Driverless trains report

[O L Leigh]

I really cant see that being a problem. Any such driverless train will have an on board camera. This will use image recognition to assess what the obstruction is and decide on course of action. Image recognition is getting scarily clever. Also the picture will be sent back to control who would be alerted to an incident and can view and take action.

Wow!! An onboard camera, you say? Fantastic.

I'm sorry, but image recognition only goes so far. The train isn't filling in an I-Spy book so there's no point it simply knowing the difference between a horse and a car. It's got to make an assessment of risk and formulate a plan of action based on what it can "see", and that level of decision making requires far greater levels of intelligence than has yet been attained.

Do you know, the one thing that I really love about this forum is that there seems to be no shortage of people here who know my job better than me. Where would I be without the guidance of these wise sages? Thank goodness my manager is happy enough with my performance at work and the competencies that I'm required to maintain.

O L Leigh

 

[carriageline]

Brilliant!! So a train will be able to decide what different action it should take if both a toy car, a manakin and a trolley are on the line. Seeing the obstruction, and reacting (let's me honest 99% of the time will involve throwing the brakes in) are only a minute part of the procedure, what happens next is the most important part. Personally, I would love to know what would be different about each of them scenarios, because all I can see is that the train will be sat there waiting for it to move (unless we install a robotic arm to the front of the train to throw it aside? )

Although you can see it now, train stops due to obstruction, onboard captain/whatever jumps down to move it, train sees obstruction has moved and foxtrot oscars, onboard wotsit is now left flabbergasted

 

[Clip]

Brilliant!! So a train will be able to decide what different action it should take if both a toy car, a manakin and a trolley are on the line. Seeing the obstruction, and reacting (let's me honest 99% of the time will involve throwing the brakes in) are only a minute part of the procedure, what happens next is the most important part. Personally, I would love to know what would be different about each of them scenarios, because all I can see is that the train will be sat there waiting for it to move (unless we install a robotic arm to the front of the train to throw it aside? )

Although you can see it now, train stops due to obstruction, onboard captain/whatever jumps down to move it, train sees obstruction has moved and foxtrot oscars, onboard wotsit is now left flabbergasted

See this is what I was alluding to earlier, posters just saying it cant be done because it may be a moving object or that with it now being moved the train will just carry on - which is a load of crap because the minute a door is enabled the train will inhibit so that it cannot be moved whilst the door is open and even if closed will need the RTD signal before it moves off.

And this is the problem, theres too much closed mindedness on the subject that posters are not even thinking about all the failsafes that are already in place to stop a scenario outlined above and just stomp their feet saying 'it cant be done it cant be done'.


We'd still be stuck in the dark ages if all we ever did was say that 'it cannot be done' as it is thanks to people who are forward thinking and actually use their heads to come up with solutions to problems we happen to be in a wonderful world where we can fire small particles(atoms?) round a big ring underground to discover stuff but an ATO train is IMPOSSIBLE because people on a rail forum say it cant be done.

*sigh*

 

[carriageline]

Calm down darling, it was a joke. Of course ATO is happening, I haven't once said it wasn't. it's just the same argument going around about having no one on board the train, and that somehow cameras will always be the way forward.

Yes the technology is there, a lot of things are showing that. But technology can only go so far before a human has to interact, and someone being on the train, or in an office 50 miles away makes the difference. Does this technology mean it's going to be better than a driver in every situation? Does that mean in certain situations it would cause less delay by having no driver/staff? I'm not quite sure.

We are not talking about failsafes, there's no doubt a train can drive it self fine when everything is hunky dory, but what about the times it isn't. When someone quotes that a train can examine the line, and a controller can then watch the recorded video to see the outcome, it just shows the lack of understanding certain members of this forum have of the actual operations of the railway. The railway isn't simple, you can't plan for every eventuality, because there will be eventualities you can't plan or think of

I personally don't have a closed mind at all thank you, my job over the next 5 or so years will be completely different to what I do now, and I cannot wait to see what the new technology does and how it affects the running of trains. Infact I have welcomed it with massive open arms and can't wait to get in a position to try it out. Some people on the the other hand are too open minded and let their imaginations run just a little too wild

 

[HSTEd]

With regards to determining what kind of obstruction is in the train's path - surely the computer aboard the train would simply detect the obstruction and attempt to stop short of it just like a driver would.

