Would You Support a Driverless Railway?

[whhistle]

I was chatting about this to a friend the other day and with all this talk of "people" announcers facing computer replacements and an ever increasing amount of driver-only trains, one group of employees who are escaping any sort of change are drivers.

I am not suggesting we shoud end jobs for anyone, but it's an interesting concept to think about surely? Don't get me wrong, I realise how much knowledge drivers have to know, it isn't just physically controlling the train. I also realise the DLR is driverless and that seems to work well. Some lines on the Underground could also look at having automatic trains.

The problem that was put fourth was an obstruction on the track. A driver would obviously be able to take action, where as an automatic train may not. But how does the DLR work? I'm not saying we should have trains running around with passengers only either. An employee would be on every servce under the title Train Manager or something.

Another case to look at is merging the Train Manager position and a Driver position, so the TM can leave the "cab" to do ticket checks, but they are also trained to be able to monitor the track ahead and take overriding action if needs be.

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

 

[Mike395]

Whilst I think it works OK for metro systems such as the London Underground and DLR, I think that it will never work fully for mainline trains - there are just so many variables that could happen on an open railway running at up to 140mph.

From what I know, the DLR can detect blockages, and very rarely goes faster than about 40-50mph so stopping distances are far lower than mainline rail. (The top speed its trains are capable of is 62mph, but there are very few long straight stretches of track).

 

[yorkie]

I am not sure what you are asking. Is this a thread asking if we support the DLR? Or are you asking if we should convert existing Network Rail metals to driverless?

The DLR works well. If you are asking whether I "support" it, I do. It's a fantastic system. My only criticisms are that the hosts are not as visible and customer-focussed as the hosts on many trams that I have been on, so there could be improvements in that area, and that they do not take bikes. But that's not a criticism of the way it works!

If you are asking if we could extend this to the entire railway system, the answer clearly is no. It works very well for a self-contained relatively small system, but the idea of having freight trains, express passenger trains, stopping trains, all interacting in such a system over thousands of miles, is flawed.

I am not answering the poll. I considered answering with "Yes - Driverless Could Work" but that is false, as it clearly DOES work! I suspect if I voted yes, you'd be counting my vote toward converting the national railway system into driverless, which would be ludicrous! So no vote from me.

 

[mallard]

I think that the use of driverless (or at least ATO) systems could be extended. Metro/commuter services are the obvious ones and it is already the plan for Thameslink and Crossrail. Extending it over a larger part of the former Southen Region would make sense.

Maybe make the whole Thameslink route ATO, along with the rest of the Brighton line as a sort of trial. A bonus would be allowing EMT to run ATO south of Bedford to see how well it works with Intercity-type services. It could even be trialled with freight to the cement terminal near St. Pancras.

I doubt that it would be well-suited to complete long-distance services, but for metro/commuter lines and certain portions of Intercity routes, it could work.

 

[Metroland]

I suspect it would be too expensive to be universal, and you would still probably need someone on board/in the cab in case of breakdown, to keep a good look out etc

The reason ATO tends to be applied to metros is you can get more trains per hour along a given section of track, and for this reason it will apply to parts of the Thameslink network.

For that reason I can see it being extended (where cost effective) on some heavily trafficked lines around major cities. But not on the Whitby branch any time soon.

I think it's inevitable automation is going to creep into jobs where computers are as good or better at doing some things than humans, it's been happening in the aviation industry for years and is starting to creep into road vehicles. Don't forget modern intercity trains have cruise control which maintains speed, I suspect when ERTMS comes along, this will be integrated with the signalling, and the driver will be monitoring it and dealing with things computers are not so good at, rather like ARS in signalling (which allows signallers to handle much bigger areas than they otherwise could, leaving them to do stuff like dealing with incidents/communications).

I believe quite a few European inter-city trains pretty much drive themselves anyway, and increasing metros.

 

[FGWman]

Whenever I get a driver getting too big for his boots telling me now essential he is for the railway I tell them GOO (Guard only operated) is the future.

