Advantages and Disadvantages of driverless trains?

[Dave1987]

It is quite amazing what some believe. The real world is a wild mysterious beast as many many designers and engineers have found out when their designs that worked on their computer screens in their office, fail repeatedly in the real world.

 

[mark-h]

A point for discussion, sorry if it's already been covered - how easy would it be to safely dispatch a driverless (an unstaffed) train from an unstaffed station, without being certain that there was no-one trying to get on last minute, no-one was stuck in the doors

The airport transit systems, mentioned above, do not have station dispatch staff or on-board staff. The Gatwick system has 12 doors per train, all of which open at each station. It has the advantage of having platform screen doors which would be blocked by somebody with their arm trapped in a train door- forcing the doors to reopen.

They also have a permanently manned control centre which will monitor the platforms via CCTV and able to override the automation when required.

----

With advances in AI realtime automated monitoring of CCTV will become possible. This could potentially result in all the doors being monitored simultaneously which is more than a driver/guard or platform dispatcher can do.

 

[driver_m]

Another point would be, if there's no one to manage. That puts the managers out of work too. And their managers.

Now we can't be having that can we?!!

 

[XDM]

When I was young, most lifts on the Underground and in department stores had a human operator, who, in the stores, called out "Going up" or "Going down" at each floor. Most people have now got over any fears about the mistakes an automatic lift might make and can interpret a downward- or upward-facing arrow to work out which way the lift is going.

A very old BR director told me the very last UK lifts to have full time lift operators were, believe it or not, on BR at the old Marylebone Board HQ. The operators were NUR(now RMT) members & BR did not want to face a national rail strike to be rid of them. They were paid off when it closed as the Board HQ. Vastly modernised it is now a grand hotel. Typical that RMT were up to their threats then. And I was told by this famous old BR veteran Director that the NUR used the spurious safety argument when he (rightly) moaned about today's RMT.

 

[Bletchleyite]

The maintenance costs will be huge.

Evidence?

 

[InOban]

The airport transit systems, mentioned above, do not have station dispatch staff or on-board staff. The Gatwick system has 12 doors per train, all of which open at each station. It has the advantage of having platform screen doors which would be blocked by somebody with their arm trapped in a train door- forcing the doors to reopen.

They also have a permanently manned control centre which will monitor the platforms via CCTV and able to override the automation when required.

----

With advances in AI realtime automated monitoring of CCTV will become possible. This could potentially result in all the doors being monitored simultaneously which is more than a driver/guard or platform dispatcher can do.

Trains won't be driverless. It's just that the driver will be in the Control centre, not on the train.

 

[Dave1987]

Evidence?

Im a little surprised I would have to explain this to you. How much railway operating experience do you have? In your completely automated system you are relying entirely on computer systems to control hundreds of tonnes metal tube travelling at very high speed with people on board. Computers rely on the data they are fed be that from sensors or whatever data feed they have going to them and then they scrutinise that data. They are wholeheartedly reliant on the data they are fed. So to ensure you always get reliable data you need to ensure your sensors maintained to the highest standards. Maintaining a whole load of sensors fitted to trains would be a pretty big and expensive task. Not to mention the enormous costs associated with the initial start up, exhaustive testing that would need to take place etc etc.

Hey I’m really not that bothered. I’m currently a net contributor to the treasury via taxes. If you or the Government wish to spend billions of pounds to automate me out of a job then I’m fine with that. I will get to sit at home reading lots and watching TV. Im fine with becoming a burden on the taxpayer as certain taxpayers seem extremely keen to automate others out of work.

 

[3141]

Four examples of driverless trains are
1) The Victoria Line from 1968, using a system of coded electrical signals to tell the train when to start and stop. There was a train operator who pressed two buttons to start it and to operate the doors.
2) The DLR. The train captain controls the doors and (I think) presses a button to tell the train it can proceed.
3) The Victoria Line now, the Jubilee Line, and new signalling systems being installed on other Underground lines. The trains receive signals telling them when to start, accelerate and stop.
4) The Lille metro, which is completely automatic.

