Driverless trains - why limited progress on the national rail network?

[Roast Veg]

If the rail industry does not act fast enough and automate quickly enough, then it will collapse. Automated road vehicles are years away - but not decades, and they spell a much more convenient alternative for both freight and passengers than rail can in its present form. Why would a business or passenger choose rail instead, when they can get door to door delivery from a machine that takes no breaks, makes no mistakes, and can arrive at the start location at any given time? Rail has to win out against automated road vehicles by being faster and cheaper in large quantities, which are only achievable with capacity increases and automation - something that only automation can do in places where the mechanical infrastructure isn't the limiting factor.

 

[LAX54]

There's a very good reason that no Level 3 ETCS installations exist yet. It works well on paper, but in real life - well, see my signature quote.

bit like TMS then

 

[najaB]

If the rail industry does not act fast enough and automate quickly enough, then it will collapse. Automated road vehicles are years away - but not decades, and they spell a much more convenient alternative for both freight and passengers than rail can in its present form.

I take it you haven't read the whole thread? This was discussed somewhere around post #120 onwards.

 

[NotATrainspott]

and look how reliable that is...not !

GSM-R is not reliable for a perm connection, ETRMS is still quite a few years off being used big time, even in its basic form !

What happens when you switch the OHL off ? which is most nights in many many places

Automation is no the be all and end all, the more automation the more to go wrong and affect a much larger area......Oh it already does ! lol

There's a very good reason that no Level 3 ETCS installations exist yet. It works well on paper, but in real life - well, see my signature quote.

The important part is that ETCS Level 3 is going to happen eventually, thereby proving beyond doubt that wireless communication can in fact be used as a crucial part of railway safety. It isn't here yet, but neither are most of the things that I'm talking about. I'm discussing the future and how technology is developing in ways that will make some currently challenging things quite a lot easier.

I know power transfer can be done in the LF-HF bands quite easily using this method with ferrites at the receiver, but I honestly don't know how feasible it would be at 50Hz.

What kind of infrastructure sensing do you have in mind for such a system, given the lower power levels which would be available.

I think there will be lots of different kinds of sensing at different scales and different power requirements to suit. The phantom 25kV power idea would be for the most massive deployments of the cheapest, most basic sensors. I'm thinking along the lines of strain or vibration gauges effectively printed every few metres onto the side of new rails. These will require very low amounts of power and would simply not be financially feasible if you did have to install cables to power them.

Larger, more complex sensor arrays such as the wireless CCTV poles would have their own power sources as appropriate due to their increased power requirement, cost and relative scarcity. These poles could supply their own wireless power to nearby sensing equipment if need be.

 

[NotATrainspott]

If the rail industry does not act fast enough and automate quickly enough, then it will collapse. Automated road vehicles are years away - but not decades, and they spell a much more convenient alternative for both freight and passengers than rail can in its present form. Why would a business or passenger choose rail instead, when they can get door to door delivery from a machine that takes no breaks, makes no mistakes, and can arrive at the start location at any given time? Rail has to win out against automated road vehicles by being faster and cheaper in large quantities, which are only achievable with capacity increases and automation - something that only automation can do in places where the mechanical infrastructure isn't the limiting factor.

I wouldn't go as far as to say that it would collapse, but it could put quite a bit of pressure on the railways if they don't try to keep up. My example of the Uber bus coming to pick up train commuters wouldn't be able to scale up to all passengers due to the fundamental limitations of road transport, but it would be able to cream off enough to reduce profitability. It would mean that the railways couldn't get away with sardine can conditions and high fares for commuters, which they often can at the moment as people have no option but to use that train service to get to the best-paying employment that they can find.

 

[najaB]

The important part is that ETCS Level 3 is going to happen eventually...

It may happen. Or it may turn out that it's just not possible to achieve the necessary level of reliability in the real world.

The annuls of technology are littered with ideas which worked well on paper.

 

[LAX54]

It may happen. Or it may turn out that it's just not possible to achieve the necessary level of reliability in the real world.

The annuls of technology are littered with ideas which worked well on paper.

I do not think there is any bit of 'technology' currently in use that is reliable, be that in the home or in the workplace.

