Robots invade the construction site

A new generation of machines is automating a tech-averse industry.



Read the whole story

 

[philbot01]

I'm surprised we didn't see a mention of FBR, their robot looks very well advanced in the construction of brick homes though the machine isn't yet in commercial production

https://www.fbr.com.au/

 

[Chuckstar]

The construction industry cannot really be painted with a broad brush like that.

Single-family home construction is still made up of a variety of individual tradesmen who often do work much the same way they did decades ago. Improvements in processes over that time have often been limited to things like laser-levels, nail guns and advances in materials (example: engineered wood flooring that is both more durable and easier/cheaper to install than traditional hardwood).

Big construction (high rises, suburban apartment complexes, office buildings, warehouses) have incorporated all kinds of technology. Warehouse floors poured by laser-guided of robotics, RFID tracking of materials, prefab building parts, etc.

I have yet to see a totally paperless construction site, though, but giant blueprint printouts are pretty sturdy and ergonomic, compared to something like maybe carrying around a tablet or laptop. But you do see a lot of things like RFID/bar-code scanners as well as tablets for tracking progress. It’s just that the big blueprints you can unroll and really examine at scale are pretty hard to replicate in something portable.

 

[Fatesrider]

That robot "drywaller" might be ok for open areas like mentioned in the article, but be useless in any apartment complex construction or anywhere that has small rooms.

While this one might not be too impressive, It's a step on the way to more advanced robots. Gotta start somewhere.

It's... not that easy.

Tight spaces, stairwells, crawl spaces, non-flat surfaces (which are all the rage in high-end homes/offices) and other such things are part of the issue.

The biggest issue is that most people are not as large as a stove.

About the size of a kitchen stove, the four-wheeled robot navigates an unfinished building carrying laser scanners and a robotic arm fitted to a vertical platform.

The problem that I see is that you can't really make it much smaller and still do the job. A human can balance by a variety of means to position themselves to apply paste and sand a drywall. A robot has a much harder time doing that, unless it's a human-shaped robot. Then you get into a hell of a lot of OTHER issues regarding balance and mobility while carrying and using the necessary supplies.

It's simpler, and cheaper, to grab a day worker from the Home Depot parking lot and show him how to do the job. Doing the job fast isn't a hallmark of most contractors. The only advantage I see to this creation is for large scale, box-like construction that takes the robot's limitations into account in the design for maximum efficiency.

That will probably not be well received by the person who's paying for the construction.

This shows a proof of concept at best, but in that proof shows a LOT of issues involved in trying to get that proof of concept squeezed down to something that can compete with a Human.

I'm not against automation. I'm against shit that won't work in real life pretending it's science's gift to an industry. Lay people may ooh and aah, but the guys who learned not to leave dickmarks in the wood are going to look at it and laugh, then stop by Home Depot for some hired temp help to do the job, instead, because they're the only ones who will get the WHOLE job done.

As I mentioned before I expect automation for the industry to come from prefab/3D printing and on site robotic assembly. That will be the cutting edge of light construction for the future instead of trying to shoehorn badly designed and limited robots into the ill-fitting environment that's conventional light construction today. Unless they design robots to literally go where no man can go before and do the job better and faster, it's not going to be more than a curiosity.

 

[Ten Wind]

I'm surprised we didn't see a mention of FBR, their robot looks very well advanced in the construction of brick homes though the machine isn't yet in commercial production

https://www.fbr.com.au/

That was really interesting, thanks for posting!

 

[r3loaded]

I am surprised that construction is named "tech adverse". This is not what I experience here (Switzerland), they have plenty of machines, more or less automated, to facilitate their work - tunnels are not bored they way they used to be bored 50 years ago, bridges are built using laser controlled machines... roads are build using remote controlled asphalt machines.


Of course there are still a lot of hard workers around these machines, to do all the stuff that cannot (yet) be automated. But a big part of now made with "smart" construction machines.

The key word in your comment is "Switzerland". High labour costs naturally leads to more investment in capital and technology to utilise that labour more quickly and efficiently.

Visit another country with much lower labour costs and you'll quickly see how tech averse construction projects become and how many more workers they employ to accomplish the same job.

 

[Gregorius]

Instead of building robots to do drywall (which, as been pointed out, this one doesn't really seem to do), how about we spend some money figuring out a construction material for walls that ISN'T drywall. It's an absolutely horrific building material. There's got to be a better material and methodology for building interior walls.

As others have also said, residential building construction needs a fundamental rethink. Probably a computerized manufactured building, whose pieces are all fabricated offsite and then shipped to the location and just have to be assembled. I saw a house done this way on This Old House (in Vermont I think). Seems it was more expensive, but everything actual fit together EXACTLY as it was supposed to since a CNC machine did ALL the cutting.

