Why isn't the Elizabeth Line Core Driverless? (Ditto HS2)

[DynamicSpirit]

I've seen various discussions on railforums over the years about the advantages of driverless trains (cost, and you have more flexibility because of not having to deal with driver availability when dealing with disruption or even when planning normal timetables), and the consensus seems to be that while driverless trains might be good in principle, there's no chance of them happening on the existing railway because our rail network wasn't built for driverless and the work to adapt it would be too great.

But the Elizabeth line core is completely new build between Paddington and Abbey Wood/Stratford and runs completely new custom-built trains. So presumably it could've been designed from the outset for that part of the route to be driverless. But it wasn't. I'm curious to know why.

Same logic applies to HS2: It's all new-build - so why is it apparently being designed without consideration for driverless trains?

 

[swt_passenger]

Elizabeth line core is already automatic train operation (ATO) - similar idea to a number of the tube lines, but like them you still need someone to check safety when the doors close. But I suppose you’re really asking about unattended operation. That won’t happen easily. Besides on a route with different signalling technology like Crossrail you still need someone sitting there ready to drive the train under normal signalling when leaving the core section.

 

[FOH]

There is the unattended reversing outside of Paddington bit. Not sure if it's yet in use though.

 

[Geeves]

If it was driverless in the middle would you not need two drivers now at either end to drive it on the conventional sections?

 

[DynamicSpirit]

If it was driverless in the middle would you not need two drivers now at either end to drive it on the conventional sections?

You'd need fewer drivers overall since - for example, someone driving a train from Shenfield would presumably get off at Stratford and could then quickly get back to work driving an Eastbound train that has just arrived from the core. The fact that a different driver would take over the train once it arrives at Paddington doesn't change that the total number of hours of a human driving the train would have gone down, so fewer drivers are required. And besides, trains heading to Abbey Wood would presumably in this scenario not need a driver at all East of Paddington.

Elizabeth line core is already automatic train operation (ATO) - similar idea to a number of the tube lines, but like them you still need someone to check safety when the doors close. But I suppose you’re really asking about unattended operation. That won’t happen easily.

Not necessarily unattended. It could be; or (at least, off peak when the train isn't too crowded) you could have someone walking through the train in order to provide the assurance to passengers of a staff presence plus deter anti-social behaviour (and deal with door operation if appropriate). Similar I guess to how the DLR works - where they have on-board staff at certain times when it's judged necessary.

Besides on a route with different signalling technology like Crossrail you still need someone sitting there ready to drive the train under normal signalling when leaving the core section.

Why would that person have to be on the train in the core? Why couldn't they simply board the train once it reaches Paddington or Stratford?

 

[stuu]

Unless you are going to go for unattended operation then it doesn't really make much sense for onboard person to not be the driver, the trains can and do get pretty rammed these days so there is a limited amount that an onboard member of staff can do. Also it must simplify the whole thing to have the same working practices everywhere, and continuity is a key to safe working as well

Having the driver onboard already avoids any issue with there not being a driver available at what are critical points on the route. Let's be honest, that is going to go wrong sometimes

(also the DLR never operates without onboard staff, although it technically can)

 

[800301]

There is the unattended reversing outside of Paddington bit. Not sure if it's yet in use though.

It is and seems to work pretty well

 

[MCR247]

You'd need fewer drivers overall since - for example, someone driving a train from Shenfield would presumably get off at Stratford and could then quickly get back to work driving an Eastbound train that has just arrived from the core. The fact that a different driver would take over the train once it arrives at Paddington doesn't change that the total number of hours of a human driving the train would have gone down, so fewer drivers are required. And besides, trains heading to Abbey Wood would presumably in this scenario not need a driver at all East of Paddington.



Not necessarily unattended. It could be; or (at least, off peak when the train isn't too crowded) you could have someone walking through the train in order to provide the assurance to passengers of a staff presence plus deter anti-social behaviour (and deal with door operation if appropriate). Similar I guess to how the DLR works - where they have on-board staff at certain times when it's judged necessary.



