Driverless trains - when?

[ainsworth74]

As a counterbalance, how many computers have won George Crosses?

I'm sorry but what on earth has that got to do with anything?

 

[Pugwash]

What cost savings? You're talking about a MASSIVE initial outlay for the kit (not to mention the safety case for using it), and yet you still need a driver for when it goes wonky. And this isn't an attendant we're talking about, but someone who still needs to be able to drive the train, to know the rules and be conversant with the routes and traction. In short, all the skills that the existing generation of drivers currently possess.

And why pay them less? Airline pilots don't get paid less for flying an Airbus.

I'm not saying that technology doesn't have a place, but frankly there seems to be a chasing after technology simply because it's new. Trains cover vast distances at all hours of the day and night safely and efficiently using the skills of the drivers alone. What real benefits are there to changing things around simply for the sake of it?

O L Leigh

The cost outlay is not prohibitive, even if it is a million pounds per train it pays for its self in 5 years

Once the software is written and the system is developed, it can be rolled out around Europe, the cost per unit falls drastically.

Trains will not need back up drivers, if the system has a problem the train simply comes to a halt.

Trains could theoretically run from med ports to terminals in scotland with no need for lots of differently trained drivers etc.

 

[driver_m]

Here is a scenario that seems far fetched. Assuming these driverless trains are able to magically spot obstructions and yet be able to tell if a pigeon or pheasant is flying about, what happens if the train happens to come across a drunk dressed as big bird who is stumbling all over the place? How does it make that call? Does it understand or is it something that requires CTRL-ALT-DEL?

It may seem ridiculous, but someone in the govt/ORR or something similar has to put their name onto a law that allows a technology that can't possibly cover all eventualities. So when a driverless train does something bizarre that couldn't be forseen causing injury or fatality, that person is going to be asked some probing questions by a smart lawyer. Given the risk averse culture that we live in, the costs to prove the technology are not going to be worth pursuing. It would also have to be tested on a european union basis to ensure fair access for all countries too, adding to the cost.

 

[O L Leigh]

Trains will not need back up drivers, if the system has a problem the train simply comes to a halt.

And then what...?

I'm sorry, but you're talking about pie in the sky. It won't happen like that for the same reason that you won't find airliners without pilots.

**EDIT**

Right, I'm off now. Got a train to drive.

O L Leigh

 

[KA4C]

The cost outlay is not prohibitive, even if it is a million pounds per train it pays for its self in 5 years

Once the software is written and the system is developed, it can be rolled out around Europe, the cost per unit falls drastically.

Trains will not need back up drivers, if the system has a problem the train simply comes to a halt.

Trains could theoretically run from med ports to terminals in scotland with no need for lots of differently trained drivers etc.

So how come it hasn't happened yet?

Same as pilotless airliners

 

[Pugwash]

And then what...?

I'm sorry, but you're talking about pie in the sky. It won't happen like that for the same reason that you won't find airliners without pilots.

**EDIT**

Right, I'm off now. Got a train to drive.

O L Leigh

Wait till driver less cars are on the roads and people are sitting there reading the paper or watching tv while the car drives them about and they will soon wonder why there is a driver at the front of their bus or train.

 

[TheGrew]

I think driverless mainline trains are only really a pipe dream until we can get moving block signalling (as was intended for the WCML) to work effectively. I was told by one of my former lecturers (who specialised in safety critical software) that the main problem was getting it to a sufficiently high SIL level. With the main issue being the system response in a communication failure. I think much like aircraft, mainline trains will probably grow in complexity and automation but will never completely loose the human in the loop (Even if the human's title changes from driver to operator).

 

[KA4C]

I think driverless mainline trains are only really a pipe dream until we can get moving block signalling (as was intended for the WCML) to work effectively. I was told by one of my former lecturers (who specialised in safety critical software) that the main problem was getting it to a sufficiently high SIL level. With the main issue being the system response in a communication failure. I think much like aircraft, mainline trains will probably grow in complexity and automation but will never completely loose the human in the loop (Even if the human's title changes from driver to operator).

Don't tell Pugwash, he is on a driver removal mission!