Edit: ninja'd by justin150 - should have read all the comments! Doh!

Here, here. Gypsum is also usually full of horrible chemicals that off-gas for years (especially if sourced from dubious countries with lax health and safety protocols....cough cough China).

Perhaps we can move toward more 3-d printing all-in-one solutions (like used in Germany), avoid dry wall altogether by utilizing green cement and bricks, integrated insulation ect...

Actually, drywall isn't full of chemicals that off-gas for years. It's an inert building material... the gypsum core is wrapped in paper. There's nothing more to it.

Back in the day, Chinese drywall used a toxic additive that deteriorated over time and caused huge issues. This additive was never used in North America. I'm a drywall contractor based in Toronto, my main supplier (and all their competitors) only used drywall manufactured in North America by massive, reputable companies (Saint-Gobain, GP, etc).

I've been in drywall business for decades and haven't seen a sheet of drywall (on our projects) that hasn't been sourced in North America.

Despite the drawbacks, there is no competitive material for interior finishing. Lath and plaster isn't price competitive and is too time-consuming to use. Same for plastering over brick or block... you can use any kind of wood product on your walls, but then you have to pay for that wood and find a way to finish it. Not realistic unless you have an unlimited budget for your project.

To the topic of automation... this robot can automate the simplest drywall task. Wow. My typical projects are so unique and complex that there's literally no way to have a robot do our work. Eventually we will probably see Terminator-style robots come along that can handle the most intricate types of manual labour... but we won't be seeing that technology in our lifetimes.

 

[justin150]

Instead of building robots to do drywall (which, as been pointed out, this one doesn't really seem to do), how about we spend some money figuring out a construction material for walls that ISN'T drywall. It's an absolutely horrific building material. There's got to be a better material and methodology for building interior walls.

As others have also said, residential building construction needs a fundamental rethink. Probably a computerized manufactured building, whose pieces are all fabricated offsite and then shipped to the location and just have to be assembled. I saw a house done this way on This Old House (in Vermont I think). Seems it was more expensive, but everything actual fit together EXACTLY as it was supposed to since a CNC machine did ALL the cutting.

Edit: ninja'd by justin150 - should have read all the comments! Doh!

I have done the same so many times!!

As for replacing drywalls I can think of a couple of "modern" ideas - wattle & daub and lime mortar. Wattle & Daub is about a 6000 year old technique, very environmentally friendly. Lime mortar dates also dates back about 6000 years, far better environmentally than cement and plaster.

More seriously 3d printing and things like SIPs panels are the way forward: environmentally much better than concrete & plaster. Steel beams can be replaced by engineered timber beams which are not only better environmentally but also beautiful

 

[BurntToShreds]

I think that modular robots like Spot are going to be the primary machines on construction sites for a long while, as well as effective tools for construction companies to use pre-build. Using a Spot on the ground with a LIDAR on it in conjunction with aerial drones to survey and map sites would be a boon for the GIS wing of any engineering firm or construction company. Spot can also be really useful for ferrying small supplies or tools from one floor to the next since it can climb stairs.

 

[Penguin Warlord]

Instead of building robots to do drywall (which, as been pointed out, this one doesn't really seem to do), how about we spend some money figuring out a construction material for walls that ISN'T drywall. It's an absolutely horrific building material. There's got to be a better material and methodology for building interior walls.

As others have also said, residential building construction needs a fundamental rethink. Probably a computerized manufactured building, whose pieces are all fabricated offsite and then shipped to the location and just have to be assembled. I saw a house done this way on This Old House (in Vermont I think). Seems it was more expensive, but everything actual fit together EXACTLY as it was supposed to since a CNC machine did ALL the cutting.

Edit: ninja'd by justin150 - should have read all the comments! Doh!

Here, here. Gypsum is also usually full of horrible chemicals that off-gas for years (especially if sourced from dubious countries with lax health and safety protocols....cough cough China).

Perhaps we can move toward more 3-d printing all-in-one solutions (like used in Germany), avoid dry wall altogether by utilizing green cement and bricks, integrated insulation ect...

Actually, drywall isn't full of chemicals that off-gas for years. It's an inert building material... the gypsum core is wrapped in paper. There's nothing more to it.

Back in the day, Chinese drywall used a toxic additive that deteriorated over time and caused huge issues. This additive was never used in North America. I'm a drywall contractor based in Toronto, my main supplier (and all their competitors) only used drywall manufactured in North America by massive, reputable companies (Saint-Gobain, GP, etc).

I've been in drywall business for decades and haven't seen a sheet of drywall (on our projects) that hasn't been sourced in North America.

Despite the drawbacks, there is no competitive material for interior finishing. Lath and plaster isn't price competitive and is too time-consuming to use. Same for plastering over brick or block... you can use any kind of wood product on your walls, but then you have to pay for that wood and find a way to finish it. Not realistic unless you have an unlimited budget for your project.

