Autonomous Forklift May Eat Up Warehouse Jobs (technologyreview.com) 122
Jamie Condliffe reports via MIT Technology Review: Seegrid, a provider of material-handling equipment, takes the kinds of forklifts that move 8,000-pound loads around warehouses and makes them autonomous. It does that by popping five stereo cameras on top of the vehicles, having a human drive them around to map a space, and then using image recognition systems similar to those in autonomous cars to navigate the facilities. (Unlike autonomous cars that use sensors like radar and lidar, Seegrid can use just cameras, because lighting conditions in warehouses are more consistent than those on the open road.) But while it's easy enough to have a forklift move objects from one side of a factory to another, reliably loading and unloading them poses a bigger challenge. Other robots designed to haul loads like this tend to pick things up from below, rather than spearing pallets with forks. So autonomous forklifts usually require humans to be present during pickup and dropoff to make sure nothing goes wrong. Seegrid's new GP8 Series 6 forklift has been engineered to reverse its forks into pallets, pick them up, and set them down without a human in the loop.
Thanks Seegrid! (Score:4, Insightful)
Another small but vital step in getting a UBI in place in this country. :-)
Re: Thanks Seegrid! (Score:5, Informative)
See grid is late to the game. At my former employer, I was part of a team who helped implement fully autonomous warehousing using human-less forklifts.
A video of them in operation [vimeo.com]
It wasn't about the labor savings. The ROI was far out compared to payroll of forklift drivers. It was the perfect loading of trucks to balance the loads on the trucks, the reduction (practically the elimination) of damaged goods, and the accuracy in knowing where the product is and how much was in stock at all times, with no errors.
Also, with this system, the downtime is spent "housekeeping". We could front the product that has an upcoming scheduled pickup time and get it close to the relevant dock door. This reduced loading times, reducing "accessorial charges" that trucks make you pay if you keep them for over a certain amount of time, and allowed the distribution center to ship more product in a crunch than humans could possibly hope to achieve.
Oh and they turn up for work more consistently, take fewer breaks, and operate at a steady calculable rate, so planning knew how many trucks they could get shipped, emphatically!
Re: (Score:2)
the last one was clearly a setup.
Re: (Score:2, Interesting)
Like SeeGrid, does your solution use stereo cameras?
I've heard bad things about autonomous vehicles basing their navigation on cameras + software processing (a la Tesla):
https://news.ycombinator.com/i... [ycombinator.com]
That is, as compared to Uber, Google and practically everyone else (who use LIDAR and other more 'reliable' sensors).
Re: Thanks Seegrid! (Score:5, Informative)
No. It uses lidar, with mapping. There are 360 degree reflectors around the warehouse. The lidar reflects off these and form. Pseudo star-navigation field for reference points. Filling in a truck is done my counting steps on the wheel drives.
Re: (Score:2)
Ah, swell - thank you.
If I understand correct, the lidar is mounted on the vehicle, not at a fixed spot - right? How do 360 degree reflectors interact with it?
Re: Thanks Seegrid! (Score:2, Interesting)
Seegrid uses LiDAR for safety only, so the robot doesn't hit anything or anybody. Their localization and route following is entirely vision based, with no reflectors or other special infrastructure like lines, fiducials, wires, or bar codes. The vision sytem uses landmarks that are already in the facility, like racking, lights, and support columns. That's Seegrid's real strength, the vision software that Hans Moravec has been working on for decades.
Re: Thanks Seegrid! (Score:4, Informative)
They really don't "reboot". They do lose track of the star map, but they roll forward to reacquainted them. The edit I think is just shotty video editing. The worst thing that would happen is that the forklifts would get into an error condition because it's return path was obscured by debris or a malformed pallet would collapse. Probably happened about once a month, and maintenance would hand clean up the mess.
These units have a hand-control on their back so you could take over them and move them along, if you needed something obscure done. However, we programmed them to do just about EVERYTHING from retrieving raw materials to fetching pallets. The only thing you really had to handle by hand is when a part came for a machine that had to be forked to the machine so it could be installed. They were vastly more reliable than the Linde lifts they replaced, and those were Cadillac quality.
Re: Thanks Seegrid! (Score:4, Insightful)
The reason being is that every year during the xmas peak times the forklift drivers would strike and demand higher wages. That forced them to get in part time workers to at least keep stock moving, of course the strikers don't like that so often there is violence and they have to get private security to protect the part time workers (often school kids on summer break - yeah it's summer here over xmas).
Also theft (or "shrinkage" in retailer talk) is bad enough that it was also a contributing factor in removing humans from the loop. Some of the stock still needs humans, but all the dry goods are automated, and they are looking at automating the rest.
