Will driverless cars save lives? (more than 130 collisions)
Discussion
ATG said:
FMOB said:
Everything the human race has achieved or discovered has been done before by mother nature all except software.
Does mother nature make teapots? Has mother nature made any stainless steel recently? And I was referring to underlying fundamentals, a bit of code gets written but the behaviour of said bit of code is not 100% predictable. As an example take a flame and set light to a bit of wood, the chemical process of combustion is completely predictable and can be repeated as it has been through the centuries.
If software were 100% reliable it wouldn't need watchdog timers or heartbeat signals and other mitigations for failure modes to keep it working.
FMOB said:
Of all the lives saved, how can tell that the saving bit is down the driver not reacting at all and the system intervenes and gets a point for really saving a life and those where the system reacted first but if the system didn't intervene the driver would have.
All you can say is the safety system reacted before the driver (hardly a surprise) but cannot say the safety system saved a life everytime it reacted because you just down know, this is just misusing faulty statistics.
What is your proposal to create accurate statistics.All you can say is the safety system reacted before the driver (hardly a surprise) but cannot say the safety system saved a life everytime it reacted because you just down know, this is just misusing faulty statistics.
Perhaps install mandatory 'cannot be turned off' LKA on 50% of new cars and not on the other 50% and to ensure people cannot argue they are "faulty statistics" that you don't know which version you will get when you buy it and you can't sell the car for three years if it is the one you don't want so you are forced to drive it.
And then sit back and see which cars kill the fewest people.
PF62 said:
FMOB said:
Of all the lives saved, how can tell that the saving bit is down the driver not reacting at all and the system intervenes and gets a point for really saving a life and those where the system reacted first but if the system didn't intervene the driver would have.
All you can say is the safety system reacted before the driver (hardly a surprise) but cannot say the safety system saved a life everytime it reacted because you just down know, this is just misusing faulty statistics.
What is your proposal to create accurate statistics.All you can say is the safety system reacted before the driver (hardly a surprise) but cannot say the safety system saved a life everytime it reacted because you just down know, this is just misusing faulty statistics.
Perhaps install mandatory 'cannot be turned off' LKA on 50% of new cars and not on the other 50% and to ensure people cannot argue they are "faulty statistics" that you don't know which version you will get when you buy it and you can't sell the car for three years if it is the one you don't want so you are forced to drive it.
And then sit back and see which cars kill the fewest people.
FMOB said:
Maybe acknowledge to yourself you are using dodgy statistics rather than quoting as gospel something that has more potholes in it than your average bit of tarmac.
And yet you have absolutely no statistics whatsoever, dodgy or not, to back up your views, so perhaps you might acknowledge that turning off LKA every time you get in the car isn't a sensible thing to do.Admittedly this is old but still interesting nonetheless.
Waymo simulated fatal collisions that happened in the past and simulated what would have occurred had their 'Waymo Driver' been involved. I haven't read the full paper but this is the headline they publish.
https://waymo.com/blog/2021/03/replaying-real-life...
'In total, the simulated Waymo Driver completely avoided or mitigated 100% of crashes aside from the crashes in which it was struck from behind, including every instance that involved a pedestrian or cyclist (20 simulations in total). This is the first time an autonomous technology company has shared its evaluation for how the system might perform in real-world fatal crash scenarios.'
I accept this may be some PR / spin going on but I think this could be a way to demonstrate safety to people to make them more accepting. It's not exactly showing how many lives it has saved but does potentially demonstrate it in a way that is easier to perceive than 'x times safer than a human driver'
Waymo simulated fatal collisions that happened in the past and simulated what would have occurred had their 'Waymo Driver' been involved. I haven't read the full paper but this is the headline they publish.
https://waymo.com/blog/2021/03/replaying-real-life...
'In total, the simulated Waymo Driver completely avoided or mitigated 100% of crashes aside from the crashes in which it was struck from behind, including every instance that involved a pedestrian or cyclist (20 simulations in total). This is the first time an autonomous technology company has shared its evaluation for how the system might perform in real-world fatal crash scenarios.'
I accept this may be some PR / spin going on but I think this could be a way to demonstrate safety to people to make them more accepting. It's not exactly showing how many lives it has saved but does potentially demonstrate it in a way that is easier to perceive than 'x times safer than a human driver'
PF62 said:
FMOB said:
Maybe acknowledge to yourself you are using dodgy statistics rather than quoting as gospel something that has more potholes in it than your average bit of tarmac.
And yet you have absolutely no statistics whatsoever, dodgy or not, to back up your views, so perhaps you might acknowledge that turning off LKA every time you get in the car isn't a sensible thing to do.I now have even more motivation to keep turning the sodding thing off because it seems to annoy you!
FlabbyMidgets said:
Admittedly this is old but still interesting nonetheless.
Waymo simulated fatal collisions that happened in the past and simulated what would have occurred had their 'Waymo Driver' been involved. I haven't read the full paper but this is the headline they publish.
https://waymo.com/blog/2021/03/replaying-real-life...
