Train drivers role disappearing?

[Rail Quest]

People have been saying that since the DLR opened 35 years ago, and probably since the Victoria line opened over half a century ago.

This was before a lot of technologies hit the scene. The rate of acceleration in the tech industry has been unprecedented since the turn of the millennia and this continues to be true with AI and ML(machine learning). As stated later in my post, I see it as unlikely that such a technology would actually be deployed on the UK network any time soon but as for the actual creation of a technology, its almost inevitable.

 

[43066]

This was before a lot of technologies hit the scene. The rate of acceleration in the tech industry has been unprecedented since the turn of the millennia and this continues to be true with AI and ML(machine learning). As stated later in my post, I see it as unlikely that such a technology would actually be deployed on the UK network any time soon but as for the actual creation of a technology, its almost inevitable.

It seems to me the question isn’t so much as whether the technology exists to get the train from a to b (we know it has done so in various forms since the sixties or before), it’s whether the expense of upgrading legacy network infrastructure, old rolling stock fleets built decades ago etc. can be justified. LU clearly think it cannot. Even with network wide intelligent ATO (and how long will that take - just look at the issues Crossrail faced), I can’t see mainline long distance trains running as UTO, so where exactly is the saving?

I could see bus drivers and lorry drivers being automated out of existence first, simply because there’s a much better financial business case to automate road vehicles: there are so many more of them, vehicle fleets are replaced more frequently and they’re far less infrastructure dependent. Even there, despite massive financial investment in developing the tech, widespread application of self driving road vehicles seems to be perpetually five years away.

 

[Rail Quest]

It seems to me the question isn’t so much as whether the technology exists to get the train from a to b (we know it has done so in various forms since the sixties or before), it’s whether the expense of upgrading legacy network infrastructure, old rolling stock fleets built decades ago etc. can be justified. LU clearly think it cannot. Even with network wide intelligent ATO (and how long will that take - just look at the issues Crossrail faced), I can’t see mainline long distance trains running as UTO, so where exactly is the saving?

I could see bus drivers and lorry drivers being automated out of existence first, simply because there’s a much better financial business case to automate road vehicles: there are so many more of them, vehicle fleets are replaced more frequently and they’re far less infrastructure dependent. Even there, despite massive financial investment in developing the tech, widespread application of self driving road vehicles seems to be perpetually five years away.

Fully agreed. I'd expect, if there ever is an implementation of such a technology it would be a really phase-heavy project that would hit key routes and perhaps only new or newish traction first, though that would take a pretty decisive government to even try and push forward.

 

[Fincra5]

Back in the 80s didn't they predict we'd all be in self-driving fly cars?

 

[Krokodil]

GoA5 or GoA6

- GoA1 is ATC/AWS/ATP/TPWS and equivalent train protection systems.
- GoA2 is self-driving with someone watching from the front and manual control of doors etc.
- GoA3 is fully automated with a human supervisor onboard 'just in case'.
- GoA4 has no onboard staff at all.

So what on earth are GoA5 and GoA6? Have you just made them up?

It's been acceptable on the DLR since 1987; the DLR is very popular with passengers and I've never seen any evidence of anyone put off using it.

Don't all DLR trains have a "Train Captain" onboard? So passengers aren't rattling around on a completely unmanned train. I highly doubt that the public would be happy with being left completely on their own in an emergency.

In order to have GoA4 you need some way to safely evacuate passengers without the assistance of a member of staff. Crossrail for example has a catwalk throughout the core. There is no way that legacy infrastructure will ever be economically converted to GoA4, never mind the signalling there's no room for said catwalks.

 

[sprunt]

Don't all DLR trains have a "Train Captain" onboard?

Yes. And there's a severely limited DLR service today because, erm, they're on strike. So that's one of OP's arguments knocked for six.

 

[yorkie]

Don't all DLR trains have a "Train Captain" onboard?

Yes they do.

So passengers aren't rattling around on a completely unmanned train.

This is the first time there's been any hint of unstaffed trains operating; that concept was beyond the scope of what I thought we were discussing.

I highly doubt that the public would be happy with being left completely on their own in an emergency.

I didn't think anyone was suggesting this?

 

[Krokodil]

This is the first time there's been any hint of unstaffed trains operating; that concept was beyond the scope of what I thought we were discussing.

I didn't think anyone was suggesting this?

What else is the OP proposing below? No wage bills and no strikes rather implies unstaffed trains (given that Train Captains and Train Operators aren't much cheaper than Drivers).

I think it will be 15 to 20 years before full automation of all lines, however long it will take it will come in eventually, I can see the attraction of it for TOCs saving wage bills, no strikes, no need to train anybody to drive...