Obviously, none of these main line operations travelling at 100 mph or more, and they don't generally have complex junctions or have to fit in with other types of services. I don't think their maintenance costs are huge. The evidence so far doesn't suggest that large numbers of sensors have to be fitted to trains. It seems likely that in the course of time computer systems will be developed that can handle the more complex information required to control the operation of faster trains over a larger geographical area. There are arguments against driverless trains but I doubt that the maintenance costs will be one of them.

 

[Dave1987]

Four examples of driverless trains are
1) The Victoria Line from 1968, using a system of coded electrical signals to tell the train when to start and stop. There was a train operator who pressed two buttons to start it and to operate the doors.
2) The DLR. The train captain controls the doors and (I think) presses a button to tell the train it can proceed.
3) The Victoria Line now, the Jubilee Line, and new signalling systems being installed on other Underground lines. The trains receive signals telling them when to start, accelerate and stop.
4) The Lille metro, which is completely automatic.

Obviously, none of these main line operations travelling at 100 mph or more, and they don't generally have complex junctions or have to fit in with other types of services. I don't think their maintenance costs are huge. The evidence so far doesn't suggest that large numbers of sensors have to be fitted to trains. It seems likely that in the course of time computer systems will be developed that can handle the more complex information required to control the operation of faster trains over a larger geographical area. There are arguments against driverless trains but I doubt that the maintenance costs will be one of them.

All of those examples are closed systems. None of them come close to the mainline and the variables you get on the mainline. Very poor argument.

 

[bramling]

All of those examples are closed systems. None of them come close to the mainline and the variables you get on the mainline. Very poor argument.

Actually, the Central Jubilee and Northern aren’t so far removed from the mainline. The Jubilee reaches 60mph in the open, and the Central would were it not for unrelated train issues, which is about the same as a 158 trundling up the Settle & Carlisle.

Nonetheless, the automated systems on both lines do not cope well with the variables faced in the open. Either running is dumbed down to assume the worst conditions, or lots of incidents are tolerated - each of which may well cause delay too. I predict we will see more of the same on the sub surface lines.

There’s a long way to go yet before the technology is there. Yes the driver could be abolished now, but not without introducing a whole new set of disbenefits. Time will tell if the scales will tip.

 

[3141]

All of those examples are closed systems. None of them come close to the mainline and the variables you get on the mainline. Very poor argument.

I actually stated that they're not main line systems. My main point was to state that driverless operation doesn't mean huge maintenance costs. Seems like you didn't bother to read it properly before weighing in.

 

[NotATrainspott]

Im a little surprised I would have to explain this to you. How much railway operating experience do you have? In your completely automated system you are relying entirely on computer systems to control hundreds of tonnes metal tube travelling at very high speed with people on board. Computers rely on the data they are fed be that from sensors or whatever data feed they have going to them and then they scrutinise that data. They are wholeheartedly reliant on the data they are fed. So to ensure you always get reliable data you need to ensure your sensors maintained to the highest standards. Maintaining a whole load of sensors fitted to trains would be a pretty big and expensive task. Not to mention the enormous costs associated with the initial start up, exhaustive testing that would need to take place etc etc.

Hey I’m really not that bothered. I’m currently a net contributor to the treasury via taxes. If you or the Government wish to spend billions of pounds to automate me out of a job then I’m fine with that. I will get to sit at home reading lots and watching TV. Im fine with becoming a burden on the taxpayer as certain taxpayers seem extremely keen to automate others out of work.

Autonomy on the railways is going to use train-mounted equipment that doesn't require any meaningful changes to the physical infrastructure. The same sensors that can read traffic lights from a car can read any age of colour light or semaphore signalling. That autonomy in the 70s and 80s required special hardware along the route doesn't mean that autonomy in the 2020s will.

Each train driver costs the railway upwards of £50k a year, and you need one for every train you want to run. Train sensors are not going to fail every week. Even if they increase maintenance workload, the saving from not requiring the drivers would massively outweigh the increase in maintenance cost. In any case, very little active maintenance could ever be done. If a sensor failed, all a depot technician would be able to do would be to unscrew it and plug in a replacement from a box.

Trains cost more than a million pounds per carriage. The cost of fitting all of the sensors and processing hardware needed for full autonomy will be a tiny fraction of that - I would guess a maximum of £10k per carriage. The components themselves are going to end up being commodity parts shared with buses and lorries, so there will be plenty economies of scale.