'In total, the simulated Waymo Driver completely avoided or mitigated 100% of crashes aside from the crashes in which it was struck from behind, including every instance that involved a pedestrian or cyclist (20 simulations in total). This is the first time an autonomous technology company has shared its evaluation for how the system might perform in real-world fatal crash scenarios.'
I accept this may be some PR / spin going on but I think this could be a way to demonstrate safety to people to make them more accepting. It's not exactly showing how many lives it has saved but does potentially demonstrate it in a way that is easier to perceive than 'x times safer than a human driver'
One word. Boll*cks.Waymo simulated fatal collisions that happened in the past and simulated what would have occurred had their 'Waymo Driver' been involved. I haven't read the full paper but this is the headline they publish.
https://waymo.com/blog/2021/03/replaying-real-life...
'In total, the simulated Waymo Driver completely avoided or mitigated 100% of crashes aside from the crashes in which it was struck from behind, including every instance that involved a pedestrian or cyclist (20 simulations in total). This is the first time an autonomous technology company has shared its evaluation for how the system might perform in real-world fatal crash scenarios.'
I accept this may be some PR / spin going on but I think this could be a way to demonstrate safety to people to make them more accepting. It's not exactly showing how many lives it has saved but does potentially demonstrate it in a way that is easier to perceive than 'x times safer than a human driver'
/
As before direct experience of getting a manufacturer early demo car onto our fleet. At the end of its term with us the debrief with the rep various issues were raised, including several occasions where the brakes were slammed full on at a most inappropriate and unnecessary moment.
Manufacturer engineers agreed this was undesirable and even showed where the possibility of this would be mentioned in the owner manual.
Noticed this issue is still being raised by owners of that make some years later, so claims systems improved now somewhat moot.
TL:DR They all do that sir!
FMOB said:
I now have even more motivation to keep turning the sodding thing off because it seems to annoy you!
LoL, quite.As before direct experience of getting a manufacturer early demo car onto our fleet. At the end of its term with us the debrief with the rep various issues were raised, including several occasions where the brakes were slammed full on at a most inappropriate and unnecessary moment.
Manufacturer engineers agreed this was undesirable and even showed where the possibility of this would be mentioned in the owner manual.
Noticed this issue is still being raised by owners of that make some years later, so claims systems improved now somewhat moot.
TL:DR They all do that sir!
Edited by FiF on Wednesday 21st February 08:28
FiF said:
FMOB said:
I now have even more motivation to keep turning the sodding thing off because it seems to annoy you!
LoL, quite.As before direct experience of getting a manufacturer early demo car onto our fleet. At the end of its term with us the debrief with the rep various issues were raised, including several occasions where the brakes were slammed full on at a most inappropriate and unnecessary moment.
Manufacturer engineers agreed this was undesirable and even showed where the possibility of this would be mentioned in the owner manual.
Noticed this issue is still being raised by owners of that make some years later, so claims systems improved now somewhat moot.
TL:DR They all do that sir!
Southerner said:
One word. Boll*cks.
What makes you say that? I haven't read the whole paper so not sure of their methodology but I think the idea of simulating an autonomous vehicle in a real life situation could have more impact on public opinion than statistics. In terms of practical value I don't think it has much in itself, although these types of collisions should be factored into scenario databases
The human road fatality rate is not the criteria for us when considering a driverless car.
We are responsible, sober drivers, who always wear seat belts.
So we need to remove from the human road deaths those that involved:
Drink drive, no seatbelt, reckless, stolen etc.
And motorcycle deaths need to be removed too.
There are also around 3 times the number of suicides each year than road deaths.
It is likely that some (many?) suicides by car crashes are not reported as suicide.
There's little or no data on that.
Furthermore, there are different driver skill levels.
It may be that there are a small number of drivers causing most of the serious crashes.
If we are not one of those, our crash rate may be much lower than the stats suggest.
Getting reliable evidence of the fatality rate that we face when driving is difficult enough before we even start to look at the fatality rate of driverless cars.
The one thing I've learnt about road safety, is that we cannot trust the authorities to do the job properly.
If we want to know if driverless cars save lives, we will have to do the research ourselves!
We are responsible, sober drivers, who always wear seat belts.
So we need to remove from the human road deaths those that involved:
Drink drive, no seatbelt, reckless, stolen etc.
And motorcycle deaths need to be removed too.
There are also around 3 times the number of suicides each year than road deaths.
It is likely that some (many?) suicides by car crashes are not reported as suicide.
There's little or no data on that.
Furthermore, there are different driver skill levels.
It may be that there are a small number of drivers causing most of the serious crashes.
If we are not one of those, our crash rate may be much lower than the stats suggest.
Getting reliable evidence of the fatality rate that we face when driving is difficult enough before we even start to look at the fatality rate of driverless cars.
The one thing I've learnt about road safety, is that we cannot trust the authorities to do the job properly.
If we want to know if driverless cars save lives, we will have to do the research ourselves!
Dave Finney said:
The human road fatality rate is not the criteria for us when considering a driverless car.
Getting reliable evidence of the fatality rate that we face when driving is difficult enough before we even start to look at the fatality rate of driverless cars.
The one thing I've learnt about road safety, is that we cannot trust the authorities to do the job properly.
