France looking to have driverless TGVs by 2023

[Pandora]

The pitfall  will be  "to whom can we offload  the blame" ?

 

After the major  incident,  so  convenient   to  throw the book at " the  deceased driver" .

 

Otherwise management face the discomfort and  rigour  of examination  of  the  driverless safety system.

Edited July 13, 2017 by Pandora

[Gerald Henriksen]

The pitfall  will be  "to whom can we offload  the blame" ?

 

After the major  incident,  so  convenient   to  throw the book at " the  deceased driver" .

 

Otherwise management face the discomfort and  rigour  of examination  of  the  driverless safety system.

 

 

The third party who make and maintain the driverless system (aka the driver) - there is after all still a driver, its just that it will be a computer, and somebody will have designed and implemented it.

[jjb1970]

I think some are under estimating how far autonomous systems have already advanced and just how much effort is being poured into technology development. As well as cars, autonomous aircraft and ships are also advancing quickly. Rolls Royce are pumping money into autonomous ships and the general consensus is not "if" but "when" they enter service, and that "when" is now increasingly agreed as being in the near future. The IMO is already working on the necessary legal instruments to accommodate autonomous ships. Again, as with cars, ships operate with more degrees of freedom and are subject to more variables than trains. All of this will profoundly affect societal attitudes to autonomous technologies.

[olivegreen]

At the risk of sounding like a Luddite, I fail to see the need for autonomous ships, driverless (even unmanned) trains - or buses, or cars and so on. As many have said, we live in a human-dominated world and humans need employment in the broadest sense of the term.

 

I am an Engineer and am acutely aware of the wonders of technology, but pointless application of it 'just because it can do it' gives way to commercialism.

 

The fact that I could, if I wished, use a smartphone in Penzance to control the oven in my kitchen in Inverness demonstrates an outstanding combination of technologies but how does that actually improve my life - or that of anyone else?

 

What much of the world needs is water to drink, sanitation and decent food to eat.

 

I think I'm getting old….

 

(Minor edit)

Edited July 14, 2017 by olivegreen

[david.hill64]

At the risk of sounding like a Luddite, I fail to see the need for autonomous ships, driverless (even unmanned) trains - or buses, or cars and so on. As many have said, we live in a human-dominated world and humans need employment in the broadest sense of the term.

 

I am an Engineer and am acutely aware of the wonders of technology, but pointless application of it 'just because it can do it' gives way to commercialism.

 

The fact that I could, if I wished, use a smartphone in Penzance to control the oven in my kitchen in Inverness demonstrates an outstanding combination of technologies but how does that actually improve my life - or that of anyone else?

 

What much of the world needs is water to drink, sanitation and decent food to eat.

 

I think I'm getting old….

 

(Minor edit)

Society is getting less tolerant of people's mistakes: look at the claim culture that is developing. Therefore organizations need to reduce risk as far as practicable. Given that humans are the principal cause of 'accidents', getting them out of the control loop on a day to day basis is a sensible thing to do.

 

There is also the need to get more out of less: for example running more trains on existing infrastructure. But signallers and drivers are not able to respond to the requirements to control trains to the degree necessary to operate 24 trains per hour on a congested railway where minor perturbations will have a knock-on effect throughout the system.

 

So I think that taking humans out of routine control functions is inevitable. This leads to all sorts of concerns about keeping skills available for when everything has gone pear shaped when the technology has tripped up.

 

I think that your example of being able to use a phone in Penzance to control your oven in Inverness is an excellent example of being able to do something without any apparent benefit, but have you considered that it might be nice to be able to control the oven to prevent it turning on at the preset time when you are caught in a traffic jam on the way home and you don't want your dinner burnt? I suspect that there are all sorts of applications that we didn't know we wanted until they are available, just as though there are many that (for me) are mostly useless or even a menace. I certainly take your point that commercialism is a real risk, and I certainly agree about drink, sanitation and food.

[DY444]

I think some are under estimating how far autonomous systems have already advanced and just how much effort is being poured into technology development. As well as cars, autonomous aircraft and ships are also advancing quickly. Rolls Royce are pumping money into autonomous ships and the general consensus is not "if" but "when" they enter service, and that "when" is now increasingly agreed as being in the near future. The IMO is already working on the necessary legal instruments to accommodate autonomous ships. Again, as with cars, ships operate with more degrees of freedom and are subject to more variables than trains. All of this will profoundly affect societal attitudes to autonomous technologies.