There will be no training cost. The autonomy hardware will be fitted to trains and set to learning mode. The only question is when the systems have learned enough to be allowed to run on their own. That might take a few years, but in the meantime there is no cost involved other than the marginal increase in electricity usage to power the systems.

 

[irish_rail]

Makes me laugh when people bang on about the cost of a train driver , well I've got news for u, we give a large proportion of our wages back to the government in tax which goes into reinvesting in the railway etc. Machines do not pay taxes, and who exactly will then fund all the out of work people's benefits.

So my wage for driving a train from plymouth to london is say for arguments sake about 150 quid (of which a large proportion goes to the treasury anyway) . Just two passengers travelling on that train for the journeys length, of say 400 people, will pay my wage. Puts things in perspective a bit .

The ONLY advantage where driverless trains may happen is on very intensively used routes / sections - mainly metro type operations ....

 

[Dave1987]

I actually stated that they're not main line systems. My main point was to state that driverless operation doesn't mean huge maintenance costs. Seems like you didn't bother to read it properly before weighing in.

I did read your post. Even attempting to compare them is a bit ridiculous IMO.

 

[Dave1987]

Autonomy on the railways is going to use train-mounted equipment that doesn't require any meaningful changes to the physical infrastructure. The same sensors that can read traffic lights from a car can read any age of colour light or semaphore signalling. That autonomy in the 70s and 80s required special hardware along the route doesn't mean that autonomy in the 2020s will.

Each train driver costs the railway upwards of £50k a year, and you need one for every train you want to run. Train sensors are not going to fail every week. Even if they increase maintenance workload, the saving from not requiring the drivers would massively outweigh the increase in maintenance cost. In any case, very little active maintenance could ever be done. If a sensor failed, all a depot technician would be able to do would be to unscrew it and plug in a replacement from a box.

Trains cost more than a million pounds per carriage. The cost of fitting all of the sensors and processing hardware needed for full autonomy will be a tiny fraction of that - I would guess a maximum of £10k per carriage. The components themselves are going to end up being commodity parts shared with buses and lorries, so there will be plenty economies of scale.

There will be no training cost. The autonomy hardware will be fitted to trains and set to learning mode. The only question is when the systems have learned enough to be allowed to run on their own. That might take a few years, but in the meantime there is no cost involved other than the marginal increase in electricity usage to power the systems.

Reading your idea of maintenance explains a huge amount about why stuff fails so often when introduced.

Ow and you will need to add in the cost of all that tax revenue not going to the treasury plus all the benefits claimed.

 

[Dougal2345]

For those making the "tax" argument, sadly I don't think that one will save the drivers.

I doubt the train companies are thinking "I can't automate that train, think of the tax revenue HM Government will lose when I make the drivers redundant".

The Government are unlikely to make a law saying "you must keep those drivers pointlessly employed so we get our tax".

No, the drivers will go the way of the coal miners, and the Government will get its tax some other way.

(I take no pleasure in saying this by the way)

 

[irish_rail]

For those making the "tax" argument, sadly I don't think that one will save the drivers.

I doubt the train companies are thinking "I can't automate that train, think of the tax revenue HM Government will lose when I make the drivers redundant".

The Government are unlikely to make a law saying "you must keep those drivers pointlessly employed so we get our tax".

No, the drivers will go the way of the coal miners, and the Government will get its tax some other way.

(I take no pleasure in saying this by the way)

Yes the government will get its tax "some other way". Magic money tree maybe. And while we are at it will pilots and nurses and office workers and just about every other job go the way of the coal miners???!!
I don't think so .......

 

[Dougal2345]

Yes the government will get its tax "some other way". Magic money tree maybe.

As I see it, automating trains will make one group of people poorer (redundant drivers) and one richer (train companies, shareholders etc.). So you reduce tax on the poorer group and increase it on the richer.

And while we are at it will pilots and nurses and office workers and just about every other job go the way of the coal miners???!!
I don't think so .......