 

[NI 271]

Anti-lock/anti-slip technology can deal with grip problems far quicker than a human can, as anyone who uses traction control or ABS on their road car will tell you.

My car has such a device. When the roads are covered in snow, as happens regularly up here in winter (and often to severe levels) thanks to my local council's abhorrent policy of simply not gritting except during peak hours, I always switch it off, because it competes with my driving and makes it FAR more difficult. If these things were so wonderful, this would not be the case.

I accept that the overwhelming majority of car drivers (and let's not forget, they aren't trained in how to use traction control mechanisms anyway) are absolutely clueless in snow - but for those who are not, such things are actually a hindrance in such conditions. Although not in the grade myself, I doubt train drivers are ever going to be as bad at driving trains as most car drivers are at driving cars, and furthermore, they ARE trained in how to use adhesion aids.



I have to say, I'm surprised this topic keeps being re-raised on here with such regularity, as new trains which WILL need drivers are still being ordered, and they're not going to be phased out for presumably at least 30-40 years, there are a huge number of posters on here (and I'm one of them) who won't be around to see these trains being withdrawn from service - therefore there will always be drivers, at least throughout the remainder of my lifetime.

 

[SPADTrap]

The cost outlay is not prohibitive, even if it is a million pounds per train it pays for its self in 5 years

Once the software is written and the system is developed, it can be rolled out around Europe, the cost per unit falls drastically.

Trains will not need back up drivers, if the system has a problem the train simply comes to a halt.

Trains could theoretically run from med ports to terminals in scotland with no need for lots of differently trained drivers etc.

Hogwash from Pugwash.

Just empty words based upon assumptions really. You mange to sound factual while conveying purely misguided personal opinion.

Do you work in local government?

And we don't run trains theoretically either.

 

[driver_m]

I'm sorry but what on earth has that got to do with anything?

I responded to a silly post by making a silly point in return. The poster clearly has an agenda as I'm sure many people can see.

 

[TheGrew]

Don't tell Pugwash, he is on a driver removal mission!

Yes so it seems.

The cost outlay is not prohibitive, even if it is a million pounds per train it pays for its self in 5 years

I think the reality is that on a railway the train will only form an incredibly small part of the cost. I think the likely greater expenditure will be on the surrounding infrastructure, including but not limited to complete grade separation, Cab signalling (which admittedly is being rolled out in some places I believe) and many other joys

Once the software is written and the system is developed, it can be rolled out around Europe, the cost per unit falls drastically.

Might I ask you the question, have you ever had the joy of writing safety critical software? It is incredibly time consuming and complex. You have to attempt (and fail!) to catch every single exception that a system could potentially incur. You have to think about the 1 in a million and depending on the SIL level 1 in a billion chance or something happening outside of standard operating parameters. It is INCREDIBLY expensive...

Trains will not need back up drivers, if the system has a problem the train simply comes to a halt.

I can see the headlines now "Computer sensor error causes meltdown on morning commute as Waterloo fast tracks are blocked causing delays to thousands of people..."

 

[redbutton]

I'd also point out that too much automation tends to make human performance worse, because it comes at a cost of experience.

For example, the recent Asiana air crash at San Francisco- turns out the pilots were extremely proficient at handling the automation, but as a result have dangerously little hand-flying experience so they failed when the instrument landing system was down for maintenance.

Personally, when it comes to trains or planes, I'd prefer to have a human pilot with just enough automation to reduce fatigue and warn him/her if a mistake is made.

 

[Moonshot]

I'd also point out that too much automation tends to make human performance worse, because it comes at a cost of experience.

For example, the recent Asiana air crash at San Francisco- turns out the pilots were extremely proficient at handling the automation, but as a result have dangerously little hand-flying experience so they failed when the instrument landing system was down for maintenance.

Personally, when it comes to trains or planes, I'd prefer to have a human pilot with just enough automation to reduce fatigue and warn him/her if a mistake is made.

And that is the best post on here.......

 

[fsmr]

Wait till driver less cars are on the roads and people are sitting there reading the paper or watching tv while the car drives them about and they will soon wonder why there is a driver at the front of their bus or train.

