Slashdot Mirror
Australian Autonomous Train is Being Called The 'World's Largest Robot' (sciencealert.com)
schwit1 shares a report: Mining corporation Rio Tinto says that an autonomous rail system called AutoHaul that it's been developing in the remote Pilbara region of Australia for several years is now entirely operational -- an accomplishment the company says makes the system the "world's largest robot." "It's been a challenging journey to automate a rail network of this size and scale in a remote location like the Pilbara," Rio Tinto's managing director Ivan Vella told the Sydney Morning Herald, "but early results indicate significant potential to improve productivity, providing increased system flexibility and reducing bottlenecks." The ore-hauling train is just one part of an ambitious automation project involving robotics and driverless vehicles that Rio Tinto wants to use to automate its mining operations. The company conducted its first test of the train without a human on board last year, and it now claims that the system has completed more than a million kilometers (620,000 miles) of autonomous travel.
What the hell is so challenging about automating trains? I can't believe train conductors are still a thing, and they're still crashing trains. What's simpler to automate than a train? The tracks are fixed. There are very few tracks or trains in any system. The trains can only go two directions on the tracks. Why aren't all trains automated by now?
I don't respond to AC's.
While a million automated kilometers (of automated operation) may sound impressive in a headline, I doubt that the statistic really means anything. For a train running on a track, the distance traveled isn't very interesting, particularly if the track runs through the middle of nowhere. I would think statistics on numbers of automation decisions made would be more useful, particularly decisions that would have otherwise been made by a human operator. Perhaps statistics on number of grade crossings (where roads cross the tracks) traversed; counts on how many times the train sounded the warning horn, slowed down or stopped, because there were animals or people on the tracks; statistics on how the automation handled other abnormal events such as sticking breaks, loss of cargo, or other mechanical failures.
Everything else is just semantics. Do a mechanical robot have to look like a human, or is it enough if the robot autonomously does the job he was designed for?
It's kind of funny. While American companies fuss with the uphill battle and endless stream of edge cases that self-driving represents, the Australians will be collecting reams of useful test data because it is a case that is fully realizable and profitable. Seems they are more interested in building on top of demonstrable successes rather than making empty promises.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
A "semi autonomous" large ore train had to be deliberately derailed in November, because it was actually less destructive than letting it continue driving and come close to the "real" rail network or civilisation.
More info at https://www.abc.net.au/news/20...
So it's probably too early to claim success for autonomous trains, even though, as stated by earlier posters above, an autonomous train in the outback is a much easier challenge than one in the city. Far fewer level crossings, obstacles or pedestrians.
sustainable living
You may be joking with that comment, but with Australian unions this could actually be the case. For example Boeing had to redesign the two-man cockpits on their 767s to contain an unnecessary third person because the Australian unions demanded it, the only three-man 767 cockpits ever shipped.
The trains are being drive remotely, rather like the London Tube trains have been for years. There's still someone at the controls, but no one on the footplate. This means you can change staff half-way through a trip, without requiring a staff-car attached.
Given that in the news too, is where BHP derailed a train that ran away from the driver (who was inspecting the train), they used remote signalling control to throw a set of points and run it into a passing loop with no escape, it would be interesting. But this is a run-away train, the sort that has happened many times before on suburban networks.
I imagine that to forfill the full function of the driver, one needs to deal with the likes of hot axle boxes, cracked and broken tyres, and all sorts of other things before it would be fully remote.
OS/2 - because choice is a terrible thing to waste.