Introducing Multipick for Stretch | Boston Dynamics
HTML-код
- Опубликовано: 16 окт 2023
- With our multipick functionality, Stretch is unlocking the next level of automated unloading. Stretch can now move multiple boxes with a single swing of the arm. In typical shipping containers filled with thousands of boxes, the robot is hitting significantly higher rates of productivity.
To learn more about multipick with Stretch, visit: bosdyn.co/3PWXaee
To lean more about Stretch, our case handling robot, watch now: • Tackling the Toughest ... 
Наука
Interesting to see BD going into more mature segments where ABB, Kuka etc have a presence.
They're more focused on the software than the actual arm.
Definitely a start. I worked at UPS and I hated unloading the truck. Was such hard continuous work and I hurt my back twice. If this could be automated in a more efficient way, it will definitely be helpful and preferred.
Also worked as an unloader at UPS, and I respectfully disagree. I loved unloading, it became like a game and I truly embraced the "getting paid to workout" mindset. With hindsight, I was mad at myself for moving to the sort aisle for a $1 raise.
Having done similar tasks for several warehouses over the last 20 yrs, and currently doing it for Amazon, I still thoroughly enjoy these types of jobs. I dislike the stereotypes that come with them, "you're just a deadbeat stoner warehouse worker". But I've always enjoyed the actual work itself.
@@richskater hope for you that you'll be able to work cheaper than the robot
@@HolyOllie TRANSLATE (IF YOU WANT):
ir estendendo é a melhor coisa.
1 RUclips
2 OLHAR DIGITAL
3 ROBÔ QUE UTILIZA ANTENAS DE GAFANHOTO
4 KADU OLIVEIRA
@@user-hh2is9kg9j no there will always be jobs. We are actually 5 million workers short in the U.S. it is jus that robots will take the jobs no one wants to work. And let’s be real, who wants to work at UPS? They do so because of the pay but if equal pay is found elsewhere with a different job they wouldn’t care.
And if it ever does get to the point there is no work, the government will just enforce universal based income
@@user-hh2is9kg9jthere are other jobs out there bro. It's called using your brain
By “lifting” using suction, the integrity of the corrugated cardboard limits how much the robot can pick
My thoughts exactly. I just can't get rid of these hands for tentacles yet
I'm also surprised the best method all those engineers could think of was gluing a bunch of suction cups to a robotic arm
@@waamintion5653 "gluing a bunch of suction cups to a robotic arm" You really know what you talk about
Cardboard is a lot stronger than you'd think
Stretch's gripper is what makes it best-in-class for warehouse automation.
When I worked at FedEx, I learned that a frighteningly large number of people thought that one strip of packing tape across the bottom of a big, heavy box would be plenty to secure it until it reached its destination. I'm afraid those wouldn't go well for Stretch until he is fitted with a heavy-duty spatula.
Same thought. These guys from B.D. should go to a real warehouse where, sea containers are loaded or unloaded. This is completely different from what they trying to reproduce here. I worked as loader/unloader for 15 years.
I guess Stretch is more aimed at loading at production sites with multiple similar and standardized boxes. Moving boxes of different size, weights and materials like in classic package delivering scenarios needs a few more iterations.
@@thomasbemme I get that, but that greatly reduces its real-world applications. Basically, it couldn't unload anything that we dealt with at my former job. It would have to be trucks coming out of specific factories that have a limit number of products all in similarly-shaped boxes that don't weight too much.
Most valuable comment right here. Looking forward to seeing a spatula as a flip out appendage on future designs!
@@misfittoytower Agreed. Still waiting to see it split a pallet of beer crates and respect a "Fragile / This side up" label. Once it can handle a truckload of delicious alcohol without spilling a drop, I might be impressed.
"The people who are unloading the trailer [...] are now the people operating and using Stretch."
That has got to be the single dumbest thing, I've heard today.
Who'd buy that robot worth half a million dollars, if it doesn't save money anywhere?
Operating that thing probably only requires one human per 5 machines or something. That means 3 operators + 5 machines could replace 15 whole unloading crews for a day.
It made me laugh. Ah yes, I'll buy this expensive ass robot to do you work but don't worry I'll totally keep paying you to press the start button.
