It does seem slightly ironic that one of the projected benefits of automatic driving road vehicles is that they will be able to communicate with each other and form 'trains' where several are following one route, thus benefiting from aerodynamics (and freeing up more road capacity, which is a double-edged sword). Unlike the detection, steering and braking systems this has AFAIK yet to be demonstrated in practice, but let's assume it will happen some time in the next couple of decades. Would a combination of ATO and ERTMS allow something similar to happen on the rails? For instance, two trains one 'slipstreaming' the other till the second turns off at a junction, or stops at a station which the first one passes through? Whether such a thing will ever be possible, I have no idea...
-
Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Train Driving and Automation
- Thread starter DerekC
- Start date
- Status
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R
RailUK Forums
Possible? Yes. Economic? Less certain. As stated above road conditions (specifically acceleration and braking) don't change significantly in wet vs dry conditions unless you have enough water on the road for aquaplaning to be a risk. Also, road vehicles can normally stop within sighting distance.
Neither of these two statements is true for trains, which seriously complicates the development of self-driving systems.
thenorthern
Established Member
- Joined
- 27 May 2013
- Messages
- 4,712
There are other issue that would need to be factored in such as trains would need to be calibrated regularly for the routes that they use so that they stop in the platform at the right location.
Another issue would be platform changes and alterations and platforms that take 2 trains.
Even though certain systems such as the Docklands Light Railway, Victoria Line and the Northern Line use Automatic Train Operation they still either have a Driver or a Train Captain meaning there isn't any money saving so its not very cost effective.
Another issue would be platform changes and alterations and platforms that take 2 trains.
Even though certain systems such as the Docklands Light Railway, Victoria Line and the Northern Line use Automatic Train Operation they still either have a Driver or a Train Captain meaning there isn't any money saving so its not very cost effective.
Goldfish62
Veteran Member
- Joined
- 14 Feb 2010
- Messages
- 13,163
It is very cost effective. You would not be able to run 34tph on the Vic line without ATO. The only way you could achieve the same capacity with less, manually-driven, trains would be to significantly lengthen them and just imagine the cost of those platform extensions underground. Now THAT wouldn't be cost effective.
In a metro environment when you need to squeeze every last drop out of the infrastructure then it make sense. There are very few (probably no) areas on the mainline that need that intensive a service. Aspects of the technology that makes ATO possible will probably be included as driver aids, but it will be a long time before full ATO is needed on the mainline.
Note that there's a big difference between 'possible' and 'necessary'.
thenorthern
Established Member
- Joined
- 27 May 2013
- Messages
- 4,712
The line blockage detection system I think is the biggest issue, The Gerrards Cross Tunnel collapse in 2005 is the biggest example of why drivers are still safer in most cases. The line blockage detection will get better in the future though I think.
An interesting one will be nuclear trains, would you want a nuclear flask train driven at speeds of up to 100mph driven by a computer.
An interesting one will be nuclear trains, would you want a nuclear flask train driven at speeds of up to 100mph driven by a computer.
Given that the flask is designed to survive just such an impact without breaching, I would be less worried about that than a passenger train as us bags of water and minerals are a lot less robust.
If a train can't stop in line-of-sight distance, why would a driver be safer? What do they know that a computer doesn't?
DerekC
Established Member
This has been looked at before and no doubt will be again. It may be feasible technically - the trains can communicate with each other - the front train can tell the train in rear what speed it is travelling at and when it applies the brakes. Measuring accurately the distance between leader and follower is more problematic The odometry that is used for ERTMS isn't accurate enough for very close convoying and using lidar or whatever as cars do doesn't work because it could be beyond line of sight.
The issue really is - what's the benefit? The trains would still need to separate when coming to a terminus or where they diverge at a junction so it doesn't seem likely to increase route capacity very much.
Could experience be a factor ? The Driver is prepared for potential incidents are specific locations. When your fast through stations for example you are looking ahead at the platform because your aware of the suicide potential so you watch passengers.
The benefit could be the same as on road, when HGVs slipstream each other: fuel savings. The reasons (that I can think of) why it might not make sense on railways are timetabling and that railways take safety more seriously than roads (though as has been pointed out, trains are usually driving beyond their capability to stop within line of sight anyway).
I agree with you, but being prepared doesn't change the fundamental physics - doing 100mph, if you can see it you are going to hit it.