To the topic of automation... this robot can automate the simplest drywall task. Wow. My typical projects are so unique and complex that there's literally no way to have a robot do our work. Eventually we will probably see Terminator-style robots come along that can handle the most intricate types of manual labour... but we won't be seeing that technology in our lifetimes.

Yeah, I don't really know what they're talking about, drywall is a great building material. Easy to cut, easy to patch, surprisingly strong for carrying vertical loads, and relatively low environment impact compared to alternatives.

It's only real downside is the dust involved, but dealing with dust control while setting it up initially is a pretty small price to pay.

 

[bburdge]

Forget this part of the drywall process, hell most residential level jobs never do this step and just do a spray texture on smooth sanded jointing.

Now if this thing could automate the sanding work? That would be worth it. It's easy enough to put down the compound, but the hours of sanding needed afterwards to get everything smooth is a backbreaking process with tons of holding up heavy power sanders in awkward positions putting high stress on the body, and a lot of manual sanding in hard to reach areas.

 

[Dadlyedly]

"The arrival of more automation may also alter demand for labor in a number of building trades."

Way to quietly bury the lede.

Having done light construction for a few years, this is not a field that can be easily automated. I worked on roofing, framing and finishing, mostly. The structures were usually one or two story buildings, usually homes. I've done both metal and wood framing. Windows, plumbing, electrical and flooring were usually done by subs.

When I saw this did "drywall", I thought, "Cool, it'll cut drywall, hang it, tape it, paste it and smooth it." Well, two out of five isn't awful, and it'll probably do it faster, but at what cost? MOST contractors can't afford something like this up front.

Drywall isn't a subcontracting specialty (at least it wasn't when I was doing it at the end of last century). It's something someone can get very good at, but it's not a subcontractor's job. This would be something a major construction firm might consider, but not the general contractors that make up the majority of the folks out there doing light construction. It needs to do a lot more a lot better and faster than a human can do to be practical for most jobs out there.

From a "can we do this" point of view, it's interesting, but it's a glorified painting machine (which is used in a lot of places) for drywall smoothing. It's TOO NICHE of a tool. If it cut, hung, taped, plastered and smoothed all by itself without supervision, hell yeah, I can see even the small contractors going for that. Unfortunately, having done that job, I don't see how it can be automated. There are just too many special cases where hanging drywall in tight spaces or for unique situations is a head-scratcher. If you've never done drywall on curved surfaces, you might not know what I mean. Yes, it can be done. Yes, it's a pain in the ass. Get a machine that can do that, and you'd have contractors throwing money at it.

So far, I've seen some penetration of automation into the light construction industry, but most of it is still in heavy construction. I don't see a lot of automation happening as long as construction continues using the same techniques they were using 50 years ago. I can see them 3D printing a home or office (It seems to me they have options for that), at least in pieces, if not on-site. Prefab is another way to automate construction.

So the day may come when you select your home from a catalog, and it's printed, wired, plumbed and assembled by robots on site. I don't see that day coming anytime soon, though.

I believe someone else said it earlier, but I think we're more likely going to see a future where you head to an Ikea website, see a picture of a beautiful little Scandinavian home overlooking a fjord, select it, and get it sent in a couple of trucks as a flat pack as soon as the robots are able to cut out the pieces to your local codes. Then you and your buddies can interpret the computer generated pictogram instructions, and hopefully everything fits in the end, even if your new little home looks better overlooking a fjord than in the middle of the 'burbs. Think Sears home for the 21st century.

 

[Roonski]

Interesting choice of tasks to try and automate. Having just been through a complete home renovation, where we did the drywall from beginning to end as DIY, the drywall plastering and sanding seems like one of the last true artisan skills. Even following all the procedures, doing a good quick job depends on complex, fine motor controlled motions that humans only get after a fair amount of time. (After doing something hundreds of joins we are only at the stage of getting a decent result but often slow). I spent many hours perched on ladders in awkward positions, gently sanding away imperfections using a bright light from sides to identify them.

Seems like a terrible task for a robot, to be honest.

 

[Anna Moose]

Much of that is being done today. Those joists on a framed house? Used to be you set up your saw horses and angle, used a trig calculator if you needed to and cut the lumber with a cutoff saw. Now a flat bed truck comes up with a huge pile of precut joists ( cut, measured and assembled largely by a machine) and then dragged up with a boom truck. Wham, the roofline is up in a day rather than a week. Cabinet tops are laser cut on a machine and brought on site. Entire cabinets are machine cut and brought on site. Wham, the kitchen is done in a couple of days (waiting for the electrician, or someone like him). There is little hand finish carpentry these days.