Re: (Score:2)
It wasn't about the labor savings. The ROI was far out compared to payroll of forklift drivers.
The Seegrid solution provides much more favorable ROI - 1 to 2 years - specifically because the devices can be deployed without 'a team who helped implement fully autonomous warehousing'. In fact, the units that just "drop and return" don't require any sort of skilled person at all, a warehouse worker can just walk them around the facility and they 'learn' where they need to go and how to get there. That's a big deal because once your warehouse has a specific role for a unit like that you can just get one,
Re: (Score:2)
I wonder how many people can be impaled on the forks of a forklift while it still is used to carry the spoils of looting?
and when jail / prison is better then liveing on s (Score:2)
and when jail / prison is better then living on street with a lot less rules / paper work to get free food / board / doctors etc under all of the in place welfare systems.
Re: (Score:2)
Yeah, there will be widespread looting and crime well before a UBI happens. OP will probably be murdered by an angry unemployed man.
This is probably when the most egregious privacy violations will begin in the name of public safety. The current liberal mentality which is generally against police overreach and brutality can only exist in a world where most people of means do not fear crime. If the upper middle class in their wealthy suburbs start to fear looting and violence, a police state will soon follow.
I pay about a tenth of the cost of a police officer in property taxes each year, and I have a relatively modest home in an affluent
Re: Thanks Seegrid! (Score:2)
With automation you won't need UBi. You can just get the upcoming VW self driving electric ID van and drive around the country parking for a few days in one paper to let its solar panels charge your battery. If you don't own some land (to setup an indoor automated farm), you can forage for food or eat at the automated soup kitchens or get a fishing robot. As for showers .. hmm how do the wild animals do that? Ok you may have to trade things you foraged with people who know how to make soap. Internet? Mesh
Re: Thanks Seegrid! (Score:1)
Get this into your head. There won't ever be UBI. First off, it's stupid for the same reason that school vouchers are stupid--it will just cause prices to rise to where it's meaningless.
Second, and most importantly, people don't want to pay you to work (and traitors to the human race here celebrate and enable it). How brain dead do you have to be to figure out that they're not gonna pay you to NOT work? You think you're so valuable doing nothing that the world gives a damn about you?
No, the only cure f
Re: (Score:1)
Who decides what level of automation is too much? Using a forklift, for example, takes away jobs from a team of people who could have been hired to lift the object. What about an electric fan? In ancient rome, a human used to have the job of waving a fan to cool people down. Textile making, humans used to be heavily involved in it .. now machines do it. Do we want to tun off those machines?
Whose job is valuable and whose isn't?
Re: (Score:2)
Either UBI, or killbot-powered genocide of the working class...one of those is closer.
Re: (Score:2)
I don't understand where the money for UBI will come from. I don't have many needs other than food (cereal and milk or p&j sandwiches would be just fine), a bed, and a library card. I'd gladly quit my 6 figure salary job to get paid to lay around and read all day every day. I'm sure I'm not the only one. If far fewer people are paying taxes - who will provide the funds for my UBI?
Re: Well, it follows... (Score:2)
Camera-driven robots will need good lights at least as much as humans would, but robots can carry them around. A/C could be eliminated in a lot of the world, but most electronic assemblies are only rated to operate up to 40 (for commercial) or 50 (for industrial) C ambient temperatures. If outside temperatures get close to 50, some site-wide air handling would still be required.
Re: (Score:2)
Many warehouses don't have ACs now. As they switch from fluorescent to LED, they often use motion sensors, so the lights are only on if someone is in the aisle. One solution for SDFLs is put the lights on the FL.
and states will ban SNAP at the self checkout so t (Score:2)
and states will ban SNAP at the self checkout so that Walmart will have to keep real cashiers on the pay roll.
This is obvious hogwash (Score:3, Insightful)
Re: (Score:1)
Re: (Score:2, Insightful)
What are you? A forty year old middle manager envious of the younger kids that surpassed you into VP/director/CEO positions?
These systems are already in use. If XYZ company is incompetent and can't get it to work, they fail. If they do get it to work they either a) get to take home much higher profits or b) reduce prices to compete.
Either situation is good because now you have higher economic output per person. Either everyone buying the product gets it for less and has more money to buy other goods OR t
Re: This is obvious hogwash (Score:3)
Buggy whip makers and other cliches would like to have words with you.
The argument is essentially always that automation will lead to dislocations and role changes, but the humans in the process will be doing more productive or less common work -- managing the production line, or programming the robots (maybe by simply demonstrating the pattern, or entering the pattern on a computer), or installing and repairing robots, or something that humans do better than robots.
