SoylentNews is people
SoylentNews
from the building-a-more-robotic-tomorrow dept.
Hadrian is not the first large-scale industrial robot that can complete a whole build from start to finish. It's not even the first outdoor construction robot.
What's remarkable is it's both. As Mike told me, "Anything you can build inside a factory ... we're getting really, really good at. Trouble is, nothing's happening outdoors."
That's because environmental factors like wind and temperature variations can make life difficult for robots outdoors.
Most robots can't adjust to small, quick changes in wind or temperature fast enough to keep up.
That's fine if little wobbles won't make a big difference. But when you're working on something as large-scale as building a house and a light breeze could lead to bricks being laid way out of position, it can get very dangerous.
So up till now, any robot building on such large scales had to be indoors in minutely controlled environments.
Hadrian has overcome this problem using the precision technology Dynamic Stabilisation Technology (DST). DST was developed in Perth by Mike's cousin, Mark Pivac, back in the early 2000s. The computer program measures environmental factors an astounding 2000 times per second, then accounts for them in real time.
If robots replace the construction workers, who then will wolf whistle?
(Score: 3, Interesting) by bob_super on Tuesday January 23 2018, @07:32PM (1 child)
Yes, you can make a super fancy robot that perfectly compensates for environment and material variations...
Or ... you could make a robot that assembles houses like Lego pieces (minus the gaps) at full speed, because the material itself compensates for some variability.
Guess which one will make the most economic sense?
(Score: 2) by acid andy on Tuesday January 23 2018, @09:12PM
Yes couldn't they cheaply print some UV or magnetic dots onto each brick that the robot can align its sensors against to orient itself? That could be combined with some kind of optical fail safe or some push switches to feel the edges of the wall. Then, so long as the dot is still triggering the sensor, it knows it hasn't moved with respect to the brick. It could even build a wall of bricks with alternate types of marking on, so that there's less risk of it being confused by an adjacent brick.
That said, I would have thought computer vision algorithms would be up to doing the whole thing optically by now (with some simple touch / pressure sensors for the final grip). I'm thinking it's just resistance to change / aversion to risk / lack of funding that's holding this stuff back rather than any fundamental technical limitations, but robotics really isn't my area of expertise.
If a cat has kittens, does a rat have rittens, a bat bittens and a mat mittens?