Driverless trains - why limited progress on the national rail network?

[najaB]

If the computer can't cope, brakes on and "Will the guard please proceed to the control panel at the front of the train" on the PIS.

There are some situations where that might not be the best idea - for example if the train has slid and is partially blocking a junction. It might be that a human driver would recognise that proceeding forward to clear the fouled junction is actually a better course of action.

Not to mention that the guard may be at the other end of packed ten-carriage train.

 

[miami]

Only another 8 years then I will be 56 & can happily retire with 40yrs in the pension scheme Can't see any movement apart from the partial XR & TL before then.

I think that's a safe assumption. As to the long term viability of your defined benefits pension scheme though, that's another assumption.

 

[Bletchleyite]

There are some situations where that might not be the best idea - for example if the train has slid and is partially blocking a junction. It might be that a human driver would recognise that proceeding forward to clear the fouled junction is actually a better course of action.

Not to mention that the guard may be at the other end of packed ten-carriage train.

There will be a few cases where a human may "save the day" with a bit of human judgement, but as those cases are relatively few it is likely that overall safety would be the same, or improved, over a case where a human is making a decision and may make the wrong one.

 

[Bodiddly]

...and I'd argue a human can react to a developing situation better than a camera and computer would.

How so? Can you give a situation where an AI system would not perform as well as a person? I can't.

 

[daikilo]

So, on the question of "can it be done" my answer is probably yes, but maybe not everything imediately.

One fundamental issue to be addressed is the rules under which fully automatic trains will operate as it will probably make no sense to try to mimic what we have today with crew operation. Are signals needed by the trackside.

There is then the question of whether we want to do it or do we prefer to keep drivers jobs.

Finally there is the question of where to start? In my opinion certainly not on one of our major trunk routes but more likely on a self-contained sub-network.

 

[miami]

There will be a few cases where a human may "save the day" with a bit of human judgement, but as those cases are relatively few it is likely that overall safety would be the same, or improved, over a case where a human is making a decision and may make the wrong one.

I would envisage a future where the train could be controlled centrally via remote control in those situations.

 

[NotATrainspott]

Also worth mentioning is the fact that automation doesn't limit you to the sensors you can have on a single train. The entire railway system becomes the robot, tasked with getting people and things from A to B. Tree fall detection, for instance, could be performed by drones sweeping the route ahead of a train looking for a problem. Every level crossing or high-risk area would be fitted with sensors and CCTV, just as new high speed lines are to avoid a Selby rail crash situation. While it might currently be extremely expensive to fit this all along the network, technology will only make it cheaper and cheaper.

 

[fredk]

The technology already exists to make trains fully automated with no need for external changes to infrastructure whatsoever.

The biggest barrier at this point are the unions.

The strikes over guards will seem quite pathetic in a few years time.

 

[NSEFAN]

How so? Can you give a situation where an AI system would not perform as well as a person? I can't.

A perfect AI can learn anything, but sufficiently powerful AI was not around in the 1960s when the Victoria line was introduced with ATO. The Victoria line is a closed environment, so used a comparitively simple control loop. An AI on the other hand would be highly non-linear and would learn about its environment in order to adjust its driving style according to the location and railhead conditions (in the same way that train drivers learn to handle trains).

The technology already exists to make trains fully automated with no need for external changes to infrastructure whatsoever.

The biggest barrier at this point are the unions.

The strikes over guards will seem quite pathetic in a few years time.

The driving may be automated but human oversight would still be required at this stage. It's simply cheaper to pay a human to sit up front to monitor the line ahead than to train an AI to do the same level of obstacle detection (for now, at least).

 

[Domh245]

The technology already exists to make trains fully automated with no need for external changes to infrastructure whatsoever.

It does? Go on...

 

[thomson787]

also those bloody unions would throw a hissy fit and not allow it to happen, The southern situation shows how militant the unions are, until the unions are forced to close I sadly can't see anything like this or any other good ideas for railways progressing.

 

[IKB]

The technology already exists to make trains fully automated with no need for external changes to infrastructure whatsoever.

The biggest barrier at this point are the unions.

also those bloody unions would throw a hissy fit and not allow it to happen, The southern situation shows how militant the unions are, until the unions are forced to close I sadly can't see anything like this or any other good ideas for railways progressing.

Riiiiiiiight....ok, what insight there is there.

--- old post above --- --- new post below ---

It does? Go on...

He'll prob quote you the DLR...wait for it, wait for it...!

 

[miami]

One fundamental issue to be addressed is the rules under which fully automatic trains will operate as it will probably make no sense to try to mimic what we have today with crew operation. Are signals needed by the trackside.

Are signals by the trackside in all cases (in cab signalling for example)?

There is then the question of whether we want to do it or do we prefer to keep drivers jobs.