Pilots - yes, nurses - no (although doctors - yes), office workers - yes (already happening )

 

[NotATrainspott]

Makes me laugh when people bang on about the cost of a train driver , well I've got news for u, we give a large proportion of our wages back to the government in tax which goes into reinvesting in the railway etc. Machines do not pay taxes, and who exactly will then fund all the out of work people's benefits.

So my wage for driving a train from plymouth to london is say for arguments sake about 150 quid (of which a large proportion goes to the treasury anyway) . Just two passengers travelling on that train for the journeys length, of say 400 people, will pay my wage. Puts things in perspective a bit .

The ONLY advantage where driverless trains may happen is on very intensively used routes / sections - mainly metro type operations ....

On a busy train the cost of the driver is indeed not significant. However, for all of the times when there are many passengers, there are many more times when there are rather fewer. Autonomy will make it much more financially viable to run off-peak trains and to do other efficiency measures like splitting and joining.

Yes the government will get its tax "some other way". Magic money tree maybe. And while we are at it will pilots and nurses and office workers and just about every other job go the way of the coal miners???!!
I don't think so .......

There are many options available. Taxes on wages are not the only way a government can fund itself.

 

[edwin_m]

Four examples of driverless trains are
1) The Victoria Line from 1968, using a system of coded electrical signals to tell the train when to start and stop. There was a train operator who pressed two buttons to start it and to operate the doors.
2) The DLR. The train captain controls the doors and (I think) presses a button to tell the train it can proceed.
3) The Victoria Line now, the Jubilee Line, and new signalling systems being installed on other Underground lines. The trains receive signals telling them when to start, accelerate and stop.
4) The Lille metro, which is completely automatic.

The Underground examples are automatically driven but are not driverless, as they are not allowed to operate unless a driver is present in the cab. Truly driverless systems not only have to automate the driving function, but have the more difficult task of working out how to handle various failure and emergency situations with no driver at the front of the train, or in the case of Lille and similar "unattended" systems no staff presence anywhere on the train. Unattended systems are normally also expected to have platform edge doors to replace the human role in train dispatch.

 

[irish_rail]

Crikey why is there so much dislike of drivers on here??? I could understand of we were forever on strike but that just isn't the case anymore and hasn't been for years, indeed I've never been out on strike.
Just find it surprising there is so much ill feeling towards drivers here....

 

[janahan]

Look, I am a software consultant, and a general technophile at heart. I personally would LOVE to have driverless trains throughout the UK mainline, however, I can tell you it's NOT going to happen anytime in the near future.

For driverless trains to work, the network and trains needs to have:

• a consistent and known infrastructure. The more varied the infrustructure (curved track in platforms, gradients, split platform faces, effect of leaves, rain, etc) the harder it is to program computers to account for it all. Just stopping a train at the right point at a platform can be affected by many factors.
• An adequate train detection and signalling system. We still have semaphores, a person in a box with a telegraph and tokens in a purse in places on the UK network. Until all of it is upgraded to a suitable system driverless trains are simply a no go.
• Stock that can electronically speak to each other. Each train would need to be able to know that nearby trains are doing to achieve the full benefits of increased capacity, and short headways. Today we have cases where even sets of the same class cannot talk to each other, let alone different classes.
• Incident management. The "wetware" neural network that exists between the ears of every average human is still far superior to the best learning AI available currently and the near future. Therefore humans are still able to detect and handle incidents far better than a computer. And lets not get started on the ability to handle passengers if things are seriously wrong.
  
• the elephant in the room, freight trains. I am not going to start on how these complicate matters, but if anyone wants to know more, look at the failure of the ability to upgrade the WCML with Moving Block Signalling for 140mph operation, due to the presence of Freight.

Rather than Driverless trains, I would rather like, and promote Automatic Train operation throughout as much of the network as possible together with a driver. Use technology to GIVE the driver as much information as possible, to assist them to use as many routes as possible. Maps, advance signals, suggested speeds, etc. This is already available in isolated areas, such as parts of the London Underground.

This way you can get most of benefits of Driverless trains without loosing the safety aspects of the real driver at the front.

----
Edited to correct typos and spelling

 

[irish_rail]

Look, I am a software consultant, and a general technophile at heart. I personally would LOVE to have driverless trains throughout the UK mainline, however, I can tell you its NOT going to happen anytime in the near future.

