Rail drivers' 28% Pay Rise to end Strike

[DY444]

I've said nothing to the contrary, but I still think it's naive to think a machine can't out perform a human in all of those tasks, whilst making fewer mistakes. People are generally the weak link. Of course it'll need sophisticated AI, I'm not saying we could just shoehorn an old Pentium 2 into the cab of an Electrostar and crack on. It's decades away, but I'd be surprised if it didn't happen.

 

You can come up with as many obscure examples about dog owners on one leg chasing people off parapets, but again, cars have to deal with all of that, and far more, and they're doing a pretty good job of it so far. Not perfect, but I'm not sure trains will face many things that we won't have overcome whilst learning about cars.

Still, that's a totally different discussion, I get it's emotive, but it'll happen.

 

My view is that a driverless car is going to have to take a very defensive approach to the dynamic assessment of risk to avoid litigation.  As a result it will cripple road capacity. 

[Reorte]

My view is that a driverless car is going to have to take a very defensive approach to the dynamic assessment of risk to avoid litigation.  As a result it will cripple road capacity. 

This is really going off-topic, but I think it very much depends upon the situation. On B roads, in town centres with pedestrians everywhere, yes. On motorways and dual carriageways they should be able to drive pretty much nose to tail with demonstratably greater safety than human drivers, which will increase road capacity. Combine with not being as prone to jamming up when such a road is at near capacity (the situation where one car brakes a bit can result in the road behind it grinding to a halt as the reactions ripple out).

[PaulRhB]

I don't doubt it will at the rate AI is going but just for clarity these aren't obscure instances they are fairly common. It may not get reported widely but we deal with this stuff regularly, it's the fact that we deal with it before it becomes a major incident means you don't hear about it.

Yes cars have to deal with random obstacles but not the same suicide problem that impersonal trains do or fact trains take further to stop and cannot take avoiding action. The railway requirements are quite different as a result. Where the money comes from to develop the railway version is another matter and by far the easiest solution are 'sealed' lines like HS1 & 2.

OD crossings are the first step to adapting the Victorian legacy and an example of how technology can vastly improve safety. My concern as I said above is the human element telling the AI what to look for and how to react, you aren't going to replace what we expect of it in hindsight with AI unless a human tells it first

[phil-b259]

I too can see the dead hand of political interfearance. It is so blatant that even those who are anti trade union can see it.

 

And with another general election years away thanks to Mrs May's gamble (that didn't go to plan) Mr Wilkinson, Mr Grayling etc will have plenty of time to ensure they get what they want.

 

Give it another couple of years and despite all the anger it raised pretty much everything Mr Wilkinson said will have come to pass. If only the RMT wasn't quite so pig headed in this (I do understand why they reacted as they have done) then something could perhaps have been salvaged and Mr Wilkinsons plans blunted slightly.

[Mark Saunders]

The art is in which battles to fight and which to walk away from!

 

There are a lot of vested interests at play here!

 

Mark Saunders

[w124bob]

Some of the technology is there but it's the cost of applying it and fitting it to a safety framework. How does a computer assess if there's stuff hanging off the outside foul after an impact, how does it assess what it has hit, human or animal? These things have huge legal consequences too. If a dog is reported lineside it is often followed by a distracted, anxious owner. I've had to deal with owners crying and incredibly distraught and not fully thinking where they are in relation to the line, I've had people sitting on parapets obviously preparing to jump. Programming all the possibilities and recognising the potential is going to need very sophisticated AI to replace a driver. We have lots of monitoring for slips in known places but drivers can recognise earth movement too. Another example is bridge bashes where a computer scanning the track would possibly only pick it up once the rails are distorted but a driver can see the chaos on the road to their side. On one occasion a tractor went under the bridge with a full trailer and came out the other side with a flatbed. Gates left open that encourage the next user to just go straight across etc the possibilities are vast and you need to equip every train with the sensors to detect this. The sums are vast whether you modify train or infrastructure and it needs a very high reliability rate.

I still see a huge gulf between what's possible and what's acceptable when it goes wrong at present.

And how does the computer know whats happened to the object after you've hit it. Many years ago I hit a 30ft length of rail which had been dragged onto the line, how the 142 stayed on still surprises me. I was on the DN in the Atherton area(ex 4track main) and thankfully the rail was thrown into the very wide gap between up and down lines. What a computer would not have been able to do was check were that rail was and also note that the rail guards were just a fag paper thickness from touching the wheels. But at least the computer wouldn't need clean underwear and a couple of days off(which I was told in no uncertain terms to take by my gaffer) Edited November 12, 2017 by w124bob

[D854_Tiger]

My view is that a driverless car is going to have to take a very defensive approach to the dynamic assessment of risk to avoid litigation.  As a result it will cripple road capacity. 