The 'stopped by obstruction' signal transmitted to control would likely cause someone in the control room to switch to the live feed of the incident (or rewound using the fact that all cameras are buffered through the server room at service control) to determine what it was that triggered the alert and has either been passed or hit and what action should then be taken.

It is also worth noting that in this era of big-data that information about vehicle ride can be collected continuously and compared (in real time) to historical recordings at the same locations, allowing a computer mounted in a blade server rack in a room in a building resembling a warehouse in Slough or somewhere to register an alert if ride conditions over track change in a major way - as it sometimes does in advance of accidents (I think it was at Hatfield but I can't remember).

Additionally there has been a lot of very interesting work in automatic processing of rail camera footage (the thing on measurement trains where a camera watches the rail) which might potentially allow it to be monitored automatically in real time. Automatic Eddy Current detection is also advancing rapidly.

 

[carriageline]

You keep reiterating the same thing, yet avoiding questions that really matter. Yes, we know the technology can stop the train.

Once that train with no staff on it has then stopped, what next?

Say we use the scenario where normally the driver jumps down, moves the shopping trolley or drags it on board, delay minutes there are very few, incident pretty much self contained.

what about our train with no staff on it. With how thinly ground staff are placed around the country, it's going to take some time to get someone there, that's if they are not dealing with other things at the same time (which happens too much ), then by that time more trains are coming to a stand, and we all know how that ends up (but of course, service recovery, in theory, should be better in this scenario) and then what if passengers start self evacuating (I wonder if passengers on an ATO train would be more or less likely to go walk abouts, with the "god like" announcements from someone cosy in an office hundreds of miles away ), whilst a driver can't really do much about it, there's more of a problem there.

Just to reiterate, I'm not against ATO, the technology is there and it will happen one day and it will work. I'm just against the idea of having NO ONE on board, be it an operator at the controls or a customer service agent inside the train. It brings too many whatifs and with the way our railway is it won't work.

With regards to vehicle ride reporting, I believe some pendolinos are fitted with similar equipment. Network Rail have also released an app (internally) for staff which monitors ride quality while staff commute to work etc, which is brilliant (similar to what some councils do with regards to poor road surfaces)

 

[Rational Plan]

Automated cars will change everything, but not in the way some are predicting. At the end of the day they will still need fueling and still need to be maintained and taxed and occupy road space.

Until all manual cars are banned they won't be able to form conveys on the motorways, and any way once the get into cities they will have to mix with pedestrians and cyclists and so can not be crammed in there.

Only people with money to burn will have their cars drop them off at work and pick them up again, to avoid paying parking charges, think of the petrol costs. Taxis though will be much cheaper and therefore threaten bus services in rural and smaller towns. But in bigger places there is not the space for everyone to have their own cab, and remember most vehicles will be empty most of the day, where will they park? They can't endlessly circle town, there won't be the road space.

People assume that they just pick up other fares, which would be great if travel demand was even, but it's not. No cab company is going to make money with their vehicles sitting unused most of the day, so their will be shortages, rationed through higher prices at peak times (currently illegal, accept some places allow slightly higher fares at night). To guarantee a vehicle at their disposal most people will need to own their own vehicle or still use public transport.

I suspect buses may be the first to go driverless to compete with auto taxis and taxi companies may offer mini buses that link half a dozen trips together in one ride.

What could happen is that automated buses start to compete on regional trips. The buses don't need to be fast or large, if five or ten people want to travel from a small town to another medium sized town there maybe a direct mini bus rather than a couple of bus and rail trips.

What maybe more important in the future are intelligent transport agents that aggregate demand between towns and offer their client prices and time options that either off direct options or combine auto cab with rail if the price is right and the speed is important.

 

[21C101]

Wow!! An onboard camera, you say? Fantastic.