Don't be that asshole who claims victory
Re: This is obvious hogwash (Score:4, Insightful)
Buggy whips were not replaced by electrically powered artificial humans who cost less than $40,000 per year per shift.
Re: (Score:2, Insightful)
You're right; they were replaced by nothing, because automobiles made horse-drawn buggies essentially obsolete. What was your point? Mine was that rsilvergun presented such a distorted version of the march-of-progress argument that it's hard to see any good faith behind presenting the distorted version.
Re: This is obvious hogwash (Score:5, Insightful)
No no.. you missed the metaphor...
Humans are not buggy whip makers. They are horses.
Re: This is obvious hogwash (Score:1)
Horses make great glue!
Re: (Score:1)
Re: (Score:3)
No no.. you missed the metaphor...
Humans are not buggy whip makers. They are horses.
Neigh!
Re: (Score:2)
I don't think the economic arguments are necessarily wrong about the macro effects of change.
The problem lies in the clinical language of "dislocation and role changes" of labor, as if labor just gets a slip of paper that reassigns them to another job in a different place. This is a major gloss over the fact that these are real people, often at later stages of careers, who practically can't "just go get another job" doing something completely different.
It's compounded by the fact that the profits from thes
Re: (Score:3)
Maybe you think "dislocations" is a clinical term. I don't. It's merely shorter than writing that a lot of jobs will go away, and people will have to change careers or fields, with a lot of uncertainty and possibly retraining. It is obvious enough what online shopping (exemplified by Amazon.com) is doing to retail; it is not at all obvious what people who used to work in retail should do instead.
There are a lot of hard questions about how to handle advancing automation and radical changes in technology,
Re: (Score:1)
But of course in today's world a majority of the "owners of capital" aren't Daddy Warbucks sitting in their mansions sucking up profit from the work of the little people, it's workers with individual retirement plans that invest in mutual funds that own stock. That makes them the "owners of capital". About half of all workers in the U.S. own stock, either directly or through their retirement plans.
So rants about "owners of capital" taking advantage of the rest of us are pretty hollow.
Re: (Score:2)
The problem lies in the clinical language of "dislocation and role changes" of labor, as if labor just gets a slip of paper that reassigns them to another job in a different place. This is a major gloss over the fact that these are real people, often at later stages of careers, who practically can't "just go get another job" doing something completely different.
This reminds me of a conservative radio piece I heard where the guest suggest that travel agents, who are becoming obsolete, should just become app developers, because that's the hot new thing where they need people.
Re: (Score:3)
skyrocketing productivity has had no negative impact on wages or employment.
Productivity is not "skyrocketing". It has stagnated [qz.com].
When demand for labor goes down it actually _increases_ its value. I know, crazy, right?
Nobody believes that. You are being obtuse. What economists believe (with plenty of evidence) is that rising productivity does NOT reduce demand for labor, it increases it. This is known as Jevon's Paradox [wikipedia.org], but it really isn't a paradox at all. If you are a factory owner, and you are installing machinery that can double the production of each worker, and double your profits from each worker, would you fire half of them, or hire more?
Re: (Score:2)
Is demand for my product expanding at the same time as the amount of product I can supply? If not, then keeping my existing workforce is going to result in a whole bunch of excess inventory that will need to be warehoused.
So, I would fire half of them at first, and then hire them back again when I have a need to be able to produce more.
Or, you know, just upgrade the robots to version 2.0
Re: (Score:2)
Is demand for my product expanding at the same time as the amount of product I can supply?
Historically, that hasn't mattered, becaus productivity improvements happened broadly across the economy. So even if demand for some particular product is fixed, there will be many more that see increased demand as production costs fall, and there will also be new products introduced based on the new technology.
Increased demand for labor in the face of rising productivity is not some ivory tower theory. It is based on historical reality.
Re: (Score:2)
Is demand for my product expanding at the same time as the amount of product I can supply?
You found the demand-side argument! Euer Gegener made the trickle-down (supply-side) argument.
So, I would fire half of them at first, and then hire them back again when I have a need to be able to produce more.
This is why we need welfare.
Also ShanghaiBill got the more-complete argument [slashdot.org], but you're both on the right page. Slashdot is actually doing pretty well in economics this morning.
Re: This is obvious hogwash (Score:2)
And that, ladies and gentlemen, is why we should not analyze the economy as if it were a GDP factory.
Re: (Score:2)
What if ... there's less risk ... because it's less-expensive to make a thing ... and so you can either expand a luxury good into a broader market, or you can undercut your competitors and capture more of the market for massive profits?
How do you think the prices got set in the first place? Do you think consumer demand sets a maximum price and is fixed regardless of what prices at which a product is available?
