given the limited view from high speed stock, in cab signals and the impossibility of stopping in time at eped if you see an object on the track, would drivers rather sit at the back (or in the middle?) of a train and drive on camera views in HS trains
-
Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Driverless trains - when?
- Thread starter orpine
- Start date
- Status
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R
RailUK Forums
Clip - what I was asking was if this technology was installed on trains and the end result was that is discouraged people from travelling, would it be worth installing it? As mentioned earlier, the systems which do have driverless technology are of the "closed " network type....however as far as I m concerned it has its limitations based on public confidence and perception. Nuclear power plants are pretty much fully automated nowadays, but I dont see any in the world which have absolutely no staff on site....
carriageline
Established Member
- Joined
- 11 Jan 2012
- Messages
- 1,897
FWIW, computers cannot, and do not always react as fast as humans.
And then what about things drivers would normally report, adjacent line blocked by XY, overheads down, trespassers, animals on the line, rail defects, subsidence, broken rails, defect with passing train. How would it spot pway workers, and give them a toot, and know that they have acknowledged it? What if a train hits someone, and the computer doesn't "see it", and it obviously needs reporting. How would a computer do any of the above?
What about when equipment fails, which it always does (trackside and on train?) signal failures, point failures, track circuits showing occupied.
Train fails in tunnel, what now?
Sent from my iPhone using Tapatalk
And then what about things drivers would normally report, adjacent line blocked by XY, overheads down, trespassers, animals on the line, rail defects, subsidence, broken rails, defect with passing train. How would it spot pway workers, and give them a toot, and know that they have acknowledged it? What if a train hits someone, and the computer doesn't "see it", and it obviously needs reporting. How would a computer do any of the above?
What about when equipment fails, which it always does (trackside and on train?) signal failures, point failures, track circuits showing occupied.
Train fails in tunnel, what now?
Sent from my iPhone using Tapatalk
David Barrett
Member
- Joined
- 9 Mar 2013
- Messages
- 554
Correct, having spent half a lifetime maintaining some of this equipment which never has bad days I can only agree with your short reply to the O.P. I have lost count of how many times, in the event of problems, somebody has said,
"That's not supposed to happen".
Theatrically I can only offer a couple of lines from a song performed by Hazel O'connor sometime around 1980 called Eighth Day,
"On the Eighth Day a machine just got upset,
A problem Man had not forseen as yet".
These words are never far away from me when the likes of Politicians and people with something to sell are making all sorts of claims about this or that product with inbuilt infallibility.
Meanwhile an unsinkable ship sinks, one train runs into the rear of another, an aircraft falls out of the sky and the financial markets go into a nosedive because of one event leading to a massive computer based wobble. Technology, or our misplaced faith in technology, has often played a part in events such as these.
Never mind they, whoever they are, as the O.P. asserts will come up with something.
BravoGolfMike
Member
- Joined
- 21 Jan 2011
- Messages
- 150
If automation is seen as the future, how come ERTMS, which actually is the future, is designed to be used by a driver? Even the fabled level 3 is not automated, let alone driverless. Put it this way, I have at least 25 years before I retire and I am not worried in the slightest. It is almost comical that DLR is used as an example of driverless trains being the future, when you can have driverless heavy freight mixing with high speed passenger, mixing with local commuter services all on the same line, over the same junctions and through the same stations, then I might start to worry.
Last edited:
This comes up from time to time on here and basically it all comes down to costs. Drivers are expensive but nowhere near the cost of setting up and maintaining automation. It is cheaper to pay thousands of drivers nationwide to do the job than upgrade infrastructure and trains to be driverless. Plus someone needs to be brave enough to remove the driver and the first incident after that which a driver could have prevented will see that persons career over...
Look realistically of the cost of driverless trains and you would see that even scratching the surface of getting the infrastructure up to date (ie not even looking at the actual technology or trains themselves yet) will cost a fortune, much more than paying a fleet of drivers.
Also, other countries have pondered driverless trains, the new French TGVs looked into it but the outcome has always been (as with aircraft) that far from removing the human being it is better to develop technology to aid the operator more. That gives the best of both worlds.
The cost alone means that I doubt we will see driverless trains in any of our lifetimes.
Look realistically of the cost of driverless trains and you would see that even scratching the surface of getting the infrastructure up to date (ie not even looking at the actual technology or trains themselves yet) will cost a fortune, much more than paying a fleet of drivers.