You are talking utopia, it is one thing in a sterile environ maybe Australia hauling coal to and from a power station in the middle of no where with no chance of Joe public getting near, but in the real world someone even in a car or truck needs to be watching out for the real world hazards, not just in front but above and to the sides

What about that kid whizzing along out of control on a bike down the hill next to the road that is going to pop out right in front of you. You reckon an all seeing chip is going to work that lot out from the chaff of everything else going on the side of the road

The algorithms are billions of combinations

It is like Amazon and their barmy automated copter delivery's. Ok till the real world i.e local dog or kids get in the way

 

[1018509]

If it's a vehicle controlled by google it'll only go where it wants you to go and repeatedly offer you suggestions on where it wants you to think you want to go.

If it's a choice between a human driver or a google automated vehicle give me the human driver, flaws and all, to google automation every time.

 

[bb21]

The cost outlay is not prohibitive, even if it is a million pounds per train it pays for its self in 5 years

So how much will it cost? Do you have a figure in your head?

Once the software is written and the system is developed, it can be rolled out around Europe, the cost per unit falls drastically.

What about the cost of upgrading signalling and other associated infrastructure? You do realise that rolling stock is not the only thing in the equation?

Trains will not need back up drivers, if the system has a problem the train simply comes to a halt.

Everytime there is a fault? The country will grind to a halt. No, thank you.

 

[Moonshot]

It is like Amazon and their barmy automated copter delivery's. Ok till the real world i.e local dog or kids get in the way

Lol ....I can imagine some local scroat taking potshots at that with a football !!

 

[irish_rail]

Seem to remember a crash with multiple fatalities in Germany a few years ago , involving a driverless maglev (?) train that careered into a vehicle on the track.

IIRC about 25 dead.

 

[Tomnick]

At first it was 2 back up drivers, they are down to one back up driver.

http://mashable.com/2012/08/07/google-driverless-cars-safer-than-you/

Given the potential cost savings to the railways, it should definately at least be trialled on the railway.

You speak as though a meaningful trial is even a remote possibility in the near future, even within the next decade. What about these computer controlled signals that you mention, directly linking to the brain of the train? That's not an immediate prospect on the majority of the network, and certainly not as straightforward as you might think - the ERTMS rollout is, after all, currently limited to the relatively simple Cambrian lines and a test site on the Hertford Loop. The rest of the network, barring the slightly more advanced ATP supervision and the ATO systems on some LUL lines (and soon the Thameslink core?), has no interaction directly between signals and trains beyond AWS and TPWS.

How would, in the real world's often rural railway, a computer cope with - for example - examining the line because of a track circuit showing occupied when clear (looking for such subtleties as a broken rail), stopping to put a phone back on the hook at a user worked crossing, taking account of railhead conditions to brake earlier and lighter as appropriate...and so on?

 

[notadriver]

Great post. Earlier today I was just on the approach to a station - adopting a defensive braking curve. Brake step 1 produced no retardation and only step 2 eventually got the train under control. I tried brake step 1 again and could still hear and feel signs of slight wheel slip protection activity so elected to approach the station very cautiously. Would a computer have done the same ?

 

[A-driver]

A lot of rubbish is now being spouted on here, took 3 pages but got there.

I don't think that anyone can deny that the technology is certainly available to run driverless trains. Nothing which a human driver can do day to day isn't possible to do by computer (and possibly better as a computer won't get fatigued). Even leaf fall is little obstacle, you can fit relevant sensors to do exactly what a human does. Basically there is nothing technologically preventing driverless trains. Or driverless aircraft, cars, lorries, space ships, cruise liners, oil tankers etc etc.

BUT, as I have said already, the cost is astronomical. One poster claims that if it cost 1million per train then it would be recouped in 5 years. Absolute rubbish. It would cost a huge amount to update the existing Infastructure before going near signals or trains. Plus every train will need staff on board, either that or you position lots of staff at very regular intervals who can get to trains very quickly. Why would these staff not also be paid well? They would need the same route, rules and traction knowledge as a current driver and would have similar union protection so may perhaps be on10-15k less a year than a current driver. So you wouldn't be looking at cutting all staff costs relating to drivers.

