Jump to content


Photo
- - - - -

driverless cars to hit california ( sorry!)


  • Please log in to reply
410 replies to this topic

#251 mariner

mariner
  • Member

  • 1,374 posts
  • Joined: January 07

Posted 25 October 2012 - 19:49

Greg, thnak you for your answers but they leave me one very big question on SBW

1) "Drive by wire will need either force sensors in the tie rods, or yet another little model,"

2) Now, given that manufacturing cars profitably is insanely difficult "

given 1) and 2) why,why , why would you set about installing an expensive electronic sensor on the tie rod , which is just about the most debris hit, water soaked, vibration battered location on a car then plan for it to work at 100% perfection for a 15 year life that OEM'sare expected to design to?

What makes SBW so much better as to justify such a cost and reliability engineering challenge ?




Advertisement

#252 Grumbles

Grumbles
  • Member

  • 326 posts
  • Joined: September 09

Posted 25 October 2012 - 20:28

Something that really irritates me (to be honest though, everything irritates me these days...) is half-asleep drivers at traffic lights. Whenever there's a lineup of cars at a red light you can bet that someone towards the front of the queue will be slow to take off when the lights turn green. So instead of say 15 or 20 cars passing through it's only 6 or 7. If all the cars could accelerate simultaneously I think it would make a massive improvement to traffic flow.

#253 Dmitriy_Guller

Dmitriy_Guller
  • Member

  • 4,044 posts
  • Joined: July 01

Posted 25 October 2012 - 20:30

Something that really irritates me (to be honest though, everything irritates me these days...) is half-asleep drivers at traffic lights. Whenever there's a lineup of cars at a red light you can bet that someone towards the front of the queue will be slow to take off when the lights turn green. So instead of say 15 or 20 cars passing through it's only 6 or 7. If all the cars could accelerate simultaneously I think it would make a massive improvement to traffic flow.

Around here, honking is mandatory immediately after the light goes green. Even if you're 10th car in line.

#254 Tony Matthews

Tony Matthews
  • Member

  • 17,499 posts
  • Joined: September 08

Posted 25 October 2012 - 20:37

Something that really irritates me (to be honest though, everything irritates me these days...) is half-asleep drivers at traffic lights.

At the end of my road, and the only way out into the Big Wide World, the traffic lights are green for less than 9 seconds. All it takes is the lead car to go throught the "Oh, they've gone green! Er, am I in neutral? Is this neutral, I think it is - left and up. No, back into neutral, can't be too careful - left, up, I think that's in - Oh, they've gone red!" routine, and you feel your life slipping away.

#255 Grumbles

Grumbles
  • Member

  • 326 posts
  • Joined: September 09

Posted 25 October 2012 - 21:32

Around here, honking is mandatory immediately after the light goes green. Even if you're 10th car in line.


That's a bit rude isn't it?

I prefer to just wind down the window and shout witty observations about their driving skills and intellect, and remark upon their uncanny resemblance to certain genitalia...

#256 Bloggsworth

Bloggsworth
  • Member

  • 7,470 posts
  • Joined: April 07

Posted 25 October 2012 - 21:46

Something that really irritates me (to be honest though, everything irritates me these days...) is half-asleep drivers at traffic lights. Whenever there's a lineup of cars at a red light you can bet that someone towards the front of the queue will be slow to take off when the lights turn green. So instead of say 15 or 20 cars passing through it's only 6 or 7. If all the cars could accelerate simultaneously I think it would make a massive improvement to traffic flow.


As long as Grosjean isn't halfway down the queue...

#257 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 25 October 2012 - 22:06

You guys crack me up! :lol:

#258 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 25 October 2012 - 22:08

What makes SBW so much better as to justify such a cost and reliability engineering challenge ?


packaging, crash, weight, gee-whiz factor, ability to do some seriously weird stuff. For instance by steering each wheel independently, well I don't know, but you could do some seriously odd things. But the main advanatage is that you no longer need a steering colum or rack, so the engine bay can suddeny be full of engine and batteries etc, instead of needing to leave long inconvenient packaging volumes for the intermediate shaft and rack. Also your steering wheel becomes part of your crash solution rather than being the main problem.

#259 bigleagueslider

bigleagueslider
  • Member

  • 855 posts
  • Joined: March 11

Posted 26 October 2012 - 02:49

Greg, thnak you for your answers but they leave me one very big question on SBW

1) "Drive by wire will need either force sensors in the tie rods, or yet another little model,"

2) Now, given that manufacturing cars profitably is insanely difficult "

given 1) and 2) why,why , why would you set about installing an expensive electronic sensor on the tie rod , which is just about the most debris hit, water soaked, vibration battered location on a car then plan for it to work at 100% perfection for a 15 year life that OEM'sare expected to design to?

