DEMO: NABi and ALPHRED (Dennis Hong, UCLA) | TC Sessions Robotics 2018

[@jarekf3083]

That actuator which controls not only movement but also strength is fantastic idea.

[@pyalot]

Jarek F controlling strenght isnt the difficult part, you just use more or less juice. the difficult part is that to get high torque you needed gears, but gearing makes something non compliant and reduces speed. how you get lots of torque and yet avoid high gearing is the mystery part that apparently the bear actuator solves. though i am sceptical of that claim, you cannot cheat electromechanics. if you are willing to use way more power (i.e. be inefficient) you can have a high torque motor with little or no gearing. of course that would be a problematic approach for something supposed to run off batteries (which is probably the reason they didnt).

[@MrGarkin]

Stop this bullshit about "cheating" electromechanics. Any engine has a wide torque and speed range it could be applied to. If you want, you could manipulate this diagram by adding a gear ratio. But there are a tonn of cases when you don't need to, like electric cars are not using transmissions.

[@pyalot]

MrGarkin the cheating electromechanics part is getting high torque, no gearing, low power and high speed in one small package. you can usually get 2 or three of these, not all of them. electric cars are not different.

[@alexyoung6418]

Having a high gear rate doesn't necessarily make it impossible to backdrive an actuator. All you gotta do is spinning the motor backwards when enough resistance is applied to the output side. The trick is having an accurate and quick feedback.

[@pyalot]

Alex Young this is not advisable because you would be fighting motor inertia with high power requirements to prevent your linkage/gearing from breaking under loads your kinematics can not comply to. in any case that is not what they did, you would have heard the motor whine if it was geared.

[@MonkeyKong21]

"humanoid robots are too expensive and complex and they fall over :'( :'( :'( " ...meanwhile, in boston they have an untethered humanoid robot that can backflip and 360 no-scope from a ledge

[@TheMaidenOnes]

But they still fall over, but they do not share those videos.

[@MonkeyKong21]

some humans would fall over in conditions such as being shoved, running on rough or icy outdoor terrain, or endlessly being experimented with to see what it takes to make you fall over. Personally, i'd probably break my neck trying to do a backflip

[@Yusuf13100]

The 360 noscope is so important! In your comment! Luv it! 👌😂

[@josecarlo5825]

spot can get up, atlas can also get up, and i ain't sure of de robot with wheels from boston dynamics can get up but i think it can

[@enchiladaplatter1]

@@MonkeyKong21 it's a software problem, yours just sucks

[@thesteaksaignant]

Come on guys stop the hate ! We know that boston dynamics owns the game but these guys are doing a great job. It's like laughing at Rocket Lab just because SpaceX does better

[@williamlee7119]

Steak Saignant, idk if thats an accurate comparison! rocket lab is pretty cool!

[@sphagtere]

Elon Musk vs Jeff Bezos would be a much better comparison

[@GB3770]

neither of which gets old :))

[@brainmind4070]

@@EngineeringTechnikcom *Funded by DARPA aka the American taxpayers.

[@brainmind4070]

@@EngineeringTechnikcom I said nothing about the value of the program, just that it was funded by DARPA (before Google ever threw their hat in the ring). No need to be defensive. I like Bruno Mars, too.

[@Zoza15]

Well, Boston Dynamics is way ahead of the curve, but i like their ideas too. maybe for different applications..

[@mikstratok]

I beg to differ, these actuators are really impressive. I wouldn't be surprised if Boston dynamics started using them in their robots

[@williamforbes6919]

Mik Wind The actuators that Boston Dynamics uses already offer active compliance, which is the only selling feature these have.

[@lovelycars1]

Morph Verse I was just about to say that how can he forget Boston Dynamics does that mean Boston dynamics.isnt real or.....

[@lovelycars1]

Bainsworth they're buying everything.

[@devrim-oguz]

I believe Boston Dynamics already use force control instead of position control. Otherwise, it wouldn't be possible for their robots to behave like that.

[@지영채-b6x]

a truly innovative actuator !!!!! Nice work!!! Can't wait for it to be used in many different applications~

[@Pyroblack76]

Cant wait till this machine is tossing me off. Anticipation for the next gen robots is unreal. Women will be become obsolete

[@acopernic]

BEAR is a real good idea with a ton of applications.

