Would You Support a Driverless Railway?

[Turbocharger]

At 125mph, drivers have no chance of stopping for an obstruction on the line.

That may be true, but hitting the brakes even a second or two before something is spotted can make a difference.

I would only support trains without drivers if the track and platforms were 100% secure.

 

[aspierail]

I would say ATO would work well with short distance/metro services however i cannot see it working on long distance services though.

 

[Nonsense]

I would say ATO would work well with short distance/metro services however i cannot see it working on long distance services though.

Whats the difference?

As others have stated, planes can land and take off without any human intervention. Trains already operate everyday without a human driver. Even cars have been demonstrated to not just drive themselves, but to do so at speed and in close convoy. Train movements are already controlled remotely, and have been for centuries.

The technological constraints are long gone.

Of course, none of that makes it easy. Just not impossible.

 

[Metrailway]

Would it? The computer is only as good as the person that programmed it and the data points in contains and I dare say it wouldn't exactly be like a NASA computer. I'd like to see a computer cope with poor railhead conditions.

Poor railhead conditions are common on the Central line. It has ATO. Technology is not the issue. Most lines on the national network just wouldn't warrant the cost of conversion.

 

[Zoe]

Most lines on the national network just wouldn't warrant the cost of conversion.

ATP will be installed as part of ERTMS so that will already have been paid for. I don't see any reason why once ERTMS in installed ATO shouldn't be introduced. A computer setting the speed on the train based on input from ERTMS with a driver there in case of emergency is not much different to a human reading the ERTMS display and setting the speed.

 

[GB]

Poor railhead conditions are common on the Central line. It has ATO. Technology is not the issue. Most lines on the national network just wouldn't warrant the cost of conversion.

But does the automatic system cope with such circumstances better than a physical driver?

 

[tirphil]

ATP will be installed as part of ERTMS so that will already have been paid for. I don't see any reason why once ERTMS in installed ATO shouldn't be introduced. A computer setting the speed on the train based on input from ERTMS with a driver there in case of emergency is not much different to a human reading the ERTMS display and setting the speed.

ERTMS is experiencing many problems with reliability and may not be the panacea some had hoped for. Peter Leppard of ATW has apparently said as much. It seems that the whole of the Cambrian ground to a halt on Tuesday morning due to a main ERTMS computer crashing because of a shunt that took place in Barmouth. If ERTMS struggles with one shunt on the Cambrian it cannot be fit for purpose on a multi track mainline and certainly will not be lead to the development of driverless trains in the near (10-20 years) future.

 

[Greenback]

ERTMS is experiencing many problems with reliability and may not be the panacea some had hoped for. Peter Leppard of ATW has apparently said as much. It seems that the whole of the Cambrian ground to a halt on Tuesday morning due to a main ERTMS computer crashing because of a shunt that took place in Barmouth. If ERTMS struggles with one shunt on the Cambrian it cannot be fit for purpose on a multi track mainline and certainly will not be lead to the development of driverless trains in the near (10-20 years) future.

I think some people's faith in technology may be misplaced!

 

[whhistle]

Some really good discussions on this

I thought of it because of the report about some parts of Thameslink being ATO. I wonder if the new 2009/S-Stock has ATO capabilities?

Computers fly planes all the time, and can take off and land; the pilot is there for passenger reassurance.

If the train systems did detect a problem, a live video link could be made back to rail HQ where a group of drivers would be available on 'fly by wire' to advance the train under caution. I would expect a pool of 6 to 10 drivers could handle the entire WCML and ECML that way.

These are the sorts of things we were talking about at work. Drivers would not be out of a job, they'd just be sitting in a control type set up like signallers. Whether they would see the train through to the end destination or it would be passed over is another question.

 

[hairyhandedfool]

....So, would you support it [on mainline trains] if ATO was done properly?

I've seen an episode of Thunderbirds where they go on an automatic train, if you have seen it you'll understand if I prefer human drivers on the mainline

 

[TDK]

ATP will be installed as part of ERTMS so that will already have been paid for. I don't see any reason why once ERTMS in installed ATO shouldn't be introduced. A computer setting the speed on the train based on input from ERTMS with a driver there in case of emergency is not much different to a human reading the ERTMS display and setting the speed.

Why would they need ATP within an ERTMS area? AFAIK ERTMS does the ATP bit for you - have you a link to your information Zoe?
--- old post above --- --- new post below ---

Whats the difference?

As others have stated, planes can land and take off without any human intervention. Trains already operate everyday without a human driver. Even cars have been demonstrated to not just drive themselves, but to do so at speed and in close convoy. Train movements are already controlled remotely, and have been for centuries.

The technological constraints are long gone.