Nobody I know in construction is averse to automation or just straight out mechanization. The trick is paying for it.

That's because construction is still in the "tools make the employees more effective" phase rather than the "tools make the employees obsolete" phase. As long as it requires a lot of supervision/interaction it's like a supercharged power tool, letting you deliver even more value/hour.

The backlash will come when you reach the tipping point where it's starting to make the specialties redundant, instead of a carpenter or electrician or plumber you have a bot herder that's moving them around and making sure they're all stocked while calling in the pros become rarer and rarer.

Isn't that how every industry has automated though? When people attempt to eliminate (almost) all humans in one go, its almost always a total failure. If its currently taking 20 people to get a job done, expecting it to go to one is just not realistic.

For example, Elon Musk made this mistake of trying to fully automate the model 3 production... but it had too many problems and they had to dump it. They've been trying to re-implement pieces of it incrementally, but drastic changes to the process and automating it all at once just never goes well, there is just too much going on.

I work in construction, and seen lots of automation attempts fail. One recent one that looked like it was going well and lots of people switched to it before they realized that the automation attempt ended up costing more once it was implemented was the Prefabricated Wall Panels. It looked clever, the walls were prebuilt in a truss plant, designed to save every penny on hardware and lumber (usually things were more overengineered as there is an expectation for a certain amount of mistakes in the field) and delivered onsite in trucks. They were numbered so you just line up the numbers and strap them together, and it looked really popular for a couple years, especially as we were at the end of the great recession.

Nearly everyone around me (including those who I worked with) abandoned them for a bunch of reasons:
1) Delivery ended up costing more, as while all the lumber fit on one truck to build a house, now it was taking 3 or 4 deliveries to get these prefabricated walls on site.
2) If they got damaged, it took way too long to get a replacement wall piece, or for an onsite fix, we needed a licensed engineer to sign off on the variation, making that much more expensive than just overengineering it. Overengineering everything also got us a better product.
3) The truss company used it as part of a way to justify a multi-million dollar mostly automated machine to build the trusses that also built these prefabricated wall panels, but when the construction industry started booming, that machine hit capacity building normal roof trusses, which were much more complicated to build and a better use of the machine. It was obvious that spending millions on another machine to make those walls would further increase the cost to the point where pre fabricating them offsite just made no financial sense, so at least in my area, that pretty much completely died.

Construction people love their tools, and tools keep getting better and more multifunction so less people can do more jobs. Countertop templating used to be done by making actual templates with multiple people, transporting them to the fabrication shop, and using the physical template with their cutters to make the countertop. Now one person comes in a prius with a laptop and a tri-pod mounted laser that measures everything and sends it straight to the machine to cut it and polish it. Instead of a person only being able to template maybe 2 houses in a man-day average, now they can do 20.

The sitework company has over the years got better equipment, instead of bringing out 3 or 4 machines (roller, excavator, track loader, and sweeper) and 2 or 3 people (2 operators and a spotter on the ground) instead they're now down to one person with 4 attachments for the machine, and to increase productivity they have laser leveling along with GPS guidance, and they can be there own spotter and operator as they can now control them from their phones. Thats how real automation is done, they keep incrementally improving things to eliminate people and allow things to be done faster. Robots that can do every step of the job that are like humans just aren't realistic in the near term future.

 

[test6554]

I like that we are seeing more automation in the construction industry. Housing is expensive and making it cheaper and more efficient will ultimately lead to better quality of life for all.

I also like the promise of this bot, but it's clearly in early stages of development vs the vision. Also, it seems much more efficient to automate tasks rather than people. Spraying, sanding, lifting, nailing, etc. are all tasks that could be the one job some robot or robot tool does. Then you can build workflows that combine those tasks to get whole jobs done.

 

[johnsonwax]

This is the wrong approach. It's perilously hard to introduce robots into environments designed around human labor. This would work much better if the nature of housing design changed to a more modular approach. Wiring and plumbing still consists of drilling through studs, pulling by hand, punching holes in drywall, etc. The US has been resistant to flexible plumbing, still mostly relying on soldered copper rigid plumbing.

I mean, we can transition to a future where we have a nice seam - akin to a well assembled panel gap on a car, every meter or two on our walls, if that comes in exchange for a dramatic reduction in cost/speed or ease of renovation in homes.

Mind you, none of this has anything to do with the high cost of housing. My 3BR single family home has a market value of about $1M and an insurance replacement value of under $200K. I have an $800K foundation, lawn, and driveway. It's just rent seeking - either by developers, or by homeowner groups blocking zoning changes, new development, etc. Construction is pretty cheap. Trades are pushed hard, a bit too hard in my opinion given how common shoddy mistakes are in new construction. And they aren't paid that well. If robots are looking attractive, it's because general contractors suck at their job.