For driverless trains to work, the network and trains needs to have:

• a consistent and known infrastructure. The more varied the infrustructure (curved track in platforms, gradients split platform faces, effect of leaves, rain, etc) the harder it is to program computers to account for it all. Just stopping a train at the right point at a platform can be affected by many factors.
• An adequete train detection and signalling system. We still have semaphores, a person in a box with a telegraph and tokens in a purse in places on the UK network. Until all of it is upgraded to a suitable system driverless trains are simply a no go.
• Stock that can electronically speak to each other. Each train woudl need to be able to know that nearby trains are doing to achieve the full benefits of increased capacity, and short headways. Today we have cases where even sets of the same class cannot talk to each other, let alone different classes.
• incident management. The "wetware" neural network that exists between the ears of every average human is still far superior to the best learning AI available currently and the near future. Therefore humans are still able to detect and handle incidents far better than a computer. And lets not get started on the ability to handle passengers if things are seriously wrong.
  
• the elephant in the room, freight trains. I am not going to start on how these complicate matters, but if anyone wants to know more, look at the failure of the ability to upgrade the WCML with Moving Block Signalling for 140mph operation, due to the presence of Freight.

Rather than Driverless trains, I would rather like, and promote Automatic Train operation throughout as much of the network as possible together with a driver. Use technology to GIVE the driver as much information as possible, to assist them to use as many routes as possible. Maps, advance signals, suggested speeds, etc. This is already and available in isolated areas, such as parts of the London Underground.

This was you can get most of benefits of Driverless trains without loosing the safety aspects of the real driver at the front.

Thankyou for an informed common sense post, makes a change from the crap spouted by armchair "experts"

 

[NotATrainspott]

Look, I am a software consultant, and a general technophile at heart. I personally would LOVE to have driverless trains throughout the UK mainline, however, I can tell you its NOT going to happen anytime in the near future.

For driverless trains to work, the network and trains needs to have:

• a consistent and known infrastructure. The more varied the infrustructure (curved track in platforms, gradients split platform faces, effect of leaves, rain, etc) the harder it is to program computers to account for it all. Just stopping a train at the right point at a platform can be affected by many factors.
• An adequete train detection and signalling system. We still have semaphores, a person in a box with a telegraph and tokens in a purse in places on the UK network. Until all of it is upgraded to a suitable system driverless trains are simply a no go.
• Stock that can electronically speak to each other. Each train woudl need to be able to know that nearby trains are doing to achieve the full benefits of increased capacity, and short headways. Today we have cases where even sets of the same class cannot talk to each other, let alone different classes.
• incident management. The "wetware" neural network that exists between the ears of every average human is still far superior to the best learning AI available currently and the near future. Therefore humans are still able to detect and handle incidents far better than a computer. And lets not get started on the ability to handle passengers if things are seriously wrong.
  
• the elephant in the room, freight trains. I am not going to start on how these complicate matters, but if anyone wants to know more, look at the failure of the ability to upgrade the WCML with Moving Block Signalling for 140mph operation, due to the presence of Freight.

Rather than Driverless trains, I would rather like, and promote Automatic Train operation throughout as much of the network as possible together with a driver. Use technology to GIVE the driver as much information as possible, to assist them to use as many routes as possible. Maps, advance signals, suggested speeds, etc. This is already and available in isolated areas, such as parts of the London Underground.

This was you can get most of benefits of Driverless trains without loosing the safety aspects of the real driver at the front.

I think you're conflating two different issues. Autonomy is essentially just the idea of trains being able to run without a driver. Then there's massively upgrading the signalling systems so that you get DLR-style moving block signalling to maximise capacity. The latter essentially requires full or partial autonomy as the humans become the bottleneck.

The former type of autonomy doesn't fundamentally require anything other than kit fitted to the train. Using the same phalanx of sensors that are going to be fitted to buses and trucks on the roads, a train can read the line ahead like a human driver and combine the things it sees with the knowledge it has been trained on. If a computer can describe a complex scene, they can easily identity if it's a green or a red light, or whether the semaphore is up or down. If a computer couldn't determine this, then neither could a human.

