Australian Autonomous Train is Being Called The 'World's Largest Robot' (sciencealert.com) 135
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.
Re: (Score:2)
AutoHaul? (Score:1)
And it has a built-in autonomous washing system called AutoWash, though the engineers refer to it as "Leeloominaï Lekatariba Lamina-Tchaï Ekbat De Sebat" for some reason.
Re: (Score:1)
Damn you, Forman! You stole the joke I wanted to write!
Re: (Score:2)
Why is everything a robot? Because we've entered the future where they are becoming ubiquitous. This train is far more autonomous than the pre-programmed "robot" arms of the 1960s.
Re:Why is everything a robot? (Score:4, Informative)
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?
Re: (Score:2)
Re: (Score:2)
Yeah, if you want to go back to origins, his "robots" were artificial biological life forms engineered to serve humanity - so we have no robots by that definition.
Re: (Score:2)
Rabota is the Slavic root for 'work' in general.
Re: (Score:2)
Apek's robot looked like a human (because it was one) but the first real robots in the 1940s were basically clunky Roombas and CNC machines.
"Android" is an even older term that specifically refers to more or less human-appearing robots. Wikipedia notes that the first known use of the term was describing a wooden toy that looks like Pinocchio riding a bicycle.
Re: (Score:1)
Re: (Score:2)
Perhaps you should read a bit about the history of the term "robot?"
The Dark Tower... (Score:2)
Re: (Score:1)
Why is biggest interesting? (Score:1)
I'd be more impressed if it was the smartest. But that's probably harder to prove.
And it's a train, right? On rails? So it does't even have to steer. Just start and stop and the right places.
Re:Unions ... (Score:4, Informative)
"challenging"? (Score:5, Insightful)
Re: (Score:2)
Nit to pick: "Conductor" is the person opening the doors and taking tickets, or "managing" a freight train. "Driver", "engineer", or "motorman" is the person actually driving the thing.
As far as automation, it doesn't hurt to have a set of eyes out front if there's a person on the tracks or a car about to get stuck in a level crossing.
Re: (Score:1)
Re: (Score:2)
Also, engineers do more than starting and stopping the train. Traditionally, they're responsible for taking care of th
Re: (Score:2)
Unions, from the original Article [smh.com.au]
"The Autohaul program has drawn the ire of unions worried about its impact on train driver jobs but the company said it had not made any forced redundancies and did not expect to make any in 2019."
Re: (Score:2)
Driving a train is literally just "speed up" and "slow down". That's it. Why is taking so long to automate train drivers away?
It's knowing when to speed up and more importantly slow down. It's detecting that there is something on the tracks that shouldn't be (and slowing down if so.) Granted obstacle detection is likely simpler than for a car, since the train follows a limited path and you don't need near as many "maps" that define what "normal" is as you would for automobile automation, but you still need to detect if you should perform an unscheduled stop. Driving a train also involves monitoring the mechanical performance - rep
Re: (Score:2)
Trains can't actually stop for stuff on the tracks. They stop too slowly. They stop after hitting something regardless of when you start applying the brakes. They have to stop so the authorities can do the required investigation, but the stopping distance/reaction time rarely matters.
Knowing when to speed up and slow down is done by reading the speed signs along the route, and hopefully following them. Accidents happen because they were ignoring the signs; either due to being macho assholes, or being distra
Re:"challenging"? (Score:4, Informative)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
An object weighing gigatons has gigatons of weight on its wheels to help it stop, assuming that it has decent brakes on each axel. It is not harder or easier for a big object to stop than a small one.
Now, a faster object is quite a different matter. Ships are a bit different due to the nature of their friction.
Re: (Score:1)
Brakes don't work by applying weight to the axles, you know that right?
Re: (Score:1)
Re: (Score:1)
There are actually a shit load of reasons. obviously a fully loaded train brakes massively differently to an unloaded one, but I am assuming you mean length providing the difference in weight. Braking between carriages will not be uniform/consistent due to a range of factors (wear/heat/weight distribution etc), this results in additional force being transmitted to the other carriages. The more carriages you have the less efficient your brakes becomes thus extending the braking distance. These trains are ove
Re: (Score:1)
The amusing part of your comment is saying ships are different, It is the same fucking problem. The friction between the train wheels and track is limited and therefore the amount of braking you can apply is also limited as mass and momentum increases so does the distance for braking as the surface area between the tracks and wheels does not change with weight.
Re: (Score:3)
You don't actually know anything about trains, and talking based on intuition isn't working. Big freight trains have ridiculously long stopping distances. A 20k ton ore train will need damn near 2 miles to stop from 55 mph on level ground. The notion that a train driver could see something on the tracks and stop or even slow down enough to make a difference is entirely incorrect. A stuck brake (spelling matters) is not going to make vibrations that a driver can feel. A broken coupling will cause the train t
Re: (Score:2)
Said the Slashdot AC suffering from the Dunning-Kruger effect.
