Robotic Fabricator Could Change the Way Buildings Are Constructed | Hacker News Reader

Robotic Fabricator Could Change the Way Buildings Are Constructed

technologyreview.com

103 pointsitamarst28 comments

I admit i am out of my depth here. Correct me if i am wrong.

I dont think construction needs to be made faster. I needs to be of higher quality. ( Ignore the land issues for a moment ) You could built a 60 floor Skyscraper in less then a month.

What we need is a much faster, semi / automatic way of internal renovation in each floors. Door, Glass Fitting, Pipes and Electrical Wiring etc.

And all of these needs to be of high quality finishing. Taking into account of Temperature variation, lighting. Making building with the precision level of a Smartphone.

Even today, you could see building with Windows not aligned perfectly, floor line running slight wide, floor not perfectly level. Or even Windows with Leaking Water.

May be all of these are specific to Hong Kong, where we paid the highest price per Square Feet in the world and get absolutely dismal quality of building.

My grandparents just moved into a new home, only after the deal was closed did anyone notice that the washer and dryer water tap was built directly above the washer and dryer electrical outlet.

I've never seen such a thing, it's basically the same as an outside garden hose tap, installed directly above the electrical outlet.

that's not something I'd normally know to go looking for, so I doubt I'd have seen it myself if I was with them looking at the place.

Surprised that it would be up to code and pass inspection really.

I thought the same thing. What can be done though?

They have a home warranty that's good til next year, but I doubt it covers incompetence, and I don't think they could survive another move financially or physically.

It covers shit like garage door openers (which was installed backwards, as we found out after moving in) but we'd have to pay a service fee we can't afford just yet.

The guy they sent out to help with our plumbing emergency doesn't give me high hopes in who they'll send to fix the garage door.

The guy who built this place stripped it to the frame and rebuilt it with a team of the lowest bidding crackhead contractors, from what I can tell.

That depends what your goal is. If the goal is to have the cheapest building possible, you want to automate construction as much as possible. You are correct though that the internal set-up is a huge part of the cost.

This reads exactly like a press release, especially the way it addresses the limitations of the robot.

Makes me think of PG's essay Submarine: http://paulgraham.com/submarine.html

Thank you, very interesting essay.

"...I didn't realize you could write as candidly and informally as you would if you were writing to a friend..."

His essays are definitely worth diving into if you haven't done so already. I don't agree with everything he writes, but I often find myself thinking back to points he's made or just ways he looks at problems. The one on Schlep Blindness is particularly insightful.

Also good, thx :)

Why "must" it be connected to the Internet for the architect (or more realistically the contractor) to make changes? Why can't they do it on site from a console?

Because then they can't get shafted by the company's crazy storage and service costs and they'd miss out on a few bucks at the expense of usability. And I mean who cares about that while developing? You aren't the sucker using it!

The photo kind of makes me laugh. There's the robot building this nifty spiral rebar sculpture when we just want it to do some regular framing.

Spiral rebar layouts are very common in circular columns, especially in area of high seismic activity. It's a bit more difficult to manufacture than the alternative (evenly-spaced "hoops"), but it gives the column higher ductility during a failure event so that the affected people inside/on the structure have more time to escape.

http://www.insteel.com/media/8177/Spirals%20Brochure%20wo%20...

Edit: I have to admit, I didn't look too closely at the photo until just now. That is a weird thing they're doing, but I could see it possibly being useful for fancy architectural columns, even decorative ones. (I suppose that sort of application is where they're likely to find early adopters.)

The point is that the robots can build things that would otherwise be too time consuming to build, on site, with higher precision than previously possible, and with a feedback mechanism that can adapt to the environment. So yeah, surely the robot can build a rectangular rebar structure. But, it can also make other shapes which could be structurally interesting in addition to aesthetic.

Exactly what I was thinking - it's interesting, but not very practical or useful.

Onsite automated construction is likely a long ways off, but they do it right now for prefab house production: https://www.youtube.com/watch?v=d-g0Ubp-Qng

Video: https://www.youtube.com/watch?v=jjcXJ8EU48o

Brick-laying (even with general purpose arms) robots are already in production use.

https://www.youtube.com/watch?v=G_Pj2GI6-xc

https://www.youtube.com/watch?v=fpD5H1EH0go

Combining it with more comprehensive vision so it can figure out where to build the wall might be novel though.