 

Driverless cars bring other problems for road capacity.

 

All I will say is that when they first brought out TV remotes to change channels, rather than getting off your fat arse to do so, channel hopping was invented.

 

What's the betting driverless cars, getting you into town, are then promptly sent back home again to save on the parking charges or getting you to work are then sent home again for the school run.

 

I can foresee lots of empty mileage but what else to expect when you invent a car that can do it.

[D854_Tiger]

And how does the computer know whats happened to the object after you've hit it. Many years ago I hit a 30ft length of rail which had been dragged onto the line, how the 142 stayed on still surprises me. I was on the DN in the Atherton area(ex 4track main) and thankfully the rail was thrown into the very wide gap between up and down lines. What a computer would not have been able to do was check were that rail was and also note that the rail guards were just a fag paper thickness from touching the wheels. But at least the computer wouldn't need clean underwear and a couple of days off(which I was told in no uncertain terms to take by my gaffer)

 

I believe it's very easy to get carried away with the concept of driverless cars.

 

They have defined three stages towards automation and they haven't got past stage one yet, which is where the car provides automated driver aids but still requires supervision.

 

Then when considering we've had automatic gearboxes for well over fifty years and they still can't persuade the majority of car buyers to buy them .... you do wonder who will buy these.

[PhilJ W]

Driverless cars bring other problems for road capacity.

 

All I will say is that when they first brought out TV remotes to change channels, rather than getting off your fat arse to do so, channel hopping was invented.

 

What's the betting driverless cars, getting you into town, are then promptly sent back home again to save on the parking charges or getting you to work are then sent home again for the school run.

 

I can foresee lots of empty mileage but what else to expect when you invent a car that can do it.

The idea is that you call up a driverless car (on an app?) only when you need to make a journey. I can see a fortune being made for returning lost property.

[D854_Tiger]

The idea is that you call up a driverless car (on an app?) only when you need to make a journey. I can see a fortune being made for returning lost property.

 

So a taxi without a driver.

 

Should be fun, at two in the morning, getting in one of those after a drunk has vacated.

[woodenhead]

Driverless cars bring other problems for road capacity.

 

All I will say is that when they first brought out TV remotes to change channels, rather than getting off your fat arse to do so, channel hopping was invented.

 

What's the betting driverless cars, getting you into town, are then promptly sent back home again to save on the parking charges or getting you to work are then sent home again for the school run.

 

I can foresee lots of empty mileage but what else to expect when you invent a car that can do it.

That's an interesting perspective, not thought of that but i could see it happen, just like a driverless car circulating a city whilst it waits for you to call it back

[woodenhead]

I believe it's very easy to get carried away with the concept of driverless cars.

 

They have defined three stages towards automation and they haven't got past stage one yet, which is where the car provides automated driver aids but still requires supervision.

 

Then when considering we've had automatic gearboxes for well over fifty years and they still can't persuade the majority of car buyers to buy them .... you do wonder who will buy these.

They want taxis to be stage two (autonomous with a set area) whilst personal cars they are all going straight to Stage 3 development because they realise any half way house is likely to more dangerous because drivers will forget when they need to take back control

[Dunsignalling]

I've seen it reported that all SWR drivers have DOO in their contracts and have done for years.  If true then the RMT will not stop it there either.

 

PS.  I believe the deals RMT want involve guaranteeing a 2nd person no matter what and the TOCs are not going to accept that.  Also I believe RMT were excluded from the Southern talks at the request of ASLEF!

SWT certainly used to make ECS moves under DOO and had done so for several years before I left the job in 2012.

 

The problem with having get-outs from providing a second person on passenger trains is that, unless the conditions under which it is allowed are tightly defined, it will rapidly become standard practice. Unfilled vacancies must definitely not be one of them.

 

The government and the TOCs will fight tooth and nail to ensure the requirements are as loose as possible. All they want is a fig-leaf big enough to cover a period of a couple of years during which all trains can be de-staffed.

 

John

Edited November 12, 2017 by Dunsignalling

[royaloak]

The government and the TOCs will fight tooth and nail to ensure the requirements are as loose as possible. All they want is a fig-leaf big enough to cover a period of a couple of years during which all trains can be de-staffed.