 

[Western Lord]

If the rail industry does not act fast enough and automate quickly enough, then it will collapse. Automated road vehicles are years away - but not decades

You have that the wrong way round. Automated road vehicles are decades away - not years. Please do not confuse the current attempts at "self driving" cars, which only work under the most favourable conditions and hand control back to the driver when they can't cope, with fully automated autonomous vehicles.

 

[najaB]

You have that the wrong way round. Automated road vehicles are decades away - not years.

I think the truth is somewhere in the middle. It's likely more than ten but less than thirty years before the majority of road vehicles are capable of fully autonomous driving.

 

[AM9]

I do not think there is any bit of 'technology' currently in use that is reliable, be that in the home or in the workplace.

What a bizarre statement. What do you mean by the words 'technology' and 'reliable'. Given that absolute infallibility is impossible in humans, or natural or man-made products, your statement does really need to be qualified in some way.

 

[NotATrainspott]

You have that the wrong way round. Automated road vehicles are decades away - not years. Please do not confuse the current attempts at "self driving" cars, which only work under the most favourable conditions and hand control back to the driver when they can't cope, with fully automated autonomous vehicles.

I think the truth is somewhere in the middle. It's likely more than ten but less than thirty years before the majority of road vehicles are capable of fully autonomous driving.

drive.ai Rainy Night Autonomous Drive

What's shown in this video simply was not possible 12 months ago. 12 months from now, we'll be able to do even more. Tesla says they'll have a full self-driving car trip from coast to coast in the US by the end of this year.

You don't need 100% autonomy to have a problem. Nissan believes that remote call centre employees will be there to take over control in the rare event of a disengagement in future. However, every single thing that the call centre people do to remedy the situation would be recorded and every vehicle on the road would learn from it. Given the number of strange situations that the millions of road vehicles would encounter each day, it would not take long to learn how to deal with basically every single one of them. Once learned, it can never be un-learned.

 

[najaB]

You don't need 100% autonomy to have a problem. Nissan believes that remote call centre employees will be there to take over control in the rare event of a disengagement in future.

I'm sorry, but if I'm doing 70mph on a motorway (or even 30mph on a busy urban road) then "All our operators are busy. Please stay on the line, you call is important to us." just isn't an option!

 

[NotATrainspott]

I'm sorry, but if I'm doing 70mph on a motorway (or even 30mph on a busy urban road) then "All our operators are busy. Please stay on the line, you call is important to us." just isn't an option!

What Nissan mean here is a non-hazardous situation which the car simply doesn't know how to handle. For instance, if there's roadworks which should have temporary traffic lights or a man with a Go/Stop pole but neither is there or working.

 

[najaB]

What Nissan mean here is a non-hazardous situation which the car simply doesn't know how to handle. For instance, if there's roadworks which should have temporary traffic lights or a man with a Go/Stop pole but neither is there or working.

Which is fine if that's the situation that has caused the disengagement. But until you have at least a five-nines level of confidence that the disengagement won't happen in a situation such as I've highlighted above, the likes of Tesla's Autopilot is as far as ARV technology will go: the automation provides driving assistance but there will have to be a competent driver to take over if needs be.

 

[Dave1987]

The important part is that ETCS Level 3 is going to happen eventually, thereby proving beyond doubt that wireless communication can in fact be used as a crucial part of railway safety. It isn't here yet, but neither are most of the things that I'm talking about. I'm discussing the future and how technology is developing in ways that will make some currently challenging things quite a lot easier.

How can you possibly say that it will happen eventually?

 

[Western Lord]

I think the truth is somewhere in the middle. It's likely more than ten but less than thirty years before the majority of road vehicles are capable of fully autonomous driving.

Please explain to me how a fully autonomous car can be directed into a farmers field being used as a temporary car park at some country event and park itself in the place directed by the person supervising the field. To whom is that person supposed to give instructions? As I understand it at the moment all these "self driving" cars have to be able to see road markings, which most of our roads don't have and even if they do they are often so poorly maintained as to be invisible.

 

[najaB]

As I understand it at the moment all these "self driving" cars have to be able to see road markings, which most of our roads don't have and even if they do they are often so poorly maintained as to be invisible.

