Originally Posted by GarryP

Although I get painted as gung-ho about autonomous vehicles, what I'm actually doing is looking at the money and what they were doing in the DARPA challenges over a decade ago. A packed commuter road is a hard scenario, but in the end there are very few options as to actions you can take (turn, speed up, slow down) and the competition is pretty bad at doing these. Beating the human is perfectly possible.

I agree that there will be accidents with automation - but

• the vehicles will be covered in cameras, and every accident will be full recorded, no debates. And I think most will come down to either other human drivers stuffing up, or poor road design (which the big companies will have the lawyers to pin on the councils)
• that will fairly rapidly push to deal with the problem on the roads - human drivers. When the first city bans human drivers and the accident rate falls, it's game over
• those circumstances where the automation screws up will be dealt with quietly and the stats will be fudged
• everyone gets to benefit from autonomous vehicles, except the drivers (who don't have a legal army and are going to get stomped).

Now, maybe the first pilots that roll out next year will be quickly followed by wider rollouts in 2021, or maybe it will be slower to 2025, but either way, it's in the medium term planning horizon (or should be) of every business because it's going to be a landrush. When you look at the ramifications of this, and the opportunities to make money, it's going to be over bar the shouting within the decade.

Originally Posted by Wol

Garry: I can respect alternative thoughts and opinions in this sort of discussion, so I'm not being dogmatic so much as concerned about the way that the proven statistics about crashes due to driver error seem to be put against largely theoretical benefits of autonomous cars.

I've just used driverless trains at two airports in the last week: I have no problem with them. In the same way I would have few concerns with vehicles on dedicated roads/tracks. What does worry me is that for all the hype surrounding autonomous cars, owned by ordinary people (the same ones, presumably, who cause the above crashes), maintained by your average garage mechanics, programmed by the same sort of people who wrote Windows Ten and driven over ordinary roads with all the issues that arise every minute, there's an awful lot of "ifs", "perhaps", "predicted" etc coming into it.

It's not just the safety aspect: we all drive on congested, poorly maintained and confusing roads filled with other fallible drivers. Unlike in the sky, where aircraft have a rigidly controlled environment, strict safety levels, regulated maintenance, hard licencing and so on the political thinking seems to be that firms such as Google and Facebook will automatically be safer than the present. That might - might - be true at present with a few on being tested on specific roads, but I can't see it being so if there are millions of these things driving themselves on our road systems as they are now.

Autonomous driving is arguably far more difficult than airport trains and aircraft operations.

Just my thoughts, anyway.

Originally Posted by Shard

It's a false analogy to compare AV software to Window 10. Even if a glitch were to occur, the software could just pull the car over and shut it down. And is the issue any different that brakes failures that we sometimes have?

Once AV is deployed en masses the vehicles will be communicating with themselves anyway, essentially networked together. There is opportunity to minimise driving risks and isolate faulty vehicles.

Driverless trains were seen as "an accident waiting to happen" when they were introduced 30 years ago, and now, as you say nobody gives them a blind bit of notice.

Originally Posted by Shard

It's a false analogy to compare AV software to Window 10. Even if a glitch were to occur, the software could just pull the car over and shut it down. And is the issue any different that brakes failures that we sometimes have?

Once AV is deployed en masses the vehicles will be communicating with themselves anyway, essentially networked together. There is opportunity to minimise driving risks and isolate faulty vehicles.

Driverless trains were seen as "an accident waiting to happen" when they were introduced 30 years ago, and now, as you say nobody gives them a blind bit of notice.

Originally Posted by Wol

The Windows Ten analogy was a bit tongue in cheek but nonetheless valid, thinking about it. The "most advanced" Windows OS has left innumerable people with huge update problems, BSODs etc etc despite being rolled out with all the hype that marketing is capable of. (I've gone "back" to W7.)

I took part in TRRL Bracknell driving experiments in the sixties, where the scientists were confidently expecting self driving cars in a decade or so. The task is a lot harder than they think, still, IMO.

Originally Posted by Beoz

You keep coming around with this "Windows 10 update problem". You obviously have your own axe to grind. Maybe buy a new PC. I am on my 3rd PC since Windows 10. Never an issue. Maybe it's just you.

Originally Posted by Wol

Yes, sorry about that but I have had hell with W10 over the last couple of years, like millions of others (Google it). Gone back to Seven.

.

Originally Posted by Wol

Having seen how automatic landings went through years of development and trials before being licenced it's inconceivable that cars without steering wheels are going to be on the roads any time soon.

Originally Posted by GarryP

Google have been driving public 'early riders' around, with nobody in the driving seat, for years/months now.

https://www.youtube.com/watch?v=QqRMTWqhwzM

https://www.youtube.com/watch?v=qtApzKnGU94

So, not only conceivable, but pregnant.

Originally Posted by Wol

All very impressive. But licenced on public roads, no manual controls, outside the US? Care to make a guess at when? Not in my lifetime.

Originally Posted by ozzieeagle

China will achieve mass implementation first for sure.

Originally Posted by Beoz

Not so sure on that. Given a driver in China gets about $20 per week there would be very little return on investment to have the machine do it for you.

Originally Posted by ozzieeagle

Plenty of articles along these lines on the net. They are on a mission.

https://www.scmp.com/magazines/post-...ck-take-global

Originally Posted by Amazulu

You party on with your Chinese made autonomous cars

I'll give them a miss thanks

I flew in a Chinese made airliner once - never again!