Also, bear in mind that the uk rail network is very diverse. You have the busy mainlines and commuter branches into large cities. You then also have very remote, barely used stretches. Would you really look at spending all that money to upgrade and then protect the lines at the north of Scotland which see 3-4 trains per day? Otherwise you end up with a half and half system which causes other problems.

Basically, the technology is there but there is no cost argument at the moment. You would never recover costs and so there is little point paying for all the work as it has no benifit. Adding technology to existing drivers cabs however has undeniable benifit on grounds of both performance and safety and will continue to happen, even if more lines become ATO it won't remove the driver as there is a huge, huge difference between ATO and driverless, something people on here seem to be unable to understand.

The cost of running the railway won't fall either. You save with one hand and spend with the other. The technology, once installed, won't maintain itself. Ticket prices won't fall as directors will take more and you will need a huge supply of support staff to ensure the technology keeps performing safely and properly.

Also, very few major incidents in the uk are down to train driver human error. There are a w examples but they get rarer as time goes on. No other country has seriously considered driverless trains so why the uk? Spain and the USA have both recently had driver caused major accidents (well that's part speculation but you know where I am coming from...) and neither of those countries are screaming that they need to get rid of drivers...some posters on here seem to implying that human drivers are deadly and a serious danger and therefore need replacing by a computer. I see no evidence to suggest humans need replacing on the footplate either cost, performance or safety wise.

 

[TheGrew]

So how much will it cost? Do you have a figure in your head?



What about the cost of upgrading signalling and other associated infrastructure? You do realise that rolling stock is not the only thing in the equation?



Everytime there is a fault? The country will grind to a halt. No, thank you.

It made me smile that you brought up the same counter arguments that I did

 

[DT611]

Why should it put off people from travelling? Most people know the DLR doesnt have drivers and they love that fact - seeing how they like to sit at the front pretending to be drivers. And do most people know about the ATO aspect of some of the tube lines? Probably not which is why you get people tweeting and such like saying they saw a driver doing something else rather than driving it.



Oh I do see the issues - but I dont see them in the ssame way that you do, as you are a driver and see your job as under threat from ATO - which is fair enough and I respect you for that. The DLR only has 3 tunnels on its system and quite a lot of it now runs on normal street level and is subject to exactly the same problems the mainline does - albeity without deer - in fact theres been a few instances this week already where they have had to shut a section due to unathorised access on the track. And it deals with hitting someone by stopping and then clearing it up. How is that any different to the mainline then when they dont have a driver?


In fact, the possibility of mainline runnig and the subject of obstructions could easily be solved - albeit expensiveley - by using the existing OHLE infrastructure to have some form of infrared grid system between each gantry looking at the track between the two all the way up the line. Any obstruction which has not moved for say more than 10 seconds would lead to say an amber aspect on the preceeding signals so the train slows down, the person on board moves to the front to take up driving position and to inspect the track ahead and then take necessary action.


The solutions to all your arguments for not having it are out there but for some, because of the loss of a drivers job, it is too hard to actually sit and think of solutions and its easier for them to slag anyone off who dares to question or discuss the future.

And lets face it, how many times on here and in the press have you seen stuff about drivers getting paid too much blah blah blah whilst sat on their arse's, or as one poster claimed ot have seen - looking at their iPod and then you have all had to justify it? Many times.

What you may not actually understand is that those very same passengers are the ones that come up to me and whinge about it and your salaries and trains being late - the majority of them blame it on you lot - and wonder why we cannot go to ATO on a lot of lines. Public perception may not be clued up on these things but time and time again they point to the DLR as the prime example. And once public opinion is going one way then the bosses and the Govt of the day will take this on board as they will claim it will make the railway cheaper and thats what the public want mainly - along with it being safe.


I agree, your job is more about just being sat there pushing buttons to open doors and all that - the public dont and more and more of them speak to people like me all over the country about it an dhow we could cut the cost of their ticket by getting rid of you lot.

but the public would be the first to kick up blue murder should an accident happen with a driverless train, i would suggest that its something you can't win either way with, their will always be a need for a trained person to be aboard the train to deal with any problems should the automation fail, what works for the dlr may not work for other rail systems.