I'm sorry, but image recognition only goes so far. The train isn't filling in an I-Spy book so there's no point it simply knowing the difference between a horse and a car. It's got to make an assessment of risk and formulate a plan of action based on what it can "see", and that level of decision making requires far greater levels of intelligence than has yet been attained.

Do you know, the one thing that I really love about this forum is that there seems to be no shortage of people here who know my job better than me. Where would I be without the guidance of these wise sages? Thank goodness my manager is happy enough with my performance at work and the competencies that I'm required to maintain.

O L Leigh

My friend the abilites of image recognition and decision software are increasing exponentially and scarily so. This sort of thing is chicken feed to them.

No one sensible is suggesting de staffed trains, just that many drivers will be replaced with train captains as has already happened on the Docklands line railways. The down side of that for drivers is a lowering of pay and status and having to do revenue protection duties. Things like over inflation pay offers or demands 32 hour weeks bring that day ever closer in the same way as the printers killed their golden goose. However the timescales for deployment are such that few if any existing drivers need fear for their future and as I said earlier it will more likely occur in high density fairly self contained lines like Merseyrail and probably never on intercity services as the econmics wont stack up for long distance trains.

 

[HSTEd]

Once that train with no staff on it has then stopped, what next? Say we use the scenario where normally the driver jumps down, moves the shopping trolley or drags it on board, what do you do? With how thinly ground staff are placed around the country, it's going to take some time to get someone there, that's if they are not dealing with other things at the same time (which happens too much ),

Well I imagine a shopping trolley would be met with a snort of derision and the train just edging up to it and shoving it out of the way at the minimum speed available (like the speed used by MGR trains).
If there is something that is more difficult to move then it would have to await assistance - unless this is one of the unmanned trains that is not actually unmanned.
You can put a lot more trackside staff in cars scattered over the shop if you aren't paying drivers and have computer controlled infrastructure allowing more intensive use. (For example using quad track where six track would previously have been necessary or similar).

 

[21C101]

You keep reiterating the same thing, yet avoiding questions that really matter. Yes, we know the technology can stop the train.

Once that train with no staff on it has then stopped, what next?

No one sensible is advocating wholly unstaffed trains, but that train drivers are replaced with DLR style train captains. wholly unstaffed trains is just a straw man.
--- old post above --- --- new post below ---

Well I imagine a shopping trolley would be met with a snort of derision and the train just edging up to it and shoving it out of the way at the minimum speed available (like the speed used by MGR trains).
If there is something that is more difficult to move then it would have to await assistance - unless this is one of the unmanned trains that is not actually unmanned.
You can put a lot more trackside staff in cars scattered over the shop if you aren't paying drivers and have computer controlled infrastructure allowing more intensive use. (For example using quad track where six track would previously have been necessary or similar).

The train would stop and the train captain get out and remove it once control have confirmed the third rail is de-energised. He would be well aware of what it is from the picture on his mobile phone app sent to him by the on board camera and processor which will have also told him what it is and what action to take as well as activating an isolation of the third rail.

 

[radamfi]

Driverless taxis should be cheaper than owning your own car as the fixed costs are shared among many users. Therefore many people will choose not to own their own car as a taxi would be available pretty much instantly in most cases and there would be no parking issues. At the moment most cars lie unused most of the time so there would be a greatly improved vehicle utilisation and far fewer vehicles required.

 

[21C101]

Automated cars will change everything, but not in the way some are predicting. At the end of the day they will still need fueling and still need to be maintained and taxed and occupy road space.

Until all manual cars are banned they won't be able to form conveys on the motorways, and any way once the get into cities they will have to mix with pedestrians and cyclists and so can not be crammed in there.

Only people with money to burn will have their cars drop them off at work and pick them up again, to avoid paying parking charges, think of the petrol costs. Taxis though will be much cheaper and therefore threaten bus services in rural and smaller towns. But in bigger places there is not the space for everyone to have their own cab, and remember most vehicles will be empty most of the day, where will they park? They can't endlessly circle town, there won't be the road space.