Why would that person have to be on the train in the core? Why couldn't they simply board the train once it reaches Paddington or Stratford?

As far as I know, there isn’t a metro system in the world that operates with a driver for part of the route, and without for the rest of the route, and I’d imagine for good reason.

For a start, would you really save that many drivers? How ‘quickly’ can the driver get back to work driving his next working back the way he came? 15 minutes later? By that time their original train would be about to exit the other end of the core, assuming there was a new driver ready to pick it up on time. Anything less than that is going to be a big performance risk.

So, if the number of drivers isn’t really reducing, you’re introducing a massive risk of trains blocking Paddington and Stratford due to the driver being delayed on its inbound service.

It would also potentially limit capacity, as dwell times at the changeover stations would likely be longer to allow the driver to set up/shut down.

I also think you misunderstand how the DLR operates. There is always a member of staff onboard in passenger operation. It isn’t part-time.

As for HS2, I don’t believe at any point of the plan it was going to be completely segregated from the rest of the network, so it definitely isn’t “all new build” in that sense.

 

[LNW-GW Joint]

The majority of the trains on HS2 will run through to the classic WCML at Handsacre after about 40 minutes.
There's no chance of driverless trains on the WCML, and you wouldn't want to stop to pick up a WCML driver en route.
The line will be run under ETCS and probably ATO, with supervision by a driver.
Now the line will go no further than Handsacre there is even less reason to go for driverless trains.
I don't think any overseas high speed line is driverless, except the short Shanghai Airport maglev line.

 

[dk1]

There is the unattended reversing outside of Paddington bit. Not sure if it's yet in use though.

Definitely in use. One of my GA colleagues was route learning the Electric lines from Shenfield to Stratford with a very friendly EL driver the other week & he invited him to join him through the core & back too. Said it was a very strange experience walking the train whilst it performed the move automatically.

 

[Bald Rick]

Why would that person have to be on the train in the core? Why couldn't they simply board the train once it reaches Paddington or Stratford?

So, if the number of drivers isn’t really reducing, you’re introducing a massive risk of trains blocking Paddington and Stratford due to the driver being delayed on its inbound service.

It would also potentially limit capacity, as dwell times at the changeover stations would likely be longer to allow the driver to set up/shut down.

Is the reason.

 

[Fincra5]

Don't forget, is the ATO system fails, someone needs to move the train.
If the Train breaks down, someone needs to deal with the train.

 

[Bald Rick]

Don't forget, is the ATO system fails, someone needs to move the train.
If the Train breaks down, someone needs to deal with the train.

Whilst this is true, there are plenty of systems around the world, including in this country, where those risks are managed withouhaving staff on the train.

 

[eldomtom2]

As far as I know, there isn’t a metro system in the world that operates with a driver for part of the route, and without for the rest of the route, and I’d imagine for good reason.

At least some of the Japanese metro lines that through-run onto other commuter lines at their ends swap crews at these points, which from a staff logistics point of view is the same thing as the Elizabeth Line operating driverless on the core.

 

[DynamicSpirit]

As far as I know, there isn’t a metro system in the world that operates with a driver for part of the route, and without for the rest of the route, and I’d imagine for good reason.

For a start, would you really save that many drivers? How ‘quickly’ can the driver get back to work driving his next working back the way he came? 15 minutes later? By that time their original train would be about to exit the other end of the core, assuming there was a new driver ready to pick it up on time. Anything less than that is going to be a big performance risk.

So, if the number of drivers isn’t really reducing, you’re introducing a massive risk of trains blocking Paddington and Stratford due to the driver being delayed on its inbound service.

It would also potentially limit capacity, as dwell times at the changeover stations would likely be longer to allow the driver to set up/shut down.

I understand the point about the performance risk due to a driver not being there at Stratford, but why would it take take time to swap from automatic to driver-controlled. Surely with todays technology, it ought to be possible to do that by the driver simply flicking a switch?