Of course, none of that makes it easy. Just not impossible.

yes but you still have pilot and co pilot and to be honest if you need a driver to monitor the systems would it cost effective in the first place? Pendos have cruise control I suppose that is a basic auto pilot sort of thing etc. With GPS the way it is currently with only a certain life span I cannot see it happening on the main line

 

[DaveNewcastle]

I'm not going to comment on the technical issues, nor on the safety implications, but I will comment on a factor I strongly recommend:-

I would prefer to see more people employed in the physical operation of the railways. Not less.

 

[sprinterguy]

I've seen an episode of Thunderbirds where they go on an automatic train, if you have seen it you'll understand if I prefer human drivers on the mainline

I don't think that Thunderbirds can be relied upon to provide an accurate depiction of the future Shame really, as I would love to see a Fireflash aircraft in service I would assume that ATO would be set up to detect something as major as a collapsed bridge, so I think we'd be safe from any such Thunderbirds-esque situations.

I'm not keen on the idea of driverless trains being rolled out nationwide, as having witnessed trains becoming ever more reliant on computer systems to function over the last decade or two, then a fault with any driverless system would just increase the number of reasons for a train to go into shutdown mode following detection of a minor fault, rather than a human driver being able to assess the situation and decide whether an on-train fault precludes the train being able to "limp on" to the next station to allow passengers to de-train conveniently or not.

I think that trains need a human element to make decisions on the best course of action take in any given situation, rather than leaving passengers at the mercy of a completely automated system that would feel no duty whatsoever to ensuring that passengers (and the train generally) make it to their destination

 

[tirphil]

ERTMS should act as ATP is designed to although 5 instances of trains passing Marker Boards without a movement authority in the ERTMS trial on the Cambrian suggest that the either ERTMS does not prevent SPADS or ERTMS has not been installed correctly.

 

[Zoe]

Why would they need ATP within an ERTMS area?

I was referring to the ATP that will be provided by ERTMS, not to either of the currently used ATP systems.

 

[Metrailway]

ATP will be installed as part of ERTMS so that will already have been paid for. I don't see any reason why once ERTMS in installed ATO shouldn't be introduced. A computer setting the speed on the train based on input from ERTMS with a driver there in case of emergency is not much different to a human reading the ERTMS display and setting the speed.

But then it would no longer be ERTMS

But does the automatic system cope with such circumstances better than a physical driver?

It copes, although certainly the braking/acceleration is less smooth than a human, there is no doubt about that. If there are any problems, the driver can intervene and drive the train themselves at full speed in 'Coded Manual' (ATP).

I wonder if the new 2009/S-Stock has ATO capabilities?

They do indeed. All stock built since and including the 1967 stock have been fitted with ATO capabilities. The 1995, 1996, 2009 and S Stock are fitted with moving block capabilities as well. The Victoria line was the first railway in the world to be fully ATO, and this system has been replaced recently to a newer ATO moving block system. The Central has fixed block ATO since the 90s, the Jubilee line has been converted this year to ATO and moving block, and the Northern is planned to be the next one.

ATO is the future for the Undergound due to higher frequency obtained, as human drivers are more cautious in driving when compared to computers, which use the full capabilities of the train. But that certainly doesn't mean the end of someone being in front of the train.

 

[Zoe]

But then it would no longer be ERTMS

Why not? ERTMS would be providing the ATP, nothing to prevent ATO from doing the actual driving. If you read http://www.ertms.com/news/ertms-news/thameslink-re-signalling-programme-to-include-ertms.aspx you will see that Thameslink will be using both ERTMS and ATO.

 

[313103]

I was quite surprised that the class 378s were not moded for ATO operation yet the S Stock for the met line was. These trains were almost built side by side.
ATO doesnt mean we get rid of the driver (unless of course you are an accountant) he/she still plays a usefull role.

 

[hairyhandedfool]

I don't think that Thunderbirds can be relied upon to provide an accurate depiction of the future Shame really, as I would love to see a Fireflash aircraft in service I would assume that ATO would be set up to detect something as major as a collapsed bridge, so I think we'd be safe from any such Thunderbirds-esque situations....

A computer is only as good as the (wo)man that programmed it

 

[Class172]

I don't think that Thunderbirds can be relied upon to provide an accurate depiction of the future Shame really, as I would love to see a Fireflash aircraft in service I would assume that ATO would be set up to detect something as major as a collapsed bridge, so I think we'd be safe from any such Thunderbirds-esque situations.

The way to solve long distance intercontinental travel - Thunderbird 2!

I do like the way that we ought to be having clubs in space as well as a new London airport with its own dedicated motorway, the M21

 

[Metrailway]

Why not? ERTMS would be providing the ATP, nothing to prevent ATO from doing the actual driving. If you read http://www.ertms.com/news/ertms-news/thameslink-re-signalling-programme-to-include-ertms.aspx you will see that Thameslink will be using both ERTMS and ATO.