Also, other countries have pondered driverless trains, the new French TGVs looked into it but the outcome has always been (as with aircraft) that far from removing the human being it is better to develop technology to aid the operator more. That gives the best of both worlds.
The cost alone means that I doubt we will see driverless trains in any of our lifetimes.
This thread just reminded me of an incident on a virgin pendolino service i was on a few years ago from Euston to Manchester. The train had to stop for 20 minutes just 2 mile into its journey, so as the driver could re boot the technology in the cab !!
starrymarkb
Established Member
I think you'll find it doesn't. Early build A320/30/40 do have analogue dials but newer ones haves have small standby monitors powered and fed from a different system in case there is a failure in of the main instrument displays (and it would require all 6 to fail as the crew can transfer the Primary Flight Display onto any of the displays). As for the controls there is a reason that there are 7 Independent Flight Control computers, the chances of all 7 failing together is pretty remote!
This is the standby system on a Boeing 737NG ('Buses have a similar unit)
Last edited:
A computer could be much better at looking for hazards, with the bonus no loss of concentration etc etc.
--- old post above --- --- new post below ---
But how many trains have been cancelled through lack of crew ?
I find some of the posts making absolute claims about the limitations of computers a little strange. Computer and artificial intelligence technology is improving all the time, and I can see no theoretical reason why any of the judgements being brought up couldn't be performed by technology one day.
I think A-driver has really hit the nail on the head. Train drivers do difficult tasks and make decisions that aren't easy to automate. Designing and training a computer to do such tasks, and proving they are sufficiently reliable is currently somewhere between prohibitively expensive and impossible.
I would hope that in my lifetime one would see driver aids implemented to help with some of these tasks.
A good analogy would be signalling systems. We've now reached a point with centeralised automation, and automatic route setting is possible. But there are still situations where a human's intuition and manual intervention are better than a computer, due to the fact that the choice of which route to assign, or which train to delay, can have complicated knock-on effects several hours in the future.
I think A-driver has really hit the nail on the head. Train drivers do difficult tasks and make decisions that aren't easy to automate. Designing and training a computer to do such tasks, and proving they are sufficiently reliable is currently somewhere between prohibitively expensive and impossible.
I would hope that in my lifetime one would see driver aids implemented to help with some of these tasks.
A good analogy would be signalling systems. We've now reached a point with centeralised automation, and automatic route setting is possible. But there are still situations where a human's intuition and manual intervention are better than a computer, due to the fact that the choice of which route to assign, or which train to delay, can have complicated knock-on effects several hours in the future.
I could ask the same of how many trains have been cancelled because of a technology failure.......
One of the drivers on this forum posted a comment about how he once stopped his train inbetween stations to pick up a dog out of the cess which was causing operational problems.
carriageline
Established Member
- Joined
- 11 Jan 2012
- Messages
- 1,897
Don't you love it when you post a list of things a driverless train cannot do, and it's ignored by the "anti driver" squad. Nothing changes
Sent from my iPhone using Tapatalk
Sent from my iPhone using Tapatalk
Tetchytyke
Veteran Member
I can't imagine a driver can see a great deal at 186mph through the front window of a Eurostar set and, even if he could, if you're close enough to see an obstruction on the line you're too close to stop. That deer the driver sees is going to get smeared down the train regardless of whether the driver is in control or a computer is.
Similarly, I (think) the Pendolinos are speed restricted when the in-cab signalling technology fails because it is understood the driver can't react to problems at full speed.
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.
I think many of the basic issues drivers deal with manually will be automated in the future, even on high speed trains. Computers can do it better and quicker than the human eye. Computers are less likely to go through red lights by mistake, or to be packing their bags instead of watching the track ahead.
But I can't see trains ever going fully automatic, at least not in my working lifetime. As with planes, I think the human desire to see a person at the front, as a failsafe, will override any economic arguments that can be made. A computer can't evacuate a crashed train, or set detonators on the track, or do any number of things you really need a person for.
For the same reason, I don't see the Tube ever going fully automatic. ATO is installed because you can run more trains automatically than you ever can using manual signalling. I can see the Tube heading towards the DLR model of having a passenger service agent. But I don't think people would ever want a tube train with no staff at all on board. For one thing, 7/7 and the vital role drivers took in that incident is still too fresh in the mind.
Similarly, I (think) the Pendolinos are speed restricted when the in-cab signalling technology fails because it is understood the driver can't react to problems at full speed.
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.
