've been working on as and, uh, yeah, yeah.
What's your most exciting like that in the open source, not about stuff.
I'm always intrigued to find who's doing what.
So sort of the final question was mad.
You got something?
Uh, without going 30.
It was What is your most, uh, Well, what project strikes your fancy the most right Now, um, in the open source community.
Well, what is it that you're sort of digging into?
It sounds like, uh, the open TV rappers, But outside of that domain, outside of computer vision, what sort of what project is interesting to you?
Um I absolutely love watching Johnny five grow.
Uh, just because not only in my friends with Rick, but just because I love seeing people just kind of turned tuning and turn onto it.
And it's a great low level starter, you know, library, uh, for things.
Um, otherwise, I'm really just kind of playing around with stuff.
I mean, I be remiss to say, like, you know, there's some interesting stuff happening here at sauce labs.
Um, happy, um, for mobile testing.
One thing that I will eventually be doing is bringing in robots into the mix for automated testing, which is super cool because a lot of times weeping about robots is doing kind of silly Millie.
Their frilly type stuff.
Um, and I wanna have robots that actually do stuff that's useful and worth your time is a really interesting project.
Um, but otherwise I don't really pay too much attention, Toe.
Like, I just look at the enormity of all of the projects.
I don't focus on any one project in particular, but I do love.
I like to watch Twitter and see what people tweet about In terms of egg I built.
This is amazing.
Are remember whoever was tweeting those she's going all that's cute.
No, no, I gotta sound.
Clip that.
Whenever we see a note about street, it's gonna say that, but yeah, I think a man was going questions, I think.
No, I think that's it for this week.
So, uh, thanks for very much for coming.
Thank you guys.
So much for doing this.
I was like, I need a weekly ish type thing that has robots.
It is hopefully recorded.
Yes, exactly.
But I think the intent of this yes, to bring the software people the hardware and bring the hardware people to software.
So for ah, reaching at least five people.
I'm satisfied.
It's definitely more than five.
We have a viewer count.
Okay, Okay.
Forward touches me.
I all right.
Thank you very much for joining us tonight.
Thank you.
See you next week.
See you next week.
All right.
Thank you.


RobotsWeekly #3 -- Raquel Vélez and NodeRobotics

林宜悉 2020 年 3 月 30 日 に公開
  1. 1. クリック一つで単語を検索


  2. 2. リピート機能


  3. 3. ショートカット


  4. 4. 字幕の表示/非表示


  5. 5. 動画をブログ等でシェア


  6. 6. 全画面再生


  1. クイズ付き動画


  1. クリックしてメモを表示

  1. UrbanDictionary 俚語字典整合查詢。一般字典查詢不到你滿意的解譯,不妨使用「俚語字典」,或許會讓你有滿意的答案喔