Well one of the major points of ERTMS is for inter-operability for trains from different countries, by standardising and removing redundant signalling equipment. This supposedly makes costs cheaper. If you add a national ATO signalling system on top of ERTMS, you get rid of the benefit of inter-operability as well as increasing price.
Thameslink and Crossrail are special cases as they carry no freight, the rolling stock operating on the ATO section will be of the same class and new, and that ATO is beneficial due to the proposed frequencies on that short section, which would not be possible with ERTMS.
Most lines on the mainline network are quite different, as freight and different older passenger train classes operate, making an ATO system more harder (but not impossible) to implement. ERTMS would provide more than enough capacity for these lines meaning that ATO's main benefit of an increase in capacity is pretty much redundant.

ERTMS is experiencing many problems with reliability and may not be the panacea some had hoped for. Peter Leppard of ATW has apparently said as much. It seems that the whole of the Cambrian ground to a halt on Tuesday morning due to a main ERTMS computer crashing because of a shunt that took place in Barmouth. If ERTMS struggles with one shunt on the Cambrian it cannot be fit for purpose on a multi track mainline and certainly will not be lead to the development of driverless trains in the near (10-20 years) future.

Yeah you are right, it pretty much has been a failure for the Cambrian line. Costs were massive compared to RETB and reliability has significantly decreased. ERTMS is not as technically advanced as previously thought. Thus, I doubt we will see a fully ERTMS network in 40+ years let alone a driverless railway network. Switzerland are only going to install a limited ERTMS Level 1, as there is very little international/high speed running to warrant a level 2 system. Many would argue that the UK is similar but we have TPWS which does the same as your limited level 1 system

 

[jon0844]

If we're really looking into the future, I can see that technology could remove the need to drive on all trains, which is progress - and may allow trains to run very closely together and much faster.

But, I can't see anyone ever accepting there being no driver. By that I mean a driver sat at the controls, keeping a watch for when things go wrong and you need to take manual control - just like on a plane in fact.

As mentioned, unlike the DLR, there are too many things that could need manual intervention. Sure you could have cameras to detect obstructions, radar and other technology that will be invented in the next 5-10 years (or perfected) but you'll always need that real person.

I don't believe a train will ever get rid of the driver and employ a guard who wanders around the train, only to be alerted by HAL when something goes wrong - or the train simply stops.

Don't DLR staff manually run trains on Sunday to keep their knowledge going? If so, I'd expect drivers will continue to be highly paid and trained, but possibly on less money than now. Sadly, most jobs will be probably lower paid in the future if we're to remain competitive in a global economy that will be probably far worse than it is now - unless we opt out and cut ourselves off from the rest of the world!

 

[Zoe]

Well one of the major points of ERTMS is for inter-operability for trains from different countries, by standardising and removing redundant signalling equipment. This supposedly makes costs cheaper. If you add a national ATO signalling system on top of ERTMS, you get rid of the benefit of inter-operability as well as increasing price.

There would not be a national ATO signalling system as the signalling/ATP is provided by ERTMS. This would not prevent interoperability as ERTMS would be used in all areas, the only difference is how the train is driven.

 

[Jonny]

From ERTMS to driverless is pretty easy, after all the ideal candidate for the job is almost an automaton (since cybermen are pure fiction) and since computers can be trained to just about anything nowadays without the problems of emotions and trade unions coming into play, including object detection which can be done better than by a human.

 

[whhistle]

It would be a benifit during disruption as the computer would work out what trains are where and the quickest way to route everything into London (or whatever). I'm suprised we don't have "model" type systems that help signallers do this anyway, to minimize trains running out of PPE. If the computer works out that if one train was delayed by a further 10 mins, but three more would slide back into PPE then delaying that one train would surely be worth it?

I think that trains need a human element to make decisions on the best course of action take in any given situation, rather than leaving passengers at the mercy of a completely automated system that would feel no duty whatsoever to ensuring that passengers (and the train generally) make it to their destination

This started off a large discussion at work about the film i.Robot and how the guy in it was saved because the robot detected he had the best chance of survival, where as a human would have saved the child. But then putting ethics aside and looking at the facts, if there is an 80% chance Subject A will survive, a 60% chance Subject B would survive, but a 20% chance if both were saved, logically it would make more sense to save Subject A.

But then this is it, what happens if the three laws of robotics are mutated by robots into a different understanding - IE, by ensuring humans continue in life, they are not allowed out after 9pm and of course, some humans would have to be killed to save the others.