I think many of the basic issues drivers deal with manually will be automated in the future, even on high speed trains. Computers can do it better and quicker than the human eye. Computers are less likely to go through red lights by mistake, or to be packing their bags instead of watching the track ahead.
But I can't see trains ever going fully automatic, at least not in my working lifetime. As with planes, I think the human desire to see a person at the front, as a failsafe, will override any economic arguments that can be made. A computer can't evacuate a crashed train, or set detonators on the track, or do any number of things you really need a person for.
For the same reason, I don't see the Tube ever going fully automatic. ATO is installed because you can run more trains automatically than you ever can using manual signalling. I can see the Tube heading towards the DLR model of having a passenger service agent. But I don't think people would ever want a tube train with no staff at all on board. For one thing, 7/7 and the vital role drivers took in that incident is still too fresh in the mind.
I am yet to hear a credible argument.
Which just goes back to the point I was making about passenger confidence......
--- old post above --- --- new post below ---
Like I told you .....if the public doesnt like it ( ie they refuse to get on trains ), its actually irrelevant whether the technology is available or not. Current driverless trains run on networks where the network itself has limited impact from outside influences , I dont see leaves on the line in the Tube for example.
--- old post above --- --- new post below ---
Like I told you .....if the public doesnt like it ( ie they refuse to get on trains ), its actually irrelevant whether the technology is available or not. Current driverless trains run on networks where the network itself has limited impact from outside influences , I dont see leaves on the line in the Tube for example.
I guess with most things, the best scenario is likely to be one that marries technology and human interaction. There are different areas of the driving role that require different skills. Some would be better with human decisions, some with that of computer decisions.
But I don't see this happening any time soon, as like others have already said, the technology that would be required just isn't available yet.
But I don't see this happening any time soon, as like others have already said, the technology that would be required just isn't available yet.
Well, I don't consider myself anti-driver, but to address your previous comment:
I'm unsure what you mean by this? Computers can react to data observed quicker than a human. If they don't, that's probably a limitation in the design of the software, or skrimping on hardware, rather than an inherent fact. We now have computers that can always 'beat' humans at rock paper scissors by taking advantage of the small lag humans take to react.
Object detection? No reason why a computer couldn't detect all the things you list.
We're talking about driverless, not unmanned, trains, I believe. We could have remote driving failsafe. Line-of-site driving by the manned presence. Duplication of failsave equipment.
Of course, it's possible that whatever redundancy you build in, everything could still fail, but the game is just about minimising risk, it can never be removed totally.
I agree we are nowhere near driverless operation yet. It's also possible that the skills required of the person remaining on the train (managing crowds/emergencies, going onto the track, etc.) might need to be basically the same as those of a driver.
O L Leigh
Established Member
What sort of credible argument would you accept?
There seems to be two camps here. Those that support ATO seem to be those that are not in the driving grade while those that oppose it are those who are in the driving grade. What annoys me is that the opposition that people like me are putting up seems to be solely based on the perceived threat to our job. Can I just say at this point that this is not the objection I have. As has already been pointed out, ATO still requires the presence of a driver to deal with failures and other unforeseen difficulties. Even the DLR has an on-train presence who can take control if needed and, because that presence isn't in the front of the train, there have to be additional safeguards on the open sections of the route.
Yes I'm sure a computer can detect an obstacle, but does it know what to do? I brought deer into the discussion because they are about the same size as a person. Obstacle detection may be able to detect a deer on the line, but can it distinguish between it and a human? I doubt it. Hitting one is a real emergency where the other is a mere inconvenience.
This is not the be-all and end-all of my objection but merely illustrates that there are shortcomings in the technology that cannot deal with the subtleties of the job. Yes there are some things that a computer could do better than a human, but that doesn't mean that the computer will do everything better. Just look at how any problems there have been with the obstacle detection equipment on the crossings between Ely and Norwich.
O L Leigh
As a counterbalance, how many computers have won George Crosses?
Silv1983
Member
As A-Driver said the cost is the biggest issue. Why invest hundreds of billions (optimistic!) in a project patching up every aspect of an infrastructure that won't return a sniff of profit in the same century.
Drivers' aren't that expensive whoever said it. They can easily pay for themselves on one busy shift if you look at it from the perspective of revenue generated ferrying passengers. I can't think of many other professions that do the same in the private sector - bankers accepted.
Too much technology and I can imagine we'd go full circle. With lifeless 9ft concrete walls encasing the rail network, CCTV cameras every 30 yards, GPS aerials sticking out of every carriage. Cancelled services because the monitoring equipment cannot safetly monitor and react to the sporadic nature of track work and track workers. People would soon say: "Can't we just stick a driver at the front of the train and be away with this utopian blur?" - "Oh good idea!".