Not all trains are going 100mph. Although a computer driving my train at 100% scares the bejebus out of me.
One of the arguments against automation is that you have to convince the public. I remember a weird day whilst out with my little monsters. We were discussing trains (as always
) and they asked about the DLR.I told them it was automatic and the reaction I got from the coach was hilarious. I was stunned some people didn't know it didn't have a Driver. Back on your point. It's not just the high speeds and blind corners that are an issue. It has already been mentioned about object detection. Its the human element of a potential risk. How fast would an automated unit "drive at caution" ?
gtr driver
Member
- Joined
- 24 Apr 2015
- Messages
- 144
As a driver, I'm not dismissing automation because it makes me redundant. I have absolutely no doubt that automating trains has many advantages. However it is 60 years since the modernisation plan in Britain and we are still struggling to electrify the trunk routes and eliminate semaphore signalling and manual level crossings. Britain doesn't have a good record of heavy investment in its railway. Unless a very cheap way of converting the mostly Victorian infrastructure emerges I don't anticipate it being automated any time soon. New lines of course, like the Victoria line and the DLR, are a different matter - you build them as automated from the word go.
Automation doesn't necessarily eliminate the 'issues' caused by a human presence either. The trains will still need someone up front to respond to suicides, accidents, fires, passengers taken ill, concrete blocks on the track etc etc. Those people may not need as full a route and traction knowledge as current drivers, but they'll need some idea of where they are, probably have to be able to move the train at caution to get it out of trouble, and still be subject to working time regs, sickness, leave, and rostering issues as discussed on other threads. Industrial action by DLR staff still brought the trains to a halt.
As for driverless cars, there's little point comparing the two forms of transport. Cars are designed, marketed and sold to private individuals. They have always represented status and freedom for individuals, and are used for the vast majority of journeys made, and therefore politicians have more willingly invested in the road network and favoured lower costs of motoring. No one worries much about efficiency - no one stops you from owning a tank like car that only you ever inhabit despite the fuel premium and the draw on the Earth's resources. Whereas the screws are always on the railway. The trains have to try to be matched to several markets with little room for individual satisfaction (look at the difficulty in providing a Thameslink fleet that pleases everyone) and the infrastructure is always lagging behind as it's politically harder to justify the investment ie I never use the train why should my taxes fund electrification.
It's telling that even on the continent where electrification and high speed lines are way ahead of us there's been no large scale adaptation of driverless trains - maybe it's just possible that the costs really do outweigh the benefits.
Automation doesn't necessarily eliminate the 'issues' caused by a human presence either. The trains will still need someone up front to respond to suicides, accidents, fires, passengers taken ill, concrete blocks on the track etc etc. Those people may not need as full a route and traction knowledge as current drivers, but they'll need some idea of where they are, probably have to be able to move the train at caution to get it out of trouble, and still be subject to working time regs, sickness, leave, and rostering issues as discussed on other threads. Industrial action by DLR staff still brought the trains to a halt.
As for driverless cars, there's little point comparing the two forms of transport. Cars are designed, marketed and sold to private individuals. They have always represented status and freedom for individuals, and are used for the vast majority of journeys made, and therefore politicians have more willingly invested in the road network and favoured lower costs of motoring. No one worries much about efficiency - no one stops you from owning a tank like car that only you ever inhabit despite the fuel premium and the draw on the Earth's resources. Whereas the screws are always on the railway. The trains have to try to be matched to several markets with little room for individual satisfaction (look at the difficulty in providing a Thameslink fleet that pleases everyone) and the infrastructure is always lagging behind as it's politically harder to justify the investment ie I never use the train why should my taxes fund electrification.
It's telling that even on the continent where electrification and high speed lines are way ahead of us there's been no large scale adaptation of driverless trains - maybe it's just possible that the costs really do outweigh the benefits.
You've added a lot of time onto a journey and thrown 100s of coaches into already congested cities. People will just use Virgin Trains on the classic lines instead.
As the driver can't stop in the distance he can see to be clear, does it matter if that changes.
Most people are happy with computers flying their planes at 500mph, including landing. It amused me that train fans are so protective over why the railway (with one variables, go/stop, on dedicated infrastructure) is so much harder to automate than roads (with go/stop and left/right) or planes (with go/stop, left/right and up/down).