> It can build stuff using a range of tools with a precision of less than five millimeters

So a precision of 10 millimeters, or worse than that? Obvious lexical ambiguity aside, I wonder what they're doing to drop the precision of that ABB arm so far. It looks like it's built around an off-the-shelf IRB2400 to me, which comes from the factory with a repeatability of 0.06mm. 5mm might be an approachable number to a civil engineer, framing carpenter, or bricklayer, but it's not impressive at all in robotics.

Also, the Thinkpad T430 and W540 made me smile. So much more appropriate for this bot than the Macbooks or iPads you see in so many videos.

Can it still provide that kind of repeatability along all degrees of freedom at full load though? Last time I worked with robotic arms (Staubli) the real world precision was much worse than the specs claimed. With something as heavy and imprecise as a construction brick, I wouldn't be surprised if there was at least a few millimeters of wiggle room close to the edge of the arm's reach. In my experience, the tolerances on parts used in automated assembly lines don't have such high variability as those in construction (and this seems like a very general purpose robot capable of some complex architectures).

> a repeatability of 0.06mm

That is in a controlled environment where the robot can be programmed to know where it stands in a coordinate system, and the surfaces it works on aren't moving relative to that point.

Take any robot in an industrial setting and put it's base on a soft pillow, and it will be a more difficult problem. It will have to use computer vision rather than just precise stepper motors to get where it wants.

This thing places things like bricks which are probably +/-5 or 10mm in shape, on a wall that sags, and does so standing on e.g. flexible wheels or tracks, on a construction site where the floor might be a temporary plywood one. I'd guess it can move its arm, if the load is light, to within 0.06mm of somewhere. It just can't know where that somewhere is because it's using a computer vision system with much higher tolerances to do so.

Thanks for the video, looks really interesting. The one thing i noticed was that those bricks look suspiciously perfect. If it gets to a stage where it can handle non-perfect standard bricks, discarding broken ones etc, that would be interesting.

Still a great proof of concept though.

I think they're metal, and it's using an electromagnet to pick them up.

...I've seen 3D printed faux-sandstone back in, um, 2012 or so? And news reports show that Dubai has just announced a 3D printed office, although I can't find any pictures of it in construction.

Here's a (moderately terrible) video of a 3D printed building in the Phillipines under construction:

https://www.youtube.com/watch?v=bTSakUtxXYY

And here's a much better one of what looks like similar technology in a testbed:

https://www.youtube.com/watch?v=XgPJdpGV8mE

No and no. Those are just standard (albeit colored) bricks and it was using a vacuum clamp to pick them up.

If you're thinking, "there's no way you can pick up a brick with a vacuum; it's too porous." Think again. The hottest thing in woodworking right now is Festool's new vacuum clamp. I've personally witnessed it holding on to a cinder block so strongly that I could not pull it off with all my might.

Interesting that you call it new. In industrial textile manufacturing, vacuum tables have been around for a long, long time. It doesn't matter if your leather is perforated or your upholstery is somewhat porous, or even if your foam is open-cell: A differential of just a few PSI, multiplied by width, multiplied by height, results in a lot of holding power.

Textile materials have a much lower density (which usually means a much higher ratio of exposed surface area to mass) and suction is a lot easier when it is in the same direction as gravity. As you point out, vacuum tables have been in use by a variety of machines for a long time, including in metal fabrication and construction. This is novel (or at least interesting) because using vacuum pumps to hold things up instead of down is usually reserved for low density, well balanced, flat objects like electronic parts.

I mainly thought they were metal because of being suspiciously shiny and the 'clang' they made as they fell down! The grasper didn't seem to have much of a vacuum hose, either. But metal bricks would be an odd choice --- very heavy and not really good for demonstrations, so I'm quite willing to be wrong.

It could still be the case that it's interesting if it only deals with near perfect bricks. If that reduces the labor needed to complete a project by X amount, then a construction company may still be interested even if the bricks cost X/2 more, in total, than other available bricks.

Interesting read. I question though the need for a robot: why don't we use said sensing tech to empower Bob the Builder to "position brick within a seven mm" tolerance, and have the added advantage of well honed human heuristics judging a much wider variety of situations?

From personal experience, the constraints that robotic manufacturing introduce in the design of an architectural object are huge: you end up designing specifically for the affordances of these tools, to the extent that you sometimes get nonsensical results.

Reply on news.ycombinator.com