Deltic said:
"This is not a discussion about the rights and wrongs of driverless trains but the technology which they require"

To answer his specific question - I believe the technology exists (or is under active field testing), however there's little financial incentive at the moment, and there's little movement by the rail companies as a large conservative oil tanker of an industry. However technology constantly gets better and cheaper, and the same applies to automation - it's inevitable -- even if the railway doesn't adapt, successful fleets of automated taxis that are cheaper per mile than a train will cause many branch lines to wither. A taxi to Crewe from my house is £15. The train is £1.50 each way, but with a network of automated cars it could also be £1.50, or even less. At that price, why would I bother with the train, even if it were free?

I don't see long distance high speed trains vanishing any time soon, but the Abbey line being paved over as a peak-flow automation-only car lane is something I could see happening in 20-30 years.

To answer the rights and wrongs, that's a far larger societal question, drivers are not a special category of person, people have lost jobs to automation through the ages, it's inevitable, be they secretarial pools, travel agents, claim adjusters. Even when there's no direct job losses, you find that automation allows a company to do more with the same number of people, where before they'd have to employ more. In days of yore there's be 2 or 3 people driving a train and half a dozen people on each station. That doesn't happen now.

How we deal with that as a society is something that could range from a post-apocolypic hellhole to a utopian society. It makes no sense to me to have people spending time jobs that could be automated, far better to increase the time people have to perform leisure. Currently the UK has a population of 65 million, of which 30 million work say 40 hours a week, so 1.2 billion hours a week. If we can automate society so 10 million of those people don't need to work, but there's 2 million new jobs looking after the new technology, that is the same output with 22 million working 40 hours a week, so 880 million hours a week. That means we have 320 million more leisure hours as a society. That's a good thing.

The trick is to ensure that work
1) gets done
2) is equitably shared

(1) is usually by reward for many of these automatable jobs. Who would do their job if they got paid regardless?
(2) is a "fairness" thing. Why should I get up at 6AM to go to work when my neighbour doesn't have to. Reward feeds into this too.

Finally there is the question of where to start? In my opinion certainly not on one of our major trunk routes but more likely on a self-contained sub-network.

We've already started. Heathrow T5 for example is an automated train that doesn't have a drive. Neil already suggested Stourbridge.

 

[Dave1987]

Always makes me laugh when this subject crops up, and the whole "driverless car" argument which must mean driverless trains are easy. For a start with cars they are still having a safety driver behind the wheel with their hands hovering over the wheel to take over at a moments notice. During a test in California with the Beeb in one of the test cars the safety driver had to repeatedly take over control. The car industry admits they are at least a decade away from car drivers being able to take their hands off the wheel but drivers will still be required to take control if necessary as the computers are being programmed to give control back to the human driver if it gets into a situation it simply cannot handle.

Thats before we get to all the arguments regarding trains.

 

[miami]

While it might currently be extremely expensive to fit this all along the network, technology will only make it cheaper and cheaper.

This is the key thing. And it's not just like-for-like replacement. As technology becomes possible and cheaper, old technology stops being useful. Just as branchlines shut down when the road network improved, the same continues now. Helicopters aren't automated, but there's less need for helicopters as many (not all, or even most, but a fair number) of the jobs they did can now be performed by drones.

 

[talltim]

As I see it there are two ways to automate trains, do the full system with a measure of centralised control, an updated version of the DLR if you will, or autonomous trains that react to th same inputs a human driver does, which is the route autonomous cars are going.
The advantage the second route has is that it requires minimal infrastructure changes, can be introduced gradually in parallel with human driven trains. Making an autonomous train should be easier than an autonomous car as the guidance is is removed

 

[NSEFAN]

As I see it there are two ways to automate trains, do the full system with a measure of centralised control, an updated version of the DLR if you will, or autonomous trains that react to th same inputs a human driver does, which is the route autonomous cars are going.
The advantage the second route has is that it requires minimal infrastructure changes, can be introduced gradually in parallel with human driven trains. Making an autonomous train should be easier than an autonomous car as the guidance is is removed

A driverless train would still have to inspect the line ahead as it travels, otherwise potential problems that a human driver would spot will be missed. Given the speed at which trains travel and the large stopping distances, I would expect sophisticated obstacle detection to be a feature of mainline driverless trains.

 

[dk1]

I think that's a safe assumption. As to the long term viability of your defined benefits pension scheme though, that's another assumption.

Nah! It's all looking very healthy indeed. No worries there.

 

[notverydeep]

It is interesting that most of the posts to date in this thread take it for granted that the rail network will continue or expand irrespective of what happens in vehicle automation across all modes. Despite being a life long rail enthusiast and having worked in the industry for 20 years, I rather suspect that this may not be the case. I would suggest that many of today's train drivers will never be replaced by automated trains, but rather their passengers and freight will be lost to autonomous road vehicles. If that technology takes off, and very large well funded companies are spending an enormous amount of cash to make it happen, then low traffic railways will be in real trouble.