Not to mention with his references to "GoA5" and "GoA6" which one must presume go beyond GoA4 where trains are unstaffed.

 

[yorkie]

What else is the OP proposing below? No wage bills and no strikes rather implies unstaffed trains (given that Train Captains and Train Operators aren't much cheaper than Drivers).


Not to mention with his references to "GoA5" and "GoA6" which one must presume go beyond GoA4 where trains are unstaffed.

I just took the title at face value; I've no idea what things like 'GoA5' mean but it's clearly not a well thought out idea so I didn't really think too deeply into it.

I think if someone is suggesting unstaffed trains, that really ought to be made absolutely clear; it's not the debate we were having until you picked up on it, and it makes it even more of a bizarre suggestion.

 

[Krokodil]

I just took the title at face value; I've no idea what things like 'GoA5' mean but it's clearly not a well thought out idea so I didn't really think too deeply into it.

Grades of automation are detailed here:

Automatic train operation - Wikipedia

en.wikipedia.org

I think if someone is suggesting unstaffed trains, that really ought to be made absolutely clear; it's not the debate we were having until you picked up on it, and it makes it even more of a bizarre suggestion.

I believe that the comments of "passengers won't accept this" that you were replying to were in reference to unstaffed trains. I infer this because - as you say - automated but supervised trains are long-established, since public testing was done on a branch of the Central Line in 1964. If said posters wish to come back and say that they instead meant "passengers won't accept any non-human form of train control" then I await their move.

 

[Vespa]

- GoA1 is ATC/AWS/ATP/TPWS and equivalent train protection systems.
- GoA2 is self-driving with manual control of doors etc.
- GoA3 is fully automated with a human supervisor onboard 'just in case'.
- GoA4 has no onboard staff at all.

So what on earth are GoA5 and GoA6? Have you just made them up?

When I said GoA5 and GoA6 this would be the hypothetical future classification of automatic dynamic running with no human input, the next level would definitely be AI related.

 

[Krokodil]

When I said GoA5 and GoA6 this would be the hypothetical future classification of automatic dynamic running with no human input, the next level would definitely be AI related.

So you have just made them up? And you think that there will not only be unstaffed trains, but an unstaffed railway to? Whatever you have been smoking with Grant Shapps, I wouldn't book on for a safety-critical job with it.

Let's head back to GoA2, where a human remains in the front of a train (as per Victoria Line, Elizabeth Line, Thameslink etc.). In order to implement this you need ETCS. Do you seriously think that the entire UK network will be operating on ETCS in 10-15 years time? We still have many sections of Absolute Block which will still be in operation then, never mind the likes of Plymouth Panel which have no end date planned. Nationwide installation of ETCS will make HS2 look like a bargain.

Then let's consider GoA3, where a human is present on the train but their operational role is mostly limited to emergencies. Therefore there's no one looking out of the window ahead. In order to make this safe, you need the railway to be effectively secured from anything that may get in the path of a train. That means that the line must be properly fenced (P-Way regularly have to chase off sheep up here because the dry stone walls are falling down), trees have to be cut back so that they can't fall onto the line, and level crossing risk must be eliminated, including the closure or upgrade of all occupation and automatic level crossings. How much do you think this will cost to do?

GoA4. No human onboard at all. In the event of an emergency the passengers are on their own. In order to evacuate them safely without staff assistance you will need a catwalk along the length of the track, widening the formation where necessary. How much do you think that it will cost to expand the bore of our Victorian tunnels? You're probably coming close to doubling the national debt now. There is no way that labour savings are going to come close to paying for all of this.

And that's apart from the question of whether the public will accept hurtling along at 125mph with no human onboard, let alone one watching ahead. If cyber security experts have said that they wouldn't trust an unstaffed train, I certainly wouldn't either.

 

[Vespa]

So you have just made them up? And you think that there will not only be unstaffed trains, but an unstaffed railway to? Whatever you have been smoking with Grant Shapps, I wouldn't book on for a safety-critical job with it.

Let's head back to GoA2, where a human remains in the front of a train (as per Victoria Line, Elizabeth Line, Thameslink etc.). In order to implement this you need ETCS. Do you seriously think that the entire UK network will be operating on ETCS in 10-15 years time? We still have many sections of Absolute Block which will still be in operation then, never mind the likes of Plymouth Panel which have no end date planned. Nationwide installation of ETCS will make HS2 look like a bargain.