 

[Veritas super omens]

I demand that the robot programmers code in a "I, for one, swear fealty to my human overlords" in the assembler code for the device .....

 

[Killjoy]

It doesn't do corners or angles, doesn't actually place the drywall. So it's a spray paint robot. For likely a hefty sum.

Doesn't really add up. If this is the pointy edge of construction automation then we're not all that far into the next paradigm.

This is not the pointy edge of construction automation.

This reminds me of that story of electrification during the industrial revolution, when electric motors first came on the scene to replace coal fired furnaces. Because of how coal fired engines work, most factories would have one massive one in the basement, with an enormous amount of machinery, crank shafts, gears, interlinks, etc. all connected back to it running all the various machines and conveyor belts upstairs. It made sense cause all your coal would get delivered to the basement and workers would be constantly shovelling it into this massive engine that ran everything.

When electrification hit a lot of factories and factory owners said "oh hey these new electric engines are the future, let's swap out my massive coal fired engine for a massive electric one". And even though these owners seemed more forward thinking then the owners who were still holding onto their coal engines, they by and large were not the ones that succeeded. They still had all the same problems as previous coal powered factories where if one machine went down the whole factory would essentially have to shut down since everything was still connected back to one massive motor. The owners failed to understand the process changes that were possible with a new technology. Sure you could gain some minor benefit by just straight swapping coal for electric in an existing process, but it was all the other aspects of electric motors and electricity (like the fact that tiny, flexible wires can carry enormous amounts of power, electric switches are very easy to create to isolate sections of circuits, etc.) that meant that the most successful factories were the ones that rethought their whole process from the ground up, redesigned with the new possibilities presented by electric motors.

This robot is basically the same thing. Let's take super advanced robotics, and dry and just swap out a person drywaller for a robot while not roboticizing or automating anything else about the processes around it.

The real pointy edge of construction automation, is stuff like modular, and off site construction. Where buildings are built in pieces in a factory and then shipped to site and assembled incredibly fast. It requires way more precision than traditional construction processes, because you can no longer afford to have mistakes in your drawings that a contractor will catch or that you can discuss with them and make a change during construction. Everything basically has to be perfect because all the pieces need to fit together perfectly on site or you risk delaying everything. It also allows for greater automation since everything is being built in a more controlled factory. This is the kind of stuff that I think is exciting about the architecture and construction industry's future, not ham fisted robot construction workers.

Much of that is being done today. Those joists on a framed house? Used to be you set up your saw horses and angle, used a trig calculator if you needed to and cut the lumber with a cutoff saw. Now a flat bed truck comes up with a huge pile of precut joists ( cut, measured and assembled largely by a machine) and then dragged up with a boom truck. Wham, the roofline is up in a day rather than a week. Cabinet tops are laser cut on a machine and brought on site. Entire cabinets are machine cut and brought on site. Wham, the kitchen is done in a couple of days (waiting for the electrician, or someone like him). There is little hand finish carpentry these days.

Nobody I know in construction is averse to automation or just straight out mechanization. The trick is paying for it. As noted previously, drywall is typically done by very low wage folk. It will be interesting to see if this machine is anywhere near competitive with that. I have my doubts. Construction folk are pretty smart. They can run spreadsheets to see what makes financial sense.

A purpose build robot to do a single simple task doesn't seem like a particularly strong example.

Sorry, I'm tweaking on the terminology here

Trigonometry, to cut a joist? Maybe you meant "rafter, " or "truss."

Joists are horizontal framing members between storeys, studs are vertical framing members, there's not much call for trig there.

 

[johnsonwax]

Isn't that how every industry has automated though? When people attempt to eliminate (almost) all humans in one go, its almost always a total failure. If its currently taking 20 people to get a job done, expecting it to go to one is just not realistic.

For example, Elon Musk made this mistake of trying to fully automate the model 3 production... but it had too many problems and they had to dump it. They've been trying to re-implement pieces of it incrementally, but drastic changes to the process and automating it all at once just never goes well, there is just too much going on.

I think that's the wrong lesson.

The kind of mass automation you are describing fails because it's done in the context of larger systems that are designed for humans. You need to go one step further back for that kind of automation to work. In Teslas case, they took a factory designed for conventional assembly (the old Toyota/GM plant) and a car designed for conventional assembly, and tried to automate it.

If you want to succeed, start with a plant designed for full automation on a car designed for full automation. To that - the cybertruck may look ugly as hell, but my eye sees panels that don't need to be stamped, completely eliminating a pretty massive part of the production process, which leads to a vehicle that is shaving time to market time off, large steps in the manufacturing process, and almost certainly an easier automation effort (simply grabbing a flat panel is easier and more reliable than a curved one). I can't tell if it's designed for automated assembly - I doubt it - but it might be a little closer.