Many of the rural and cross country routes are heavily reliant on two categories of passenger, both of whom would switch to road if cars were automated and cost no more than rail fares. These are the young, priced out of cars by high insurance costs and older people as they lose the confidence to drive longer distances. The trips these groups make are highly dispersed (from many origins to many destinations) and so usually require one or more interchange by rail. That and slow line speeds on such routes mean that rail has little speed advantage. Autonomous Road Vehicles would have other advantages, they won't have the fixed costs that go with 'ownership', they would be door to door, they could depart whenever the passenger chose and they could stop for refreshments pretty much whenever the occupant wanted.

There are quite a few cross country routes that may seem quite busy, e.g. two trains per hour in each direction, but those trains are two carriage sprinters carrying say 140 people. That said, I think 'mass transit' routes with either a good speed advantage (high demand intercity) or capacity advantage (metro or dense commuter routes) will survive - and with drivers at least for the next couple of decades. I think it is safe bet that taxi drivers are at more immediate risk of being put out of work than train drivers, because the technology in cars is likely to be easier and cheaper to implement in towns and cities first and it is there that the market is largest.

 

[AM9]

I know that the op declared that the post was "not a discussion about the rights and wrongs of driverless trains but the intoduction of a systemtechnology which they require", but the system's introduction would need a proven embodiment of the principle that in the event of a loss of human life, injury or other irreplaceable loss, there must be:

a means of identifying those responsible (be they an individual, a commercial organisation, an official body etc.), unless it was proven in a court of law to be an act of god.​

It is likely that there would be massive public resistance to the introduction of comprehensive ATO. In the event of an event involving serious loss to passengers/public that would otherwise be avoided in a manned driver environment could result in calls for the abandonment of the practice and reversion to something with a lot more human intervention.
Getting a viable system that met those objectives could take decades so I don't think that the current enthusiasm for an early introduction of a comprehensive implementation is realistic.

 

[miami]

A driverless train would still have to inspect the line ahead as it travels, otherwise potential problems that a human driver would spot will be missed. Given the speed at which trains travel and the large stopping distances, I would expect sophisticated obstacle detection to be a feature of mainline driverless trains.

That will be no harder than it is for automated cars.

However there is a big advantage in a robotic system -- a train upahead spots an obstruction across the track (say a tunnel caves in ala Watford), immediately all trains in the vicinity know about it -- no delay while a signaller is contacted etc.

It's all technically possible.

It is likely that there would be massive public resistance to the introduction of comprehensive ATO. In the event of an event involving serious loss to passengers/public that would otherwise be avoided in a manned driver environment could result in calls for the abandonment of the practice and reversion to something with a lot more human intervention.
Getting a viable system that met those objectives could take decades so I don't think that the current enthusiasm for an early introduction of a comprehensive implementation is realistic.

Far less of a problem than with automated cars.

There are quite a few cross country routes that may seem quite busy, e.g. two trains per hour in each direction, but those trains are two carriage sprinters carrying say 140 people. That said, I think 'mass transit' routes with either a good speed advantage (high demand intercity) or capacity advantage (metro or dense commuter routes) will survive - and with drivers at least for the next couple of decades. I think it is safe bet that taxi drivers are at more immediate risk of being put out of work than train drivers, because the technology in cars is likely to be easier and cheaper to implement in towns and cities first and it is there that the market is largest.

Fully agree, although there will be pressure on many commuter lines. Automated cars at 1/second would allow 3,600 per hour per track, it's a tempting way to increase road space.

 

[Dave1987]

That will be no harder than it is for automated cars.

However there is a big advantage in a robotic system -- a train upahead spots an obstruction across the track (say a tunnel caves in ala Watford), immediately all trains in the vicinity know about it -- no delay while a signaller is contacted etc.

It's all technically possible.

Far less of a problem than with automated cars.

I think you are overstating exactly where the technology is, how reliable the mass of sensors required are and how reliable the data is from those sensors. When simple obstacle detection was put in place at level crossings they had massive problems with sensors providing false readings. They now have to be cleaned regularly. You are talking about hundreds of tonnes of train travelling at very very high speeds with hundreds and hundreds of passengers on board, not a little hatchback running around town at 30mph. There are many many situations where human perception is required from a properly trained driver. Yes there will be gradual advances in signalling technology over the next few decades but truly driverless trains aren't even on the horizon on the mainline. Even the British Government are only currently discussing legislation to allow drivers to take their hands off the wheel, not sit in the back and play games on their smart phone.

 

[ComUtoR]

How does the DLR close the doors automatically ? Does it check the doors are free of passengers or just close the door ?