People assume that they just pick up other fares, which would be great if travel demand was even, but it's not. No cab company is going to make money with their vehicles sitting unused most of the day, so their will be shortages, rationed through higher prices at peak times (currently illegal, accept some places allow slightly higher fares at night). To guarantee a vehicle at their disposal most people will need to own their own vehicle or still use public transport.

I suspect buses may be the first to go driverless to compete with auto taxis and taxi companies may offer mini buses that link half a dozen trips together in one ride.

What could happen is that automated buses start to compete on regional trips. The buses don't need to be fast or large, if five or ten people want to travel from a small town to another medium sized town there maybe a direct mini bus rather than a couple of bus and rail trips.

What maybe more important in the future are intelligent transport agents that aggregate demand between towns and offer their client prices and time options that either off direct options or combine auto cab with rail if the price is right and the speed is important.

Agree with a lot of what you say but I would be surprised if manual driving is still legal in thirty years time.

The first automated driverless "taxis" are due to start running in Milton Keynes very soon http://www.bbc.co.uk/news/uk-england-beds-bucks-herts-24849948

As to where to park them. If far fewer people have cars then there will be plenty of roadside parking spaces. I would expect operators to space them out strategically based on usage patterns in much the same way as paramedic vehicles are now.
I agree though that they may not all be sole use. Although that brings the same problem as wholly unstaffed trains. People shut in a box with strangers and no staff at all onboard.

 

[Rational Plan]

Agree with a lot of what you say but I would be surprised if manual driving is still legal in thirty years time.

The first automated driverless "taxis" are due to start running in Milton Keynes very soon http://www.bbc.co.uk/news/uk-england-beds-bucks-herts-24849948

As to where to park them. If far fewer people have cars then there will be plenty of roadside parking spaces. I would expect operators to space them out strategically based on usage patterns in much the same way as paramedic vehicles are now.
I agree though that they may not all be sole use. Although that brings the same problem as wholly unstaffed trains. People shut in a box with strangers and no staff at all onboard.

People share lifts though. Mini buses could be under real time camera surveillance. Bus drivers are unable to stop attacks on buses at the moment. I'm not sure if that will be as big a problem as no doubt the media will try and portray it.

 

[Tomnick]

My friend the abilites of image recognition and decision software are increasing exponentially and scarily so. This sort of thing is chicken feed to them.

No one sensible is suggesting de staffed trains, just that many drivers will be replaced with train captains as has already happened on the Docklands line railways. The down side of that for drivers is a lowering of pay and status and having to do revenue protection duties. Things like over inflation pay offers or demands 32 hour weeks bring that day ever closer in the same way as the printers killed their golden goose.

I won't disagree that the technology will, one day, be developed sufficiently to allow this thing to meet most of the demands of the real world. I don't agree that the driver's status will be reduced in the way that you suggest though - at present, it's when things go wrong that they really earn their money and justify their very thorough rules, traction and route knowledge, and I'm not convinced that there'll be any change to that. Whether there'll be the opportunity to ge the driver carrying out revenue duties, I don't know - so yes, there might ultimately be a bit of a cost saving there.

Incidentally, and I know this is old ground again, but how many pay deals have seen an increase in pay (in real terms) without a similar increase in productivity? Not many at all, I suspect - it's not all about pay. The 32 hour week is being pushed by the RMT, not (as far as I'm aware) ASLEF. Don't let the facts get in the way of that though!

 

[DarloRich]

it's when things go wrong that they really earn their money and justify their very thorough rules, traction and route knowledge

Sadly none of that matters to many "experts” - they have no idea/care about what happens operationally during degraded working or during the leaf fall season.

 

[Albatross]

Might just as well talk about flying cars. We drive 142's for gods sake. None of you on here will ever see driverless mainline trains.

 

[Kneedown]

Well I imagine a shopping trolley would be met with a snort of derision and the train just edging up to it and shoving it out of the way at the minimum speed available (like the speed used by MGR trains).

The last shopping trolley I hit clean knocked off the main reservoir drain and ripped out all the wiring below the solebar.
If a train were to "nudge" it out of the way it could well go underneath, maybe not knocking off air drains but certainly causing severe damage to more delicate equipment.

