Would you go driverless?

[Meatyfool]

There's also the issue that many people want their cars on the road at the same time. Hence rush hours. So I don't think you would get away with significantly fewer vehicles overall.

gryff

I failed to get my point across last time I posted, so may not do so again!

There will be less cars on the road. Massive pools of driverless cars will be the death of the railway. Instead of being dictated to about when you can travel (and not get a seat), you inform the taxi company when you need to leave home/work and they plan your journey around those times and requests of other travellers.

Don't want to travel with others? Well, technically you won't. The driverless "car" that turns up at your door has "X" separate "pods"** in it, each one with an office chair, screen, Wi-Fi etc. You start work if you wish when you leave home. If you do, your employer factors that into your 8 hour day and you get to leave at three. Rush hour? No, start and finish hours will spread out.

Have to work late? No worries, inform the hire co, and the algorithm sorts it all out.

As an aside, those who do want to carry on driving may well find that they can't afford to IF the stats show that the driverless cars really are safer. Insurance may well rocket!

** Think about it, your present car is your own personal little pod, that is unproductive in that you have to drive it. A present day taxi is your own personal little pod that you can be productive in. Why a big issue if there are two personal little pods on a single vehicle? How about a coach size driverless "car" with thirty pods on? Does it really matter?

Meatyfool..

[Watis]

Dickhead cyclists and pedestrians will soon learn to behave in the appropriate manner.

The appropriate manner being (in their view of course) to pull right out in front of driverless cars, in the sure knowledge the car will give way to them.

I don't see why software should be made at huge expense to accomodate them.

Fortunately, it will be made at huge expense to accommodate them.

Until of course the software can distinguish a dickhead from an innocent child.

Scott.

Even software cannot override the laws of physics, so, after a few 'dickheads' have met their maker, normal behaviour will resume.

Watis

[Watis]

There's also the issue that many people want their cars on the road at the same time. Hence rush hours. So I don't think you would get away with significantly fewer vehicles overall.

gryff

I failed to get my point across last time I posted, so may not do so again!

There will be less cars on the road. Massive pools of driverless cars will be the death of the railway. Instead of being dictated to about when you can travel (and not get a seat), you inform the taxi company when you need to leave home/work and they plan your journey around those times and requests of other travellers.

Don't want to travel with others? Well, technically you won't. The driverless "car" that turns up at your door has "X" separate "pods"** in it, each one with an office chair, screen, Wi-Fi etc. You start work if you wish when you leave home. If you do, your employer factors that into your 8 hour day and you get to leave at three. Rush hour? No, start and finish hours will spread out.

Have to work late? No worries, inform the hire co, and the algorithm sorts it all out.

As an aside, those who do want to carry on driving may well find that they can't afford to IF the stats show that the driverless cars really are safer. Insurance may well rocket!

** Think about it, your present car is your own personal little pod, that is unproductive in that you have to drive it. A present day taxi is your own personal little pod that you can be productive in. Why a big issue if there are two personal little pods on a single vehicle? How about a coach size driverless "car" with thirty pods on? Does it really matter?

Meatyfool..

How are we paying for all these driverless cars?

If it is per mile, it will be annoying to have a journey cost more than expected because a diversion was necessary.

If per minute, it will similarly mean that a delay - which was not your fault - costs you more than anticipated.

And in the rush hour, when demand is at its highest, will surge pricing come into play? Will it be acceptable to be late for work because you waited for the price to fall? Or because no cars were immediately available at the time you leave for work?

Watis

[swill453]

Even software cannot override the laws of physics, so, after a few 'dickheads' have met their maker, normal behaviour will resume.

No, it won't I'm afraid.

The programming will always have to take the cautious approach. All other road users will be deemed sacrosanct, and will be able to change their behaviour to take advantage of the driverless cars.

The law might be able to intervene, but technology won't.

(How can it be sure to distinguish a temporary traffic light from a big red torch in the dark?)