Re: (Score:2)
A sensor breaks down or just reports wrong values somewhere or some 'simple' logic in the programming doesn't cover an infinite number of 'unforeseeable circumstances' the new better and cheaper than a human robot will cause an accident as well.
This is nothing new in engineering, human failures in imagination have caused innumerable disasters over history
Cue someone coming in and blithely saying 'but bounds checking and error control you eediot11!!', and the answer is you cannot foresee all the possibilit
Re: (Score:2)
Was just thinking about this some more, and I dont know if the average commercial or custom robot control system is hardened to external bit shifts, that happens often enough to be a worry.
Re: (Score:2)
the answer is you cannot foresee all the possibilities, its not a controlled lab, even then surprising problems can arise that were not even considered in planning.
There is nothing you can do about possibilities that were unforeseen. Trains don't react fast enough for that to even be a thing. You would absolutely need planning for the human in the train to be able to do anything to solve some sort of problem.
Re: (Score:2)
the answer is you cannot foresee all the possibilities, its not a controlled lab, even then surprising problems can arise that were not even considered in planning.
There is nothing you can do about possibilities that were unforeseen. Trains don't react fast enough for that to even be a thing. You would absolutely need planning for the human in the train to be able to do anything to solve some sort of problem.
And I consider that a perfect example of lack of imagination, you are suggesting that there is 'nothing ever' that could be done? Certainly if it was the best robot 'AI' today then I agree, but a human can do orders of magnitude more than the best AI today let alone a good expert system robot.
https://en.wikipedia.org/wiki/... [wikipedia.org]
https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:2)
Again, complete lack of imagination, the fact that you can put forward of 1 scenario out of an infinite supply of them is nonsense, the point of what i'm saying is that a human is much more capable of handling random problems, certainly we can get robotic control to some number of 9's, for the routine operation 99.99% who knows but something out of the ordinary happens 1/10000 then the computer at best can just brake, report a failure and stop.
And pay attention, i'm not arguing against computer controlled t
Re: (Score:2)
But you won't have engineers exceeding the speed limit causing the train to derail such as this one [cnn.com] where the train was going 80 mph in a 30 mph zone.
Re: (Score:2)
Yes, this has all been worked out before, and there are still train crashes and derailments. And now you want to start almost from scratch and redesign all the safety critical systems for trains to be fully automatic with no human intervention, and you think that will automatically solve problems that cause crashes and dera
Re: (Score:2)
Most safety critical systems on trains ARE fully automatic.
About 75% of train accidents are due to human error. The situation is similar with aircraft.
Re: "challenging"? (Score:2)
Because it's the real world, that's why.
People fear automation as the next big thing that will take their jobs. But it isn't. It's been the thing that can take their jobs for atleast the last 50 years. It just hasn't for non-technical reasons. Technical ability has the most say in why something isn't automated but the least say in why something IS automated.
When I audit business processes, I see atleast 3/5 positions or stages that could have been replaced by automation 10-20 years ago.
Forget my personal e
Re: (Score:2)
Try 250 years. Used to be almost everybody worked in agriculture, because that was the only way mankind could feed itself, and even then it wasn't enough now and then. Farm machinery eliminated the vast majority of those jobs, making it so that only a small percentage of the population could easily keep the rest fed. (yes, there are still famines
The future aint what it used to be (Score:2)
Machines traditionally only replaced very repetitive jobs. But over the next 50 years they are going to become much smarter.
Sure, over the next 20 years most of the truck drivers etc. will find some sort of underpaid work elsewhere. But over the next 50 there will be very little unskilled work.
But, as per Parkinson, bureaucracies will grow and grow to take up the slack from those with mediocre intelligence.
And then, maybe in 200 years, computers will be able to program themselves, and will no longer need
Re: (Score:2)
What the hell is so challenging about automating trains?
The effect of failure causes an immediate and very high risk to safety and lives. Especially an iron ore train. Rio Tinto's competitor BHP has only recently shown what needed to be done when a the automation fails: https://www.abc.net.au/news/20... [abc.net.au] After forcing the train to derail the result was 1.5km of damaged railway and a huge mess to clean up, combined with a lot of luck that in this remote part of Australia it's possible to derail a train without injuring people.
There are many automation projects for
Re: (Score:2)
Why aren't all trains automated by now?
Your question is a microcosm for all the idiocy behind the self-driving push.
By the way, the answer to your [rhetorically-intended] question is "Exactly."
Re: (Score:2)
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?
One crashed earlier this year. [abc.net.au]
Having lived up there the challenge is in both the distance they have to travel, the sheer length and weight of the train as well as the method of loading.
The trains are travelling through the most inhospitable places on earth. Unlike most other automated trains these are not well fenced in self contained units within easy reach of a control team. The trains will travel 400+ KM though areas that can get in excess of 40 degrees in the day and can have a temperature vari
Better metrics needed (Score:4, Insightful)
Re: (Score:2)
A robot operator does all that a lot better. Detecting obstructions and mechanical failures is a lot better with the right sensors. And even so, for a train it doesn't matter much what is on the tracks, by the time it's detected it's probably too late.