Image recognition software is truly awesome these days, but I don't believe the technology to recognise all manner of objects or obstructions, at all possible angles to the sensor, is there yet and I don't believe it will be for many years yet. Software designers are truly awesome people as well, but they are not railway staff. A computer is only as good as the software it runs on and any problem in the coding that doesnt become apparent straight away, or if the designer forgets to program the "sensor" to apply the brake when it spots a refrigerator end on, could potentially lead to trouble. Alternatively, if the software was overprotective, how many delays to journeys would occur everytime a pigeon flew out of the trees at the very last minute?

TBH I don't know why we are all arguing about this. One day it will happen, but I don't believe it will be in any of our lifetimes. I think issues will arise during the tests of driverless cars that will set back the timescale for them taking over from human operated cars.

Road or rail, in the big, bad, open world there are just too many variables for current technology to overcome IMHO.

 

[devinier]

I would quote Walter Reuther, a former leader of the Union of Automobile Workers in the US of A :
"In 1951, the Ford Motor Co. opened up a new engine plant in Cleveland, Ohio, adjacent to the municipal airport. It was the first fully automated engine plant. … I went through that plant many years back…
So they said to me, “Aren’t you worried about how you are going to collect union dues from all of these machines?”
I said, “the thought never occurred to me. The thought that occurred to me was how are you going to sell cars to these machines?” You know you can make automobiles, but consumers are still made in the good old fashioned way.” [Laughter.]"

 

[southern442]

And I thought the government were trying to get people into work, not take them out of it!

 

[carriageline]

The last shopping trolley I hit clean knocked off the main reservoir drain and ripped out all the wiring below the solebar.

If a train were to "nudge" it out of the way it could well go underneath, maybe not knocking off air drains but certainly causing severe damage to more delicate equipment.



Image recognition software is truly awesome these days, but I don't believe the technology to recognise all manner of objects or obstructions, at all possible angles to the sensor, is there yet and I don't believe it will be for many years yet. Software designers are truly awesome people as well, but they are not railway staff. A computer is only as good as the software it runs on and any problem in the coding that doesnt become apparent straight away, or if the designer forgets to program the "sensor" to apply the brake when it spots a refrigerator end on, could potentially lead to trouble. Alternatively, if the software was overprotective, how many delays to journeys would occur everytime a pigeon flew out of the trees at the very last minute?



TBH I don't know why we are all arguing about this. One day it will happen, but I don't believe it will be in any of our lifetimes. I think issues will arise during the tests of driverless cars that will set back the timescale for them taking over from human operated cars.



Road or rail, in the big, bad, open world there are just too many variables for current technology to overcome IMHO.

Leave it, I have come to the conclusion that he is absolutely deluded, and frankly doesn't have a clue. It obviously shows that the knowledge the general public has, is better than the experience of the rail staff!

 

[Rugd1022]

The last shopping trolley I hit clean knocked off the main reservoir drain and ripped out all the wiring below the solebar.
If a train were to "nudge" it out of the way it could well go underneath, maybe not knocking off air drains but certainly causing severe damage to more delicate equipment.

Image recognition software is truly awesome these days, but I don't believe the technology to recognise all manner of objects or obstructions, at all possible angles to the sensor, is there yet and I don't believe it will be for many years yet. Software designers are truly awesome people as well, but they are not railway staff. A computer is only as good as the software it runs on and any problem in the coding that doesnt become apparent straight away, or if the designer forgets to program the "sensor" to apply the brake when it spots a refrigerator end on, could potentially lead to trouble. Alternatively, if the software was overprotective, how many delays to journeys would occur everytime a pigeon flew out of the trees at the very last minute?

TBH I don't know why we are all arguing about this. One day it will happen, but I don't believe it will be in any of our lifetimes. I think issues will arise during the tests of driverless cars that will set back the timescale for them taking over from human operated cars.

Road or rail, in the big, bad, open world there are just too many variables for current technology to overcome IMHO.

The last one I hit happened on Coalville level crossing as I was passing over it light engine, it became well and truly wedged under the second traction motor with part of the handle wrapped around one of the sand pipes, it took over an hour for two of us (I had a shunter with me) to get it out and for me to get moving again. Luckily it happened in the middle of the night so only a few motorists were delayed.