There's really no other way it can happen.

Scott.

[gryffron]

There will be less cars on the road. Massive pools of driverless cars will be the death of the railway. Instead of being dictated to about when you can travel (and not get a seat), you inform the taxi company when you need to leave home/work and they plan your journey around those times and requests of other travellers.

You start work if you wish when you leave home. If you do, your employer factors that into your 8 hour day and you get to leave at three. Rush hour? No, start and finish hours will spread out.

Nice dream. I don't see much of it as realistic.

Traffic density on railways is hugely greater than cars - even if they are all driverless. Trains are STILL twice as fast and twice as safe as cars. (And in fact, if we could get rid of the drivers, trains would be safer still. Far too many train crashes still caused by train drivers skipping red lights. But try getting that past the unions)

We ALREADY have the technology for huge numbers of people to work from home, or flexible hours. Managers don't accept it. How do you manage staff you cannot see? Or when said manager is not "at work".

gryff

[saechunu]

The programming will always have to take the cautious approach. All other road users will be deemed sacrosanct, and will be able to change their behaviour to take advantage of the driverless cars.

The law might be able to intervene, but technology won't.

(How can it be sure to distinguish a temporary traffic light from a big red torch in the dark?)

There's really no other way it can happen.

Scott.

Correct. The technology's primary focus is likely to be, and remain on, ensuring human safety. An insurance model that shifts to a manufacturer product liability approach for the autonomous systems will ensure this focus is relentless.

The novelty of shooting down drones or other types of interference with autonomous vehicles will quickly lose their novelty once the novelty of the vehicles themselves passes, and after the law has punished the handful of idiots who do choose to test that law. Few people currently waste their time interfering with existing aspects of our transport or communications infrastructure so there's little reason to expect a persistent upsurge in that behaviour.

Vehicles are used in a very wide range of situations, some of which will be extremely difficult to achieve Level 5 autonomy in. Single track rural lanes with infrequent passing places, requiring semi-off-road antics, are one everyday example of a very tricky problem domain where judgement, cooperation and other complex decisions are required. This will not be easy stuff to solve. I'd like to be pleasantly surprised here, but expect that full Level 5 autonomy will not therefore be achievable in all situations for something in the order of 40 years or more (ie. sufficiently distant in time as to be able to ignore it for now). Of course, (much) sooner for simpler environments.

It should be interesting to see how we manage this period between limited and eventual full autonomy everywhere: many journeys will involve both autonomous and manual sections, with the former increasing over time. The transition, on-the-fly, between the computer and human piloting presents a lot of safety problems in terms of engaging the human's attention, and over time, ensuring the human's skill levels are up to the job as the periods where they are required to pilot diminish.

Tricky, but eventually solvable, problems.

[swill453]

The novelty of shooting down drones or other types of interference with autonomous vehicles will quickly lose their novelty once the novelty of the vehicles themselves passes, and after the law has punished the handful of idiots who do choose to test that law. Few people currently waste their time interfering with existing aspects of our transport or communications infrastructure so there's little reason to expect a persistent upsurge in that behaviour.

You're correct in general, but some specific examples already push beyond the limit of the law, and I'd expect them to get worse once they know they can take advantage of the safety programming. I'm thinking, of course, of those red-light-running cyclists or bike couriers we love to hate. Not all cyclists, I'll quickly add.

It should be interesting to see how we manage this period between limited and eventual full autonomy everywhere: many journeys will involve both autonomous and manual sections, with the former increasing over time. The transition, on-the-fly, between the computer and human piloting presents a lot of safety problems in terms of engaging the human's attention, and over time, ensuring the human's skill levels are up to the job as the periods where they are required to pilot diminish.

This of course means that the time where we can use driverless cars to take us home, drunk, from the pub* is, as you say, "sufficiently distant in time as to be able to ignore it for now".

* - or send the kids off to school in one, or any situation where a licensed driver isn't on board.

Scott.