Re: (Score:3, Insightful)
It's kind of an argument between "knives don't cause wounds" and "any contact with a knife will leave a permanent, bleeding wound and you die!"
Technical progress reduces labor. It boots people out of jobs. This also reduces the cost and, thus, risk of entering and operating in a market. For higher-cost goods, you expand your market, moving luxuries down to commodities. The same pressures that set the price point before (competition, consumer interest) create a new price point. Consumers will tolerate
Re: This is obvious hogwash (Score:1)
Who cares about the supply/demand for labor? As long as the GDP remains constant, there's enough to go around. The problem is figuring out how to distribute it fairly when the robots do all the work.
Re: (Score:1)
Re: (Score:1)
Sad! (Score:2)
Re: (Score:3)
Nope, just videos of hacked forklifts chasing down and forking the boss and the rest of the board.
Re: (Score:3)
Nope, just videos of hacked forklifts chasing down and forking the boss and the rest of the board.
Ah, so you see the good in everything. 8^)
Re: Doubt it (Score:2)
Re: (Score:2)
The more you automate all of this though the less likely you are to have such problems. The robot fork lifts won't try to overfill a staging area as they'd just stop once it is filled. Or maybe you do program them to automatically extend the staging area, at which point they'd know where the boundaries of that area are, while stacking pallets neatly within it.
Re: (Score:2)
Re: (Score:1)
Read upthread of already implemented and successful warehouse automation projects and you'll see humans create more problems (theft, striking at just the perfectly wrong time, violence) than the ones you propose.
The "weak" point becomes the non-automated trucks.
Re: (Score:1)
You're damn right! A robot made to read Slashdot comments would have stopped at "adapatble" but we still understood your badly written comment!
Re: (Score:3)
Re: (Score:1)
And amazon is well down the road towards developing robots that pack things and put them on shelves.
Re: (Score:2)
Saw a show on Dubai airprt. The bagage handling system is awesome
https://m.youtube.com/watch?v=... [youtube.com]
name of the system (Score:2)
I hope they name the autonomous forklift system "Klaus".
Re: (Score:3)
Automation of warehouses is amazing (Score:3)
But I'm having a hard time imagining 8000 pound loads. I mean I only worked with supply chain management for about 20 years and it was amazing to me that automated systems could store boxes in carousels (and retrieve them as well) and even drop boxes of pills into totes for drug stores but I don't think that even the pallets being loaded on trucks weighed anything close to 8000 pounds.
Maybe they were - meat is heavy but even when I worked for a protein provider (otherwise known as an abattoir) a cow only weighs about 1000 pounds (actually less I think especially after being disassembled and put into boxes).
I don't think I ever saw a forklift carrying 8000 pounds. We were usually more concerned about how much space it took up.
Of course the trucks that they were loaded on to carried much more than 8000 pounds. Wake me up when those are automated.
Labor standards were a big issue both for our customers and the unions though. We had engineers who mapped warehouses and determined how much time it should take someone to pick all the product that was being received or shipped out. We calculated the shortest path, determined how much time someone should take to traverse it and how much time it should take for them to pick an item.
Complete automation was always the dream and I'm sure it still is. The fewer human hands that have to touch something in a warehouse, the more efficient it is and the fewer mistakes that will be made - unless us developers totally screw up. (And we sometimes did)
But at least robots don't steal products off the shelves (or do they?)
And for reference I looked up how much a pallet can hold.
https://greenwaypsllc.com/how-... [greenwaypsllc.com]
4700 pounds,but I'm sure most pallets don't actually need to carry anything near that weight.
But forget weight, the automation is the exciting aspect of this, but even in the '90s there were automated picking machines that could go down an aisle in a warehouse and grab pallets off shelves 50 feet in the air.
I'm sure there is some need for pallets that can hold 8000 pound loads - that link I just used shows a pallet of brick for example.but your typical retailer like a grocery store or a drug store or Best Buy isn't shipping things that weigh that much.
A warehouse without people - that is the dream.
Re: (Score:1)
Re: (Score:2)
I was working a dock in Greensboro, NC before college, and had to load rolls of carpet that night. The forklift had a 20ft steel poll sticking out that was slid down the center of the 4ft high roll. I had to sit way back in the seat to get enough weight on the rear tires to steer. The trailer did not line up with the dock, but was about 6in higher. I had to get a running start to get up the slight ramp, and when I hit the top, precariously balanced on the front wheels, the forklift was bucking as hard a
Disaster waiting to happen (Score:2)
A few useful benefits (Score:2)
One good thing that would come out of this is that the autonomous forklift wouldn't use somebody's crates as an alternate set of brakes. You'd also eliminate forklift operators' propensity to practice jousting on crates.
Not news (Score:1)