The DLR still has drivers and they still operate the trains. There is an issue somewhere and the technology isn't capable or reliable ?

ATO is happening through the core section so there is technology that will still drive the units but it still requires a Driver up front to stop the unit in emergency, deal with faults etc and close the doors. Is there technology that could remove the guy/gal up the pointy end ?

 

[Domh245]

How does the DLR close the doors automatically ? Does it check the doors are free of passengers or just close the door ?

The DLR still has drivers and they still operate the trains. There is an issue somewhere and the technology isn't capable or reliable ?

ATO is happening through the core section so there is technology that will still drive the units but it still requires a Driver up front to stop the unit in emergency, deal with faults etc and close the doors. Is there technology that could remove the guy/gal up the pointy end ?

The onboard staff close the doors on the DLR. They also drive the trains from time to time and in emergency situations.

I agree that when the time comes, the ATO trains will still need someone onboard for door closing, fault fixing, degraded working and passenger service. However we may find ourselves in the situation where they don't have to be in the cab.

 

[Dave1987]

How does the DLR close the doors automatically ? Does it check the doors are free of passengers or just close the door ?

The DLR still has drivers and they still operate the trains. There is an issue somewhere and the technology isn't capable or reliable ?

ATO is happening through the core section so there is technology that will still drive the units but it still requires a Driver up front to stop the unit in emergency, deal with faults etc and close the doors. Is there technology that could remove the guy/gal up the pointy end ?

Nope DLR door close is performed by the "train captain". There are sections where they have to take over as well. I was traveling on one when there was a points failure, it seems the automated driving systems where stoped straight away and manual driving taken over. I don't really understand how their signalling system works so couldn't tell you what conditions require manual control just that it's not an unusual occurrence for the "train captain" to drive manually.

 

[trash80]

Technically its possible of course, the interesting questions to me are should we and why?

Years ago i had an interesting discussion with a friend about the S'pore metro which was going driverless. He mentioned about the drivers losing their jobs, i said they could retrain. He replied "But what if they like their jobs?"

Is there room for such a sentiment in this world? Probably not but maybe there should be.

 

[J-2739]

This is true. HS2, OTOH, is new-build - is there any reason that should not be fully automated from day one?

HS2 is literally a hybrid of new track and Victorian infrastructure, so wouldn't really work. Add to the fact that ATO has never been applied to a service that fast (in terms of speed anyway)...

 

[Dave1987]

I agree that when the time comes, the ATO trains will still need someone onboard for door closing, fault fixing, degraded working and passenger service. However we may find ourselves in the situation where they don't have to be in the cab.

I think it all depends on what the real capacity benefits of ETCS signalling are. In a MR article Roger Ford has said he believes the 40% figure is wildly optimistic and it will probably be closer to 10%. Real benefits of digital signalling lie in higher linespeeds like on the East Coast to enable 140mph working. Widespread ETCS level 2 is still decades away. It all comes down to cost. Are the benefits of introducing full ATO with the potential problems, pitfalls and huge cost worth the increase in capacity it may give?

 

[bramling]

Nope DLR door close is performed by the "train captain". There are sections where they have to take over as well. I was traveling on one when there was a points failure, it seems the automated driving systems where stoped straight away and manual driving taken over. I don't really understand how their signalling system works so couldn't tell you what conditions require manual control just that it's not an unusual occurrence for the "train captain" to drive manually.

Any situation where people are on/about the track, if the train loses communication with the system, or during any degraded failure mode where there is some kind of signal failure (with control room, track or train borne equipment).

 

[IKB]

The onboard staff close the doors on the DLR. They also drive the trains from time to time and in emergency situations.

I agree that when the time comes, the ATO trains will still need someone onboard for door closing, fault fixing, degraded working and passenger service. However we may find ourselves in the situation where they don't have to be in the cab.

Which begs the question...WHY. Nobody has sufficient addressed this yet.

There will still be the salary outlay for the person on the train (as per DOO now).

The capacity benefit won't be a great as some expect. Take trains into South East terminals...there are only so many platforms and trains need minimal turnarounds. You can't have trains 2 minutes apart at intermediate stations, only for 10 of them to be queued in a line waiting outside a terminal station. A ten car in the peak will take a good five minutes to empty out and for passengers to walk the length of the platform, before the new passengers start to board. This is not the Underground.

The tech on the train may drive even more defensively than a human with worst-case scenario braking distances to account for weather conditions etc.

It will cost many millions or billions to implement (at todays prices) although may get cheaper in the future.

How is the capital expenditure recouped? And if operating overheads are not reduced significantly, how is the vast expenditure justified?

Lots of people stating what "can" be done now and in the future, but not many people setting out the benefits as to "why" it needs to be done. The fact that something can be automated does not in itself justify it.