driverless trains on the network

[stewartingram]

The Piccadilly Line is still driven conventionally.  It's the Northern Line which has the same system as the Jubilee and DLR.

The Victoria was the first to go automatic in 1968, and was recently re-equipped during the massive refurb which also saw the trains replaced.

The Central was equipped with another system later. On Sundays the drivers are encouraged to drive manually to keep up the knowledge. 

The Jubilee was equipped with the Thales system during the recent upgrade this century, followed by the Northern. At this time, the sub-surface lines, (Met, H&C, and District) are being done with the same system; as I've now retired I'm not sure how much is actually in use though I know some sections have been accepted.

The remaining lines - Bakerloo, Piccadilly, Waterloo & City - should also follow these upgrades in the future.

(The Circle is not actually a separate line, being just a "cuckoo in the nest" service on the other sub surface lines, so is actually included in the conversion as well).

The separate DLR has also been automatic since it first opened. The on-board staff perform a driver role in emergency situations.

So the Underground has 3 different systems, and the DLR has another.

 

Stewart

[njee20]

 

 

I thought the proponents were on about conventional lines and doing away with me and my colleagues? 

 

 

Yes, ultimately, through use of radar/lidar. Which have the potential to be significantly more effective than a human.

It's an emotive subject (as your post shows), but people make a lot more mistakes than computers. If done right then the railway capacity and safety would both increase.

 

There will now ensue 50 "ahhh, but how would a computer deal with x scenario", which is a waste of breath. I'm not talking about putting C3PO in the cab of an Electrostar and having away, it's a long term thing, but naive to assume that a) it can't be done and b) it can't be done well. 

 

 

Edited March 28, 2018 by njee20

[Reorte]

Yes, ultimately, through use of radar/lidar. Which have the potential to be significantly more effective than a human.

It's an emotive subject (as your post shows), but people make a lot more mistakes than computers. If done right then the railway capacity and safety would both increase.

I'm starting to fear the societal impact of making humans increasingly irrelevant, and viewing them as a nuisance to be done away with as soon as possible, more than any safety concerns. Taken individually the cases for most of this technology looks sensible but the cumulative effect, I'm really having doubts.

[The Stationmaster]

 

 

Yes, ultimately, through use of radar/lidar. Which have the potential to be significantly more effective than a human.

It's an emotive subject (as your post shows), but people make a lot more mistakes than computers. If done right then the railway capacity and safety would both increase.

 

There will now ensue 50 "ahhh, but how would a computer deal with x scenario", which is a waste of breath. I'm not talking about putting C3PO in the cab of an Electrostar and having away, it's a long term thing, but naive to assume that a) it can't be done and b) it can't be done well. 

 

 

It would have to some remarkably clever radar, or anything else, to actually know the route, recognise staff lineside in HV clothing and sound a warning, be able to recognise any emerging hazards such as a bank slip at the lineside, and various other things.  I think on the generally curvaceous British railway network with various train in close proximity on curved track such a system is going to take an awful lot of developing probably using various techniques and methodology that don't yet even exist.  A degree of automation might well reduce the incidence of SPADs (although even then increasing human reliance on it could be a double-edged sword as it encourages the Driver to pay even more attention to his mobile 'phone than the road ahead).

 

To be honest we are going into realms which are, in the broad picture, way beyond the sort of thing we I have thus far seen on an ordinary mainline railway in Britain and I suspect taht as ever the biggest problem might not necessarily br staff objections but public fears which will not accept unmanned train driving cabs.  And it's all a lot further away than some might like to think

Edited March 28, 2018 by The Stationmaster

[Zomboid]

I doubt it'll happen on any great scale for a long time. Simply because on most of the network there would be little to no benefit.

 

I could forsee some of the highly trafficked sections of trunk routes benefitting (Waterloo - Woking, for example), but not everywhere, and not particularly soon.

[phil-b259]

 

I find it difficult to believe that the central core of Thameslink has such a low maximum speed, but it is possible.

 

The relevant sectional appendix containing line speeds south of Faringdon can be found here.http://archive.nr.co.uk/browse%20documents/sectional%20appendix/sectional%20appendix%20full%20pdf%20copies/kent%20sussex%20and%20wessex%20sectional%20appendix.pdf

 

Scroll to pages 346 & 347. They confirm 20 / 30mph to be the maximum.