[@williamforbes6919]

Andre Cabezas And late to the game. Equivalent devices have existed for a decade.

[@acopernic]

Matthew McKellar let's see the price tag and the real application to judge.

[@williamforbes6919]

Andre Cabezas a good specific example of a recent robotics company to do this is RobotiQ, who offer active compliance on their actuator's, but the concept originates with the SCARA arm as designed by NEC back in 1981. A link to RobotiQ talking about active complance is here: www.google.com/amp/s/blog.robotiq.com/bid/69962/How-Do-Industrial-Robots-Achieve-Compliance%3fhs_amp=true#ampshare=blog.robotiq.com/bid/69962/How-Do-Industrial-Robots-Achieve-Compliance I have however actually used a Reis robotics arm before which was programmed using a puppet mode under active compliance.

[@acopernic]

Matthew McKellar thanks for the comment

[@Abricos4440]

Bear bldc motor and magnetic pole stop and encoder ...

[@TommyLikeTom]

I watched this after the boston dynamics videos and I can't stop laughing

[@vertigo2893]

Funny, I also watched it after the boston dynamics video, but Im not laughing at all. I think these guys are on to something. Its simple, cheap, effective and relatively versatile. OTOH, those BD 4 legged robots do look super impressive, but we can only guess what they cost and IMO are a solution looking for a problem.

[@tobifoong8025]

I watched it too after boston dynamics. yeah it looked unsophisticated at first but then I liked the simple ideas here that use the natural properties derived from structural design and materials.

[@jenkem4464]

That natural spring and dampening is pretty fucking cool. Don't underestimate the difficulty of achieving that mechanically.

[@prince87x]

@@vertigo2893 Well I suppose money/budget isn't something that concerns the military :p

[@jasonbaxter3658]

The real thing that will make them rich and famous is those compliant actuators

[@McJaews]

When comparing NABi and other UCLA bots to Boston Dynamics' stuff, keep in mind that these are designed from a goal of being low cost. It's like comparing a luxury car to a bicycle. Both will do the job they're meant to, but with a difference in mechanics and cost. Your Boston Dynamics dog-bot will NOT be cheap.

[@OxOMRK005HOxO]

To be able to construct something that dampens pressure but at the same time can spring load a lot of force In literally the size of a joint for the machine is an incredible feat in robotic engineering. Very proud they’ve created this.. it has now revolutionised the way robots will move. With perfect precision.

[@Dogbertforpresident]

It is easy to see that one day humans will be surrounded by many different types of robots. We are creating a new culture. Exciting!

[@decode110]

Guys, great job.. Not using just gear motors you did a great job..this kind of motion control can give us almost more than 1000 types of motion controls. Very useful for space science. As you shown single arm can be used for in different ways..it's really amazed everyone. Best luck

[@georgemissailidis3160]

1:43 Humanoid robots: "They're too expensive Too complex ...And too expensive"

[@radian2323]

Lol

[@Alexzw92]

I noticed that too

[@LotusStar48]

Basically. The only commercial bot that does all this is a dog robot from Boston and its costs tens of thousands of $$$

[@dagrizzbee303]

It's tough to make a walking bipedal robot if you neglect the importance of the dinamic strength in bipedal feet. That said you guys still have done an amazing job! I personally can't wait to see what else you guys do with that B.E.A.R actuator! Mad probs!

[@WatDoino]

The motion of these robots will actually be revolutionary. It can adapt to any type of environment and change its form/motion for the intended purpose. You put two finger like claws and this could do anything you want or hang from the ceiling and change electric cables.

[@mho...]

great talk, one of the best speakers! and that actuator really is breathtakingly "organic",great job!

[@khs5534]

온라인 수업 듣다가 왔어요!

[@GeneralSzadek]

oh wow, love the BEAR motor! That's perfectly solving the stiff look and behavior of robots.

[@kael13]

I laughed to start with but that actuator is very cool. It’s like taking robotics back to first principles, which I honestly think has been needed.