I was also under the impression that Thameslink trains drove themselves between Blackfriars and St Pancras, albeit with a driver still sitting in the cab? If that's correct, that doesn't seem to cause delays at the transition points.

As for HS2, I don’t believe at any point of the plan it was going to be completely segregated from the rest of the network, so it definitely isn’t “all new build” in that sense.

Well yes, obviously I wasn't imagining HS2 trains running driverless when they are running on existing lines! I meant, for the sections that actually are new build - so, for what's actually being built, basically between London and Birmingham (which by the way encompasses the entire route for London-Birmingham HS2 trains, so the issue of a driver joining en route wouldn't have arisen).

 

[Meerkat]

Unless you are going to go for unattended operation then it doesn't really make much sense for onboard person to not be the driver, the trains can and do get pretty rammed these days so there is a limited amount that an onboard member of staff can do. Also it must simplify the whole thing to have the same working practices everywhere, and continuity is a key to safe working as well

Having the driver onboard already avoids any issue with there not being a driver available at what are critical points on the route. Let's be honest, that is going to go wrong sometimes

(also the DLR never operates without onboard staff, although it technically can)

It doesn't make much sense for the staff member to be a driver locked away if they aren't needed!
Whereas if there is room to move along the train a steward would offer reassurance and travel advice, and if there isn't room the steward can sit in a cab/office and monitor the CCTV for issues (and things like 'much more room at the front','keep feet off seats please') etc

 

[gordonthemoron]

For precisely the same reason that Munich S-Bahn trains have a driver, the Stammstrecke uses ATO, but drivers are required outside the Stammstrecke, and it is easier to have them onboard the whole time

 

[DanNCL]

As far as I know, there isn’t a metro system in the world that operates with a driver for part of the route, and without for the rest of the route, and I’d imagine for good reason.

About a decade ago it was considered for the Tyne & Wear Metro and very quickly ruled out. Nexus themselves proposed it and noted at the time that nowhere else in the world did such a thing. The fact nowhere else appears to have even considered it since I think is quite telling.

 

[MCR247]

I understand the point about the performance risk due to a driver not being there at Stratford, but why would it take take time to swap from automatic to driver-controlled. Surely with todays technology, it ought to be possible to do that by the driver simply flicking a switch?

The driver has to open the cab door, get on board, close the cab door, take their coat and bag off and store them, adjust the seat to ensure they can see everything they need to see, perhaps adjust the climate control, and possibly more things I’ve not considered before we even get into logging into any systems/setting up any safety systems on the train - I’m not a train driver so I can’t comment on that aspect - but all of these small things will add up. I’m not saying every train would need 10 minute dwells at Stratford and Paddington, but any time taken above the dwell time is capacity lost. Which in turn reduces the benefit of going for systems such as ATO in the first place.

I was also under the impression that Thameslink trains drove themselves between Blackfriars and St Pancras, albeit with a driver still sitting in the cab? If that's correct, that doesn't seem to cause delays at the transition points.

Because the drivers stay in the cabs and therefore aren’t leaving or joining the trains mid journey?

Well yes, obviously I wasn't imagining HS2 trains running driverless when they are running on existing lines! I meant, for the sections that actually are new build - so, for what's actually being built, basically between London and Birmingham (which by the way encompasses the entire route for London-Birmingham HS2 trains, so the issue of a driver joining en route wouldn't have arisen).

Everything has a cost. To ‘upgrade’ from the signalling system currently specced, to one suitable for driverless would increase the costs of HS2. To make this worthwhile, there would need to be a corresponding increase in benefits. Most of the time, going for higher spec signally is done for increasing capacity and that is where the benefits are seen. I’m not sure capacity was expected to be a concern for HS2 on opening, and with it since being truncated to its current form, perhaps even less so.

It isn’t always a case of can we do something, it’s a case of why should we do it.

At least some of the Japanese metro lines that through-run onto other commuter lines at their ends swap crews at these points, which from a staff logistics point of view is the same thing as the Elizabeth Line operating driverless on the core.