Usually shuts them up

 

[Class172]

Although it works well on light rail systems such as the DLR, I'm just not convinced that it would work on the larger scale at much higher speeds, not to mention the fact that the passengers would be freaked out at the fact there is no driver. That is why I have voted no.

 

[TDK]

I voted no as I would be out of a job if I did

 

[exile]

I've said yes - but that's in principle. I've used one (DLR).

I suspect there will always have to be one staff member who CAN drive the train in an emergency (as is the case with DLR AFAIK) and with long trains there will probably always have to be someone at each end of the train to check doors are closed etc before setting off. But do they actually have to drive the train for the entire journey? Could they be like airline pilots who actually only fly the plane for a part of the journey (takeoff and landing) and if something goes awry?
I've seen videos of TGV drivers at work and once they get onto the dedicated LGV track they seem to have virtually nothing to do unless a cow gets onto the track (in which case I think it's minced beef in the dining car)

Given you'd have to have someone qualified to drive in an emergency I'm not sure what the savings would be.

 

[gordonthemoron]

U2 Nuremberg is driverless, but is driven remotely. ICE3s are computer controlled most of the time, it is feasible that the driver could do other tasks when not required (as opposed to texting, reading the newspaper, chatting to colleagues, having a coffee)

 

[Via Bank]

On mainline railways I would support ATO in principle, but only if there was a driver/operator at the console to apply the brakes manually if something goes wrong. There are far too many variables to consider a DLR-type arrangement where the guard or another service agent is trained to operate the train in an emergency.

Until you perfect artificial intelligence, the safest way to run a mainline railway will always be with someone eagle-eyed peering out of the front, even if the only control they have is a big red "EMERGENCY STOP" button.

 

[Failed Unit]

You need to remember the DLR / Victoria line are fairly self contained and very little interface between the outside world!

Imagine if you had a driverless ECML, what would stop the 125mph when some @£$%head puts something on the line, at least the drive may see it and hit the breaks the computer will just be full speed ahead. I have nothing against computers assisting the driver but not replacing!

Hope the OS isn't Windows - blue screen of death in the middle of a cold winters night!

 

[SS4]

No, computers are good at simple, repetitive tasks with no variation. They're a great idea for calculating all the prime numbers up to 1000 for example or for arranging huge lists.
The human brain is more adaptable and more powerful when it comes to variable tasks as well as being more reliable.

The DLR is slow and doesn't really get extremes of weather like a mainline. Then of course there is the psychological side of it, to know someone, not something is taking care of you
--- old post above --- --- new post below ---

Hope the OS isn't Windows - blue screen of death in the middle of a cold winters night!

Or running on IE6

 

[FGWman]

You need to remember the DLR / Victoria line are fairly self contained and very little interface between the outside world!

Imagine if you had a driverless ECML, what would stop the 125mph when some @£$%head puts something on the line, at least the drive may see it and hit the breaks the computer will just be full speed ahead. I have nothing against computers assisting the driver but not replacing!

Hope the OS isn't Windows - blue screen of death in the middle of a cold winters night!

Computers would probably see a fixed obstruction on the track just as well as a driver would. The latest technology at Level Crossings will shortly include obstruction detection which can put the signals at danger if there is an obstruction on the track.

Its probably like pilotless planes. The technology is there it could be done but the travelling public will not accept it.

 

[yorkie]

Imagine if you had a driverless ECML, what would stop the 125mph when some @£$%head puts something on the line, at least the drive may see it and hit the breaks the computer will just be full speed ahead. I have nothing against computers assisting the driver but not replacing!

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

 

[142094]

DLR has its problems as well - one person who fell onto the track at a DLR station was not picked up by CCTV and subsequently died when the train ran over him.

 

[Via Bank]

Hope the OS isn't Windows - blue screen of death in the middle of a cold winters night!

The present-generation Desiro TMS runs on Windows 3.11. The Electrostars run Windows 95.

It's old software, it's had time to mature enough so that it will be suitably reliable to run in such an embedded environment.

 

[Failed Unit]

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

True Yorkie, but at least they will be attempting to stop. I know a small object such as a sleeper or a tree they will hit irrespective, but a larger object (or failed unit) at least we have a chance.