That is incorrect. While they may use road markings to supplement their navigation systems the primary source of information is a LIDAR unit which builds up a three-dimensional model of the immediate environment.

The US Army has units which can successfully navigate both unmarked and dirt roads:

[youtube]x6ePtBPwZE0[/youtube]

 

[Emblematic]

Which is fine if that's the situation that has caused the disengagement. But until you have at least a five-nines level of confidence that the disengagement won't happen in a situation such as I've highlighted above, the likes of Tesla's Autopilot is as far as ARV technology will go: the automation provides driving assistance but there will have to be a competent driver to take over if needs be.

You just need to follow the money to realise that the technology won't just stop at any point where you still need the driver; the payback comes when the professional driver is no longer needed at all. Logistics operators want the cost savings, as well as the ability to use vehicles without the restrictions a human operator imposes.
There's already a worsening shortage of qualified HGV drivers, the average age in the profession in the UK is 53, and very few young people look to driving as a career. 4% of drivers are over 65, only 1% under 25.
Most people who can drive are indifferent or negative as to whether their next car is going to be autonomous. However, that's not where the real money is going to be made.

 

[najaB]

You just need to follow the money to realise that the technology won't just stop at any point where you still need the driver; the payback comes when the professional driver is no longer needed at all.

I agree. But it's my opinion that we're closer to 15 years than five away from that point.

 

[Western Lord]

That is incorrect. While they may use road markings to supplement their navigation systems the primary source of information is a LIDAR unit which builds up a three-dimensional model of the immediate environment.

The US Army has units which can successfully navigate both unmarked and dirt roads:

[youtube]x6ePtBPwZE0[/youtube]

And the farmers field scenario?

 

[najaB]

And the farmers field scenario?

Navigate to the farmer's field as per usual and follow hand signals once they get there.

If ARVs are common it's likely that there will be positioning beacons set out. Parking is actually trivial compared to other tasks that ARVs will have to accomplish.

 

[Emblematic]

I agree. But it's my opinion that we're closer to 15 years than five away from that point.

Yes, could be anywhere between, or beyond those figures - very hard to predict when any change like this will happen, and even harder to predict the eventual impact.
My computer research colleagues from university days in the 80s predicted (and worked on) technologies such as voice recognition, video and audio streaming, AI. We knew that computing would become ubiquitous, and everything would become connected eventually. At a time when the CD was the next big thing, though, it was hard to foresee when all these things would be commonplace. Even when the IPhone was launched, it wasn't immediately obvious that it was the catalyst that would put all of these technologies in the average man's pocket.
All I foresee is that, once someone manages to monetise all this research, and the regulatory/legal/moral issues are sorted or glossed over, things will change very rapidly.

 

[Emblematic]

Navigate to the farmer's field as per usual and follow hand signals once they get there.

If ARVs are common it's likely that there will be positioning beacons set out. Parking is actually trivial compared to other tasks that ARVs will have to accomplish.

Or more likely - get out of hired autonomous vehicle at drop-off point. Car departs to next booking - no longer your problem - parking unnecessary.

For those still stuck with the ownership model, still get out of car at drop-off point, but send car to nearest autonomous-capable parking (and, chances are, charging) bay.

 

[Western Lord]

Navigate to the farmer's field as per usual and follow hand signals once they get there.

If ARVs are common it's likely that there will be positioning beacons set out. Parking is actually trivial compared to other tasks that ARVs will have to accomplish.

So every vehicle has to be directed by hand to a space! And country fair organisers have to invest in beacons! What absolute nonsense. The proponents of driverless cars seem to believe that cars are used like trains, running on a well maintained and well defined permanent way. Cars are not used like that.
There is no doubt that we will see an increase in the number of cars with some sort of self driving capability, but the totally autonomous vehicle with no control possible by the occupant is a long way from being anything other than a party trick.

 

[emil]

Driverless car crashes in race, but the winning car slows down to avoid a dog in Argentina

http://www.bbc.co.uk/news/technology-39027477

 

[najaB]

So every vehicle has to be directed by hand to a space...