I stand corrected about the existence of these lines. Do these lines have two changeover points though, as I’d argue that logistically that still makes it quite different. Having a train run by itself from A to B and then having a driver pick it up to drive it to C, is different to having a train going from A to D dropping its driver at B and picking a new one up at C.

 

[Fincra5]

Whilst this is true, there are plenty of systems around the world, including in this country, where those risks are managed withouhaving staff on the train.

But not in a Heavy Rail capacity. I can only think of Airport People Movers, in this country, with no-one onboard. Very much a fixed system.

I understand the point about the performance risk due to a driver not being there at Stratford, but why would it take take time to swap from automatic to driver-controlled. Surely with todays technology, it ought to be possible to do that by the driver simply flicking a switch?

I was also under the impression that Thameslink trains drove themselves between Blackfriars and St Pancras, albeit with a driver still sitting in the cab? If that's correct, that doesn't seem to cause delays at the transition points.



Well yes, obviously I wasn't imagining HS2 trains running driverless when they are running on existing lines! I meant, for the sections that actually are new build - so, for what's actually being built, basically between London and Birmingham (which by the way encompasses the entire route for London-Birmingham HS2 trains, so the issue of a driver joining en route wouldn't have arisen).

The within the Thameslink Core (London Bridge area/ Elephant & Castle to Dock Jnc or Canal Tunnels) the Class 700 can be driven in ATO... Certain hours of use apply.
There is always a driver onboard, as ATO requires someone to do the Platform-Train-Interface check, shut the doors and tell the train to "Go". - Said Driver must also be train in ETCS Level 2 signalling.
-
Could the train drive itself without someone, probably, would it? No.
You run into the same issues at Crossrail, there needs to be a Driver on conventional railways.

 

[Meerkat]

But not in a Heavy Rail capacity. I can only think of Airport People Movers, in this country, with no-one onboard. Very much a fixed system.


The within the Thameslink Core (London Bridge area/ Elephant & Castle to Dock Jnc or Canal Tunnels) the Class 700 can be driven in ATO... Certain hours of use apply.
There is always a driver onboard, as ATO requires someone to do the Platform-Train-Interface check, shut the doors and tell the train to "Go". - Said Driver must also be train in ETCS Level 2 signalling.
-
Could the train drive itself without someone, probably, would it? No.
You run into the same issues at Crossrail, there needs to be a Driver on conventional railways.

Why? Even the PTI could be done by someone else if it can’t be automatic, arguably more safely.

 

[eldomtom2]

I stand corrected about the existence of these lines. Do these lines have two changeover points though, as I’d argue that logistically that still makes it quite different. Having a train run by itself from A to B and then having a driver pick it up to drive it to C, is different to having a train going from A to D dropping its driver at B and picking a new one up at C.

Quite a few do, like the Toei Asakusa line, which connects to the Keikyu and Keisei railways at its ends - a typical northbound train on it starts out with a Keikyu crew, swaps to a Toei crew when it joins the Asakusa line at Sengakuji, then finally switches to a Keisei crew at Oshiage, the northern end of the line.

 

[MCR247]

Quite a few do, like the Toei Asakusa line, which connects to the Keikyu and Keisei railways at its ends - a typical northbound train on it starts out with a Keikyu crew, swaps to a Toei crew when it joins the Asakusa line at Sengakuji, then finally switches to a Keisei crew at Oshiage, the northern end of the line.

But again, is that not still different to the proposal here?

 

[eldomtom2]

But again, is that not still different to the proposal here?

As I said, it's not different in a crew logistics sense.

 

[MCR247]

As I said, it's not different in a crew logistics sense.

Right, ok I’ve looked back through your posts and now I think I understand what you’re getting at. The section of my first reply that you quoted was specifically talking about there not being a system which has trains operating driverless for part of the route and driver operated for the remainder. When first reading through your reply I thought you were pointing out places where that did happen, but now I believe you were talking about having crew changes part way through the route, which isn’t really what I was talking about at all.