 

John

Post 01 Jan 2020 when the PRM TSI takes effect and the TOCs are sued by the disabled passengers who are unable to use the network as freely as they should be able to, necessitating the return of the onboard staff.

[Dunsignalling]

The idea is that you call up a driverless car (on an app?) only when you need to make a journey. I can see a fortune being made for returning lost property.

No need, with no driver to supervise matters, anything left behind that's worth having is likely to depart along with the next customer. 

 

John

[Dunsignalling]

Post 01 Jan 2020 when the PRM TSI takes effect and the TOCs are sued by the disabled passengers who are unable to use the network as freely as they should be able to, necessitating the return of the onboard staff.

Yes, but they want to get rid of everyone on decent conditions/pay before that happens so they can recruit the necessary "wheelchair wranglers" on zero-hours contracts and minimum wage.

 

IMO, the primary purpose of the strikes has to be ensuring that the current on-board staff are still in place when your scenario comes to pass.

 

John

Edited November 12, 2017 by Dunsignalling

[PhilJ W]

No need, with no driver to supervise matters, anything left behind that's worth having is likely to depart along with the next customer. 

 

John

Now were did I put those kippers that I found on the reduced to clear counter a few hours ago?

[brack]

Automation is likely the goal, and the future of train driving, it's hard to see otherwise, but it'll take some time yet.

There is another possible step - if a bloke sat at a pc in a warehouse in the Midwestern US can remotely pilot a tiny aircraft around sandy hot places during missiles at suitably shifty looking bearded gentlemen without any technical issues (no comment made on any ethical or moral issues - that's way off topic), then I could see remotely driven trains as a sort of halfway house. You can lose cab space and associated safety/noise standards, all the 'drivers' work nice shifts in one place, get 2 10 minute breaks and a half hour lunch, then go home, no worrying about getting them back to their starting point. One man gets ill or gets to home time, a new man takes control. Public wouldn't like it, but I could see freight being moved that way first, then passengers - they'd still employ someone to do the guards role to make people feel slightly better. After a few years of that you can transition into remotely AI driven trains.

 

Not saying I like it, but it looks plausible.

[bimble]

Thinks of the large automated model railways (such as Minitur Wunderland in Hamburg) where all the trains (and indeed cars, boats & planes) are run automatically with minimal manual input from "real" people...

[BG John]

Thinks of the large automated model railways (such as Minitur Wunderland in Hamburg) where all the trains (and indeed cars, boats & planes) are run automatically with minimal manual input from "real" people...

It made me think that once I've got my radio control working, the next step could be a camera in the cab, so I can get a drivers eye view to operate it. But my models don't carry real people, or face the risks of colliding with them!

[eastwestdivide]

Thinks of the large automated model railways (such as Minitur Wunderland in Hamburg) where all the trains (and indeed cars, boats & planes) are run automatically with minimal manual input from "real" people...

I think you've got the answer there - glue all the passengers to the platform, and all our troubles are solved.

[The Blue Streak]

As an Engine driver in Oz, there are plenty of people here who think I earn too much for a 4 day week too.

 

But I usually point out that My 4 day week generally adds up to 10 hours a week more than most people do, if most people on the average wage did 10 hours O/T a  week and worked night shift and weekends. they would be right up there in wages too.

 

I get a higher flat rate, but don't get anything extra for shift work and Public Holidays etc. The only time I get extra is for working an RDO which I very rarely do, (despite management pressure) as I'm usually to knackered too do anything else on My days off anyway.

 

I realise that UK drivers are probably on a different set of conditions, but it's not as easy a job as people like to think.

 

Start with the fact that irregular hours are part of the mix, I might start at 2000 one day then My next shift Might be 0200 start then I might get a day shift. It's all over the place. It's not like normal shift work where you start at a regular time the whole week. Then there is being targeted by thugs and yobbos who do everything from throw stones to tie concrete blocks at head height from overpasses, the odd suicide and idiots who do their absolute best to get cleaned up in their cars.

 

Plenty of people put there hands up to start the job, but a high percentage drop out pretty quickly too. Especially after a visit or two to the coroners court.

 

You need to pay a decent rate to retain decent decent drivers.

You aren't going to put up with the stress and miss out out on weekends with the family or sporting events or birthdays or public holidays or all the things that the rest of the world seems to take for granted and accept getting paid a handful of gravel for it.

 

As for driverless trains, they are being trialed in NW Australia, where it's all power to the mines and all Dynamic back to the coast. Without putting those drivers down, I have driven Iron Ore trains - big long trains are dead easy to drive as they even out over lumps and bumps.