 

I would argue there's a substantial difference between ships, aircraft (and trains) on the one hand and road vehicles on the other.  It seems to me that the number of variables relating to road vehicles is substantially greater and random in nature making it a totally different proposition.  I still maintain that an acceptably competent automated vehicle for general use by Joe Public is a long way off not just because of the technical challenges but because of the regulatory and legal ones too. 

 

All software has to cater for unknown cases with one or more sets of default actions which may or may not be appropriate to the unknown case (because if you knew what to do for the unknown case it wouldn't be unknown!).  How happy is a driver going to be to take unconditional personal responsibility for the actions of an automated vehicle controlled by software written by some unknown college kid?    I say not happy at all and I think that will take a very long time to change. 

[toffee]

Pilotless planes are all very well until something goes wrong. AF 447 was caused by the autopilot disengaging because of a technical glitch and the 2nd and 3rd pilots who were then in charge being so poorly trained that they crashed it. And when the railway version of this decides that it doesn't want to play ? And then there is the possibility of terrorists gaining control of a driverless train by hacking it and smashing it into the buffers ? Unlikely ? We could well find out.

[w124bob]

You might have driverless trains/cars etc but the elephant in the room is the ability to generate tax revenue. At the moment the average pay for a train driver is £50k plus, if you reduce the the driver to (lets say) operator at some future point the pay scale will fall either by time or deliberate regrading. The driverless car model seems to suggest that individuals may not be the owners, again having impications for tax generation. Not forgetting the alternative fuel debate. If governments move to replace one tax option (wages) with taxes on business profits, automation of many jobs may just not happen. As a self confessed petrolhead I feel my driving days will become increasingly restricted and expensive.

[admiles]

I once read that German ICE drivers always drove with the sun blinds fully down as if they were going to hit something, they didn't want to know about it. There is some merit in that approach!

 

I can confirm that to be BS of the highest order!  Seen literally hundreds of ICEs over the years and never once one with the sun blind all the way down on the leading cab.

In the ICE3 (and ICE-T too I think) you can sit directly behind the driver and look through the glass cab bulkhead and through the windscreen.

[Joseph_Pestell]

I think a game changer for driverless trains will be autonomous cars. Technically there has been no reason why trains couldn’t be fully automated for many years, as evidenced by a number of automated light railways there are, even if they sometimes retain a “driver” in the cab. Traditionally there has however been some societal resistance to the idea as people often feel uncomfortable about entrusting their lives to a computer, even though if they consider the question most will quickly realise just how much they already rely on computer controlled systems to lead their lives and stay safe. As people get used to the idea of autonomous cars I expect that attitude will change, after all if a car can operate autonomously (and yes, I know automating trains is not the same as using computers to control an autonomous car) which faces far more challenges in terms of degrees of freedom, congestion, interaction with other road users and potential obstructions than a train then why not automate trains too? I’m not saying that is right or wrong, nor am I advocating driverless trains per se (for what it’s worth I think my preference would be to retain a cab operator alongside an automated train control system for mainline trains) but I do think that wider societal changes will mean that we will soon see a culture where it is human control of trains which is questioned. And politically, I think the RMT have already done a lot of the groundwork to prepare people to embrace automated trains in this country, which is a rather unfortunate by-product of their behaviour, sadly.

 

Braking performance of cars is so much better than that of mainline trains travelling at 200+mph.

[Joseph_Pestell]

I don't use TGVs very often these days as it is so much cheaper and easier to fly with Ryanair or Easyjet. If they become driverless, I will be even more inclined to stick with the airlines although it can only be a matter of time before Michael O'Leary decides that pilots are an expensive extra.

[black and decker boy]

I don't use TGVs very often these days as it is so much cheaper and easier to fly with Ryanair or Easyjet. If they become driverless, I will be even more inclined to stick with the airlines although it can only be a matter of time before Michael O'Leary decides that pilots are an expensive extra.

But airplanes can already fly themselves and do take off and landing. The pilot can be 'just in case' even now. I think many / most do actually manually fly the things but don't need too and you can't tell the difference when you are strapped in so you won't know who was in charge if they don't fess up

[griffgriff]

but... who will open the doors??