Re: (Score:2)
A robot operator does all that a lot better. Detecting obstructions and mechanical failures is a lot better with the right sensors. And even so, for a train it doesn't matter much what is on the tracks, by the time it's detected it's probably too late.
I don't disagree. My point is that to make the argument for automation, more meaningful statistics should be used.
Big deal (Score:2)
Important Details (Score:1)
This company has 400(!) train drivers with some making $240,000/Yr.(!)
They claim to have not let anyone go(forced redundancies), yet. But, no company is going to continue paying $50million per year if they don't have to.
I'm amazed that a train driver(engineer) is paid so much. I'm amazed that they have 400 drivers for 200 trains.
I'm wondering how people will find 400 such highly paid jobs in the near future and forever after.
Re: (Score:2)
Re: (Score:2)
RIO Tinto does mining almost in the middle of nowhere, true dessert / wilderness stuff. The trek is probably at least 2 days if not 3 or 4 solid to get back to civilisation.
These guys with a little upskilling will be able to do a 'normal' train driver job (again?) which may be a downgrade of a full 50%, but it's still high paying work.
With the insanely off the chart, ridiculous immigration ponzi Australia is running, surely the major cities are hiring more and more train drivers? Right?.... Surely...
Re: (Score:2)
I think I read the ending to this (Score:1)
Meanwhile Our Governments... (Score:1)
Automated Driving to the Aussies (Score:4, Interesting)
Re: (Score:2)
What are you talking about. There's a world of difference between autonomous cars making a decision on the go on public roads, and a set of carriages able to chose between going forwards or backwards along a 1700km of straight tracks with a couple of privately controlled and operated intersections.
Re: (Score:2)
Putting it another way, painting the Mona Lisa won't make you Leonardo Di Vinci. You have to start with the basics and work up.
Re: (Score:2)
And just when should good old Leo stop playing with his basics? We have fully autonomous rail systems all over the world and have had so for many years.
Not quite ready for prime time (Score:4, Informative)
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... [abc.net.au]
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.
Re:Not quite ready for prime time (Score:4, Insightful)
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... [abc.net.au]
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.
Where in that article does it even mention semi-autonomous? The article talks bout a driver getting out of the cab and the train taking off. Nothing to do with the Rio Tinto trains.
Re: (Score:2)
OP here. Yes, that might have been where I got mixed up. I guess you could say the BHP crash was remotely instigated, although the original reason it had to be crashed at all was human (driver) error.
Re: (Score:2)
BHP crash was remotely instigated
You can't really say that either from that source. Orchestrated and instigated are two different words. But while they did instigate the crash they had no control over this train what so ever, it was not at all automated. They derailed the train movie style, by switching the tracks while the train was already passing through a junction.
That's no more semi-autonomous that cars are due to the presence of traffic lights.
Re: (Score:2)
I ride Metra every day, and they are also one of our biggest clients. I'm pretty sure none of their trains are remote controlled, even in the yards.
Re: (Score:2)
Is it cheaper to derail the occasional train (Score:2)
Re: (Score:2)
Re: (Score:2)
A "semi autonomous" large ore train had to be deliberately derailed in November
No it didn't. A non-autonomous runaway train with no controls and no driver had to be deliberately derailed in November.
You're talking about a different project, different track, all run by a completely different company, and better still a completely different topic (autonomous vs driver controlled).
Also it's the first useful blockchain (Score:1)
Transports blocks of passenger in a chain, can only add, not remove (otherwise its a split i.e. two blockchains). Also usually issues its own tockens, called tickets.
Remote, yes, Robot, no. (Score:3)
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.
Re: (Score:2)
The Pilbara routes are short enough to do the trip in a day. But runs typically in the 600 mile run get staff-cars. The Westlander and Inlander in Qld run staff-rooms in the van. The main reasons for putting on staff-cars are to be able to close remote engine-sheds, like Charleville and Cloncurry.
My brother worked coal trains, and they would depending on the company, work to Mooronbah (where quarters were provided), or crew change at points halfway between depots, so in and out of copperbella.,
It's i
Re: (Score:2)
The runs on the LU are too short to do staff-changes in this way. It's easier to just have the new crew take over at some station. Most suburbab trains here don't have staff-cars. Staff who are travelling to a point do so in the passenger cars.
In Australia, there are runs from Melbourne to Darwin, or Sydney to Perth, which are handled by sets of on-board crew, these change at various places where there is no staff-room or even town. Because there are a lot of companies operating, each company would pr
The war on miners continues (Score:2)
All you folks who used to work here? G'day, and don't let the door hit you on the way out. What will you do for a living? Sorry, we don't care about that, that's got nothing to do with ROI for our CEO.
Those silly drivers... (Score:2)
https://edition.cnn.com/2018/1... [cnn.com]
Solution: get rid of the drivers!