 

[notverydeep]

What has struck me repeatedly about all of the threads relating to driverless technology, is that almost all the posters take it for granted that the rail network will continue as is, or expand irrespective of what happens in vehicle automation across other modes. I largely agree with those who comment that outside of Metro operation, the technical challenges of fully driverless trains mean they will probably not be a threat to today's train drivers employment prospects in the next decade or so. But I fear that for the younger drivers, well represented on this forum, there is a more immediate threat to your job security from automation. This is that rail passengers and freight will be lost to Autonomous Road Vehicles (ARVs).

The development of ARV technology is being led by very large well-funded companies spending an enormous amount of cash to make it happen, and when they succeed, low traffic railways will be in real trouble. Many of the cross country routes are heavily reliant on two categories of passenger, both of whom would switch to road if cars were automated and cost no more than rail fares. These are the young, priced out of cars by high insurance costs and older people as they lose the confidence to drive longer distances. The trips these groups make are highly dispersed (from many origins to many destinations) and so usually require one or more interchange by rail. That and slow line speeds on such routes mean that rail has little speed advantage for these customers. ARVs would have other advantages, they will be available on a trip by trip basis and so won't have the fixed costs that go with 'ownership', they would be door to door, they could depart whenever the passenger chose and they could stop for refreshments pretty much whenever the occupant wanted.

Our industry has no particular right to exist. Despite being a lifelong rail enthusiast and railway worker for more than, I suspect that there is trouble ahead. Our passengers travel (or freight consigned) by rail only because they want to be somewhere else and for that journey rail is the best combination of price and convenience available now. If ARVs open up cheaper and more convenient possibilities, people will change their travel behaviour where it suits them, as they did in the years leading up to the Beeching closures with mass car ownership. For railway routes carrying fewer than perhaps an average of 1000 - 2000 passengers per hour, such a change if and when it comes to pass - will be very bad news indeed. There are quite a few cross country routes in the UK that may seem quite busy, e.g. at least two trains per hour in each direction, but those trains are often two carriage sprinters carrying say no more than 140 people.

This change isn't going to happen next year and 'Mass transit' routes with either a good speed advantage (high demand intercity) or capacity advantage (metro or dense commuter routes) will survive - and probably still with drivers at least for the next couple of decades. I think it is safe bet that taxi drivers are at more immediate risk of being put out of work than train drivers, because the technology in cars is likely to be easier and cheaper to implement in towns and cities first and it is there that the market is largest. But as ARVs spread out of urban areas the transport industry in the UK will start to look quite different.

 

[3141]

The Underground examples are automatically driven but are not driverless, as they are not allowed to operate unless a driver is present in the cab. Truly driverless systems not only have to automate the driving function, but have the more difficult task of working out how to handle various failure and emergency situations with no driver at the front of the train, or in the case of Lille and similar "unattended" systems no staff presence anywhere on the train. Unattended systems are normally also expected to have platform edge doors to replace the human role in train dispatch.

Yes, there will indeed be questions that have to be resolved in relation to train dispatch (and several other issues!), if "unattended" systems were extended to main line operation. In relation to failure and emergency situations, the DLR is interesting, because the train captain isn't normally at the front. He can go and take over if necessary, but at the moment an incident occurs he probably wouldn't be aware that it was imminent.

Crikey why is there so much dislike of drivers on here??? I could understand of we were forever on strike but that just isn't the case anymore and hasn't been for years, indeed I've never been out on strike.
Just find it surprising there is so much ill feeling towards drivers here....

I don't know if you'd include me in that, but I'm not anti-driver. I'm taking part in a discussion about the implications of driverless operation, and I expect that's true of most other contributors. There are one or two who seem to think that even discussing the idea means someone is anti-driver, but that's not true. I think there are some very interesting comments about the extent to which driverless operation is feasible, what exactly we mean by "driverless", and so on.

 

[janahan]

I think you're conflating two different issues. Autonomy is essentially just the idea of trains being able to run without a driver. Then there's massively upgrading the signalling systems so that you get DLR-style moving block signalling to maximise capacity. The latter essentially requires full or partial autonomy as the humans become the bottleneck.

I don't think I am conflating two different issues. My main point is that I do not think that driverless trains are not possible across the UK rail network, and gave a few reasons why. Headways and capacity are issues that affect all drivers, computer or human.