The Victoria line was the first railway in the world to be fully ATO, and this system has been replaced recently to a newer ATO moving block system. The Central has fixed block ATO since the 90s, the Jubilee line has been converted this year to ATO and moving block, and the Northern is planned to be the next one.

So is already in operation on the Victoria line then?

 

[DavidBusDriver]

The one thing though is that I find the Vic-line ATO system uses the power and brake like on/off switches, travelled on one standing from signal to signal (or block to block more accurately I suppose) and rather than just taking a little bit of power to slowly drag up to the next signal block like a human driver would decide to do, it seems to go full throttle to the highest speed possible before braking hard again to stop - seriously one of the most uncomfortable rides I have ever been on!

Plus I am slightly biased I certainly do not want to be out of a job I bloody enjoy doing!

 

[starrymarkb]

Why would they need ATP within an ERTMS area? AFAIK ERTMS does the ATP bit for you - have you a link to your information Zoe?
--- old post above --- --- new post below ---


yes but you still have pilot and co pilot and to be honest if you need a driver to monitor the systems would it cost effective in the first place? Pendos have cruise control I suppose that is a basic auto pilot sort of thing etc. With GPS the way it is currently with only a certain life span I cannot see it happening on the main line

Which makes me think of the Air France A330, Airspeed readings become unreliable (sensors disagree) so Autopilot can't work and disengages, only for the pilot to not follow procedures (IIRC set throttle to CLimb power and hold the nose at 5 degrees and the plane will settle at a known safe speed) and stall the plane making little attempt to recover despite the stall warning sounding all the way down!

This comes to mind
[youtube]2hMn7ZweF6s[/youtube]
The hooter is the warning that the landing gear is still raised!

 

[Metrailway]

There would not be a national ATO signalling system as the signalling/ATP is provided by ERTMS. This would not prevent interoperability as ERTMS would be used in all areas, the only difference is how the train is driven.

Isn't the plan for Thameslink is to overlay an ATO system on top of the ERTMS system? ATP signalling itself is generally insufficient to allow ATO, as an ATO box on a train can mis-read ATP indications on rare occasions, and can get confused if the train is diverted from it's pre-programmed paths due to delays etc. Thus, a central computer system monitors every train on the system, and by looking at the signals, timetables, speed of each train etc, this computer then tells the ATO equipment on board to change the speed of the train. This central computer interacts with the ATO boxes via trackside equipment. As there is no EU standard for ATO, the systems used in different countries is likely to be different preventing the ERTMS goal of inter-operability.

So is already in operation on the Victoria line then?

Yes, ATO has been in operation since the line opened in 1968! Central was next in 2001; Jubilee done a couple of months ago; and the Northern is planned by the end of 2014. The SSL (Metropolitan, Circle, Hammersmith & City, District) is likely to be at least partially ATO by 2020. It'll be interesting to see how they apply moving block ATO to the Met lines shared with Chiltern, if it goes ahead.

The one thing though is that I find the Vic-line ATO system uses the power and brake like on/off switches, travelled on one standing from signal to signal (or block to block more accurately I suppose) and rather than just taking a little bit of power to slowly drag up to the next signal block like a human driver would decide to do, it seems to go full throttle to the highest speed possible before braking hard again to stop - seriously one of the most uncomfortable rides I have ever been on!

Plus I am slightly biased I certainly do not want to be out of a job I bloody enjoy doing!

The original 1968 ATO system is pretty rough, due to the technology limitations of the era, but the newly installed system is a lot smoother. Still, it isn't as smooth as a human driver!

Don't DLR staff manually run trains on Sunday to keep their knowledge going?!

I don't know about the DLR, but the Central line drivers do operate trains manually on Sunday. Some Central line drivers also get practice by driving trains on the manual Waterloo & City line. Drivers on the Victoria line on the other hand, only get to drive manually in and out of the depot.

 

[Zoe]

Isn't the plan for Thameslink is to overlay an ATO system on top of the ERTMS system? ATP signalling itself is generally insufficient to allow ATO, as an ATO box on a train can mis-read ATP indications on rare occasions, and can get confused if the train is diverted from it's pre-programmed paths due to delays etc. Thus, a central computer system monitors every train on the system, and by looking at the signals, timetables, speed of each train etc, this computer then tells the ATO equipment on board to change the speed of the train.

This is not needed, ATO boxes actually can work with the ATP. If there was for any reason a fault with the ATO box and it misread information from the ATP and for example thought the target speed was higher than it actually was then the ATP would intervene in the same way it would with a human driver. There would be nothing to stop both human and ATO driven trains running on the same route as they would both be protected by the same ATP so interoperability is retained. A study for a national ATO standard that would use the ERTMS ATP has actually already gone out to tender.