Drivers' aren't that expensive whoever said it. They can easily pay for themselves on one busy shift if you look at it from the perspective of revenue generated ferrying passengers. I can't think of many other professions that do the same in the private sector - bankers accepted.
Too much technology and I can imagine we'd go full circle. With lifeless 9ft concrete walls encasing the rail network, CCTV cameras every 30 yards, GPS aerials sticking out of every carriage. Cancelled services because the monitoring equipment cannot safetly monitor and react to the sporadic nature of track work and track workers. People would soon say: "Can't we just stick a driver at the front of the train and be away with this utopian blur?" - "Oh good idea!".
Johnny_w
Member
It might not add anything to the thread, but there is no way on God's green earth that I would get on board a mainline train without someone 'up front'*
DLR seems a world away from Mainline to my small mind.
* - NOT A DRIVER. Not connected with the industry in any way shape and/or form - just a pleb who loves to travel by train.
JW
DLR seems a world away from Mainline to my small mind.
* - NOT A DRIVER. Not connected with the industry in any way shape and/or form - just a pleb who loves to travel by train.
JW
BravoGolfMike
Member
- Joined
- 21 Jan 2011
- Messages
- 150
That's because it is. The comparisons are ridiculous.
O L Leigh
Established Member
Yes, that's just what we need. Someone who knows better in spite of almost two centuries of operational experience and all the lessons they have taught us.
But then I guess they know best. :roll:
O L Leigh
It would take quite a bit, given that google have a computer driven car that has driven hundreds of thousands of miles on real roads, with real traffic lights, and real pedestrians, cyclists, lots of other traffic and endless other complications.
Driving a train on a linear track, with usually nothing in the way, with computer controlled signals directly to the brain of the train would appear a relatively simple task in comparison. Coupled with the fact that a computer could see in the dark with infa red and the like and with 4g / satellite link there could easily be cctv in a control room.
A good start would be with freight and then move to passenger services.
O L Leigh
Established Member
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?
O L Leigh
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.
I am all for having a driver up front augmented by latest automation to aid the driver and reduce workload, but you always have that presence available to back up the systems in the event of a technology failure, or for instance to detect abnormalities on route such as bridge damage, flooding or as recently the case, when a drill; rig came through Old st tunnels
http://www.raib.gov.uk/publications/current_investigations_register/130308_old_street_station.cfm
Without the visual skills of a twin mk1 eyeball driving the previous train, it would have resulted in a very serious accident to the following train through, in this case the job was stopped and an empty inspection train sent through, again using quadruple Mk1 Eyeball technology to identify the cause of the previous report and resulting in the train stopping clear of the damage before the augers came through
Of course this could easily have been on the LU or DLR which then leads us on to the vexed question of still having someone up front
in the Former the same outcome as on the network hopefully with just damage to the tunnel and a pending HSE court case, however the DLR would have crashed without doubt
This is just one small example of how having a professional well trained driver up front keeps the network as one of the safest in the world and no technology can ever replace the detail a driver is able to discern. For instance that group of teenagers hiding in the bushes waiting to lob stones at the next train or worse
The same questions are now being asked of replacing pilots, but I know what I would want up front when all the tech fails and you have to land on the Hudson river which apparently computer says wasn't possible
http://www.raib.gov.uk/publications/current_investigations_register/130308_old_street_station.cfm
Without the visual skills of a twin mk1 eyeball driving the previous train, it would have resulted in a very serious accident to the following train through, in this case the job was stopped and an empty inspection train sent through, again using quadruple Mk1 Eyeball technology to identify the cause of the previous report and resulting in the train stopping clear of the damage before the augers came through
Of course this could easily have been on the LU or DLR which then leads us on to the vexed question of still having someone up front
in the Former the same outcome as on the network hopefully with just damage to the tunnel and a pending HSE court case, however the DLR would have crashed without doubt
This is just one small example of how having a professional well trained driver up front keeps the network as one of the safest in the world and no technology can ever replace the detail a driver is able to discern. For instance that group of teenagers hiding in the bushes waiting to lob stones at the next train or worse
The same questions are now being asked of replacing pilots, but I know what I would want up front when all the tech fails and you have to land on the Hudson river which apparently computer says wasn't possible
O L Leigh
Established Member
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
Last edited:
- Status