 

Faringdon and places north is covered by the East Midlands route - and I cannot find their sectional appendix at present. However occasional travels on that route leaves me to believe that speeds do not change significantly till beyond St Pancras

Edited March 28, 2018 by phil-b259

[njee20]

It would have to some remarkably clever radar, or anything else, to actually know the route, recognise staff lineside in HV clothing and sound a warning, be able to recognise any emerging hazards such as a bank slip at the lineside, and various other things. I think on the generally curvaceous British railway network with various train in close proximity on curved track such a system is going to take an awful lot of developing probably using various techniques and methodology that don't yet even exist. A degree of automation might well reduce the incidence of SPADs (although even then increasing human reliance on it could be a double-edged sword as it encourages the Driver to pay even more attention to his mobile 'phone than the road ahead).

Of course it would have to be clever. Again, not suggesting you chuck a robot in the cab of a 377 with a webcam and send it up the Brighton main line tomorrow. People don’t see very well around corners, so that argument is daft. A network of trains scanning everything around them (like driverless cars) can create an infinitely more detailed picture, and share it with every other vehicle, far better than one bloke sat in the cab.

 

It isn’t going to happen tomorrow. It won’t happen in 10 years. But it’s naive to assume that it can’t happen.

[caradoc]

It would have to some remarkably clever radar, or anything else, to actually know the route, recognise staff lineside in HV clothing and sound a warning, be able to recognise any emerging hazards such as a bank slip at the lineside, and various other things.  I think on the generally curvaceous British railway network with various train in close proximity on curved track such a system is going to take an awful lot of developing probably using various techniques and methodology that don't yet even exist.  A degree of automation might well reduce the incidence of SPADs (although even then increasing human reliance on it could be a double-edged sword as it encourages the Driver to pay even more attention to his mobile 'phone than the road ahead).

 

To be honest we are going into realms which are, in the broad picture, way beyond the sort of thing we I have thus far seen on an ordinary mainline railway in Britain and I suspect taht as ever the biggest problem might not necessarily br staff objections but public fears which will not accept unmanned train driving cabs.  And it's all a lot further away than some might like to think

 

Valid points, and in addition, if there was really no human in the cab, how would incidents such as fatalities (unfortunately far too common), animals on the line and suspected track defects, to give just three more examples, be recognised, let alone dealt with ? Of course, in a perfect world, or a totally enclosed, protected and fault-free railway system, such incidents would not occur, but that is not the world we live in.

[The Stationmaster]

Of course it would have to be clever. Again, not suggesting you chuck a robot in the cab of a 377 with a webcam and send it up the Brighton main line tomorrow. People don’t see very well around corners, so that argument is daft. A network of trains scanning everything around them (like driverless cars) can create an infinitely more detailed picture, and share it with every other vehicle, far better than one bloke sat in the cab.

 

It isn’t going to happen tomorrow. It won’t happen in 10 years. But it’s naive to assume that it can’t happen.

 

Not a daft argument at all - as anybody who has ever set in a driving cab will tell you.  Railway lines curve about, they have points which take you to other lines and that also involves curving about - at times the detection system would be looking head on at walls, trains coming the other way and trains going in the same direction on apparently converging courses.  obviously the system will know if the train has been given a proceed signal but it won't know about a wall or stretch of security fence unless every last possible combination of lineside and infrastructure detail has been programmed into it.  and how does it recognise HV clothing on staff working lineside or, far more so, in the four foot - especially when they're round a long sweeping curve?

[njee20]

Look at what driverless cars are doing. It’s infinitely less complex on the railways. HV makes it easier to recognise people, not harder. How does a driver recognise a defect on a parallel track at 125mph? He doesn’t. Simple. No reason an autonomous train couldn’t.

 

How do animals and fatalities get recognised? With the same radar systems that are scanning the track anyway, that’s the point. With the added advantage that as soon as one train detects animals/people/any obstruction lineside they can immediately alert all other trains to brake if necessary, rather than having to rely on some sort of communication via signalling centres and the signals, which delays things. Detection can be 360 degree, not just what is in front of the driver for the period in which he can see and the area he happens to be looking at.

 

It’s always the same vastly flawed arguments. People are not perfect. The technology exists, it’s not in trains, it may never be, if the benefits aren’t deemed sufficient, but it could happen. The barrier will not be how much more awesome people are, that’s ludicrous.

[jjb1970]

Trains running South of Farringdon are very slow. Not helped by the very tight station spacing between Farringdon, City Thameslink and Blackfriars. I really can't see the point of City Thameslink as the platforms aren't far off touching the stations either side.