D1009
Established Member
And as has been made clear in another thread, no-one's bothered about trains being driven at line speed in thick fog.
I place the railway as a similar level of difficulty as self-driving cars, and considerably more difficult than aeroplanes. As has been pointed out numerous times above, road traffic operates on a stop on sight basis but railways do not. So while the route is determined by the rails, they require a much higher degree of predictive ability - you need to be able to react to things that cannot be seen.
Other than the terminal environment, things are a *lot* simpler for aircraft as their motion is determined by very predictable physical laws and there should never be a need to avoid another aircraft (1,000ft separation) and the mountains tend not to move either.
So how exactly does a driver get this information? Sixth sense? Gut Instinct?
Because weather and turbulence never changes.
Bletchleyite
Veteran Member
I'd say rail is fairly easy - 99.999% of the time a train just has to do precisely what the signalling system and specified speed limit says. You would still want on board staff for safety and customer service, but as with the DLR that might be better being a guard.
Last edited:
bramling
Veteran Member
That's the easy bit. The bit that's harder to crack is stopping the train accurately, without losing time needlessly or having operating incidents. It's a pretty fundamental issue, that none of the ATO systems currently in use in Britain have been able to fully address.
With the performance of modern human drivers being fairly good (I think I'm right that Southall is the most recent fatal accident caused by driver error that wouldn't have been fully protected by TPWS), ATO introduces a range of issues without providing major benefits - especially if we're still retaining a member of staff on the train. Railway engineering is a compromise, and currently the scales tip away from ATO. If performance of human drivers is unable to meet a specification for a specialised application (for example Thameslink) then the scales may tip the other way.
Some of it is gut instinct, some of it is from the signalling systems. I'm not saying that train driving can never be automated because clearly it can, but the experience implementing ATO out in the wild has been mixed.
(a) Airliners tend to avoid the worst weather; (b) the way to fly through turbulence (at least during cruise) is to set turbulence penetration speed, set the autopilot to attitude hold and then keep your hands off the controls.
D1009
Established Member
The turning point could come if someone invents a braking system which is proved to be totally reliable in all weathers.
I'd agree with that. Electromagnetic eddy brakes aren't affected by poor rail conditions, but are also not effective at low speeds.
gtr driver
Member
- Joined
- 24 Apr 2015
- Messages
- 144
Paul, again, it's not about being protective, it's whether those who control the purse strings are willing to invest in the technology to make it happen! Just because something is possible doesn't mean that it will happen.
I can see the financial argument, however financial situations change as technology progresses -- will freight for example be cheaper on the rail than on a fleet of automatically driven lorries?
A lot can change in the technology world in 50 years, even if it doesn't in the rail world. Will automatic mixtures of PRT and bus technology mean the end of some branches -- Looe for example, or Exmouth, or Greenford? Will technology mean less need for high capacity peak time trains (through increased mobile working as well as easier car-based commutes as you don't need to deal with the hassle of driving or parking)? Will there be another beeching style cut to pay for the next banking crisis?
The economy can change easilly and dramatically, and rail isn't immune. If 18th century rail technology can't keep pace with more modern technology, will it carry on regardless, or will it go the way of the horse and cart?
I suspect high speed rail will have a continuing resurgence, assuming that fear doesn't close down the railway by putting in airline style security on the rest of the network (as you currently get on eurostar), I'm not convinced about all branch lines.
Goldfish62
Veteran Member
- Joined
- 14 Feb 2010
- Messages
- 13,163
The German high speed lines use LZB, which is a form of ATO. However, it only recognises signals and line speed and the driver still has to brake for stations.
notadriver
Established Member
- Joined
- 1 Oct 2010
- Messages
- 3,737
But planes aren't pilotless and the automatic pilot is an pilot aid. The pilots must still oversee the autopilot and be ready to take over at any time if necessary.
It's also possible that fear could send almost all long-distance traffic on to trains, by closing down air travel and rendering roads too unpredictable, cf Atwood's Madd Addam series.
carriageline
Established Member
- Joined
- 11 Jan 2012
- Messages
- 1,897
Which is exactly what ATO will be, and that's how it will be for the foreseeable future. The general driving and stopping of trains can be automated, it's when you have degraded working and when things go wrong that you need a driver, but even then some of that can be automated to an extent. IE emergency speed restrictions.
- Status