 

[orpine]

Entirely unattended without any technical support and the back-up of a person to step in when it goes wonky...? Does it carry fare-paying passengers? Or is it just a proof-of-concept prototype that requires a lot of support from specialist technicians? And at what cost?

It's a proof-of-concept. But my question (as OP) wasn't about doing it today, it was about doing it in 10+ years time when it's no longer proof-of-concept for cars, but commercial. Any issue solved for cars should be able to be carried over for trains. I.e. object "avoidance" - car's have a much more active environment than a railway line so that should translate over.

There are certainly some things unique to trains, but why couldn't they be resolved? If drivers report to HQ about line conditions (well, the good drivers will - I'm sure there are bad ones just like in any profession) and the like, a computer probably could too (compare previous day footage) - bonus a computer would work just as well at night (infra-red). Lots of folks are saying computers don't currently have the vision skills, and for some tasks they don't, but in 10 years they'll be much further along.

Also, why would the infrastructure need to be upgraded especially to handle it? Roads won't be for driverless cars.

But...the bigger question is, should we want this? The fact we can doesn't necessarily mean we should. I can think of two very good reasons why we shouldn't want this: 1) Edge cases where a human driver can exercise judgement in a way a computer can't; 2) We're shooting ourselves in the foot by putting people out of work by means of automation. Increasing population, decreasing pool of jobs...

Good question. But:
1) I suspect there are more "edge cases" where the human is a detriment than a benefit. See car accidents.
2) That's the broken windows fallacy.

(Disclaimer - I'm not anti-driver, just generally pro-technology).

 

[Captain Chaos]

What is the obsession with automating everything? Because it's progress? Progress for progresses sake? It's just silly. Why do people have an obsession with automating and getting rid of people's jobs? To reduce prices? It never happens! Eventually we are going to automate ourselves into economic oblivion. You can't run an economy without employment. You need people to be earning money to spend money.

The economy needs jobs, not computer programs replacing them. Not all technology is a good idea. Nobody ever considers the human cost of things. Only the 'financial benefits' and the 'cost savings'. I can't help but feel that eventually we will all shoot ourselves in the foot with our obsession with it. Be careful what you wish for...
--- old post above --- --- new post below ---
Has anyone actually asked if we actually NEED driverless cars? I've never understood the point of it. All that effort to do something we can already do ourselves. Is it really that hard, difficult and terrible to drive your own car? Really? If we need to still keep a watch out, what possible benefit is just staring at the road doing nothing going to achieve? Won't that be more dangerous as people nod off through boredom? Am I the only one who thinks driverless cars are the most pointless thing ever?

 

[orpine]

What is the obsession with automating everything?

Because I believe it'd be safer. Human error is a (if not THE) leading cause of deaths on roads - several thousand a year, and those car accidents sometimes carry over to cause rail ones too. I don't have stats for train accident causes, but Humans do cause at least some. Computers aren't perfect, but I think they could resolve this.

The economy needs jobs

So why don't we employ people to walk around breaking windows? And then we'd also have more jobs for glaziers too. I'm not being silly, this is actually the broken windows fallacy. - https://en.wikipedia.org/wiki/Parable_of_the_broken_window - Those folks who are currently drivers could find jobs doing something else, just like the people who used to work in farming (i.e. 90% of the population some centuries ago) now work doing stuff like driving the trains or whatever.

 

[notadriver]

How does a driverless train help passengers if an emergency evacuation is required? What does a driverless train do if it hits a person ?

 

[Captain Chaos]

I'm not just referring to drivers. I'm referring to everything in general. There is a desire at the moment to automate as much as we possibly can. You say that they will find work doing something else. Perhaps you're right. But there will probably be less jobs available compared to those lost. Which means higher unemployment. Not only that but we have the population issue to contend with too. More and more people for fewer and fewer jobs. That's how it comes across to me, anyway.

 

[irish_rail]

In the UK driverless trains would NOT be any safer.

The main risk of fatal accidents in the UK now are level crossings or perhaps track defects, NOT human error.

TPWS / ATP (since 2003) mean that as drivers, it is virtually impossible for us to cause an error that leads to a fatal accident.

Presumably in the US and Spain these systems are not quite so comprehensive.