No way that automation could be used in many places where heavy shorter trains (say roughly 850m and 4500 tonnes) are used on hilly lines where the whole train can be going up or down hills almost constantly and you have to actually drive the thing. Plenty of "gun drivers" have gone to pieces trying to operate on some of the more challenging lines !

 

As for single man crewing, it got in here on many trains (and was getting in regardless of what the crews or unions said or did about it) so if you get an extra bit of cash for  the extra responsibility that came with it, well that is probably fair .

 

Don't get Me wrong, I don't mind the job at all, but it's not something everyone can or wants do. Either due to the hours or conditions or they just don't have a knack for it.

 

I could go on but I really CBA because quite frankly I think some people would still think We are overpaid and under worked whatever I say.

Edited November 13, 2017 by The Blue Streak

[jim.snowdon]

I think you've got the answer there - glue all the passengers to the platform, and all our troubles are solved.

Maybe, but then again, maybe not. What would be bebeficial is getting passngers, and the law, to understand that once the door close warning has started, any injuries resulting from getting in the way of the doors, or getting caught by them, is their fault and not that of the train driver. Door detection systems have a finite limit on sensitivity and passngers need to learn to respect that.

 

Jim

[PaulRhB]

U.K. Drivers also work shifts starting at different times every day and much of the talk about what computers do better comes as a result of the effects of fatigue at work combined with home life. Some say they should control their sleep better but they just doesn't work with real life with people mowing lawns, neighbours partying late at night, that parcel being delivered or young children. I live in a fairly quiet road on the second floor with blackout curtains and double glazing and that can't shut out everything and my hours are less random than traincrew.

Like I said before automation is fine if the system is designed for it but automating trains weighing thousands of tons vs a 45 ton lorry has to take into account hugely different dynamic performance.

Driver aids an TPWS are good and things like gps and moving block are going to increase capacity but what do you do when the automated system shuts down due to a failed power supply or connection? You can have a bod onboard to take over but then they need to be fully trained and suddenly ramp up to full performance and get all the info on what's gone wrong and then deal with it. Funnily enough they are going need to be paid a decent wage to do shifts and take on the responsibility. I think people need to realise that computers are great but not foolproof either and a combination is best. Humans are better at adapting to random situations fast and you'd need a very expensive suite of robo droids to get out and deal with the physical world like a single human can. More to the point stuff that advanced is going to cost more than the human just to build and run for many years yet

I went to a presentation on TMS signalling and it's still a long way off working round problems like a good Signaller can plus heavily reliant on input of info from various sources. TMS is going to rely on traincrew details being more accurate than currently as there are still roster mistakes, someone has to tell the system x is sick, late on certain restrictions etc.

TMS also can't currently tell the driver change ends and go via X due to Point failure, it still needs a Controller/ Signaller to take action. My concerns about ramping up to full speed fast also apply here in a critical situation.

I've fed this back in at the meetings I've been to as we aren't against it and ops staff are heavily involved in testing and development already.

The media do a good job of demonising the railways and staff but it's no different to any big organisation in reality and contrary to the image the media live to saddle us with many are trying their damnedest to get people and goods there on time.

Like everything our expectations of computers ability regularly exceed the real world reality

Edited November 13, 2017 by PaulRhB

[The Stationmaster]

A good few years back the Western Region took part in a series of ATP trials on the Badminton route.  The equipment (which subsequently appeared as a full scale installation from Paddington outwards) was tested with an HST running at various speeds and a loaded train of MGR hoppers.  Irrespective of its  speed the HST was stopped automatically every time it passed over the 'stop' 'loop' and within the estimated distance.  The MGR train, hauled by Class 56 was all over the place and didn't even achieve a consistent stopping distance from several attempts at the same speed over the loop, also it usually overran the required/calculated stopping distance - again irrespective of speed.

 

The problem with any train is that numerous factors will affect the way it performs.  While disc brakes are far more consistent than wheel tread brakes brake effect, especially on wagons, tends to vary considerably as indeed does the effect of gradients and weather conditions on any type of train.

 

Automation might sound like a very good idea but i suspect the complex algorithms needed to turn it into a reality on a mixed traffic railway with varying stopping distances between signals and variable weather and loading conditions, let alone what the Driver could observe were he/she there, is a very long way from what has been done on the Hammersley.  And if the ideas behind it are anything like certain others emerging from NR's 'digital railway' boffins (who, from what I have heard, seem totally detached from operational railway reality) I reckon we would see less dense operational patterns running at slower speeds than manned trains.