[@llificy5024]

Gotta say, these guys r killing it. Good job.

[@robertweekes5783]

Their new BEAR actuator is definitely their coolest thing. Changing mechanical properties on the fly can make robots much more nimble, flexible and fast. This can be a breakthrough in robotics

[@DeneSimpson]

Without a doubt! And not just robotics either. Suspension, bumpers(?), aviation (imagine the type of wings you could make), renewable energy, prosthetics, exoskeletons and even teenager proof unslammable doors! The possibilities are endless. I want some

[@Gravitationification]

How are people not cheering for this dude. This is amazing.

[@brianswelding]

So in a few years we'll have Atlas robots equipped with BEAR actuators (and A.I.) frickin running around and jumping 20 feet high doing flips... Robot Olympics!!!

[@GabiN64]

10:00 damn he missed his chance to say: "Let's put a smile on that face!"

[@Jorvanius]

At the moment, this is the best comment in this video.

[@yolo_xenophone]

Congrats! You made the first spider crab robot!

[@jasongooden917]

When that toddler cheered at the robot kicking the ball it was just heart warming.

[@jedimastersterling1]

I wouldn't call it "revolutionary," but it certainly looks promising. Wish you had more than two days to play with it, looking forward to seeing what you do with it in the future.

[@Exen88]

Wow.. the BEAR actuator is quite impressive!

[@AlphaMachina]

I have a Bioloid built into a 22DoF humanoid with inverse kinematics. Many of the robots in that competition were built using the same Dynamixel servos that I use.

[@galaxywolf4895]

The BEAR actuator is a great idea, I am wondering what the difference in power consumption would be vs a traditional motor and could the actuator be scaled smaller.

[@hapliniste]

Are they gonna do prosthesis with the actuator? It could be a good fit. This is also the more elegant alternative to "artificial muscles" and I totally bet on it. Robots were a bit underwhelming but let's see what they do next. I just wonder what it could achieve with the software and controls of let's say Atlas. Anyway we just have to wait and see

[@Joe-zz6oq]

That is the amazing thing about robots. Yeah a general purpose robot is great but we can design them to be application specific. This is amazing!

[@pauljmeyer1]

Highly impressed by the flexible and noiseless control.

[@PerkasaJob]

I say this is better than boston dynamics, because it is definitely cheaper to make, with open loop it is already stable, great job with the actuators, love to see how it will continue to develop

[@jaybyrdcybertruck1082]

Those actuators are gold!

[@joshuahicks7798]

Still prefer Boston Dynamics

[@codewizard5327]

These two companies are in different pages. They are not competing whatsoever -- rather complementing.

[@jonathanolson772]

I agree. Boston Dynamics has mainly a quadruped (SpotMini) and a humanoid (Atlas). Both of those are more conventional ideas for both of those types of robots, and, of course, well refined. This company is experimenting with unconventional types of locomotion which will almost certainly have a wide range of possible use cases in the future.

[@DynastySheep]

at least you get to see different prototypes and these are not so bad.. everyone has to start somewhere right?

[@harryguy76]

That jumping robot with the wheels for feet from Boston Dynamics is amazing.. As for a firefighting robot for ships a robotic arm that runs on a track on the wall or ceiling would seem better then a robot walking around with a hose that gets caught on everything imo. Have a water connection in each corner of the room or where to best allow for best coverage of the room so the robot arm on the track can hook up to water,foam ect to fight the fire... If the tracks were put on all ceilings through out the ship the robotic arm could also be used to move cargo through out the ship through the narrow passage ways that would other wise have to be done by hand....or man lol....like carrying an injury person.

[@g.agenaral6543]

That what I'm saying talk to me when ya beat Boston dynamics

[@jasonwings2967]

I like the simplicity of these robots, the Bear motors are really a breakthrough. Curious to see what will happen when these motors become available to the open source communities

[@Noonenaught]

There's going to be some mind bendingly weird robots in a few years, things we would never have imagined!

[@Wulfcry]

If the robot is a demonstration part of these actuators these guy's are on their way for an actuator business, Just selling them be interresting.