If we want to know if driverless cars save lives, we will have to do the research ourselves!
One question has bothered me since you started this thread: Are driverless cars intended to save lives?Getting reliable evidence of the fatality rate that we face when driving is difficult enough before we even start to look at the fatality rate of driverless cars.
The one thing I've learnt about road safety, is that we cannot trust the authorities to do the job properly.
If we want to know if driverless cars save lives, we will have to do the research ourselves!
There are many advantages to driverless cars and some sit in authority's camp. Controlling traffic flow, route access and remote disablement come to mind. Workload reduces for the vehicle occupants too. Improved safety is just one of many potential objectives.
Are driverless cars justified even if KSI figures do not reduce?
bigothunter said:
One question has bothered me since you started this thread: Are driverless cars intended to save lives?
There are many advantages to driverless cars and some sit in authority's camp. Controlling traffic flow, route access and remote disablement come to mind. Workload reduces for the vehicle occupants too. Improved safety is just one of many potential objectives.
Are driverless cars justified even if KSI figures do not reduce?
Yes of coarse.There are many advantages to driverless cars and some sit in authority's camp. Controlling traffic flow, route access and remote disablement come to mind. Workload reduces for the vehicle occupants too. Improved safety is just one of many potential objectives.
Are driverless cars justified even if KSI figures do not reduce?
The advantages you mention for authority, but also for us.
Set the destination, and then we can read a book, watch a film, take a nap.
But safety is the primary criteria.
If I am at the same or greater risk with a driverless car,
would I accept that risk and pay more despite it so that I can enjoy other benefits?
Plus, driverless cars cost more.
Would I actually be safer buying a car I had to drive, and spending the saving on smoke detectors for my house?
As with anything, we can only make an informed choice if we have good evidence,
and history has shown that we cannot trust the authorities to provide it!
bigothunter said:
Dave Finney said:
Yes of coarse.
Not too coarse I hope 
Dave Finney said:
But safety is the primary criteria.
Authority frequently sets-up smokescreens to hide their genuine objectives. Safety is a favourite. Are you convinced safety really is the primary objective? I'm not.The people pumping billions into solving autonomous driving are private enterprise without govt funding, and they are in it to try and make some serious money.
soupdragon1 said:
That's a bit tinfoil hat!
The people pumping billions into solving autonomous driving are private enterprise without govt funding, and they are in it to try and make some serious money.
Actually there’s plenty of government grant funding about for this sort of stuff, and regulatory toleranceThe people pumping billions into solving autonomous driving are private enterprise without govt funding, and they are in it to try and make some serious money.
The Wookie said:
soupdragon1 said:
That's a bit tinfoil hat!
The people pumping billions into solving autonomous driving are private enterprise without govt funding, and they are in it to try and make some serious money.
Actually there’s plenty of government grant funding about for this sort of stuff, and regulatory toleranceThe people pumping billions into solving autonomous driving are private enterprise without govt funding, and they are in it to try and make some serious money.
soupdragon1 said:
That is true. In relation to the context of hidden motives, do you also think the govt have alternative reasons, outside of the obvious? (less accidents, less cars in the road etc)
Certain parts of government clearly don’t like private cars. I’ve said many times on here that most mobility conferences have at least one under-qualified civil servant with too much influence spouting their dream of a utopia, for example the last one I went to it was someone from TfL saying the whole of London should be pedestrianised with everyone using public transport so they could plant trees in the road and the children could run round and play like the countryside. No sense of the reality of what it takes to build, run and feed a city, let alone the real need for private vehicles in many circumstances.FWIW I don’t have an issue with full autonomy, the prospect of a car that takes you home from the pub or steers itself for the length of a long motorway journey is great, but there needs to be a choice, and there needs to be a sense of realism about the capability and limitations of the technology.
Partial autonomy and driver augmentation is a different prospect, there seems to be very little in the way of human factors engineering and even in terms of basic safety these systems clearly are rationalising some of the risks that intervention introduces. You might laugh at my hyperbolic example earlier (actually someone else’s I quoted) but, to use a better example, inadvertent application of brake pressure on a controlled braking system is categorised as the highest level of risk due to the risk of injury of being struck from behind by a large vehicle, or being struck from the side while crossing a junction. It is what’s called an ASIL D level risk and what it means when you’re carrying out functional safety on thay type of system that it can NEVER happen. That is a fact. However there are many examples of cars with active braking slamming on the brakes down to a stop at random due to errors in interpretation of things like white lines, I’ve experienced it myself.
My belief is that relying solely on heavy handed driver augmentation systems for safety improvements is deeply flawed, you only have to look to aviation to see there was a long period where increasingly advanced autopilot systems just replace one accident for another. The bar for entry needs to be higher for driving with more regular training and regular testing.
Driving standards are at an all time low IMHO and it’s all well and good to run off down a path of statistics of particular types of accident, but I’ll bet you the vast majority of bad accidents involve at least one ignorant or incompetent driver who hasn’t respected the damage they can do with a car.
Edited by The Wookie on Thursday 22 February 08:45
Gassing Station | Speed, Plod & the Law | Top of Page | What's New | My Stuff