 

[mallard]

I don't think "driverless" is really the correct word. I wouldn't want a railway that is completely without drivers. However, an largely (or even completely) ATO railway is desirable.

Except, it'll never happen. Installing ATO costs money and the main benefit is that, in theory, drives would have less responsibility and thus could be paid less. The unions will never accept this, so, it'll only happen where the secondary benefits (enhanced reliability and capacity) can justify it on their own.

 

[valenta]

I am yet to be convinced by driverless railways. I think that they can often been dangerous and more hassle than they are work. Surely it would mean a huge loss of jobs if railway companies turned to a driverless option?

 

[Greenback]

I don't think "driverless" is really the correct word. I wouldn't want a railway that is completely without drivers. However, an largely (or even completely) ATO railway is desirable.

On what grounds would a largely or completely ATO railway be desirable? Just curious!

 

[mallard]

On what grounds would a largely or completely ATO railway be desirable? Just curious!

Improved reliability, capacity and even speed on some routes (at least a reduction in "padding" due to the extra reliability) as well as lower ongoing costs (in theory, good luck getting that past the unions as I said).

 

[Greenback]

Improved reliability, capacity and even speed on some routes (at least a reduction in "padding" due to the extra reliability) as well as lower ongoing costs (in theory, good luck getting that past the unions as I said).

I'm not sure how it would be more reliable. It seems to me that the more advanced trains have become, the more chance there is of thigns going wrong!

I'm a bit dubious about capacity and speed as well. I don't see how these are affected by whether a train is driven by a human being or a machine, but maybe I am just thick!

 

[mallard]

I'm not sure how it would be more reliable. It seems to me that the more advanced trains have become, the more chance there is of thigns going wrong!

I'm a bit dubious about capacity and speed as well. I don't see how these are affected by whether a train is driven by a human being or a machine, but maybe I am just thick!

Reliability in terms of sticking to the timetable, not the actual mechanical and electrical reliability of the trains!

The benefits come because the computer would drive the train far more precisely than a human driver would. It would always get the acceleration and braking curves exactly right, always leave a station exactly on time (unless overridden for some reason), always leave a signal the moment it clears, etc.

It would also not be subject to human difficulties with signal sighting and poor visibility. With moving-block signalling the trains can maintain the _exact_ minimum separation in a way that a human driver would never manage, increasing capacity.

 

[GB]

The benefits come because the computer would drive the train far more precisely than a human driver would. It would always get the acceleration and braking curves exactly right, always leave a station exactly on time (unless overridden for some reason), always leave a signal the moment it clears, etc.
.

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.

 

[Greenback]

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.

Indeed, I've always been led to believe that driving trains is an art, rather than a science, with human judgement required to cope with different train characteristics, weather conditions and the like. Can all this replicated by a computer? I don't know, I'm no expert by any means, but I have my doubts!

 

[amcluesent]

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

I'd expect a forward scanning radar would do better in fog, rain, darkness in detecting any objects on the track than a driver.

Plus at peak times, two trains could be computer linked and run just a few metres apart at full speed with syncronised brakes and traction.

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.

 

[Yew]

I'm not sure how it would be more reliable. It seems to me that the more advanced trains have become, the more chance there is of thigns going wrong!

I'm a bit dubious about capacity and speed as well. I don't see how these are affected by whether a train is driven by a human being or a machine, but maybe I am just thick!

I would imagine in some places route frequancy could be increased (if only slightly) If we integrated automatic trains into a computerised routing system you could probably get more paths around major stations

Also the telemetry of trains could be comunicated to others, so for example if one train performs an emergency stop, the train behind it could bring its self to a halt in plenty of time

 

[2Dogbox]

Lets not give McNaulty any ideas!

 

[Nonsense]

No, computers are good at simple, repetitive tasks with no variation.

Like driving a train?

I don't see any technological barrier to it. A few tabloid grunts might raise the odd objection but most travelers wouldn't bat an eyelid.

The only significant obstruction will be unions.