[phil-b259]

Not a daft argument at all - as anybody who has ever set in a driving cab will tell you.  Railway lines curve about, they have points which take you to other lines and that also involves curving about - at times the detection system would be looking head on at walls, trains coming the other way and trains going in the same direction on apparently converging courses.  obviously the system will know if the train has been given a proceed signal but it won't know about a wall or stretch of security fence unless every last possible combination of lineside and infrastructure detail has been programmed into it.  and how does it recognise HV clothing on staff working lineside or, far more so, in the four foot - especially when they're round a long sweeping curve?

 

It cannot - but the DLR doesn't kill people on a daily basis despite not having a driver to look out for the dangers you mention.

Once again I fear people are not thinking things through properly.

 

Lets have a little look at the DLR shall we?

 

It has:-

 

No level crossings

No freight traffic

High quality fencing (and plenty of viaducts) which reduce trespass fears

Persons cannot undertake, what is know as 'Red Zone working with lookouts' while on track - any engineering work requires trains to be stopped.

Relatively moderate maximum speeds

In an area not prone to extremes of weather or where landslides are a risk.

 

In fact everything that currently applies to the Thameslink core and large chunks of the Underground network.

 

HS1, and in future HS2 will also by their very construction and high speed operation will have most of these mitigating measures inbuilt - though as with aircraft passenger resistance to going driverless is likely to be the deciding factor.

 

Of course if you need drivers either end of the journey (as is the case for Thameslink) then there is very little point ditching them for a tiny section of the journey - but hypothetically you could see something like a train start from a London termini as a driverless one, working in auto mode till it gets to a certain place where the driver boards for the rest of the journey.

 

While I agree that the bulk of the conventional UK at present is unsuited to driverless operation, that does not mean that certain sections are not - nor that as upgrades take place and technology improves, that driverless trains will not be possible on a wider scale. As ever it is wise to keep an open mind on such things - if anything, history teaches us that automation is not going to go away.....

[letterspider]

Is driving a train such a bad occupation that the driver has to be replaced by a robot?

Or is it the cost?

I would suggest replacing much of the senior management team with robots - most of those tasks can be automated and the cost savings would be far greater.

[njee20]

Was being a telephone operator that bad? Surely replacing the executives at telecoms companies would be far better.

[phil-b259]

Valid points, and in addition, if there was really no human in the cab, how would incidents such as fatalities (unfortunately far too common), animals on the line and suspected track defects, to give just three more examples, be recognised, let alone dealt with ? Of course, in a perfect world, or a totally enclosed, protected and fault-free railway system, such incidents would not occur, but that is not the world we live in.

 

Trains do not have a steering wheel! - if a train driver sees an obstruction on the track or something that would endanger their train, then unless they are going very slowly then they WILL HIT IT.

 

I'm sure its not beyond the whit of engineers to fit suitable sensors and monitoring systems that will take the necessary emergency actions should the worse occur.

 

However what is needed in such a situation is a human to act as reassurance and to assist passengers - I don't know how about calling them something like a Guard or Conductor.

 

I'm sorry, but with the technology we have available these days, doing away with the driver and keeping the Guard (as happens on the DLR) is the long term way forward - which is kind of ironic given which grade is under attack by the DfT

[royaloak]

I'm starting to fear the societal impact of making humans increasingly irrelevant, and viewing them as a nuisance to be done away with as soon as possible, more than any safety concerns. Taken individually the cases for most of this technology looks sensible but the cumulative effect, I'm really having doubts.

Who will pay for everything when we are all on the dole because we have been replaced by R2D2 so just sit at home all day with nothing to do except sign on every fortnight?

[phil-b259]

Is driving a train such a bad occupation that the driver has to be replaced by a robot?

Or is it the cost?

I would suggest replacing much of the senior management team with robots - most of those tasks can be automated and the cost savings would be far greater.

 

Its not a bad occupation, but in the eyes of certain political commentators and think tanks they are overpaid. Getting rid of them (as with Guards) saves money and breaks the power of the train drivers trade union to bring things to a halt.

 

However it is a fact that human reactions are far slower than electronics - 24tph through the Thameslink core would not be possible without all driving functions being removed from human operation and with the driver reduced to closing the doors.