[DrFfybes]

This thread does seem to have got rather ahead of itself.

Self driving cars are here, and will become widespread in the next 5 - 10 years. However these will be an upgrade of our current cars, much as we have gained radial tyres, disc brakes, radios, seatbelts, heaters, electric windows, ABS, aircon, heated seats, parking sensors, satnav, Bluetooth, road sign recognition, self parking, adaptive cruise, lane assist, brake assist, then the next stage will be cars that self drive in certain circumstances without a requirement for driver input (the next step on from Tesla's Autopilot).

Eventually technology will be there where the cars are completely autonomous, probably not for a couple of decades, but they'll still be driver owned.

However at THAT point, the opportunities open up for driverless HGVs, pool cars, car clubs will start using them as they can self deliver instead of you going to their parking spot. Another few years down the line and carshare or poolcars will become commonplace - 6 seaters that pick a few people up on the way to a common place of employment. Possibly employer funded via tax breaks. They'll text you a with an arrival time based on crowd-sourced travel information.

Eventually the fleets of self driving taxis might exist, but a lot of people seem to be failing to see the difference between driverless and ownerless.

[TopOnePercent]

Until of course the software can distinguish a dickhead from an innocent child.

Not many innocent children wear silly dayglo spandex though, so there's an easy filter.

True 'dat.

[Slarti]

I wouldn't go driverless in cars / trucks / bus until they go driverless on trains ( a very regulated environment). Very unlikely to happen due to the (erroneous IMHO) safety case.

How long has the DLR been running?

And the Newcastle city Metro?


The problem with getting trains to run driverless is more to do with politics than safety.
See the Southern Rail dispute on conductors doing the doors, compared to the Anglia commuter trains that have been running with one man operation for decades.

Slarti

[redsturgeon]

The problem with getting trains to run driverless is more to do with politics than safety.
See the Southern Rail dispute on conductors doing the doors, compared to the Anglia commuter trains that have been running with one man operation for decades.

Slarti

In fact I have the answer to the Southern rail problem. Keep the guards but get rid of the drivers.

John

[Slarti]

Until of course the software can distinguish a dickhead from an innocent child.

Not many innocent children wear silly dayglo spandex though, so there's an easy filter.

I find that the worst dickheads don't wear silly dayglo spandex, they wear black clothes and don't have lights when out on the roads in the dark. similar to many kids.

Slarti

[Slarti]

road sign recognition, self parking, adaptive cruise, lane assist, brake assist

How many cars have those?

I've never even heard of road sign recognition?

Even if those things exist in top end new cars, how long will they take to filter down to that huge majority who drive second hand cars? Especially as 2nd hand car owners tend to keep their cars for far longer than owners of new cars, on average, or so I understand.

Slarti

[YeadonLad]

How long has the DLR been running?

And the Newcastle city Metro?

The problem with getting trains to run driverless is more to do with politics than safety.
See the Southern Rail dispute on conductors doing the doors, compared to the Anglia commuter trains that have been running with one man operation for decades.

Wrong. Driverless trains are only possible on a 'closed system' i.e. where the public can not get easy access. Ever the Docklands Light Railway wouldn't be allowed today as platform doors would be necessary (they can get away with it at present since there is a staff member on every train). In fact the biggest safety hazards on the railways today are foot and road level crossing, which is why Network Rail are trying to get rid of as many as possible.

It would be possible to convert the London Underground since there are no longer any level crossings [1] but it would require massive investment at stations to fit platform doors plus new trains. The Victoria and Waterloo & City lines would be easiest since they are both totally underground (at least for the passenger bit).

The Newcastle Metro has many level crossing and has had a number of accidents over the years.

[1] Following the closure of the Epping to Ongar line.

[TopOnePercent]

Ever the Docklands Light Railway wouldn't be allowed today as platform doors would be necessary (they can get away with it at present since there is a staff member on every train).