[@sarge27271]

Hmm? I think you guys should get in touch with Boston Dynamics. Their Bipedal robots don't fall down.......

[@username-unknown69]

Whoa... I didn't think like this amazing actuator like this... Last year, I went to your lecture in Ulsan, Korea because l'm a great fan of you and robotics! l'm so exciting what you'll lecture in this sep.22 talking in Ulsan. See you then, professor!

[@qwertyb18]

@2:47 "your left and L I G H T R E G"

[@funnygreenman6813]

Hold up

[@funnygreenman6813]

*R E G*

[@williamcarrigan1082]

engrish

[@levi2408]

Spoonerisms are very common, especially in nervous people.

[@junuhunuproductions]

Lol he isn't a native american, come on :)

[@thefrub]

It's crazy that the drone community is mainly hobbyists in workshops and they've advanced so much in just 10 years, while robotics is scientists in Universities and it seems like not a lot has changed in the last 50 years?

[@suicidalbanananana]

I love how these guys use dynamixels for most of their robots and bscly created a new actuator based on the many tricks a dynamixel has.

[@tresalas6800]

You guys with the BEAR + Boston Dynamics can seriously change robots for the better. If you changed this from a race to a team effort, that will make these robots so much more advanced! Just think about it, Boston Dynamics has robot stability and movement down, but you have the BEAR, and if you put those together, it would be great

[@matthewford937]

All valid and important research! Awesome!

[@glowytheglowbug]

this is impressive! great work!

[@JonaCorrea]

I'm very interested in see how these actuators works. (Internally)

[@esper86]

@1:43 they're too expensive, too complex and too expensive. :D

[@shmookins]

Interesting leg movement. This one and the Google one have very different walking legs to 'conventional' robots I've seen here on youtube. And both Google's and MIT's are very different from each other too.

[@Tovenaar13]

It's like they are back engineering evolution. Very cool!

[@megananoman]

So sick. These are the type of simple elegant designs that will have useful robots massively deployed. (Cheap to manufacture and repair ) global impact alphred and nabi. And BEAR omg the future is here.

[@bjk80b]

Crazy robot and really cool motion

[@sheetalpuri6913]

BEAR is amazing! Can't wait to see more applications of it.

[@AlphaMachina]

That actuator is incredible.

[@etackhelicopter]

"we can't kick the robot dog over anymore" because this act is following it up.....

[@tempname8263]

Great actuators, but robots themselves are still too raw. Looking forward to new developments.

[@fantasticphil3863]

Forget the robots, these actuators are amazing!

[@xaytana]

The bear motor, or whatever the acronym is, is the only interesting thing here. Everything else is way behind the curve, not just in relation to Boston Dynamics, but even hobbyists can create better robots. You need omni-rotational joints and bio-mimicry to have a successful bipedal robot. Boston doesn't yet implement this on Atlas, but even they have a better bipedal than what you've created. Sorry to be so critical, but you're doing what hobbyists have far surpassed, and hobbyists do it with their own money. Like I said, only interesting and new thing here is the bear motor, and I wouldn't be surprised if someone quickly re-engineers it. At least Boston gives us interesting updates, like a jogging Atlas. Can NABi or ALPHRED jog?

[@fantasticphil3863]

You are so right! I've noticed in the past the huge difference between human muscles and motors, but I thought that pnuenematics would be the answer. THIS ACTUATOR IS AMAZING, & QUITE AHEAD OF ITS TIME!

[@Lambda_Ovine]

Well, I'd say the actuators are more than interesting.

[@rahulv8882]

in my opinion Nabi is a great advancement as compared to Boston Robotics' Atlas and family only on one and most dominating parameter - the cost! with the new actuator, they can change the game!

[@Molb0rg]

basically, it is direct drive motor(maybe with some reduction). should be used for ages now for that kind of applications

[@Mighty_Deeds]

i will take this technology any day if it's gonna save me half the price.

[@zeev]

wow. that actuator is amazing. haven't seen anything ever like it on all the robot videos out there. that actuator seems like it's a big big deal ( the robot not so much but who cares ) . very very amazing actuator. wish he went into more detail about the physics and feedback of it.

