Route Learning in Simulators

[jfowkes]

A comment in the EWR construction updates thread got me thinking about route learning:

Except that a more logical timescale would have been to have been in a position to start this process so that it finishes when the assets become available for use. (Accept that route learning would need to wait until then, but I would have thought that would be a relatively small part of the process.) Unfortunately I don't think the DfT do "logical" very often, and in particular the delays in choosing a TOC to run it won't have helped.

Basically, my questions are how much route learning is done in simulators now, and how much could be done in future?

I know nothing about the topic, so I don't know if simulators are too limited to really give you the real world fidelity you need for good route learning.

I imagine you'd need a number of real-world runs anyway. But if a simulator can get you 90% of the way there, plus getting you learning in different weather conditions etc., I imagine this could substantially reduce the amount of actual runs required? Seems especially useful for new lines like EWR, where the route doesn't actually exist yet.

 

[Sly Old Fox]

No route learning is done in the simulator and that won’t change.

 

[Gemz91]

Route videos can be used to utilise route learning. Not sure using sims would be a very cost effective way of route learning.

Edit to add: I’m not actually sure if some tocs do use sims for route learning, but if they did I seems to be an expensive use of the sims.

 

[Llanigraham]

There used to be a Route Simulator at Machynlleth when the first introduced ERTMS on the Cambrian. I spent a couple of hours on it one day, but I think it was used more for ERTMS learning than route learning.

 

[jfowkes]

No route learning is done in the simulator and that won’t change.

For what reasons? Running a sim must be cheaper than actually running a train, and (I assume) you can simulate a variety of weather conditions, seasons, leaf fall etc.).

Or is the accuracy/fidelity just not good enough?

 

[dctraindriver]

There’s not enough simulators for a start. And unless it’s like a flight simulator that moves around these are just a static box. You learn the route by driving it in all weathers. They do not replicate how a slide feels in leaf fall or where you are in fog. And the train is running anyway.

 

[Gemz91]

For what reasons? Running a sim must be cheaper than actually running a train, and (I assume) you can simulate a variety of weather conditions, seasons, leaf fall etc.).

Or is the accuracy/fidelity just not good enough?

Learning a route is usually done by travelling in a train that’s running in service already. Depending on a few things, this may either be done by driver under instruction of another driver or sitting in the second persons side watching the route as the train goes along. The train is an ordinary passenger service train (usually) so no extra cost is incurred.

Route learning trains may be organised for diversion routes where passenger trains don’t run or are infrequent. Not sure of the cost of these though.

 

[LCC106]

For me personally I find it hard to learn a route by sitting in the second man’s side or reading maps. I have to drive it to feel the gradients etc. and really pick up on things like linespeeds, names of stations, junctions and so on. Even when you think you know a route you might get a yellow signal where you’ve never had one before and scratch your head as to where the red is! As above, a simulator doesn’t give you the appreciation of slippy tracks, fog etc. Good question though.

 

[MCR247]

For what reasons? Running a sim must be cheaper than actually running a train, and (I assume) you can simulate a variety of weather conditions, seasons, leaf fall etc.).

Or is the accuracy/fidelity just not good enough?

I think you’re forgetting the start up cost of getting a simulator set up with all of the routes that need to be learned to a high enough level of accuracy and detail. I’m not a driver but from what I know, drivers sometimes use landmarks or features as guide points as when to brake etc. Can you be confident that these are in the exact same place in a simulator?

 

[bahnause]

We mainly use videos and a learning platforms with maps and pictures nowadays. Unless it is a large station where you might do a "guided walking tour".

 

[InkyScrolls]

Simulators do not, and never will, make good route learning tools. As other posters have said, only by driving the route can you truly get a feel for it and 'learn' it.

 

[etr221]

I understand the latest simulators are up providing for "unusual" (emergencies, etc) situation training, and for tasks as running a league table of how efficient/economic (in use of power, etc) drivers were, so I would imagine could be useful as part of the route learning toolkit, along with everything. Whether an appropriate one is available is another question.

 

[jfowkes]

Yeah I'm definitely not suggesting simulators would ever replace actual route learning. Just to what extent they can supplement it. But sounds like the answer is "not a lot".

I can imagine some cases though, when combined with videos and other materials:

• As mentioned by @etr221, lines/routes that are difficult to experience in real-life (things like diversionary routes? rarely used platforms? single line working? that kind of thing)
• Routes like EWR/HS2 that physically don't exist yet - this was what got me started thinking about it. In these cases you can't do it in real life or even watch a video. But a very good quality digital simulation might go some way to reduce the time it takes to get drivers trained on a new route. But I guess this would be a lot of cost for a fairly marginal time saving...

BTW this is all out of curiosity, I'm not advocating for any of this!

 

[12LDA28C]

Route videos can be used to utilise route learning. Not sure using sims would be a very cost effective way of route learning.

Edit to add: I’m not actually sure if some tocs do use sims for route learning, but if they did I seems to be an expensive use of the sims.

If a simulation of a new route can be provided before the route is actually open for train running then it's a very cost effective method of route learning.

For what reasons? Running a sim must be cheaper than actually running a train, and (I assume) you can simulate a variety of weather conditions, seasons, leaf fall etc.).

Or is the accuracy/fidelity just not good enough?

The accuracy and resolution that can be provided in route simulations these days is staggering. Not just in the railway infrastructure itself but surroundings as well including local buildings and other landmarks that can be used as points of reference.

Learning a route is usually done by travelling in a train that’s running in service already. Depending on a few things, this may either be done by driver under instruction of another driver or sitting in the second persons side watching the route as the train goes along. The train is an ordinary passenger service train (usually) so no extra cost is incurred.

Route learning trains may be organised for diversion routes where passenger trains don’t run or are infrequent. Not sure of the cost of these though.

