So when can you get in the first self-driving car? GM says 2019. Mobileye says 2021. Waymo says 2018 – yes, this year After several years of hype about autonomous vehicles – cars that can truly independently drive themselves – the big question has become: when will people other than beta testers get in them? And the answer, according to General Motors' chief technology officer Jon Lauckner is, incredibly, next year: 2019. That's when the …
Don't be so sure. This potential market is not just big, it's f**king huge. This is basically the automotive industry version of what "the cloud" has done for IT infrastructure and everyone wants to be the new AWS.
Of course, there's no way I'm getting in one for a few years, especially the minute one goes wrong and locks on the accelerator at full pelt on the M5 and there's nothing to do but wait until it hits something or runs out of fuel. Or worse it drives full pelt down Regents St into Piccadilly Circus at 50mph.
...and it's pretty impressive when it works. The intelligent cruise control (not to be confused with ProPilot) that keeps your car at a fixed safe distance from the vehicle in front is 100% reliable because it uses radar. But the lane following tech that is based on cameras, can fail, and the car will start to drift to the side before it self corrects when the cameras detect the edge of the road or kerb or other object. I think it gets confused when the solid or dotted white line at the edge of the lane (on the left here in the UK) vanish, perhaps at a junction. It is all very Beta.
One has the advantage of being able to grab the steering wheel and take control, but like you, i will not be comfortable on a motorway or other fast road with no manual override.
No, they just have to work significantly more safely than human driven vehicles, which is actually a pretty low target. In the UK there were 1716 road fatalities in the UK, and 24,101 serious injuries. No automated technology that caused that level of injuries would be acceptable, and it's actually quite hard to see how we could make self driving cars as bad as that. 26000 serious injuries and deaths is 70 every single day.
If we had most cars being self driving I suspect that we could reduce that by 90% or more. For a start 13% of those fatalities involved drivers over the limit.
Data here:
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/648082/rrcgb2016-02.pdf
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/680051/illegal-alcohol-levels-provisional-2016.pdf
No, they just have to work significantly more safely than human driven vehicles, which is actually a pretty low target.
Agreed that AVs don't need to be 100% safe, and that being significantly more safe than humans is a low target. By this measure, AVs may look good even today.
But I wonder if a more important measure is injuries/deaths involving an AV that either would not have occurred were a competent human in control, or are avoided by a human (taking control or in/on another vehicle). I think that needs to be very low indeed, and I suspect is a far greater challenge.
"i will not be comfortable on a motorway or other fast road with no manual override."
Not just on a motorway, what happens when you come to blocked road with a police office work traffic director signally to take a detour or circle around? There could be a sign or transmitter that tells autonomous cars the way forward is blocked, but if an incident just happened, it may be some time before a signalling device can be put in place. Ok, so you just get out that point and cancel your trip but what if you have just done your weekly shop and are in a hurry to get home and get your perishables into the fridge.
How about "bad" neighborhoods? There are sections of the city I don't go even if it's the shortest/fastest route. Will an autonomous car service "redline" areas? I expect they will until they get caught doing it and laws are passed.
This is basically the automotive industry version of what "the cloud" has done for IT infrastructure and everyone wants to be the new AWS.
Lets not forget that one of the features of "the cloud" seems to be massive breaches of personal data left swinging in the breeze in ill-secured cloud backups and the like.
With vehicles it will be unacceptable to leave safe operation* to the customers as the risks to life and limb dwarf the severity of the risks from cloud. There won't be a "new AWS". AWS can shuffle all the responsibility for third party damage onto its customers; vehicle manufacturers won't. It's not just a huge potential market for manufacturers, it's also a hugel risk.
* The opportunities for gathering, mining and subsequently leaking personal data are the same or worse than the cloud but now only a side-issue.
Lifts just go up and down a fixed route which they don't share with any other lifts, cyclists, pedestrians or stray animals.
Same applies to the driverless transports at airports, like Stansted for example. These are nothing more than horizontal lifts. But it's fun when the thing speeds up just before the up ramp.
you guys need to be thinking about going up. Self driving cars... No we should be asking for flying cars. No roads to repair. no pot holes, no gas tax, no speed limits, no traffic jams, no more zig-zag routes to work, less time in transportation, wont rust, won't get stuck.
Flying cars would HAVE to be autonomous with no ability for the passenger(s) to have any control whatsoever over the car other than its destination.
While autonomous driving on the ground still has a ways to go (despite some irrational exuberance on Waymo's and GM's part) before it can go everywhere it needs to go, autonomous flying at low altitude (say 100-200 ft) is much easier so today's technology is already there.
That takes care of the biggest objection to flying cars, but does nothing about the biggest roadblock to them - technology to make a car fly without being very inefficient and noisy. Still, it would certainly tempt people who commute in slowly moving parking lots in many big cities...
No roads to repair - ok
no pot holes - ok
no gas tax - only if they are solar
no speed limits - you think so? I'm doubting it
no traffic jams - until you try to land that is
no more zig-zag routes to work - have you seen flight paths?
less time in transportation - until everyone is doing it
wont rust - I doubt they'd be in the air permanently. Less rust, sure.
won't get stuck - flight levels and paths are a thing.
The system will only work on freeways in the United States and Canada – and will be specifically geo-fenced to make sure it won't work outside of them\
So, driver hits the highway, turns on autopilot, takes nap. Car gets to exit with snoozing driver... and then turns it self off because of the geo-fencing. Or does it ignore the exit, pull over and shut down? I'd hope that it just doesn't keep going.
1. The [GM] system will only work on freeways...
2. Glance away any longer that five seconds and [... the car] will start slowing to a stop.
On a freeway? What can possibly go wrong?