There isn't though. Try using it late at night - definitely operates driverless. I lived in the docklands area for a decade so am very familiar with the DLR, then used it to get to work for the following 7 years.

Public access is quite easy, you just step off the platform. The first death was gang related where a young lad was pushed in front of a driverless train, which then ran him over.

[gryffron]

Wrong. Driverless trains are only possible on a 'closed system' i.e. where the public can not get easy access. Ever the Docklands Light Railway wouldn't be allowed today as platform doors would be necessary (they can get away with it at present since there is a staff member on every train). In fact the biggest safety hazards on the railways today are foot and road level crossing, which is why Network Rail are trying to get rid of as many as possible.

It would be possible to convert the London Underground since there are no longer any level crossings [1] but it would require massive investment at stations to fit platform doors plus new trains. The Victoria and Waterloo & City lines would be easiest since they are both totally underground (at least for the passenger bit).

Disagree. The ONLY time a train driver is any use these days is as the train pulls into a station at slow speed. If someone falls/jumps/is pushed in front of the train, the man in the cab just might be able to stop it in time. Cameras could easily takeover this role, but it would require significant infrastructure in the stations. The rest of the time a driver is useless. Even at level crossings. He can't see the track far enough ahead to stop the train. In fact, worse than useless since driver overridden SPADs remain a constant problem.

A guard within the train, looking at/after the passengers, would generally be far more use than a man in the cab. But the unions will not tolerate this change.

London's Victoria line is already 99% automated. The only job of the driver is to press a "go" button to shut the doors and start the train moving. A monkey could do it. Again, a job someone on the platform could do easier. If it were not for the unions insisting on a "driver".

Gryff

[YeadonLad]

There isn't though. Try using it late at night - definitely operates driverless. I lived in the docklands area for a decade so am very familiar with the DLR, then used it to get to work for the following 7 years.

There is always a staff member on board - there has to be to close the door and check that it is safe to start. They don't have to be at the front of the train but they are there.

[YeadonLad]

[/
Disagree. The ONLY time a train driver is any use these days is as the train pulls into a station at slow speed. If someone falls/jumps/is pushed in front of the train, the man in the cab just might be able to stop it in time. Cameras could easily takeover this role, but it would require significant infrastructure in the stations. The rest of the time a driver is useless. Even at level crossings. He can't see the track far enough ahead to stop the train. In fact, worse than useless since driver overridden SPADs remain a constant problem.

A guard within the train, looking at/after the passengers, would generally be far more use than a man in the cab. But the unions will not tolerate this change.

London's Victoria line is already 99% automated. The only job of the driver is to press a "go" button to shut the doors and start the train moving. A monkey could do it. Again, a job someone on the platform could do easier. If it were not for the unions insisting on a "driver".

Gryff

There is a discussion on Driverless Trains on the Rail UK forum. Many of the contributors are professional railway staff - it makes interesting reading but doing it on the main line railway system is rather more difficult than most people realise.
http://www.railforums.co.uk/showthread.php?t=139956

As for the Victoria Line - this line could be converted to fully automatic operation quite easily, just requires platform doors (i.e. as on the Jubilee Line) and modifications to the trains.

It is also planned to convert the Glasgow Underground (or Subway as it's actually called) to driverless in the next few years. http://www.spt.co.uk/subway/modernisation/

[DrFfybes]

I find that the worst dickheads don't wear silly dayglo spandex, they wear black clothes and don't have lights when out on the roads in the dark. similar to many kids.
Slarti

Black appears (or rather doesn't appear) to be the new yellow.

[TopOnePercent]

There isn't though. Try using it late at night - definitely operates driverless. I lived in the docklands area for a decade so am very familiar with the DLR, then used it to get to work for the following 7 years.

There is always a staff member on board - there has to be to close the door and check that it is safe to start. They don't have to be at the front of the train but they are there.

http://www.theregister.co.uk/2007/11/01 ... _platform/

They may usually be there, but they are not required - the train can be operated wholly from the control centre.