You're the one who introduced someone manning the field:

Please explain to me how a fully autonomous car can be directed into a farmers field being used as a temporary car park at some country event and park itself in the place directed by the person supervising the field.

It can just as easily be done by the car itself. As I said, parking is trivial - many cars on the market today can park themselves.

 

[Western Lord]

many cars on the market today can park themselves.

Not in an empty field they can't. How do you tell an autonomous car to park over in the corner by the hedge?

 

[najaB]

Not in an empty field they can't.

Only because nobody has yet decided to offer the feature. As I said, parking is possibly the most trivial self-driving function that there is.

How do you tell an autonomous car to park over in the corner by the hedge?

You bring up the map that the car has built of its surroundings on the in-car display, and tap where you want the car to park.

 

[NotATrainspott]

Which is fine if that's the situation that has caused the disengagement. But until you have at least a five-nines level of confidence that the disengagement won't happen in a situation such as I've highlighted above, the likes of Tesla's Autopilot is as far as ARV technology will go: the automation provides driving assistance but there will have to be a competent driver to take over if needs be.

High speed driving on controlled-access roads is one of the easiest things for autonomous vehicles to do and covers a huge proportion of the vehicle-miles that are currently driven. An autonomous vehicle needs to be able to slow down and stop safely, which is something that the technology doesn't currently demonstrate mainly because of the knowledge that a human can and should take over when the car demands it.

The cutting edge of cruise control is SAE Level 2, with full autonomy at Level 5. We will not see vehicles produced with Level 3 capability because it will be lethal, as humans simply cannot be trusted to remain able to take over control. Nissan's idea of call centre control is the solution to any lingering disengagement issues once the vehicles can be guaranteed to be able to come safely to a stop whenever something's gone wrong.

How can you possibly say that it will happen eventually?

Because if it doesn't, the cost of traditional signalling systems will cripple the railways financially. It might not be called ETCS Level 3 and it may not have the moving-block part but you can be guaranteed that you'll be using purely wireless control for movement authority eventually.

Yes, could be anywhere between, or beyond those figures - very hard to predict when any change like this will happen, and even harder to predict the eventual impact.
My computer research colleagues from university days in the 80s predicted (and worked on) technologies such as voice recognition, video and audio streaming, AI. We knew that computing would become ubiquitous, and everything would become connected eventually. At a time when the CD was the next big thing, though, it was hard to foresee when all these things would be commonplace. Even when the IPhone was launched, it wasn't immediately obvious that it was the catalyst that would put all of these technologies in the average man's pocket.
All I foresee is that, once someone manages to monetise all this research, and the regulatory/legal/moral issues are sorted or glossed over, things will change very rapidly.

Very rapidly indeed. The cost savings involved are so enormous that there is simply no way that a traditional transport company can compete. Uber will wipe out traditional minicabbing as soon as their autonomous taxis are unleashed. Plausibly they could bring their own taxi model to the market and pay people to drive it, so that their algorithms can learn the art of being a taxi driver in that specific location, and then one day just have their taxis go and drive themselves around instead. Autonomous taxis could become common quite literally overnight.

Not in an empty field they can't. How do you tell an autonomous car to park over in the corner by the hedge?

Tell it to find somewhere convenient to park. It'll then seek out an appropriate space using the various sensors it has. Parking as we know it today will be gone because of the dramatic efficiency improvements possible with autonomy. Your event operator would love to be able to fit 1000 cars in the space traditionally needed for 500 when you've got terrible human drivers and the need to open the doors at the side. People hate parking with a passion and autonomy means you can have valet parking everywhere for free, and there's no downside for the site owner or operator.

 

[najaB]

High speed driving on controlled-access roads is one of the easiest things for autonomous vehicles to do and covers a huge proportion of the vehicle-miles that are currently driven.

I know that, and I fully expect we will see more and more systems like Tesla's Autopilot over the next few years. However, at some point even the easy task will be more than the automation can handle - it could be something as simple as water getting into a connector causing the sensors to drop out.

The point I was making is that, if that were to happen, a remote operator in a contact centre somewhere in the world isn't going to be much use to me when I'm heading towards a bridge abutment at 70mph (or, for that matter, if the failure is in the HGV that's behind me and my car has to brake rapidly).