Amazon is trying to automate their distribution centers - with some success, but compare it to Ocado's approach. And Ocado's approach then begs some new questions - can they get producers to standardize packaging sizes to make it further easy to do. Walmart tried this with milk, which was an admirable effort, but consumers didn't like the design. Someone will get that right and it'll be transformative. Witness shelf-stable milk in Europe, or milk sold in bags pretty much everywhere but the US. These are not technical innovations as much as they are cultural ones that allow the technology to be used. That culture can be among consumers (Walmart milk) or among workers (US resistances to flexible plumbing because it reduces the need for plumbers) or industry leaders (automakers are not experts at engineering cars, but at engineering assembly lines - they don't want their expertise to lose value - so they design cars that favor their expertise).

 

[HiggsForce]

What we need at this moment are a few construction robots that can pull extra cables to stabilize and fix Arecibo (due to the risk of collapse this is too dangerous for humans to do). Where are they?

 

[The Wonderful Wizard]

. Wiring and plumbing still consists of drilling through studs, pulling by hand, punching holes in drywall, etc. The US has been resistant to flexible plumbing, still mostly relying on soldered copper rigid plumbing.

In australia steel stick framing is now the norm, because of the high price of timber and the reduced insurance cost (termites more than fire, though it is mandatory in some high fire risk regions). Those frames all have slots and holes cut in them at the factory to reduce material, but that also makes wiring and plumbing faster

 

[TMilligan]

In my previous career I was a general contractor and flipped quite a few houses. When I first started I made a point to try every trade and become somewhat proficient at it before subbing it out. That way I could better tell if the sub contractor was good or not. Drywalling is one profession I have a huge respect for that is often under appreciated. It is both extremely hard labor but also requires extreme finesse. It’s often one of the most customer forward professions where a potential buyer could easily tell a bad drywall job from a good, however a perfect drywall job is often never noticed... just expected. Hence, why I felt the profession was always under-appreciated.

Watching this robot work is kind of sad in a way. I think it would be fine for huge, long tracts of wall by saving the mudder some time but truly this machine is only doing a fraction of the job. Someone needs to hang all the drywall, then put up all the metal outside corner bead, mud all the screw holes, tape and mud, then sad, then do 1-2 more mud passes with sanding in between.

If you ever get the chance to try mudding a wall, then watch a pro do it, you would be blown away by how fast, effortless, and perfectly they move and smooth the mud.

I think replacing the human while retaining the systems and techniques we’ve developed under the boundaries of human physicality is erroneous. IMO, if we are going to automate a process we need to focus more on the end result and less on the established techniques. In that case, have the robot cut a full wall out of one sheet and fasten it. Then it would on need to mud the corners. Rethink what is possible without a human’s physical limitations.

 

[MDCCCLV]

Automation don't necessarily take away jobs, it could (as in this case) change them.

I appreciated the term "skilled workers", and its cynicism, as the main purpose of this machine seems to replace a skilled worker by a lowest cost possible worker, trained in hours, not in years!
The same way that initial industrialisation of process replaced highly skilled artisans by quickly trained low-wage workers.

Naturally this also makes these workers replaceable on a snap.

In this case everyone I've ever talked to says that doing mud for drywall is universally reviled as a tedious dull task that no one wants to do. And it isn't that hard to learn how to do, I would say months not years.

 

[Deleted member 1]

That's nice but when will it bring back the lost art of curved walls. But...

Where's the robo-electrician?!?

For all the advertisement and in person bluster I have yet to meet an electrician that can wire a data-center, new building, re-wire an existing building, or even a home, and meet both local and national codes as well as the additional standards set in the contract. Never, not once in 30+ years

Well, maybe if you spent the money and hired licensed, union labor instead of "I do it all" unlicensed non-union idiots, you'd have better results, eh?

 

[Hydrargyrum]

I'm surprised we didn't see a mention of FBR, their robot looks very well advanced in the construction of brick homes though the machine isn't yet in commercial production

https://www.fbr.com.au/

There was a hyperlink in the article referring to the "brick laying robot", but that was the extent of it.

If Ars wants to talk to someone at FBR for an article, I can arrange that.

 

[Sulla]

I don't think this "robot" worth it in any scenario. One testimonial says this thing cut a level 5 job from 7 days down to 2 days, no mention of what square footage it did. Also to get to level 5 it takes good old "mudders" (crack fillers) applying and hand sanding to get to level 5.

Excellent graphic—thank you.

 

[Andrewcw]

Even if it is only for what people are describing Level 5 Finishing is. I can see if for large construction projects like high rise offices or large structures with large walls. Will it be wide spread? Only if the project is large enough. But the machines works at night. So lets say it costs triple of what a regular worker costs per year. In construction for large projects you won't be working 24/7 because of labor/unions cost. So setup this machine and press a button and have a team remotely monitoring each of these systems for a night time shift. The time savings of getting it done within contract time without paying overtime or having to hire an additional shift of workers.