[@tdlaustralia7791]

Awesome. Lots of potential there

[@toddspeck9415]

So amazing. Such an awesome presentation!

[@heinrichwonders8861]

The backlash on their drive still leaves room for much, much improvement. Genesis LiveDrive is way ahaed of their game.

[@kenneilangelo]

boston dynamic totally beats all biped researches in the world cuz not only atlas walks very well it does parkour too like backflips and one leg jumps in such obstacles

[@JauWeek]

yea and we need robots that make parkour very importan for society by the way i was been sarcastic

[@jimwilliams1536]

Good actuators. Working on any electromuscles?

[@lxc647]

A little late on the robotics game, maybe that actuator can be used for prosthetics instead?

[@index7787]

This is so fantastic. Love the motor especially, would love to mess around with some.

[@magnvss]

Nice concept. If all goes fine and dandy I can shameless see an improved version of this being helpful and modestly popular 20 years ahead. There are always unsuspected hurdles and caveats, of course, but given that these models rely more on their inherent-by-design stabitily and functionality, whith additional improvements they could be something useful, at last. In any case, the more the better (in the early years of automoviles there were many types of cars, even electrics ones, even those that worked with steam engines...) and here we are.

[@falcychead8198]

What would really help robotics research take off would be a standardized "NATO Alphabet" of descriptive words, to facilitate the rapid prototyping of cute clever robot names.

[@Nae_Ayy]

1.they're too unstable 2. they're too slow 3. they're too expensive 4. they're too complex 5. they're too expensive

[@toamaori]

I'm sure henry ford heard the same comments xD

[@heraclitus6100]

Also, they are too expensive.

[@thegoodthebadandtheugly579]

I don't know if they are unstable, slow or complex, but I'm pretty sure they are too expensive.. In fact, they are twice as expensive as they are unstable, slow and complex..

[@autocollimator]

알프레드는 여러곳에 쉽게 적용해 쓸수 있을거 같습니다.

[@williamforbes6919]

The BEAR actuator is extremely old hat, it is cool, but equivalent devices have existed for a decade already

[@Fanaz10]

well touchscreen was also an ancient idea, yet the apple stuck it on the phone and called it revolutional

[@williamforbes6919]

:23 Apple didn't revolutionize touch screen though. They popularised the marked for smartphones. What you are saying is equivalent to "Paint was also an ancient idea, yet Henry Ford stuck it on a car and it was revolutionary"

[@Fanaz10]

Matthew McKellar, I meant that more as a joke than a direct comparison.

[@williamforbes6919]

:23 Forgive me for the misunderstanding then.

[@adisharr]

I don't think that actuator is working quite the same way as a standard feedback loop although I could be wrong.

[@julianeyt4059]

Can't you simulate the wobble effect just by software, using force or amperage sensors

[@renaissanceman5847]

an actuator that can be back driven means that it must also use significant amount of current to maintain its position. Any system that requires constant current draw to maintain even a static position will then require a higher capacity power source which then adds weight which in turn requires more power, a never ending cycle. Its a break through, but very inefficient in terms of power consumption. Better to have feedback sensors to control non back-drive actuators to simulate compliance/strength.

[@sammydistel]

Those BEAR actuators could be really useful for suspension in vehicles. I would love to see this more widespread

[@Gabtube252]

the BEAR actuator is actually really cool

[@JustforFun-cb7bo]

Yes this is amazing it should go in every robot.

[@lambdacalculus3505]

why are so many people comparing these guys to Boston Dynamics?

[@kewalbajaj837]

can the bear motor turn 360 and do continuous rotations ??

[@MrPePeLePuo]

"Left and light reg"

[@Inertia888]

🍪🍪🍪🍪🍪🍪🍪🍪🍪🍪🍪🍪🍪🍪🍪

[@screes620]

You have to carry your robots on and off the stage? Can we get boston dynamics back out?

[@alexyoung6418]

What's the difference between a BEAR and an SEA(Serial Elastic Actuator)?

[@augurcybernaut4785]

Actuators for the win.

[@geniusmodernus]

Now I can see the future of fluid motion animatronics!