I wonder - has HSTEd ever spent a full shift in the cab of a loco and seen with his own eyes what the job actually entails...? I'll wager not...

 

[carriageline]

The last one I hit happened on Coalville level crossing as I was passing over it light engine, it became well and truly wedged under the second traction motor with part of the handle wrapped around one of the sand pipes, it took over an hour for two of us (I had a shunter with me) to get it out and for me to get moving again. Luckily it happened in the middle of the night so only a few motorists were delayed.



I wonder - has HSTEd ever spent a full shift in the cab of a loco and seen with his own eyes what the job actually entails...? I'll wager not...

You obviously did it wrong. Just nudge it out the way mate, HSTEd is clearly more experienced than all of us put together

 

[Tomnick]

You obviously did it wrong. Just nudge it out the way mate, HSTEd is clearly more experienced than all of us put together

I see that you've shot to fame on Connect already! Did your young friends have any opinions on the future of driverless trains? I reckon they've got the benefit of more practical experience than some on here .

 

[carriageline]

I see that you've shot to fame on Connect already! Did your young friends have any opinions on the future of driverless trains? I reckon they've got the benefit of more practical experience than some on here .

Hahaha don't!!! I never knew it was going to end up on Connect! Photo and all!

 

[orpine]

It obviously shows that the knowledge the general public has, is better than the experience of the rail staff!

You'll find that this carries across to public levels of knowledge of all industries. And the other way works too - almost every industry thinks they're "special"; none are.

Software designers are truly awesome people as well, but they are not railway staff. A computer is only as good as the software it runs on and any problem in the coding that doesnt become apparent straight away, or if the designer forgets to program the "sensor" to apply the brake when it spots a refrigerator end on, could potentially lead to trouble.

Yes. But that's why the software engineers talk to the rail types, and when you do that properly you end up with things like one of the most efficiently run subways systems in the world - http://www.newscientist.com/article...that-deploys-hong-kongs-subway-engineers.html

 

[Kneedown]

Yes. But that's why the software engineers talk to the rail types, and when you do that properly you end up with things like one of the most efficiently run subways systems in the world - http://www.newscientist.com/article...that-deploys-hong-kongs-subway-engineers.html

The link you post descibes a computer system for programming maintenence to infrastructure, not driving trains.
I'm not an expert on the Hong Kong Metro, but a quick look gives me the impression that it is another self contained system. Once again, there is a whole world of difference between that and a mainline railway with trains that travel at high speeds through the middle of nowhere.
Like I said previously, the technology will arrive, but not in our lifetime IMO.

 

[21C101]

The link you post descibes a computer system for programming maintenence to infrastructure, not driving trains.
I'm not an expert on the Hong Kong Metro, but a quick look gives me the impression that it is another self contained system. Once again, there is a whole world of difference between that and a mainline railway with trains that travel at high speeds through the middle of nowhere.
Like I said previously, the technology will arrive, but not in our lifetime IMO.

The issue really is that there are two types of main line railway in the UK

The Southern east of and including the Waterloo - Weymouth line and the rest, with certain lines like Merseyrail counting in the same category as the Southern.

The southern is virtually all electric with fairly similar stock and has comparatively few workings off the network, mostly metro style operations like Thameslink and the East London line.

Its quite possible to spend an entire career on the Southern and this will give people different perceptions of what a main line railway is.

Southern - two or four tracks, frequent stops, frequent service, stations at close intervals, railway shoehorned into the landscape with space at a premium. Typical speed 60-90 with very occasional 100. Not that many level crossings. Typical journey time one hour to one and a half max, third rail

Rest - one or two tracks, one or two trains an hour, mixture of local and very long distance and freight, stations far apart, plenty of space and able to drive a van up the side of the track. Speed 75 to 125, riddled with level crossings. Typical journey time from half an hour to all day or all night, not electrified or that strange stuff up on poles that jumps at you like lightning if you get too close.