<B>The former type of autonomy doesn't fundamentally require anything other than kit fitted to the train. </B> Using the same phalanx of sensors that are going to be fitted to buses and trucks on the roads, a train can read the line ahead like a human driver and combine the things it sees with the knowledge it has been trained on.

If a computer can describe a complex scene, they can easily identity if it's a green or a red light, or whether the semaphore is up or down. If a computer couldn't determine this, then neither could a human.

... and this is pure Science Fiction. As a software developer who is actually very good at my job, and have been programming computers almost all my life, I can tell you with absolute certainty that NO current or near future computer system can match the absolutely awesome hardware and software that is the human brain.... period...

Sure in a deterministic logic system, where a system is specifically programmed for a specific task it may be quicker/more accurate than a human. However, if you put it in a situation outside its programming it will be not be able to try and resolve the situation. It is only as good as its programming, and trust me programming such complex behavior is VERY VERY VERY difficult.

Regarding a complex scene, trust me, i have seen computers screw up recognise a scene with a bit of snow (it failed to recognize the difference between sky and ground) Using human interaction devices (lights/semapores, etc) is difficult with computers.

 

[Dougal2345]

... and this is pure Science Fiction. As a software developer who is actually very good at my job, and have been programming computers almost all my life, I can tell you with absolute certainty that NO current or near future computer system can match the absolutely awesome hardware and software that is the human brain.... period...

I'm a software developer myself, although probably only a fair-to-middling one who has never worked on any kind of image processing.

However I do know that there are driverless cars in existence now. The problems with programming those must be an order of magnitude greater than a similar train system (At the most obvious level, you don't need to steer a train).

So driverless trains are impossible science fiction but driverless cars science fact? I don't quite understand that.

 

[janahan]

I'm a software developer myself, although probably only a fair-to-middling one who has never worked on any kind of image processing.

However I do know that there are driverless cars in existence now. The problems with programming those must be an order of magnitude greater than a similar train system (At the most obvious level, you don't need to steer a train).

So driverless trains are impossible science fiction but driverless cars science fact? I don't quite understand that.

I never said that driverless trains themselves are pure science fiction, i was replying to the assertion by the person i was replying to who said that a computer should be able to recognize a scene and interpret semaphores and signals as well as a human being, is science fiction.

I actually think that self driving cars are actually harder than self driving trains on our current roads. Try driving in the streets of London. I think in certain situations such as on a motorway where ALL cars are self driving, we may be more likely to see self driving cars.

It is not the driving part that is hard. Actually a smart kid with a raspberry pi, a few motors, can create an autonomous robot that can follow "lanes" painted on a floor pretty well. Under controlled conditions it works very well. And as you said a train appears simpler, after all its on rails. That is true too.

Coding for that is actually relatively simple, after all its pure physics and mathematics. You can using formulas and equations code it. In the sixties and seventies, using computer power less than in some calculators today we automatically put satellites in orbit, and flew to the moon and mars.

But what about the out of ordinary things? This is where unless a computer is coded to do it, it cannot. And for it to be coded, someone would have to be aware of the potential problem in the first place.

Self driving trains are ALREADY here in complete controlled and isolated systems such as the DLR, and Paris Metro. These systems are specially designed infrastructure (radio sensors, and pads instead of lights for example) that is more suited for computerization. In fact it could be argued that such systems are harder for a human to drive. Also out of the ordinary issues are controlled and understood in this limited system. But to code for the National Rail network with its greater variations, equipment designed for humans to understand better than computers (semaphores, colour lights, etc) is just that much harder.

 

[Bessie]

I think mainline train drivers can relax for a few more years. Autonomous road vehicle (ARV) technology will transform the taxi and road haulage business first. Our largest cities in the next 10-20 years will become UBER++ zones. Privately owned and driven vehicles will be banned. ARV technology will go together with electric vehicles (EVs). The recently launched TESLA truck in the US has already had orders from AB In-Bev (Budweiser) and PepsiCo. They will still have drivers to begin with but you can see how this technology will expand so trucks become autonomous. When it comes to the UK I can see supermarket fleets and the likes of Eddie Stobart switching by 2025-30.