Slowing to a stop may well be the right solution if the assumption is that the driver - er... occupant - is dead or incapacitated. Chances are though that, just like the Uber beta-tester, he/she is simply not paying attention.
I don't believe the LEDs and the buzzes will help avoid the slowing down. Lots of cars around me have MobilEye. In my experience, with a decent driver behind the wheel the alarms are close to a 100% false positives[*]. For certain there are hordes of dimwits around who should be jolted into some semblance of attention quite often, but people who have MobilEye tell me that the low S/N makes the mind learn to thoroughly ignore the alarms very quickly indeed. [Frankly, I have general doubts about a gadget that says, effectively, that it's OK to drive when you are tired or distracted because it'll warn you, but that's a separate issue.]
So, a car carefully slows down on a freeway every now and then. The dimwit notices and starts looking straight ahead and assume the car starts and accelerates to the allowed speed again. That particular dimwit stays alert and away from the funny cat movie on his phone for a while. All is well...
Assume the best case scenario. Nothing goes really wrong in any of these cases, ever. Say, because every other car will have ACC and ADAS and everything else and will also slow down and stop without any chance of a prang whatsoever. But can you imagine what those slowing down and stopping cars - and that will happen - will do to congestion on freeways?
[*] One of our cars is equipped with MobilEye (the car I usually drive, mercifully, isn't - it does have a fatigue detection system though which has never been activated, to my knowledge) that was thrown in for free. Sometimes I am a passenger, sometimes I drive it. The bloody thing beeps most often when there is already a tense situation and you move a bit sideways without indicating - no time, but normally everyone around sees/senses that something is amiss - and then the system that looks for lane separators beeps and distracts and alarms and disorients you when you need your concentration the most ("What else has just happened?!?! Ah, nothing..."). It also knows nothing about what the driver is actually doing so it beeps, early, when the distance to the car in front of you decreases, even if you are already braking (again, you are already tense since you've noticed the danger, the alarm only makes you think that something else is going wrong). It always gets confused if there is more than one set of lane markings (roadworks), etc.
My Madza 6 with 'lane assist' beeps at me constantly, telling me to hold the steering wheel, when it is firmly gripped with two hands. Apparently if I drive precisely as it thinks I should , it thinks that it is it the one doing the driving and gets worried. You get into the the habit of wiggling the wheel every now and them to let it know your only human!
> the big question has become: when will people other than beta testers get in them?
I want to know when they will be affordable for an average guy, like me?
If the first-generation AVs - ones that aren't death-traps: either for occupants or third parties, are going to be in the £ 6-figures, then they may as well not exist. But when they are at a price that is comparable to standard new models now then they become a viable option.
However, I still reckon that the financial model for domestic AVs is one of on-demand hailing. What is the point of buying such an expensive object, that depreciates faster than you can burn £50 notes and that is only used for a small percentage of time. Just so long as the previous user of an AV I call up hasn't puked in it.
"I want to know when they will be affordable for an average guy, like me?"
I doubt they'll ever be affordable to buy.
The legislative basis for their use, at least in the UK, seems, quite rightly, to put the legal responsibility for safe driving on the manufacturer. That means that the manufacturer rather than the owner will have to insure themselves. The manufacturer will, of course, pass this on to the customer. In the event of a straight sale, however, the manufacturer will only be able to have one opportunity to do that so would need to charge the customer for the vehicle's life-time insurance as part of the purchase price. That would substantially increase the price of a new car. The likelihood is that these vehicles will only ever be available for lease.
"What is the point of buying such an expensive object, that depreciates faster than you can burn £50"
To have one available when you need it. If your prime use is in the rush hour when everyone else wants a ride you'll be in competition with everyone else. If the numbers of available vehicles are such that peak demand is adequately covered they'll be mostly idle during the day and the costs per mile will go up to allow for that. If you have your own car now you'll still need your own AV. If you can manage by taxi now you'll use and AV taxi.
I want to know when they will be affordable for an average guy, like me?
Why would you want to buy one, when you can call one up at the snap of your fingers? Or even better get one to take little Timmy swimming and bring him back.
The economics for manufacturers of higher yields are compelling. The vast majority of cars are stationary for over 90% of the day and this is a huge drag on the economy because it means providing huge areas of concrete for the tin cans to wait on. I suspect we're likely to see charges for parking spaces to start rising in anticipation of this and this will favour cars that don't spend all day in the same place and/or have additional sources of revenue.
The vast majority of cars are stationary for over 90% of the day and this is a huge drag on the economy because it means providing huge areas of concrete for the tin cans to wait on.
Yeah, and the vast majority of cars are also moving during the same period of the day. So when you take into account those shared vehicles shuttling around all over the place unoccupied to their next customer, the amount of tarmac & concrete required is likely to be greater, not less.
"when will people other than beta testers get in them?"
When they do the correct term will be "guinea pigs". The unfortunate aspect of this is that while the guinea pigs who get into the cars will be volunteers those in the surrounding traffic or on foot will be innocent bystanders.
I've made the point before but it still needs reiterating: compensation for death, injury or damage to innocent bystanders and their property should not rely on them having to take on manufacturer or insurance funded lawyers in court.
maybe in certain environments but not down in my part of the World (sunny Cornwall).
My mainly country road commute of 25 miles negotiating very narrow village streets only just wide enough for a car (sets the bloody warning sensors off lots of the time as the car thinks your going to hit something) full of gormless emmets stepping out in front of you. Then country roads with badgers, pheasants the occasional peacock and the odd sheep in the road, pot holes, flooding, mud from tractors, emmets driving at 25mph, caravan, etc all need to be safely negotiated. Then in to Plymouth via the Torpoint ferry that'll tax any autonomous system getting on\off that fecking thing!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