 

[Bald Rick]

But not in a Heavy Rail capacity. I can only think of Airport People Movers, in this country, with no-one onboard. Very much a fixed system.

But the principle is the same for other, heavy rail, systems elsewhere in the world.

And as far as the risks and their controls are concerned, there’s no difference between ‘heavy rail’ and a fixed metro system.

 

[Busaholic]

As far as I know, there isn’t a metro system in the world that operates with a driver for part of the route, and without for the rest of the route, and I’d imagine for good reason.

For a start, would you really save that many drivers? How ‘quickly’ can the driver get back to work driving his next working back the way he came? 15 minutes later? By that time their original train would be about to exit the other end of the core, assuming there was a new driver ready to pick it up on time. Anything less than that is going to be a big performance risk.

So, if the number of drivers isn’t really reducing, you’re introducing a massive risk of trains blocking Paddington and Stratford due to the driver being delayed on its inbound service.

It would also potentially limit capacity, as dwell times at the changeover stations would likely be longer to allow the driver to set up/shut down.

I also think you misunderstand how the DLR operates. There is always a member of staff onboard in passenger operation. It isn’t part-time.

The whole idea is folly and madness of the highest order! I wonder if the OP has ever experienced Stratford Station in the rush hour. Any slightly added delay can throw a spanner in the works to a system where seconds count. It's the sort of stuff that some vastly overpaid management consultancy comes up with to whet the appetite of a costcutting government or transport authority, whereas giving that and other consultancies the permanent boot would save vastly more money and not adversely impact passengers.

 

[jagardner1984]

In a way, isn’t the reason why it doesn’t really function the cost cutting by linking existing services in the first place. If you build an entirely segregated route, with platform doors throughout, providing a double interlock on “nothing trapped” timer driven sensor doors which obviously auto reverse when they close on something / someone, then clearly it is possible to press go (either onboard or in a control room and for the train to proceed unstaffed). My understanding is the DLR is setup to require staff to operate doors, presumably as some form of compromise with the unions, there isn’t a reason why theoretically, it could be configured for them to auto open and auto close after a set period of time.

However, the minute you integrate with other human operated lines, lines without Platform doors, the risks to both service robustness and the procedures for auto operation become much more complex, and the human and technical factors of crossing over on route introduce so many more failure points that on a metro service, it is deemed easier, for the sake of a few minutes of “unnecessary” human driving in the core, just to stick with human driving throughout.

Whilst the DLR seems to be an overwhelmingly successful, human present but otherwise automated segregated system, aside from Airport systems, my understanding is Glasgow Subway will be the first example of Unattended Train Operation here, and that is also a segregated system, there being no examples of trains moving without humans present on the main line.

 

[Meerkat]

my understanding is Glasgow Subway will be the first example of Unattended Train Operation here

How are they doing the doors for that?

 

[whoosh]

The whole idea is folly and madness of the highest order! I wonder if the OP has ever experienced Stratford Station in the rush hour. Any slightly added delay can throw a spanner in the works to a system where seconds count. It's the sort of stuff that some vastly overpaid management consultancy comes up with to whet the appetite of a costcutting government or transport authority, whereas giving that and other consultancies the permanent boot would save vastly more money and not adversely impact passengers.

Yes! I think some people are forgetting the utter chaos that occurred when the Thameslink May 2018 timetable was introduced. Because hardly any drivers knew the full routes from Peterborough to Horsham, or Cambridge to Brighton, there were changeovers at Finsbury Park.
Cue trains stood blocking the southbound platform waiting for a late running northbound train which had the driver for the southbound train on it - and vice-versa.

Why introduce a performance risk of a crew change mid journey, when you can keep your current performance secure arrangement of having a driver stay with the train for the whole journey - which is how all Elizabeth Line driver duties are planned.

How are they doing the doors for that?

Glasgow Subway will have waist height Platform Edge Doors at all platforms.