I mean, technological innovation will reduce people working in that sector, but that just means they go somewhere else. The strength of capitalism is in its adaptability to technological advancement. You invent a new machine that can till the soil? You fire all the farmers who did that job until now. They will have to look for a new job in another sector, so I guess they wouldn't be happy, but for society it's a net positive because they have freed workers and can now be moved to other activities. Until some centuries ago some good 80% of people were farmers. As technology progressed and less farmers were needed per capita, people moved to other jobs. The same is happening here. Sucks for them, net positive for everyone else.
@@davidebic Neither of us expressed an anti-tech/automation sentiment, we're mocking the idea that the workers would still be on payroll doing a fraction of the work. It's a hollow attempt to diminish the worries of job displacement.
I do think what corpos say is just sophistry to make people believe adopting technology isn't going to get them fired. But I still do think that it's true that firing them opens up new possibilities. I do not think anybody's life aspiration is to lift boxes in a warehouse. But life is hard, sometimes you have to do things you don't want to make a living. But that's where technology could actually rescue us. Technology is replacing "boring" jobs that up until now people did because "they needed to make a living". And that is good. It should not be stopped.
What should be stopped is unrestrained capitalism in favour of more social democracies. I do not get why it's even allowed for someone to have more than 10 million dollars in their bank account.
Anyway, what I hope will happen is a shift towards more creative jobs or jobs that are intrinsically human jobs. Examples would be entertainers, singers, dancers, psychologists, judges, and plenty more. Or maybe even a society where working isn't mandatory. But we aren't capable of that, yet.
@@keepeetron Nobody wants an ass robot.
Would like to see hiw it deals with boxes of different size and weight and chaotically arranged.
I worked at inbound for the warehouse, used to do robots job and unload trailers onto conveyer belt,and I had never seen such a perfectly loaded trailer 😅
I miss Atlas videos :(
America fascinates me !
Incredible technology and entrepreneurial spirit.
Looks like we just unlocked a sorter upgrade in Factorio!
Literally stack inserter
Must be lazy there huh
Never seen a truck with boxes stacked like that. Everything is stacked and wrapped on skids to ease loading/unloading
Those boxes seem empty and the boxes are all perfectly uniform. Give it a more realistic challenge.
The boxes were, in fact, filled with concrete pavers and foam. And you can see with your own eyes that the boxes are different sizes. 👀
@@GreyKnight7777 There are like 3 different sizes and they are all together in their respective groups. And how do you know the boxes are filled?
@@Overonator I've got an inside track.
@@GreyKnight7777 Just have to take your word for it.
@@Overonator I guess so. But either way, it's objectively true!
Management isn't going to take an unloading associate who was previously breaking back for 10 hours a day $20+/hr and stick him operating a button. They'll fire 9 unloading associates and keep 1 on to monitor 10 trailers. The other 9 go work at Target or McDonalds and miss car payments.
It is becoming increasing clear to me that technological advancements never created value. It just moved value from the many to the few.
well, never was tech gonna solve a social problem
It works on any box ... that is light enough to be picked up by suction onto the side or top ... which is actually very few boxes. A large fraction of boxes must be picked up by supporting the bottom of the box.
Then you only need an additional gripper that grabs the boxes from below, redundant if the box is heavier or the suction holder fails. They could also be released using a revolver that changes the gripper. The weight of the boxes can then also be communicated in advance via QR codes or barcodes, so that the SPS system can plan which boxes with which gripper. Just a thought.
I have suction system to carry 120kg box. But you grab box from top not side
Currently I'm working in a small warehouse, in my department there are 5 ppl working 4 ppl on station A and single on station B. Station A has very light boxes and station B has much heavier (up to 38kg). Currently we switch every week for a full 5 week (about month) cycle. I can tell station B is highly disliked and if station A becomes robotized it would be unsustainable to operate. It's really a sad thing that at first roborts won't replace the worst, but the best part of job
surly the robot will do the heavy lifting isn't that the point?
Doesn't it makes more sense for it to replace station b? It could probably match the speed of a human there, but humans might be faster at station a.