 

[Sulla]

In my neck of the woods, there is still a stigmata against manufactured homes.

I sure hope you're not a native English speaker.

 

[rmgoat]

I think they would be better off automating cutting, hanging, and taping. Personally I am awful at mudding and respect anyone who can do a good job at it. Unfortunately a lot of contractors hire guys in the Lowe's parking lot for under the table day labor.

I live in an 'interim' house from the days of transition from lath and plaster to current drywall with metal mesh corners and 2'x4' drywall panels finished with a 'scratch or brown coat' with 'fine' white coat over that. When you drop chunks of it when tearing out a wall it sounds like you are dropping china on the floor. Matching this to new drywall especially on a not perfectly square wall is beyond my skills. Fortunately a master carpenter friend looked at my problem and said "You need Eugene, here's his number." Never regretted that, the man is an artist.

 

[agpob]

I think the real issue isn't- "a horse and buggy was good enough for grandpa, so it's good enough for me"... What I see is a huge waste of resources for something that lots of people actually enjoy doing. (yes, they really do) This is not dangerous or mind numbingly boring work. Put those brains, money and time to better use on something most humans HATE doing or is crazy dangerous to health, life and limb. This might have a place in assembly line modular construction, but on-site custom homes, it may always be cost prohibitive. IDK, that's my 2¢ worth anyway!

 

[Roy G. Biv]

Have been on many commercial and industrial construction sites. Automation of any consequential amount is very far away. Job sites are dirty, dark, wet, uneven, strewn with debris and stacks of supplies. There are tight angles and small spaces. Almost every building is completely custom designed, so there's little standardization. Plans and specs constantly change, even after things have already been purchased and installed. The architects and engineers regularly design things that are impossible and/or wrong in real life. Owners keep changing their minds. Inspectors demand random changes to satisfy their bizarre interpretations of code requirements. Sometimes I'm amazed the jobs ever get done. Automating the majority of jobsite work just wouldn't be possible; human collaboration in real time is utterly essential.

 

[kenja]

One hundred and five years ago my great grandparents ordered a two story farmhouse from Sears. You would never guess it was a kit by looking at it.

 

[GottaSaySomething]

This just paints and finishes and it says it can't do corners...... how fast is it?

Like I guess if a guy could set it and leave it be, then come do the corners later that's good. BUT IT'S THE REAL WORL how often is it going to break down and shit? Even if it doesn't need to recalibrate to its environment with human help

 

[Anna Moose]

Isn't that how every industry has automated though? When people attempt to eliminate (almost) all humans in one go, its almost always a total failure. If its currently taking 20 people to get a job done, expecting it to go to one is just not realistic.

For example, Elon Musk made this mistake of trying to fully automate the model 3 production... but it had too many problems and they had to dump it. They've been trying to re-implement pieces of it incrementally, but drastic changes to the process and automating it all at once just never goes well, there is just too much going on.

I think that's the wrong lesson.

The kind of mass automation you are describing fails because it's done in the context of larger systems that are designed for humans. You need to go one step further back for that kind of automation to work. In Teslas case, they took a factory designed for conventional assembly (the old Toyota/GM plant) and a car designed for conventional assembly, and tried to automate it.

Tesla's Model 3 production line wasn't using any old equipment from the Toyota/GM plant and the line was built from the ground up to be fully automated using all new equipment. Tesla designed a whole new model with that line... I don't know if they were taking into account their automated production line with the physical design of the car, but I assume they would: Elon was claiming that was a big part about how they were going to hit the $35k price point, which they never really did, as a result of the complete automation failure. Tesla even blames this failure on bringing the company to be mere weeks from bankruptcy during their "production hell" period. A TON of money was spent on creating a fully automated line, and it failed miserably.

If you want to succeed, start with a plant designed for full automation on a car designed for full automation. To that - the cybertruck may look ugly as hell, but my eye sees panels that don't need to be stamped, completely eliminating a pretty massive part of the production process, which leads to a vehicle that is shaving time to market time off, large steps in the manufacturing process, and almost certainly an easier automation effort (simply grabbing a flat panel is easier and more reliable than a curved one). I can't tell if it's designed for automated assembly - I doubt it - but it might be a little closer.