[@joannot6706]

Clever, we evolved to be like we are for survival purposes, the robots having different purposes and abilities, we shouldn't make them like us unless there is a reason, (like if we want to make "interaction" robots for instance)

[@suicidalbanananana]

While theres definitely alot to say for this approach, there also is alot to say for the humanoid approach, to avoid typing a wall of text again, just think of it like this, we humans have shaped the world to our humanoid bodies, from stairs too doorhandles too coffeecups too cars too chairs, everything is tailored to a humanoid shape, so thats why alot of companies keep trying to invent a humanoid robot, its what makes 'the most sense', also because you dont wanna have to alter an environment so that a robot can function in it, you want the robot to be able to function in an environment that makes sense for us (that bscly covers the globe already) Dont get me wrong tho, i can definitely see a valid reason to branch off into robots that 'walk' differently, but i do get why so many companies are pushing for humanoid, and i honestly feel like that will still be the design that comes out on top in the end, things like shown in this video could be great for 'simpler' robots only meant to roam certain environments, but multi purpose robots will most likely always be humanoid

[@leecaste]

Another reason is to replace damaged limbs or even get rid of biological bodies some day so we should encourage humanoid robot development

[@xaytana]

Bio-mimicry is in more than just bipedal robots, and in more than just robots in general; the bullet trains in Japan were designed after birds to be more aerodynamic. There's reasons why certain things in nature are the way they are, and that's because evolution made certain parts to be as efficient as possible. Evolution isn't just about survival, it's also about efficiency. Bipedal robots would greatly benefit from omni-directional joints, like our hips.

[@joannot6706]

Sometimes their are reason to make humanoids, sometimes there are none. Same goes for biomimetics. The examples are this robot and boston dynamic's handle (the one with wheels). Some design aren't found in biology, hydraulics aren't found in nature, wheels neither. Evolution is limited by previous iteration, by the way it is formed, and by many other factors. Sometimes shaping autonomous machines like what we see in biology is useful, sometimes it isn't because it isn't the most efficient for the problems we are trying to solve.

[@mrmexicano64]

Is it just me or are the acronyms slides a reference to "Kids Next Door"?

[@solarsynapse]

Where can you buy BEAR actuators and how much do they cost?

[@haephaestus]

They didn't see atlas ?? Boston dynamics is 2 decades ahead

[@Fanaz10]

this

[@mho...]

thats like saying a rocket is decades ahead of a airplane.... different tech for differnt applications!

[@movedaccount9958]

boston dynamics' robots aren't that efficient compared to the making price. UCLA makes useful robots with much less price.

[@jeffreytan2948]

One thing I noticed, they do love their acronyms

[@przemysawchudzik9977]

This is actually pretty damn good.

[@F00dstamp96]

Its all fun and games until somebody puts a spinning blade on top of it.

[@b1ngb0ngk1m]

존경합니다, 홍박사님! :D

[@Stanon85]

it's just a revolution. finally, robots will enter into everyday life! T-800 crying with laughter...

[@kvazimode2]

so... how does it turn?

[@MekazaBitrusty]

I wish I could find more information about the BEAR actuator.

[@dojostarfox4520]

Ok. so how does NABi turn?

[@kickinbackinOC]

BEAR actuator is GREAT! Should be suitable for force feedback, also, like the human system where our nerves tell our brain how to vary the force we use. That way, a robot can be very nimble, pick up an egg with a metal hand, etc.

[@kickinbackinOC]

However, it seems Boston Dynamics has done very well with a bi-ped called Atlas. It runs, jumps, walks on rocks, does flips. And then there's Cassie, who does extremely well with two legs that "go all the way up" and make a robot out of themselves.

[@jf_moreira]

Weird. It seems these guys had a short window for presenting the stuff and rushed it a lot. Not cool.

[@lyomon9981]

Ok these are better Than some of the Boston dynamic ones

[@janlorenzabellana4206]

This is actually a great idea for a different use. You can't laugh at these guys because the more different inventions, the better it is for science. Just until they finalize their product, im sure this cover weak points that the boston robot has. Thats how inventions works.

[@RemydeRuysscher]

Will it blend?

[@sagarwadhai7813]

what is the torque of these servo motor