Would you go driverless?

[YeadonLad]

For city folk it may well be a better model to have a car turn up only when required, but I can't see it catching on in much of the country where things are just not as close to each other and you'd have to think well ahead to call for the car.

Slarti

It would have to be a very clever car to reach a friends house on its own - it would have to work out how to open the gates to get there (and they are on a public road not his land).

My view is that making them do 90% of the job will take 10% of the effort. The last 10% will take 90% of the effort (and take years). There are still too many unanswered questions and I can't see how they would work without some user action. Take for instance a car park at a site I visit frequently - the surface is gravel with no marking so how would it find the required spot to park:
1) When I visit for a meeting - plenty of space so no need to block anyone in.
2) When there is an event on so we double/triple park so there is room for visitors to park.
3) As 2 but when all the available staff parking has been used so I have to park down the road.
4) When I have things to unload so I need to reverse into a disabled parking spot (since they are outside the door).

[redsturgeon]

For city folk it may well be a better model to have a car turn up only when required, but I can't see it catching on in much of the country where things are just not as close to each other and you'd have to think well ahead to call for the car.

Slarti

It would have to be a very clever car to reach a friends house on its own - it would have to work out how to open the gates to get there (and they are on a public road not his land).

My view is that making them do 90% of the job will take 10% of the effort. The last 10% will take 90% of the effort (and take years). There are still too many unanswered questions and I can't see how they would work without some user action. Take for instance a car park at a site I visit frequently - the surface is gravel with no marking so how would it find the required spot to park:
1) When I visit for a meeting - plenty of space so no need to block anyone in.
2) When there is an event on so we double/triple park so there is room for visitors to park.
3) As 2 but when all the available staff parking has been used so I have to park down the road.
4) When I have things to unload so I need to reverse into a disabled parking spot (since they are outside the door).

Either your friend would remove the gate or install an automatic gate, or decide to meet the car at the gate.

The car would not need lines to park, we already have cars that can parallel park by referencing the other cars around it.

The cars would not need to stay, they can drive off to deliver someone else to another destination and come back when needed.

There will need to be special instructions for unloading, again you can be delivered to the door, unload your stuff and the car drives off.

Nothing insurmountable although clearly some new infrastructure and rejigging of existing infrastructure will be needed. In the last few hundred years the challenges of of each new transport technology were greater than that required for the nest stage to driverless cars.

Canals were dug across the length and breadth of the UK, a huge engineering feat requiring viaducts, tunnels and lots of digging without the use of machines.

Ditto for the railways with millions of miles of special tracks and signalling being built and stations to handle the passengers while dealing with hitherto unheard of speeds at which it was thought people would die.

Ditto for cars with the added problems of autonomous drivers each in control of more than a ton of deadly machinery that in the wrong hands, either through incompetence or murderous intent could kill dozens of people.

Air travel has required huge airports and associated infrastructure to be built, sometimes requiring new islands to be reclaimed from the sea, along with the technology to safety move millions of passengers in and out of them, basically on flying bombs.

The problems associated with the next step to autonomous vehicles are trivial by comparison and are mostly social rather than technological in nature.

Many here have attempted to come up with problems or situations where the driverless car will not be as convenient for the individual as having a normal vehicle. No doubt similar arguments were used by horse owners to tell us the many reason why cars would never replace the horse.

"Cars can't jump fences, how could they ever reach my house like old Bess can"

"Car need petrol every few miles, where would we store all that stuff, there's nowhere to fill up between here and Truro!"

"What happens when I get a puncture, or the car breaks down, the local vet or blacksmith won't be able to fix it"

"I love my horse, no car will ever take its place"

The issue is the same, for many individuals, the motor car as we know it has not been a positive invention and one that comes with a great many inconveniences however for society as a whole it has been a boon.

Driverless cars are coming and sooner than many here think is possible.

John

[swill453]

Driverless cars are coming and sooner than many here think is possible.

I still don't believe it (the "sooner" bit), so we'll have to wait and see.

In your long post you didn't even refer to what in my opinion is the biggest hurdle - the massively difficult technical challenge of programming cars to cope with the unpredictable, random nature of other road users and road conditions.

Scott.

[redsturgeon]

Driverless cars are coming and sooner than many here think is possible.

I still don't believe it (the "sooner" bit), so we'll have to wait and see.

In your long post you didn't even refer to what in my opinion is the biggest hurdle - the massively difficult technical challenge of programming cars to cope with the unpredictable, random nature of other road users and road conditions.

Scott.

That's already been done and being refined. The computer will react much more quickly, more predictably and more reliably in any situation than human drivers who may be 17 or 77, wide awake or half asleep, talking on the phone, eating etc etc.

In some situations that reaction may be inconvenient, ie. the car may just come to a halt in an unforeseen situation. It may end up being a bit like the Asimov I Robot stories where the algorithm in use may result in the car circling a stray cat at a predetermined distance for ever but don't doubt that it can and will be done.

I have already pointed out some of the potential problems with pedestrians knowing that the car will always stop if they step in front of it but I guess new laws on jaywalking could be brought in to stop that.

John

[swill453]

That's already been done and being refined. The computer will react much more quickly, more predictably and more reliably in any situation than human drivers who may be 17 or 77, wide awake or half asleep, talking on the phone, eating etc etc.

I don't believe it's "been done" to anything like an acceptable level, and the "refining" will take years.

In some situations that reaction may be inconvenient, ie. the car may just come to a halt in an unforeseen situation. It may end up being a bit like the Asimov I Robot stories where the algorithm in use may result in the car circling a stray cat at a predetermined distance for ever but don't doubt that it can and will be done.

I don't doubt it either, we just (hugely) disagree on the timescale.

Scott.

[bungeejumper]

It may end up being a bit like the Asimov I Robot stories where the algorithm in use may result in the car circling a stray cat at a predetermined distance for ever but don't doubt that it can and will be done.

Three self-driving cars at a mini roundabout. Each of them saying "no, I insist, after you" to the other two. A scene from daily life in our local town, of course, except that in the real world somebody eventually says "soddit, I'm going first", and the deadlock is broken.

In the robo-world, I suppose, the three cars will need to play scissors-paper-stone to decide which of them gets priority?

BJ (still playing devil's advocate, mostly for the fun of it)

[redsturgeon]

It may end up being a bit like the Asimov I Robot stories where the algorithm in use may result in the car circling a stray cat at a predetermined distance for ever but don't doubt that it can and will be done.

Three self-driving cars at a mini roundabout. Each of them saying "no, I insist, after you" to the other two. A scene from daily life in our local town, of course, except that in the real world somebody eventually says "soddit, I'm going first", and the deadlock is broken.

In the robo-world, I suppose, the three cars will need to play scissors-paper-stone to decide which of them gets priority?

BJ (still playing devil's advocate, mostly for the fun of it)

Of course in our algorithmic future this will all have been decided down to your like/dislike ratio on twitfacetagram with the big hitters having a clear passage in their AutoRollsMaybach while the proles wait in their dirty little NuTrabants.

Also having fun pooh-poohing the luddites.

John

[swill453]

Three self-driving cars at a mini roundabout. Each of them saying "no, I insist, after you" to the other two. A scene from daily life in our local town

Solved ages ago for Ethernet with CSMA/CD (carrier sense multiple access/collision detect).

Scott.

[Meatyfool]

To all those who say that having a fleet of autonomous-cars-to-hire will not work in "the sticks"/"my neck of the wood"/out in the country", would you like to explain the reason why?

This is why from where I am sitting. Most people have a car because an suitable alternative doesn't exist. Period.

If autonomous cars are introduced to maximise their environmental benefit, no-one will want to own a car, autonomous or otherwise. Do you honestly think with zero or practically zero ownership of personal transport, the "to hire" companies are going to ignore these areas?

Far from! With a per mile hire price, they are going to make more money from rural to urban hires than they are for itty bitty rides in town where more "voids" will occur.

Putting vehicles where they are needed using known history of hires is not anything new - ambulances haven't been parked in ambulance stations for years - they go wherever the need to so that they can arrive at the next call out as quickly as possible.

Don't forget, the autonomous vehicle doesn't have to be a car. If 30 people in a village all book a Mon-Fri return journey to the nearby town, the vehicle turning up at their door will not be a car.

Finally, what is the most expensive part of a bus route - the driver!

Meatyfool..

[Hallucigenia]

As other have said - just as with electric cars, autonomy will bring advantages to different groups of people at different rates, and you'll see different rates of uptake in different areas - just as you do with different types of car today. But the people for whom it _does_ work provide the cashflow to allow R&D for the more difficult cases - it doesn't matter if a technology doesn't work for some people, it's the fact that it does work for others that enables a technology to develop and expand.

Mobile phones are crap in many rural areas - but that doesn't prevent mobile phones being a useful technology for many people.

And to be honest, things like getting home from the pub and transport for the elderly loom larger for rural dwellers than townies.

The three cars at a mini-roundabout only occurs because of human incompetence - the first to arrive is a cautious old lady. Self-driving cars would know to the millisecond who got there first, and could communicate among themselves. But there are other cases that have proved more challenging, and the likes of Google are already working on them. Joining a busy motorway needs a bit more aggression (and marginal law-breaking) than the cars had to start with, but they're learning where to draw the boundaries in that grey area.

"the massively difficult technical challenge of programming cars to cope with the unpredictable, random nature of other road users and road conditions"

It's difficult - but not insurmountable, certainly not to get them "significantly better than human drivers". 0% accident rate is not achievable, but at least they won't drive drunk, drugged, tired or while trying to placate screaming kids in the back. They're already on the roads, encountering those unpredictable conditions - and the fact that there hasn't been more accidents suggests current technology is already not at 10% accident rate or 1% accident rate, but a fraction of that.

It's been suggested that half the lifetime carbon emissions from a car come from building it, never mind the other environmental effects of digging up metal ores etc : https://www.theguardian.com/environment ... nt-new-car

Obviously the balance tips again with electric vehicles, but things like batteries are relatively recyclable. You also have to consider the effect of rationalising all the space used for parking cars in cities.

[swill453]

Also having fun pooh-poohing the luddites.

A Luddite wouldn't want it to happen, I just don't think it'll happen* (any time soon**).

* - by this I mean a driverless car coping with an average British high street.
** - by this I mean any time less than 10 years for the first car that can cope with the average British high street.

Scott.

[redsturgeon]

Also having fun pooh-poohing the luddites.

A Luddite wouldn't want it to happen, I just don't think it'll happen* (any time soon**).

* - by this I mean a driverless car coping with an average British high street.
** - by this I mean any time less than 10 years for the first car that can cope with the average British high street.

Scott.

No offence meant, just a slightly tongue in cheek response. I think by 2027 autonomous cars will be on Britain's high streets.

John

[swill453]

I think by 2027 autonomous cars will be on Britain's high streets.

So earlier in the thread I wrote

I'm talking getting on for 10 years here.

Yes, I'd think that's a reasonable estimate for when we'll start to see many driverless cars on (some of) our roads.

Add another couple of decades before they become anything like the majority, and we start to see some of the significant social changes people are talking about here.

I'm sticking to that, but maybe we're not a million miles (years) apart.

Scott.

[Slarti]

It would have to be a very clever car to reach a friends house on its own - it would have to work out how to open the gates to get there (and they are on a public road not his land).

Either your friend would remove the gate or install an automatic gate, or decide to meet the car at the gate.[/quote]

As it is a gate on a public road, that would not be an option for the friend.

Such gates, while not massively common, are not that uncommon in country areas.

Slarti

[redsturgeon]

It would have to be a very clever car to reach a friends house on its own - it would have to work out how to open the gates to get there (and they are on a public road not his land).

Either your friend would remove the gate or install an automatic gate, or decide to meet the car at the gate.

As it is a gate on a public road, that would not be an option for the friend.

Such gates, while not massively common, are not that uncommon in country areas.

Slarti[/quote]

Aha, sorry I missed that "on a public road " bit. That's even better then, fixing that problem will be paid for by the council.

John

[Slarti]

I have already pointed out some of the potential problems with pedestrians knowing that the car will always stop if they step in front of it but I guess new laws on jaywalking could be brought in to stop that.

They've already had problems in Sweden with "drivers" bullying test autonomous cars and so have had to stop marking them.

Slarti

[Slarti]

As it is a gate on a public road, that would not be an option for the friend.

Such gates, while not massively common, are not that uncommon in country areas.

Slarti

Aha, sorry I missed that "on a public road " bit. That's even better then, fixing that problem will be paid for by the council.[/quote]

Somehow I doubt it as they've often survived for well over 100 years separating the lands of Liz and her relatives.

Slarti

[Slarti]

If autonomous cars are introduced to maximise their environmental benefit

If you think that is the reason for introducing them, you are sadly mistaken.
Profit, pure and simple and those of us out in the sticks just won't have an autonomous vehicle within range when they want it. Certainly not at a sensible price.

Putting vehicles where they are needed using known history of hires is not anything new - ambulances haven't been parked in ambulance stations for years - they go wherever the need to so that they can arrive at the next call out as quickly as possible.

That's why it can take 4 times as long for an ambulance to get to somebody in the country as the cities.

Don't forget, the autonomous vehicle doesn't have to be a car. If 30 people in a village all book a Mon-Fri return journey to the nearby town, the vehicle turning up at their door will not be a car.

Except that the only reason people use buses is because they can't afford a car and those of us who can afford one, don't want to share or, unreasonably, want to go somewhere else.

On Thursday I had to take my son to Queens Hospital Romford. Forty minutes each way.
Public transport is 3 hours each way. If you're lucky. That won't change with an autonomous bus. especially if you have to travel at inconvenient times.

Slarti

[bungeejumper]

Except that the only reason people use buses is because they can't afford a car and those of us who can afford one, don't want to share or, unreasonably, want to go somewhere else.

On Thursday I had to take my son to Queens Hospital Romford. Forty minutes each way.
Public transport is 3 hours each way. If you're lucky. That won't change with an autonomous bus. especially if you have to travel at inconvenient times.

Oh, buses, Just the job for the journey back from Sainsburys with sixty pounds of shopping in the bags. (Most of which will be wine, beer or water, if I'm honest.) I guess that home deliveries will solve that one, at least.

But, as Slarti says, it would probably surprise a lot of people how long it can take us stick-dwellers to get from one place to another via public transport. Our second-nearest small town is six miles away, or two hours and three buses, always supposing that they turn up in the first place. And yes, ambulances don't hang around in the country lanes waiting for customers - mostly they're in the towns, queueing up outside the A&E departments.

Despite my reservations, I reckon the self-driving car will win through in the end. But I'm expecting it to be an expensive business for those of us who live in the kinds of places where we've already had to resort to the cudgels to get half-decent broadband. It's just too easy for the providers to go suddenly deaf when basic services to 'uneconomic areas' are required. I don't expect that will change, somehow.

BJ

[9873210]

If the on call fleet is a million cars (something like a 96% reduction compared to the current licensed vehicles), and they are distributed roughly by population then at any one time there will be 14 in Bradwell on Sea. That's enough that the on call system will work roughly as well there as in central London. Excepting the couple of times a year when half the village wants to go to the same wedding. Then you'll just have to hope the dispatching program reads the banns.

This is quite different from dispatching the 10,000 or so ambulances in the UK where there is on average only 1/7 of an ambulance serving Bradwell on Sea.

So while there might be problems in truly rural places, for these purposes there are no rural places in Essex, and are probably no rural places in England.