I got it

[Joseph_Pestell]

But airplanes can already fly themselves and do take off and landing. The pilot can be 'just in case' even now. I think many / most do actually manually fly the things but don't need too and you can't tell the difference when you are strapped in so you won't know who was in charge if they don't fess up

Oh yes, I know that. But would a computer have been able to "land" a plane on the Hudson River (or even to decide to try).

 

From the very bumpy landings that I regularly experience with Ryanair - their repair bills on landing gear must be quite something - I suspect that the pilots are in control rather than the computer. I will ask my neighbour who flies for them.

[jjb1970]

Braking performance of cars is so much better than that of mainline trains travelling at 200+mph.

But not affected by whether a computer or driver applies the brake. If something gets in front of a 200mph+ train then whether or not you have a driver is pretty much irrelevant.

Edited July 14, 2017 by jjb1970

[rockershovel]

Re bumpy landings, BA served up some real corkers in the later days of the LHR - Baku service: especially at LHR for some reason. I was on one where there was genuine alarm among the passengers - and these were experienced travellers!

 

Why eliminate crew on sea-going vessels? Simplez. They expect wages, require accommodation on board, need feeding and functioning sanitation. Galley and steward crew to look after it, laundry, rec room, some wifi provision these days. Flights to join vessels. Admin staff to recruit, pay the wages and administer the tax. Wages aren't the half of the cost.

[jjb1970]

I think there are three aspects to the autonomous systems argument:

 

Safety

Technical feasibility

Societal impact (loss of jobs etc)

 

In terms of technical feasibility I don't think many can seriously doubt that the technology can be developed. Especially when for airliners and trains there are already mature technologies capable of automatic control of complex systems.

On safety, it mystifies me why so many people are nervous about the safety of autonomous systems when probably the dominant causal factor in most safety reports I read or incidents I've been involved in is human factors. That's not to say automation eliminates human error, it moves it to another part of the process (design and verification), but that part of the process is subject to more rigorous engineering controls and assurance than an operator at the sharp end. If you consider shipping, probably the most hazardous offshore operations have relied on dynamic position which is fully automated ship control since the 1980's and many energy companies and regulators mandate DP for these operations now. You still have DPOs on the bridge but all they do is the initial system set up then monitoring. So for safety, if anything the arguments are against human control.

The third point is the only really valid argument against increasing automation but the process of automating processes has been in progress for many decades and is nothing new. Pneumatic PID control loops go back many decades. The problem is that society changes and efforts to freeze things tend to end up counter productive, the question isn't so much about whether the world will change (it will) but how we adapt to change. Change tends to be a traumatic experience and we don't know what the future holds but the answer isn't to try and stop change and evolution, after all the human race has been going down the tubes for hundreds of years according to conventional wisdom.

[pete_mcfarlane]

The third point is the only really valid argument against increasing automation but the process of automating processes has been in progress for many decades and is nothing new. Pneumatic PID control loops go back many decades. The problem is that society changes and efforts to freeze things tend to end up counter productive, the question isn't so much about whether the world will change (it will) but how we adapt to change. Change tends to be a traumatic experience and we don't know what the future holds but the answer isn't to try and stop change and evolution, after all the human race has been going down the tubes for hundreds of years according to conventional wisdom.

I've been to a couple of presentations recently (and watched a video of my firm's CEO) where they've talked about the impending arrival of automation (in the form of driver-less cars, machine learning etc) as if it was something incredibly new that was coming out of the blue, rather the continuation of the previous 250 years of technological development driving more automation. 

 

If nothing else, it shows how poor the teaching of the history of technology is. 

[meil]

Re bumpy landings, BA served up some real corkers in the later days of the LHR - Baku service: especially at LHR for some reason. I was on one where there was genuine alarm among the passengers - and these were experienced travellers!

 

 

And yet some of the smoothest landings I experienced was flying Azerbaijan airlines when they had their 727s in the late 90's.

[Nearholmer]

The most surprising thing about the concept of a "driverless" TGV is that anyone finds it surprising. I hope to goodness, and am pretty sure, that a computer is doing most of the job now, because it isn't a job you would want to trust to a flaky old human.