 

Its well known that outer London commuter lines are at the limit of what they can handle with conventional signalling and manual control - Converting the Brighton mainline to ATO between Victoria and Croydon would unleash lots of extra train paths by saving vital seconds here and there (and thus save expensive infrastructure capacity enhancement schemes at Clapham Junction say). Such a proposal looks very attractive in the corridors of the DfT and HM Treasury as a result....

[phil-b259]

Who will pay for everything when we are all on the dole because we have been replaced by R2D2 so just sit at home all day with nothing to do except sign on every fortnight?

 

Good question - but one the Politicians have failed to answer so far.

 

There is a saying that just because you can do something doesn't mean you have too.

 

Thus with driverless trains there are two distinct aspects. Answering the question "Can it be done safely" (and the DLR experence says that in principle, and subject to the right mitigations, the anwer is yes) is not the same thing as saying "It should be done"

[royaloak]

It cannot - but the DLR doesn't kill people on a daily basis despite not having a driver to look out for the dangers you mention.

Once again I fear people are not thinking things through properly.

 

There have been a couple of deaths where the report hinted that a human driver would have been able to stop the train quicker than the computer did and the person would have survived.

[njee20]

Who will pay for everything when we are all on the dole because we have been replaced by R2D2 so just sit at home all day with nothing to do except sign on every fortnight?

Who pays for the telephone operators, or the typing pool writers, or chimney sweeps or anyone who loses a job for any reason whatsoever? You get another one. That’s not a reason to keep people employed in jobs which cease to be required.

 

Maybe one day we’ll have a universal wage and no one will have a job. I don’t want to be the one to say that unfortunately you (and not just you, anyone) are not indispensable.

[phil-b259]

There have been a couple of deaths where the report hinted that a human driver would have been able to stop the train quicker than the computer did and the person would have survived.

 

Interesting - do you have any further details.

 

The other thing of course would be to undertake an analysis of incidents between the DLR and a section of the conventional rail network in south London. How do the stats compare - near misses, fatalities, obstructions on the line etc.

 

Obviously where stations are concerned there is a grater risk - which is why platform edge doors are favoured for all new construction as it mitigates against some of the possible hazzards.

Edited March 28, 2018 by phil-b259

[Kris]

I thought the proponents were on about conventional lines and doing away with me and my colleagues? 

 

I firmly believe driverless trains will happen across the network, the only question is when. 

[black and decker boy]

. and how does it recognise HV clothing on staff working lineside or, far more so, in the four foot - especially when they're round a long sweeping curve?[/i]

That’s easy, we all have to wear a big QR code, no need for hi-viz. train will scan us like a supermarket and immediately know who we are and why we are there.

[Reorte]

Who pays for the telephone operators, or the typing pool writers, or chimney sweeps or anyone who loses a job for any reason whatsoever? You get another one. That’s not a reason to keep people employed in jobs which cease to be required.

 

Maybe one day we’ll have a universal wage and no one will have a job. I don’t want to be the one to say that unfortunately you (and not just you, anyone) are not indispensable.

Is the assumption something else will always come along always going to hold? And a world where no-one does anything useful sounds like a pretty grim place no matter how superficially appealing it may be.

 

Going in this direction is probably inevitable. It's not desirable, and it's not beneficial to society as a whole other than when it removes jobs that no-one really wants to do in the first place. And note that many of these jobs, like driving trains, are still required. They're just getting done by fancy, complicated computers instead of people. And as more and more of the real problems get solved (we've got precious few real problems in the UK compared to a couple of centuries ago) the genuine benefit of these developments becomes increasingly questionable.

 

So tell me I've got to accept that it's going that way, well, it looks like that's unavoidable. But don't expect me to like or respect it.

 

Daily rant done

Edited March 28, 2018 by Reorte

[royaloak]

Who pays for the telephone operators, or the typing pool writers, or chimney sweeps or anyone who loses a job for any reason whatsoever? You get another one. That’s not a reason to keep people employed in jobs which cease to be required.

 

Maybe one day we’ll have a universal wage and no one will have a job. I don’t want to be the one to say that unfortunately you (and not just you, anyone) are not indispensable.

Maybe those displaced telephone operators, typists and chimney sweeps went on to be train drivers etc, eventually you are going to run out of jobs to do!

 

As for a universal wage, we have that now, its called job seekers allowance, but I dont understand how the Country will work with everyone only being given enough money to survive, nobody will have a car (no need), or a mortgage (because of no job) or holidays etc so how will it 'work'?

Edited March 28, 2018 by royaloak