@@Astra2 I would guess the major restriction with a robot like this would be how it picks up boxes. Even if the suction cups are strong enough to pick up something that may be 30kg+ the box itself may not be able to handle it. For humans, we pick up the heavy box from the bottom, which supports both the box and the object inside. I'm sure most people had situations where a box bottom fails on them at some point lol
@@Astra2 You'd be surprised at the dexterity required to pick up a very heavy box, and to support it so the bottom won't fall out if it's improperly packed. It required not just suckers, but the entire arms, hands and palms, and fingers. And amazing dexterity as well.
And even then, humans still mess it up from time to time.
Just add a tray to the bottom of the robot that extends only once the boxes are away from other boxes. Weight problem solved.
“On a case by case basis” - I see what you did there 😏
Looks like the biggest bottleneck is dropping the boxes (with the spacing). If it would drop the first box further down the belt and then the second a bit closer to itself spreading them out it would take less time. It looks like it would have the reach to do it.
at 1:29 it looks like its doing exactly that with the first 2 boxes.. but anyhow, cant wait to see how fast these things work in 1-2 years.
3 arms, more cameras
I think the more easy way would be to just make the transport band move faster.
This could probably handle most of the Amazon boxes pretty well.
Not at all. I've done some warehouse in my time and the boxes were never similar sizes and neatly packed. Many times the trailers were a wreck w/stuff all over place. How would this handle tires, weird shaped boxes, paper bags. For some companies (Amazon)maybe some but certainly not all. UPS trucks were a wreck and they have packages below and the trailers were a mess...
This will be a case where a machine could do it and I'm sure they will get better, but more time will be spent on telling the machine what and how to do in each truck, by the time they do it, a couple of humans would have done it faster and better. This machine only makes sense to a company like Amazon to maximise profit by removing humans. That's great and all for them, but what about the those humans in the workforce that lose their jobs.
@@alltheeasynamesweregone its a waste to have the most intelligent species on earth to move boxes no? Not only that but its really hard and demanding work both physically and mentally to do some of these warehouse jobs and yes off course there is people who love their job but they will find work at smaller companies, that will most likely be less demanding. Sorry but we can't just stop evolving and finding better solutions to problems just because some will lose their job...... It's not so hard to assemble these robots so maby some can do that instead.
Having worked on the warehouse. Box sizes vary greatly. Not to mention a lot of trailers come with plastic containers called totes, which will not be able to be picked up with suction
@@christianhernandez9078 The video description mentions unloading shipping containers specifically. One form of stowage in shipping containers is floor loading, which seems to be the test setup in the video (so no pallets to maximize use of the volume).. It's an obvious choice for automation, since floor loading normally requires loading and unloading each package by hand. Having said all that, it's clear that in different stowage set-ups this robot will become either much less efficient or not usable entirely (in the configuration shown in this video).
I wonder how Stretch would deal with a wet cardboard box.
probably not well. good thing the situations where it is used prevents this from being a problem it has to deal with, unless something has already gone terribly wrong
Can someone explain to me how this robot can handle packages with more weight? Those lightweigt packages are not a problem, but I'm sure the structure of the package would make it impossible to hold only on one side. Whereas the human arms work like a forklift and can handle packages with more weight better I guess 🤔
Edit thanks to @yaki996:
"he [I] meant that the cardboard itself may not be able to withstand such weight when held by the top/side instead of from the bottom"
Stretch can lift up to about 23kg. That's lightweight? I would say it's pretty decent weight. Enough to damage someone's back overtime, if they would have to unload it.
Edit: Also, in one video, stretch also easily carried an entire Spot, which is 32,5 kilograms. But it was able to lift Spot by a flat surface, not paper tho.
I presume they're gonna have them only work on light to medium weight category packages, so there's likely a lot of businesses which still are in need of something like this that would work with heavy or unwieldy stuff.
however I think Boston dynamics with this will still have a lot of potential customers such as amazon with their really high performance demand and all of that
@@superior5005 i think he meant that the cardboard itself may not be able to withstand such weight when held by the top/side instead of from the bottom
@@superior5005 what Strech can lift is not the same as what that suction gripper can lift
@@yaki996 yeah, I get that. Depens. There are good quality and bad quality cardboard boxes. I think that no one even if it won't be handled by robots would pack something heavy into a bad quality cardboard, because in truck or something it can tear. So heavy things should be always sent in good and strong cardboard boxes.