I am obviously generalising a bit but the southern is different and it would be much easier to have driverless trains on it than the lines North/West of London.

To be honest I think they messed up in 1948, it would have made more sense in some ways to either leave the southern separate or merge it with the Underground than with the LNER, GW and LMS.

 

[orpine]

The link you post descibes a computer system for programming maintenence to infrastructure, not driving trains.
I'm not an expert on the Hong Kong Metro, but a quick look gives me the impression that it is another self contained system.

I'm aware of that. My point was that software engineers can talk to rail people. My other point was that had someone said on this forum 10 years ago that the managing of a entire maintenance system would be run by computer, they'd have be laughed off the board by all the maintenance types saying how special and irreplaceable they were. It has happened to many industries before.


---

Incidentally, for this interested, Wikipedia does of course have a list:
https://en.wikipedia.org/wiki/List_of_driverless_trains
Seems there are actually *lots* of Grade 4 systems; pretty much all metro things, but it does show that the obstacle recognition is well on the way to being resolved. Still leaves the issue of obstacle removal for a larger line in no-train-captain systems.

 

[southern442]

The issue really is that there are two types of main line railway in the UK

The Southern east of and including the Waterloo - Weymouth line and the rest, with certain lines like Merseyrail counting in the same category as the Southern.

The southern is virtually all electric with fairly similar stock and has comparatively few workings off the network, mostly metro style operations like Thameslink and the East London line.

Its quite possible to spend an entire career on the Southern and this will give people different perceptions of what a main line railway is.

Southern - two or four tracks, frequent stops, frequent service, stations at close intervals, railway shoehorned into the landscape with space at a premium. Typical speed 60-90 with very occasional 100. Not that many level crossings. Typical journey time one hour to one and a half max, third rail

Rest - one or two tracks, one or two trains an hour, mixture of local and very long distance and freight, stations far apart, plenty of space and able to drive a van up the side of the track. Speed 75 to 125, riddled with level crossings. Typical journey time from half an hour to all day or all night, not electrified or that strange stuff up on poles that jumps at you like lightning if you get too close.

I am obviously generalising a bit but the southern is different and it would be much easier to have driverless trains on it than the lines North/West of London.

To be honest I think they messed up in 1948, it would have made more sense in some ways to either leave the southern separate or merge it with the Underground than with the LNER, GW and LMS.

You are generalising quite a lot! There are lots of four-track sections on the rest of the network, and lots of very frequent services on the rest of the network like in Glasgow. Stations vary on all parts of the national network, with some stations (like Culrain and Invershin) being close and others (like Petersfield and Rowlands castle) far apart, those two are examples are from the "rest" of the network and the Southern network respectively.

The Thameslink service shouldn't count as a metro as it is 140 miles long, there is currently a huge debate on the desiro city thread about this, as a result of their contraversial interior. Yes, some sections of third rail land are frequent but that is not to say it should be treated as a metro, or merge it with the underground, as an underground line that streches 150 miles to weymouth with no stops actually in London bar waterloo couldn't make less sense. I understand the point you're trying to make but the Southern region is very much in operation in and alongside the main network, with certain lines such as the West London Line, Thameslink, HS1 and the Cross Country route that couldn't be driverless as these traverse sections of third rail and overhead lines.

 

[21C101]

Regarding an "underground" line that stretches to Weymouth, we already have one that stretches 40 miles into Buckinghamshire with a four track main line and shared with diesel services to Alyesbury and soon Milton Keynes. in many ways the subsurface lines are more similar to the southern than they are to the deep tube lines. In several cases they are historically part of the southern - including an entire deep tube line!

I would think that all the lines you suggest could go driverless within the next 30-40 years. There is no reason why all trains on a line would neccesarily be driverless. Within 50 years I would be surprised if there are not driverless Thameslink trains mixed with "with driver" east midlands trains and freight services on the midland main line south of Bedford or possibly Kettering/Corby.

The issue of the config of the new Thameslink trains is an interesting one. It will never please everybody, but it looks like the seats won't be particularly comfortable. However as the staff accomodation, sorry, first class has decent seats many of the people responsible for such decisions will never have to suffer it.