 

My guess is that either SNCF or Alstom is "pushing" the topic for one or both of the following reasons:

 

- sabre rattling in the endless debate with trades unions, by pointing out, gently, that unattended train operation is perfectly feasible;

 

- making very sure indeed that everyone in the world knows that France Plc can do this, just in case anyone thinks that China Plc, or Japan Plc, or Germany Plc has got a capability that France hasn't.

 

Kevin

[Dunsignalling]

But not affected by whether a computer or driver applies the brake. If something gets in front of a 200mph+ train then whether or not you have a driver is pretty much irrelevant.

It's not affected at much lower speeds than that, purely because of the way braking of a metal-wheeled vehicle on metal rails behaves. Quite simply, all trains do "going" rather better than they do "stopping".

 

Unless one has detailed (radar?) vision of the line for a couple of miles ahead, braking itself can be more hazardous than not, depending on the nature of the obstruction.

 

Take the example of a person deciding to end it all under a train travelling at 80mph who steps out from cover when it's 100 to 200 yards away. The approved (and natural) reaction is for the driver to make an emergency brake application.

 

Even with extremely good reactions, this won't mitigate the impact upon the poor soul by more than 10 mph and is unlikely to make a difference to his/her prognosis.

 

Meanwhile, on board the train, standing passengers and any in wheelchairs, hot drinks, the conductor, anyone moving to/from the toilets, or operating a buffet trolley (and the trolley itself) are all in danger of being thrown about with unknown consequences.

 

John

Edited July 15, 2017 by Dunsignalling

[jjb1970]

It's not affected at much lower speeds than that, purely because of the way braking of a metal-wheeled vehicle on metal rails behaves. Quite simply, all trains do "going" rather better than they do "stopping".

 

Unless one has detailed (radar?) vision of the line for a couple of miles ahead, braking itself can be more hazardous than not, depending on the nature of the obstruction.

 

Take the example of a person deciding to end it all under a train travelling at 80mph who steps out from cover when it's 100 to 200 yards away. The approved (and natural) reaction is for the driver to make an emergency brake application.

 

Even with extremely good reactions, this won't mitigate the impact upon the poor soul by more than 10 mph and is unlikely to make a difference to his/her prognosis.

 

Meanwhile, on board the train, standing passengers and any in wheelchairs, hot drinks, the conductor, anyone moving to/from the toilets, or operating a buffet trolley (and the trolley itself) are all in danger of being thrown about with unknown consequences.

 

John

I agree, unfortunately it is quite common to mix questions relating to control with mechanical aspects which do not change in the slightest whether or not you automate. Momentum, friction, the laws of energy, mechanics etc are what are.

[jjb1970]

The most surprising thing about the concept of a "driverless" TGV is that anyone finds it surprising. I hope to goodness, and am pretty sure, that a computer is doing most of the job now, because it isn't a job you would want to trust to a flaky old human.

 

My guess is that either SNCF or Alstom is "pushing" the topic for one or both of the following reasons:

 

- sabre rattling in the endless debate with trades unions, by pointing out, gently, that unattended train operation is perfectly feasible;

 

- making very sure indeed that everyone in the world knows that France Plc can do this, just in case anyone thinks that China Plc, or Japan Plc, or Germany Plc has got a capability that France hasn't.

 

Kevin

I think this is a raw nerve in France. They were early pioneers of the modern concept of high speed rail (though not as early as Japan) and for a while TGV was an internationally used term. Over the last 20 years or so Alstom has seen their product slide from a global by-word for high speed trains to almost an also ran as German, Chinese, Spanish, Italian etc companies have stolen much of the limelight. And of course Japen continues to develop their Shinkansen trains and are now promoting their technology aggressively over here.

[olivegreen]

I agree, unfortunately it is quite common to mix questions relating to control with mechanical aspects which do not change in the slightest whether or not you automate. Momentum, friction, the laws of energy, mechanics etc are what are.

 

I absolutely agree with you and Dunsignalling - hence my comment in post 7 (above) regarding the stupidity of the sentence as written in the article quoted in the OP!

[Joseph_Pestell]

I agree, unfortunately it is quite common to mix questions relating to control with mechanical aspects which do not change in the slightest whether or not you automate. Momentum, friction, the laws of energy, mechanics etc are what are.

Or, as Scotty would have said, "Ye canna beat the laws o' physics".

 

Beam me up.