But thats just value engineering. As you imply, removing the stamping process cuts a massive part of the production process out, but its clearly changing the product massively, and we don't know yet how the market will react to that value engineering. Tesla's trying to figure out a way to make the price of the Cybertruck reasonable, they also have proposed eliminating painting the vehicles and having them all be the same color. These types of things aren't automating the process, they're literally eliminating it. Thats a completely different thing. Time will tell if people are willing to take unpainted cars with boxy shapes in exchange for a bit lower cost. They might, but thats not automation. Infact, the robot in this article is automating what is generally considered an optional step of drywall finishing... low cost, value engineered prejects remove this step as that layer is purely cosmetic. Many homes exclude that step entirely, especially if you're not at the million dollar price point.

Amazon is trying to automate their distribution centers - with some success, but compare it to Ocado's approach. And Ocado's approach then begs some new questions - can they get producers to standardize packaging sizes to make it further easy to do.

Amazon tried to do this before as well, and Amazon has had huge impacts on product packaging, including very early on in the history of the company. However, Amazon found out a real major problem when they were doing that: standardizing packaging to make it easier for robots to grabs often means making the packaging less efficient and more bulky, so it ended up costing more. Packaging isn't so standardized for a reason, because its optimized more per product.

Walmart tried this with milk, which was an admirable effort, but consumers didn't like the design. Someone will get that right and it'll be transformative.

Yup, because the previous reason the cartons weren't a cube still held true: that shape doesn't make it as easy to pour milk, and consumers would rather pay a small amount more money to prevent spilling any of the product. They did keep the new design in their Sams Clubs stores where price is considered even more important, and the design kept for many business to business sales where price was considered more critical.

Witness shelf-stable milk in Europe, or milk sold in bags pretty much everywhere but the US. These are not technical innovations as much as they are cultural ones that allow the technology to be used.

Shelf stable milk is available in America too in every grocery store I've been in. It has a different taste though which our culture doesn't desire as much... to make milk safe requires a much higher pasteurization temperature which changes the taste along with a belief among Americans that it reduces certain nutrients.

US resistances to flexible plumbing because it reduces the need for plumbers

I gotta ask, how does flexible piping reduce the need for plumbers? Flexible pipe fittings tend to require more expensive tools and more skill then the super cheap pipe glue and CPVC pipe that is currently America's preferred option.

I know theirs pros and cons between the different products, but CPVC is generally looked at preferably to PEX as it hasn't had issues with bad fittings. PEX's ease of use does make it dominate in renovations, and its also used in most modular structures because of the ease of linking them together on site after being assembled offsite. Use of copper piping has been almost entirely phased out for new installations in America.

 

[coslie]

Interesting choice of tasks to try and automate. Having just been through a complete home renovation, where we did the drywall from beginning to end as DIY, the drywall plastering and sanding seems like one of the last true artisan skills. Even following all the procedures, doing a good quick job depends on complex, fine motor controlled motions that humans only get after a fair amount of time. (After doing something hundreds of joins we are only at the stage of getting a decent result but often slow). I spent many hours perched on ladders in awkward positions, gently sanding away imperfections using a bright light from sides to identify them.

Seems like a terrible task for a robot, to be honest.

How is a frustrating, finicky, repetitive, tedious job a terrible task for a robot?

 

[coslie]

When placed in a room, the robot scans the unfinished walls using lidar, then gets to work smoothing the surface before applying a near perfect layer of drywall compound....

Don't know why people are claiming this thing just does the skim coat.

I tried following the company link to see what exactly the thing does, but it was pretty vague. It did however mention "Remove Dust" under the Safety heading, and the article says it smooths the surface, so I'm pretty confident it sands.

 

[deet]

Anyone who has seen Runaway knows how this ends:

https://gfycat.com/impressionableimmediatehuia

 

[void&]

When I was young and working odd construction jobs, drywall work was kind of fun, except when I had to put up a ceiling. Having to hold a sheet of drywall with one hand while the other is hammering nails upwards, that was hard. It would be easier now with nail guns. If I were to automate drywall work, I'd start there.

 

[Arkleseizure]

The future of construction isn't in robots replacing human tasks, but in robots performing tasks that humans can't do that result in the same outcome.

The nature of house construction will change to suit robotic processes.

Also, as an Englishman, what the heck is dry wall? Is it plastering?

 

[rmgoat]

When I was young and working odd construction jobs, drywall work was kind of fun, except when I had to put up a ceiling. Having to hold a sheet of drywall with one hand while the other is hammering nails upwards, that was hard. It would be easier now with nail guns. If I were to automate drywall work, I'd start there.

Panel lifter, battery screw gun.

https://www.youtube.com/watch?v=6oDF5LM ... el=ToolPro

https://www.homedepot.com/b/Tools-Power ... Selection=

 

[rmgoat]

The future of construction isn't in robots replacing human tasks, but in robots performing tasks that humans can't do that result in the same outcome.

The nature of house construction will change to suit robotic processes.

Also, as an Englishman, what the heck is dry wall? Is it plastering?

https://en.wikipedia.org/wiki/Drywall