Got a flash frame at 20 seconds in when opening the interview with Grant.
Hi Marques (he always watches every Boston dynamics video)
These guys look really happy putting people out of work.
Sick! They researched stack inserters!
What happens if one of the boxes is much heavier than the others? And suction may not work on lifting from the top because the box is relatively weak and requires lifting from both sides?
A tray that extends only once the suction has slid the box away from the boxes it was stacked on, once tray is extended, you now have the weight problem solved. Also the motion of spinning around and dropping on conveyer is wasteful. Just have a conveyer run underneath the rovotic arm so it just pulls box off then drops the box on conveyer right in front of it and goes right back to getting other boxes. Right now it's doing all the work that a conveyer should be doing when it's wastefully spinning around.
i wonder how much weight is actually in these demo boxes. they need to throw in random weights and couple really heavy ones for a better test
Good Update
GOOD WORKKKKKKKKKKKKK
Looks good, as long as governments tax automation that replaces human workers, so that we can all receive a UBI instead.
I get where folks are coming from when they mention Non-standard size, weakened, or overweight /Heavier boxes that stretch cannot handle. But consider this: boxes can be sorted out ahead of time at a distribution center where the nonstandard parcels can be placed by humans in a different container while the more standardized boxes (which shipping companies might be incentivized to use due to the fact they are machine compatible therefore possibly cheaper at scale to ship) can be loaded and unloaded by Stretch robots on either end. these can come in various sizes based on the robot's capabilities. So, theoretically, the bulk of all shipping can eventually be performed by machine. This is hypothetical, and I'm not familiar with the shipping industry but BD isn't a bunch of hobbyists taking shots in the dark here. They're a professional engineering firm whose job is solving edge-case problems like what has been mentioned in the comments and more! What we're seeing is just the worst version of what is to come. All I'm saying is, don't sleep on these guys, there's almost no task any of us knows how to do they can't build a machine to perform. it really comes down to which is cheaper in the long run.
Looks like a good robot to unpack empty boxes.
What about, say, a 30kg box? Trying to lift that from the side will just rip the side off as the box needs to be picked up from the bottom. And also, this is still far, far to slow. Instead of paying the pickers they will instead be paying the driver to sit in their cab.
The cardboard understander has logged on...
Do you take custom robot orders? I'm especially interested in the A.I. "suction" systems.
"Amazing, i love that" ~ Boss
"Fuck no, fuck that" ~ Warehouse worker, now former
Whats the maximum load that a single suction cup can handle? Seems like those boxes are full of feathers.
Thank you Boston dynamics for not creating a terrifying robot this time
It is beautiful to see how they find new solutions so as not to overexploit humans at work since they also have their many problems and needs.
why the robotarm pick/suck the boxes on top and not on side surface, for less movement and fast working
Awesome
Does Stretch have any experience, so it learns from time to time how it could do it better?
No
Does it have to drop everything?
Boston dynamics + open ai + breakthroughs in nuclear fusion = OH MY GOD ITS THE FUTURE!!!!!!
@@WwZa7 i don't think you have any idea whats going on behind closed doors
@@WwZa7 time will tell and im confident it will be on my side
it puts the boxes back in truck?
When the truck arrives, the packages inside will definitely not be stacked as straight as in this video. I also worked as an unloader and in my experience, you can't open the door without the packages falling towards you. The packages in the container are always messy
Yeah theyre not exactly training this robot on real world situations. It has to adapt to a mess, this is too neat. Programmers thinking in neat lines smh.
Ill buy it
I look forward to the day when Stretch will outperform my younger self. When I unloaded for UPS in my 20s, my unload quota was 1,200 packages per hour, and the boxes were not uniform in size and shape like they are here.
What if the boxes are open or slighty open??
Next feature idea: Topple. I've heard that human unloaders often find a weak-link box near the bottom, remove it like a bad Jenga move, and cause a whole stack of boxes to topple closer to the conveyor.
Wouldn't that damage the product....? Or was it sarcasm...
I work for a shipping company, and we do this. Well, we grab one near the top and pull the stack down, grabbing the bottom one would rarely be possible. We don't do it with fragile stuff, don't worry. It definitely makes it easier, especially for short/tall people. I'm 6'1", so I usually tip a stack over, then get on my knees to put the boxes on the Adjust-O-Veyor.
