Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I didn't realize this was underground mining. So, you're really working on driverless trains, not cars or trucks, correct?


> I didn't realize this was underground mining. So, you're really working on driverless trains, not cars or trucks, correct?

Why would one think that "underground" implies trains?

Trains require tracks. Roads are less precise and have cheaper intersections.


Maybe because of this exchange in one of the pages he posted?

AH: Where did the concept of driverless trains come from and how will it enhance efficiency at the Pilbara mines?

GG: Automated trains have been around for some time in airports but they are not used on a heavy-oil long-haul railway network of this scale. We believe the network at the Pilbara mines to be one of the largest private railways in the world, so although the technology is not new, the scale is.

The railway stretches from our Pilbara ports to the mines, which covers roughly 330km. It's a long stretch and train drivers can only physically drive so far before being replaced. We only have a single driver per train, with each train stretching over 2.5km long usually consisting of 230 carriages. We cannot find enough train drivers to maintain our expansion forward and the ones we are hiring are getting older and harder to replace.


They are diesel powered trucks which navigate by following the wall profile - slightly confusingly they are often called 'trams' because moving muck underground is called 'tramming' from the original medieval mine railways!


So that brings me to my original question -- what kind of sensing do you use on these trams, both to follow the walls and to detect obstacles?


Time of flight laser scanners (LIDAR). Scanners on each side of the truck measure the shape of the wall. you drive the truck down the tunnel once (in one system) it stores a map of the wall shape and then follows the same shape to keep the same distance form the walls.

Another laser scanner checks the tunnel ahead (and behind) for obstacles, other trucks, people, rockfalls etc.

On some systems a spinning laser on the top hits fixed prism reflectors with known position on the roof and walls, it constantly resects it's position from this.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: