Driverless Cars- What do we think
Discussion
W124 said:
I'm not worried about the tech side. I believe it already works perfectly - and has done for a while.
You have to love PR 
The google cars have been running in a small area which has been extensivly 3D mapped and as is near perfect and stable conditions for a driver less car as possibleGet a few to cover all of london, the first AL taxi it comes across will have it into a wall
Edited by Dave Hedgehog on Thursday 16th April 13:05
W124 said:
People won't need to own cars. Dial one up and there it is. Order one for the way back and there is is. School run? There it is.
You can do that now!I can certainly see autonomous cars making taxi drivers redundant, but I think people will still continue to want to own them. For many people, car ownership is about more than driving or even about having to wait for a cab to turn up. For many it's a personal space and a status symbol.
I like driving, and would lament not being able to.
That said, I like changing gear too, but that doesn't mean I don't sometimes wish I had an auto. Sometimes even motor enthusiasts are just trying to get somewhere.
I think "driverless" cars are coming, ultimately, but it will be more like auto pilot, rather than the concepts people produce of sitting in the back having a kip, or business meetings on the move.
Being able to push the "oh, drive yourself car, I can't be bothered" button on a 200 mile slog up the motorway sounds brilliant to me though. Pretty sure some cars already come with lane correct, auto park, radar cruise, so presumably would be able to keep in lane, at a set distance from the car in front, without any input from the driver.
That said, I like changing gear too, but that doesn't mean I don't sometimes wish I had an auto. Sometimes even motor enthusiasts are just trying to get somewhere.
I think "driverless" cars are coming, ultimately, but it will be more like auto pilot, rather than the concepts people produce of sitting in the back having a kip, or business meetings on the move.
Being able to push the "oh, drive yourself car, I can't be bothered" button on a 200 mile slog up the motorway sounds brilliant to me though. Pretty sure some cars already come with lane correct, auto park, radar cruise, so presumably would be able to keep in lane, at a set distance from the car in front, without any input from the driver.
Not especially interested in them if I'm honest, as like many I enjoy driving too much. Even drudging down a motorway, or being stuck in traffic, is more enjoyable and satisfying from the driver's seat than the passenger's.
Can't say I'm too convinced by the technology yet either. Ignoring the fact that self driving cars all look like larger versions of the Little Tykes car I had as a kid, all the experiments so far seem to be in the US, with it's straighter roads and intersections (and closed Google Campus as mentioned). However bring them over to Europe, and ask them to navigate round, say, the Arc de Triomphe, the Magic Roundabout in Swindon, London in general or Stafford (the worst designed road system in the history of mankind) and I can see them just sitting there like Clarkson and May's impression of Peugeot drivers trying to figure out what the f
k they are supposed to do next.
Can't say I'm too convinced by the technology yet either. Ignoring the fact that self driving cars all look like larger versions of the Little Tykes car I had as a kid, all the experiments so far seem to be in the US, with it's straighter roads and intersections (and closed Google Campus as mentioned). However bring them over to Europe, and ask them to navigate round, say, the Arc de Triomphe, the Magic Roundabout in Swindon, London in general or Stafford (the worst designed road system in the history of mankind) and I can see them just sitting there like Clarkson and May's impression of Peugeot drivers trying to figure out what the f
k they are supposed to do next. Daston said:
I worry about their ability to detect danger. For example, we see parked cars down a residential street on a sunday when the sun is out. We know through experience that there is a high chance a kid may pop out from between the cars so we are alert and driving accordingly.
Does the computer know this? Or does it just go at 30 because that is all it knows.
Plus if these cars are so clever and safe why does the google car look like a bumper car with its rubber front and pedestrian air bags etc? Shirley if it was safer than a normal car it wouldn't need all that.
Why wouldn't it know that? You've described the way to identify such a situation and the correct response. No reason a computer couldn't follow that logic. And the manufacturers have a very strong incentive to ensure they do.Does the computer know this? Or does it just go at 30 because that is all it knows.
Plus if these cars are so clever and safe why does the google car look like a bumper car with its rubber front and pedestrian air bags etc? Shirley if it was safer than a normal car it wouldn't need all that.
The economics of the two approaches are: you can either spend, what, sixty man hours per person teaching someone to drive (including both their time and the instructor's)? After which they are still, objectively, pretty s
t. They'll probably never be more than, say, 90% as good as the best driver. Many will go through at 30, or more, in that situation, maybe because they're thick, maybe because they just don'e care. Or you can employ people to create the best driver they can... and copy that driver into every single car with the only per-unit cost being the extra sensors (and if there's one guess about technology safer than just about any right now, it's that sensors will get smaller, better and cheaper).The only worries are: a)they'll be basically like android handsets in terms of needing active manufacturer support. Will you even be allowed to continue using a model with a known bug the manufacturer refuses to patch? And b)how to deal with situations with early cars that will need to give up and cede control to the human occasionally. Under fully autonomous control a software failure would be treated the same as a hardware failure - that is to say, manufacturer is clearly liable - but if the car warns the driver human input is necessary and the human f
ks up, at what point can they be expected to be liable?W124 said:
I'm not worried about the tech side. I believe it already works perfectly - and has done for a while.
Yeah, right until it drives off a cliff avoiding what it perceives to be an obstacle on the road, but is actually a small bug that just got splatted against one of the sensors.Mafffew said:
There are a fair few benefits. However, personally I would never have one. Despite how stressful it can be, I love driving.
There is also the question of, what if the systems go wrong and they do have an accident. Who is at fault, man in the car or manufacturer?
I don't think you'd ever own one. I don't think anyone would own one.There is also the question of, what if the systems go wrong and they do have an accident. Who is at fault, man in the car or manufacturer?
Owning a car will no longer be the norm. You'll simply order a car on your smart phone (Uber will eventually do this i bet). Your driver-less car will turn up, and off you go.
It's crazy really, that most cars aren't doing anything for 95% of the time. This will eliminate that problem all together.
Some of us will still have cars (for fun), but most wont mourn their less when they realise how convenient just ordering one is.
W124 said:
People won't need to own cars. Dial one up and there it is. Order one for the way back and there is is. School run? There it is. Whoever gets it working first is going to have a model T on their hands.
But surely this will mean that it will be like rush hour on the roads all day, with empty cars driving around constantly to pick people up.MitchT said:
Yeah, right until it drives off a cliff avoiding what it perceives to be an obstacle on the road, but is actually a small bug that just got splatted against one of the sensors.
The sensors are cleverer than this. And would have double/triple redundancy. If one sensor fails, you won't die instantly. Actually, the car would detect, it and book itself in to be fixed.Howard- said:
Daston said:
I worry about their ability to detect danger. For example, we see parked cars down a residential street on a sunday when the sun is out. We know through experience that there is a high chance a kid may pop out from between the cars so we are alert and driving accordingly.
Does the computer know this? Or does it just go at 30 because that is all it knows.
.
This is my concern. There are a million and one little subconscious decisions we make based on our surroundings when we drive. How can a driverless car possibly replicate that same level of thought process / decision making? Does the computer know this? Or does it just go at 30 because that is all it knows.
.
http://www.telegraph.co.uk/motoring/11382073/Roadt...
Saying
"In all of their trials - and they have clocked up nearly a million miles - there has only been one accident, when a car was rear-ended while the vehicle was in manual drive, and that was down to human error."
So it's not like it's just cruise control with an autobrake and autosteer function, it will drive the car and react to things around it.
I'd get one for a boring commute, but not for a route that I would actually want to drive!
walsh said:
I think "driverless" cars are coming, ultimately, but it will be more like auto pilot, rather than the concepts people produce of sitting in the back having a kip, or business meetings on the move.
Being able to push the "oh, drive yourself car, I can't be bothered" button on a 200 mile slog up the motorway sounds brilliant to me though. Pretty sure some cars already come with lane correct, auto park, radar cruise, so presumably would be able to keep in lane, at a set distance from the car in front, without any input from the driver.
I think there will be a gradual transition via part time autonomous operation.Being able to push the "oh, drive yourself car, I can't be bothered" button on a 200 mile slog up the motorway sounds brilliant to me though. Pretty sure some cars already come with lane correct, auto park, radar cruise, so presumably would be able to keep in lane, at a set distance from the car in front, without any input from the driver.
witko999 said:
But surely this will mean that it will be like rush hour on the roads all day, with empty cars driving around constantly to pick people up.
Not really, they'd regulate the number of cars on the road for the demand. We already do this with webservers (turn some off during off-peak periods). There would be some driving around, but many would be parked up waiting for demand to pick up.I understand some of the concerns, but I don't think they're insurmountable.
If I can find it, I'll post it, but read a report from someone who had been driven by one in California, actual road. It was dark, there was no-one behind them and the car came to a stop. He had enough time to think "why?" before a deer ran out from behind bushes and across the road. Seems the car has detected it, anticipated it was a possible collision but that there was no risk in stopping, so it did. You'd never have seen the deer.
Reality is, the systems will probably pick up on all the slight cues we do (feet under parked cars, cars that have just stopped etc) and probably is able to act on them even faster than our experienced unconscious. That seems to me like the inevitable consequence of bare minimum safety, coupled to something capable of making thousands of calculations a second.
Being able to program in a trip through Birmingham and relax before a work meeting or hospital appointment will sell me on one. The ethics are interesting though, very rarely people do encounter situations where a collision with something is unavoidable. We likely don't really consciously make those decisions in cars, even if we later rationalise them, it's probably all too fast. Someone will be making them in anticipation with driverless cars.
If I can find it, I'll post it, but read a report from someone who had been driven by one in California, actual road. It was dark, there was no-one behind them and the car came to a stop. He had enough time to think "why?" before a deer ran out from behind bushes and across the road. Seems the car has detected it, anticipated it was a possible collision but that there was no risk in stopping, so it did. You'd never have seen the deer.
Reality is, the systems will probably pick up on all the slight cues we do (feet under parked cars, cars that have just stopped etc) and probably is able to act on them even faster than our experienced unconscious. That seems to me like the inevitable consequence of bare minimum safety, coupled to something capable of making thousands of calculations a second.
Being able to program in a trip through Birmingham and relax before a work meeting or hospital appointment will sell me on one. The ethics are interesting though, very rarely people do encounter situations where a collision with something is unavoidable. We likely don't really consciously make those decisions in cars, even if we later rationalise them, it's probably all too fast. Someone will be making them in anticipation with driverless cars.
framerateuk said:
The sensors are cleverer than this. And would have double/triple redundancy. If one sensor fails, you won't die instantly. Actually, the car would detect, it and book itself in to be fixed.
It might not kill you but it strikes me as a lot of complex tech that'll kill your wallet. I have a lovely old car. Changing a headlamp bulb takes a couple of minutes and costs about a quid. I've heard stories of modern cars where half the front has to come off just to change a fecking bulb. This will be even worse, and as it's all tech rather than mechanics it'll all be concealed behind a main dealer firewall enabling them to charge whatever they want because there's no competition. Meanwhile your every move will be logged by one of millions of new civil servants employed for no other purpose, and taxes will go through the roof to pay for them all.MitchT said:
It might not kill you but it strikes me as a lot of complex tech that'll kill your wallet. I have a lovely old car. Changing a headlamp bulb takes a couple of minutes and costs about a quid. I've heard stories of modern cars where half the front has to come off just to change a fecking bulb. This will be even worse, and as it's all tech rather than mechanics it'll all be concealed behind a main dealer firewall enabling them to charge whatever they want because there's no competition. Meanwhile your every move will be logged by one of millions of new civil servants employed for no other purpose, and taxes will go through the roof to pay for them all.
You missed my other post. YOU won't own these. You'll order them like a taxi. Owning a car will no longer be the norm.Ahimoth said:
Reality is, the systems will probably pick up on all the slight cues we do (feet under parked cars, cars that have just stopped etc) and probably is able to act on them even faster than our experienced unconscious. That seems to me like the inevitable consequence of bare minimum safety, coupled to something capable of making thousands of calculations a second.
And also having a considerably enhanced array of senses compared to ours.Gassing Station | General Gassing | Top of Page | What's New | My Stuff