Not for a brand new route such as EWR as drivers will be required to learn the route well before passenger trains start running.

 

[eldomtom2]

The accuracy and resolution that can be provided in route simulations these days is staggering. Not just in the railway infrastructure itself but surroundings as well including local buildings and other landmarks that can be used as points of reference.

"Can" is the key word. That's going to cost a lot of money...

 

[12LDA28C]

"Can" is the key word. That's going to cost a lot of money...

And yet some TOCs already have that level of detail in their simulators so clearly the money spent is seen as worthwhile

 

[eldomtom2]

And yet some TOCs already have that level of detail in their simulators so clearly the money spent is seen as worthwhile

Really? Can you provide any specific examples?

 

[Ediswan]

The accuracy and resolution that can be provided in route simulations these days is staggering. Not just in the railway infrastructure itself but surroundings as well including local buildings and other landmarks that can be used as points of reference.

Looks like something AI could help with, using cab ride videos as a source. That would still require human input. For example, you would not want a landmark rock misidentified as a cow, which was then modelled as moving around.

 

[Saj8]

Even if simulators could be made to be high resolution photo realistic, or even to use actual video footage, they would not be suitable for route learning, mainly because they use 2 dimensional screens, so the only way of judging how far away an object is would be to try to judge how large it is. 2D images mean no depth perception, which is essential when learning a route. I guess some sort of VR setup could be used, but then the user would not be able to see or interact with the actual cab controls whilst wearing the VR goggles. Another barrier, as had already been mentioned, is that current simulators are completely static, as opposed to aircraft simulators, where the whole cockpit is mounted on a hydraulic platform that can be moved according to the attitude of the simulated plane. To be an effective route learning tool, a train sim user would need to feel the effects of acceleration, braking, wheelslip, traversing points, etc. Quite a few of my colleages report feeling sick when they're using a full cab simulator because they expect to experience the feeling of motion as they are driving along. That's not to say sims can't be used as a supplimentary tool, but a route learner may as well just use a real train as it is in service and running anyway.

 

[Route115?]

It must be difficult to cover all rarely used curves, relief lines, etc by actually driving the route, unless you plan a trip that reverses umpteen times ideally in non traffic hours. Can you use a route video for this, rely on a simulator or have to actually drive it? A typical example would be Addlestone - Byfleet & New Haw (I forget the name of the curve) but there are many such examples.

 

[dk1]

We have route leaning videos for all routes available online but only hours with an instructor in the real driving cab count. No route learning is done on a simulator.

 

[LCC106]

Simulated videos that you watch on a computer are made for remodelling of routes prior to works being completed in some cases. Forgot about that!

 

[zwk500]

It must be difficult to cover all rarely used curves, relief lines, etc by actually driving the route, unless you plan a trip that reverses umpteen times ideally in non traffic hours. Can you use a route video for this, rely on a simulator or have to actually drive it? A typical example would be Addlestone - Byfleet & New Haw (I forget the name of the curve) but there are many such examples.

There are a couple of strategies for this - one is to have a handful of trains that use a curve or diversionary route at a time when few people are travelling and rotate drivers onto that working so that over the course of 6 months all drivers have a chance to refresh their knowledge. TPE do this in Manchester, and I think that Byfleet curve has some services to Woking rather than Weybridge (e.g. https://www.realtimetrains.co.uk/service/gb-nr:L57738/2024-07-01/detailed#allox_id=0). You don't have to cover every curve at once.
Another is to have an out of service train or loco that has a set of paths to cover all the lines (although not all could be run on the same day), and on any given day a different set will be used depending on what the driver needs to cover. LNER has paths in at Doncaster for this (see the 0Dxx services here: https://www.realtimetrains.co.uk/se...9?stp=WVS&show=non-passenger&order=wtt&toc=GR). 0D07 can run with 0D08, but if 0D06 runs neither of the others can. Between them they cover the routes to Leeds via Hambleton, Wakefield via Pontefract, and Church Fenton via Knottingley.
Or you can just put in short-term trains for when you need a particular line for planned work. This is quite common for once-a-year diversions or major blocks.

 

[Undiscovered]

There's nothing better than doing the real route.

I know exactly where I am in a journey, based in the feel of the track and the kick of the points as we go over them.

Bear in mind, I'm a guard, and I'm often having conversations and interactions with passengers, in the saloon,with no view of the surrounding scenery at all. I have to know where we are on the route, if anything happens, and to get into position to run the service smoothly.
At night, in the saloon, I'm running the route purely on feel alone

 

[75A]

In my experience nothing beats the real thing. People say about going up gradients, but coming down the other side can be as tricky.

 

[12LDA28C]

Really? Can you provide any specific examples?

XC's simulator based at their training centre at Derby has exactly that level of detail on some routes, as one example.

 

[eldomtom2]

Is the 2016 video of XC's simulator on Youtube still accurate? It doesn't look that visually detailed...

 

[Trackman]

From what I have seen on simulators, it's not for route learning but for process learning. For example, multiple alarms going off, obstruction on track, traction failures, passcoms, signal light out etc.. what buttons to press/ processes to follow etc..

 

[dk1]

From what I have seen on simulators, it's not for route learning but for process learning. For example, multiple alarms going off, obstruction on track, traction failures, passcoms, signal light out etc.. what buttons to press/ processes to follow etc..

Exactly that. Used for training purposes but certainly not route learning or for trainee driving hours.

 

[FGW_DID]

Simulated videos that you watch on a computer are made for remodelling of routes prior to works being completed in some cases. Forgot about that!

When Reading was remodeled with the new flyover, feeder lines and depot, a DVD and supporting booklet was issued to assist in learning the new routes. I've still got a copy somewhere at work.