Then let's consider GoA3, where a human is present on the train but their operational role is mostly limited to emergencies. Therefore there's no one looking out of the window ahead. In order to make this safe, you need the railway to be effectively secured from anything that may get in the path of a train. That means that the line must be properly fenced (P-Way regularly have to chase off sheep up here because the dry stone walls are falling down), trees have to be cut back so that they can't fall onto the line, and level crossing risk must be eliminated, including the closure or upgrade of all occupation and automatic level crossings. How much do you think this will cost to do?

GoA4. No human onboard at all. In the event of an emergency the passengers are on their own. In order to evacuate them safely without staff assistance you will need a catwalk along the length of the track, widening the formation where necessary. How much do you think that it will cost to expand the bore of our Victorian tunnels? You're probably coming close to doubling the national debt now. There is no way that labour savings are going to come close to paying for all of this.

And that's apart from the question of whether the public will accept hurtling along at 125mph with no human onboard, let alone one watching ahead. If cyber security experts have said that they wouldn't trust an unstaffed train, I certainly wouldn't either.

Of course I made it up, because it doesn't exist yet remember this is a hypothetical thread !

The more frequent and faster the train is, can a human process all the visual information and act according within time, that is the question.

 

[43066]

When I said GoA5 and GoA6 this would be the hypothetical future classification of automatic dynamic running with no human input, the next level would definitely be AI related.

That doesn’t really make sense, as GoA4 is fully unattended operation. There isn’t another stage beyond that, however smart the automation is.

In any case, I think your initial suggestion that we will have driverless, let alone fully unattended trains widespread on the network within 15-20 years has been comprehensively debunked.

 

[brad465]

However much support there is for automated trains on the national rail network now, it will go down as soon as a cost is announced, and I can easily see the total cost exceeding HS2, even without the cutbacks, all things considered. Then of course the inevitable overruns and ballooning costs will further turn away support.

 

[sprunt]

When I said GoA5 and GoA6 this would be the hypothetical future classification of automatic dynamic running with no human input, the next level would definitely be AI related.

What does this even mean? What value are you expecting AI to add? It seems like you've just mentioned it because it's a bit of a current buzzword.

 

[Sorcerer]

What does this even mean? What value are you expecting AI to add? It seems like you've just mentioned it because it's a bit of a current buzzword.

One might assume that AI drivers could detect any disturbances on the rails ahead such as people or objects on the line and bring the train to a stand still. That said, even assuming that was possible or not already covered, I'm not sure most people would trust an AI robot to control a high speed train. I think the public opinion is irrelevant though to be honest because even if we could automate the national network the cost would likely be beyond labour savings of no having train drivers. I think driverless trains from GoA1 to GoA4 have their place on self-contained systems such as metros, but not much more beyond, so I don't think AI or beyond GoA4 is even worth thinking about.

 

[Sly Old Fox]

Have you seen how good they are at electrifying a railway?

 

[Krokodil]

Of course I made it up, because it doesn't exist yet remember this is a hypothetical thread !

Hypothetical? It's delusional!

 

[Vespa]

Hypothetical? It's delusional!

Not sure why you're getting worked up over it, it's a hypothetical thread in a "Speculative Discussion" thread, crayons aplenty from the box, it's perfectly valid discussion to cover pros and cons, AI technology is progressing quickly and it's been theorised that it may reach a singularity point before we know it.

You need to take a step back and chill a bit.

 

[Krokodil]

it's a hypothetical thread in a "Speculative Discussion" thread, crayons aplenty from the box, it's perfectly valid discussion to cover pros and cons, AI technology is progressing quickly and it's been theorised that it may reach a singularity point before we know it.

I'll file it alongside "build a viaduct from London to New York".

Not sure why you're getting worked up over it,

Is robust and informed critique not allowed?

 

[Vespa]

Is robust and informed critique not allowed?

To a point as long as it debative, when you get emotional you lose focus.

I can understand the fear of losing jobs and encroachment of technology potentially taking away humans from job roles.

 

[Krokodil]

To a point as long as it debative, when you get emotional you lose focus.

I can understand the fear of losing jobs and encroachment of technology potentially taking away humans from job roles.

"Emotional"? Well I suppose that finding something hilarious could be an emotion.

"Fear"? What have I got to fear? When I come to retire in 30-40 years time (never mind the 15-20 years you think "full automation" will be done in) , I am confident that GoA3 will cover less than half of route mileage in the UK (it's pretty likely that there will even be a few pockets of mechanical signalling, such is this country's inability to get anything done. There will certainly be plenty of level crossings left. GoA4 will be limited to a tiny handful of situations using newer infrastructure (such as the Crossrail Core).