However, the trucks rarely arrive as cleanly loaded as in your test setup. I have worked for a long time at DHL Germany...and also put packages from the truck trailer on the belt... there they are partly so that if you pull out the one falls a whole tower of packages over. Or they are broken and open because a liquid has leaked.
but can it throw spot across the room?
why one big arm and not 4 little octopus arms doc ock style? 2 on the front 2 on the back, that way when it spins the back side is grabbing while the front is setting?
Reinventing the octopus one tentacle at a time.
What about if the boxs are heavy?
i've done that kind of work. robots can have that job 100%
Agreed
HEY BD,
I just noticed, in this vid there are loose boxes stacked on the floor of the trailer all the way to the ceiling.
NO ONE packs a trailer like that. Nothing is shipped loose. EVERYTHING goes on a pallet & that pallet is
wrapped in clear plastic. The pallets are loaded/unloaded with a forklift or pallet jack. The pallets are
unloaded/loaded in the warehouse. A last mile delivery company ships loose boxes but not simi trailers
or box trucks or seagoing containers. Pallets only.
!
How efficient is it to have to wait to put the next box on the conveyer belt? Have it move its arm as far as it can on the belt (when no box is detected there on it) and drop the first box the furthest away then the second ext. as its pulling the arm back to grab more.🙃
I just saying if you want efficiency get a gamer with an engineering degree (idk know who😂😂but)
Can it suction well with severely dusty boxes? A lot of boxes travel from years on the shelf to years more on other shelves, dust ruins suction cups normally
I work for a shipping company. You have a point! Some stuff is REALLY dusty. I don't think suction cups will work so well
"До чего дошёл прогресс,
До невиданных чудес!..
Вкалывают роботы, а не человек..."
What if the boxes were palletized before loading into the truck, and then unloaded using a forklift or pallet jack? Are trucks commonly packed with individual boxes?
Nope, not at all. I have no clue where BD is going here; this machine is not even remotely close to handling anything in the actual trucking/warehouse scenario.
It drops them from a pretty high height? Fragile items would suffer
Now when the robots take over, they'll already have an efficient way to move our coffins.
might be a little too heavy for this model
Trying to think of a way to improve this and all I can think of is what if the box was taken out of the equation and the products were managed on their own
And it seems like object recognition could make that possible one day hopefully
The box does provide security, strcuture, privacy and protection; but it significantly contributes to waste
I do not see a replacement for the functionality added by the box sadly
What about weight in the box ? .
The bottom should be supported
This is massive!
Would have been cool to have a time lapse of Stretch unloading a full van with some time estimates
Hurry up, robot! We need more stuff! I’m still waiting for my Steam Deck so I have something to do when I visit my grandma dying of cancer in the hospital!!! MORE STUFF!!!!!
Excellent 👍
Having worked at Amazon and UPS, and unloaded trucks myself, it is MUCH faster for humans to do this because:
* You can have boxes of varying sizes
* You can add more humans to multiply the speed
* The contents of the truck could have fallen over and are no longer on a grid
* The contents can be oblong shapes, not boxes
* The objects could vary in weight
They claim this robot is fast, but unless it is moving twice as fast, can't compete with 2 people doing the same task. And I usually see 3 people unloading a truck at a time.
Worked in a warehouse as well and could see the automation coming. I've gone into maintenance and technical instruction instead. The danger here is the slippery slope that this represents. Sure, it can be brushed off now. In 10 years, however..... I feel for the people who's opportunity (like the one I had) to move from poverty into middle class will be gone. This is one more pathway to success that will be killed off.
This would be true if it was a straight swap of humans for robot, but in practice the presence of the robot would require redesigning the system to make the robot more efficient.
None of that really matters. The only thing that matters is that eventually these robots will work well enough to make the balance sheet look better by moving the costs from labour to capital (to buy the robots).
Once that happens it doesn't matter if a human does it 5x as fast, businesses will buy the robot.
Robot doesn’t want any benefits 😂
People aren’t as fast as they appear. Lots of extra movement from extra people doesn’t mean a linear increase in productivity.
So no pallets? And what about really heavy items?
People really think owners are going to pay warehouse workers to press a button...😐
10 unloading dock workers. Corporate fires 9 and keeps 1 to man 10 trailers/robots. The other 9 find part-time seasonal jobs at target or become car-hops at sonic and miss their house payments.
Humans really think they’re valuable lol
@@kiowastew Unfortunately but people need to adapt. If they want better wage with job security, they need to acquire better skills.
@@0xD1CE saw an article in Forbes recently that said 49% of all skills obsolete in two years, something like that.
@@0xD1CEand how do you get better skills? college? oh wait
That's my job! It'll take a lot for robots to do my job, it's not impossible but it'll take a lot. How are robots going to deal with irregular box shapes? Things falling over? Boxes that can't be tipped without spilling (eg, flowers in water)? Boxes ripping/breaking open? What about bags/envelopes? If it's loaded by people, it'll be very hard for robots to unload. A robot could definitely unload a perfectly loaded trailer full of boxes, but it would have a hard time otherwise. Im not against automation, I actually love it, but it'll take a hell of a lot of work for robots to replace unloaders.
What about wet boxes? Trailers can have leaks, and cold boxes will collect condensation. When boxes get wet, they fall apart. If the contents are heavy enough, you'll pick up the box and the bottom of it won't come with!
Wow.
An post mail centre need this DPH/DMC Parkwest.
It's like that inserter upgrade in Factorio.
Can it grab 36Kg?
Stack inserter!
this is pretty useful, but I think robot dancing is more challenge
can stretch move by itself or does someone need to push it further in everytime so it can grab all boxes?
ok. just saw the other video. it actually does move by itself in the container. pretty fucking sick
@@Razilation well....one less person to pay wages to then I guess.....
Stretch will soon ask for my boots my clothes and my motorcycle
well done
You take the same size boots as stretch? Nice.
I would design a robot that picks up multiple humans at a time and the humans with their multiple hands would then pick up boxes. This has the benefit of adding value to our shareholders.
Amazon 2013 - We’ll be delivering small packages such as books, dvds, & cds via drones in the future.
Amazon 2023 - We are now a grocery store selling ever increasing bulk-size items.
Why not move the box over the bot, like a "rainbow" (arch trajectory), instead of turning the base 180⁰?
Or, if the 180⁰ base rotation is reality needed, why not 2 arms working at the same base, but doing opposing functions, at the same time? (One picking while the other is releasing)
Are you not loading on pallets in US?
Many companies don't palletize loads
it still drops boxes from like 4-6 inches. why?
Why not multi-arm? Like opposing arms, or even a round-robin of arms? Shouldn’t that actually multiply the productivity by the same factor - up to 4 arms I suppose?
im sure they considered it but the cost of maintainence would far outpace the cost saved
What about oddly shaped packages?
No Package for you.
Как можно попасть к ва на работу? Я готов даже просто собирать этих роботов
They should use this to my travel box.
But you need another stretch or some sort of bot on the other end to store the cases or place them in inventory. Unloading is great, but where the box is going is also a consideration.
That is called case picking, which theoretically stretch could do at some point
yessir
does the robot come with a monthly subscription that also sells your data?
Only YOUR data.
You should make a mini one to help with washing dishes
Did they mention the maximum weight load the robot can handle?
5 pounds on a full charge is the spec, 7 times.
Lets see these suction cups do a truck full of housewares. There are almost no boxes, but 75lb+ plastic bins, usually and furniture. Neat tech, but only practical in warehouses; which are developing their own realistic solutions.
I'm still surprised nobody has made a gantry style robot yet that could just be built into the very upper section of some box truck or cargo unit. Might take up 20cm or so off the very top (where the rails and robot could be nested), but not every instance of usage would stack that high. In a truck that could also make it self loading/unloading with an extending crane bringing boxes out the back to a delivery person. Nobody would need to climb in or out of there except for cleaning or servicing.
how about the pricing? 1M USD?
Great. Now make it dual-arm so that it can pick up while unloading.
why all your boxes look empty?
Needs 2-3 arms and more eyes so it can pick, plan the next pick, and drop at the same time.
How about an opposite arm that would pick up more multiple boxes while the first one is unloading on the belt.
Never imagined will arnett was a robot engineer xD
inserter stack size research complete
"Dey took 'er jerbs!!"