What makes SBW so much better as to justify such a cost and reliability engineering challenge ?


Many cars already have what are effectively SBW systems. Stability control systems use various combinations of ABS at each wheel, electronic throttle control, electronic diff control, and electronic control of dampeners to override driver inputs and stabilize the vehicle. Software, digital controls, solid state sensors and electronic systems operating at 100Hz or more are way more capable than any human.

The steering effect produced by these systems doesn't typically require a direct sensor feedback such as a strain gauge on the tie rods. Instead they make calculated decisions that rely on several indirect forms of feedback from sensors like 3 axis accelerometers or hall effect sensors on each wheel.


Advertisement

#260 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 26 October 2012 - 03:03

However none of those systems have the "freak out" factor of a 100% Steer-by-wire system.

Personally, I am quite cool with it. We already have FBW in aircraft and numerous heavy road vehicles (eg articulated mobile cranes) have only hydraulic connection to the steering system.

Edited by gruntguru, 26 October 2012 - 03:06.


#261 Tony Matthews

Tony Matthews
  • Member

  • 17,499 posts
  • Joined: September 08

Posted 26 October 2012 - 15:12

http://www.gizmag.co...tm_medium=email

#262 Tenmantaylor

Tenmantaylor
  • Member

  • 8,317 posts
  • Joined: July 01

Posted 26 October 2012 - 19:10

That seems like it's asking for trouble. There maybe a lot of situations where driving into the back of the car infront is desirable. When there are pedestrians nearby, oncoming traffic, driving near a ditch, tree, cliff on a bridge etc. Can the system detect and assess if swerving is safer than hitting the car infront?

#263 munks

munks
  • Member

  • 344 posts
  • Joined: January 03

Posted 26 October 2012 - 19:38

FTFA: "In those cases, it would proceed check for vehicles approaching from the rear, and for forward zones that were free of obstacles such as other vehicles or pedestrians."

Again, you guys think you're the first ones to think of potential problems.

Edited by munks, 26 October 2012 - 19:39.


#264 Dmitriy_Guller

Dmitriy_Guller
  • Member

  • 4,044 posts
  • Joined: July 01

Posted 26 October 2012 - 19:50

But what if what looks like an empty stretch of road the car will swerve into is actually an active minefield? Humans would be able to read the yellow sign warning of the minefield in the left lane, but will computers have language recognition skills to realize it before they blow up the car in the avoidance maneuver?

Edited by Dmitriy_Guller, 26 October 2012 - 19:52.


#265 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 26 October 2012 - 21:47


And of course a skilful driver will realise that by driving over the edge of a cliff on the other side he'll land on top of the speeding train below and will survive.

#266 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 26 October 2012 - 23:51

http://www.gizmag.co...tm_medium=email

The level of sophistication required for that system is approaching what is required for a fully driverless system. My guess is Nissan (and others) figure the market isn't ready for that yet and will only develop/offer what is palatable for now. Or could it be that Google have the jump in terms of patent protection?

#267 desmo

desmo
  • Tech Forum Host

  • 12,988 posts
  • Joined: January 00

Posted 27 October 2012 - 02:00

And of course a skilful driver will realise that by driving over the edge of a cliff on the other side he'll land on top of the speeding train below and will survive.


:lol: Roughly as likely a scenario as most proffered here to make the point that human drivers are insuperable.


#268 munks

munks
  • Member

  • 344 posts
  • Joined: January 03

Posted 01 November 2012 - 01:51

Not sure if anybody posted this yet:

race between self-driving car and pro driver

#269 Lee Nicolle

Lee Nicolle
  • Member

  • 5,853 posts
  • Joined: July 08

Posted 04 November 2012 - 02:06

With the electric power steering I honestly can not see any advantages of it, and is as known a whole lot of disadvantages.
A driver does need to know the conditions underneath him, damping it down is generally advantagous though a lot of cars with mechanical steering have either too much feedback, or often overdamped or slow steering.
Suspension design often comes into this too, a double wishbone is generally nicer and the cheaper strut is usually more nervous and twitchy. While these days none are bad some a mentally more demanding if not physically.
And in inclement conditions, broken up roads or loose surfaces feedback is essential.
The joke is so many cars have huge wheels, wide low profile tyres and then wonder about ride quality!!! With the huge amounts of steering feedback that comes from such dumb set ups. Manufacturers are making vehicles for best case scenario and most countrys are worst case these days. That is b efore we even talk about dirt where these stupid wheels are downright dangerous, an accident waiting too happen.
Then taking away the road feel is just another degree of stupidity. A mechanical system, correctly damped will have at least some feel, an electric motor seldom does, and must be far more prone to failure in adverse conditions too.

#270 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 04 November 2012 - 09:06

With the electric power steering I honestly can not see any advantages

Here ya go! You only have to look.


packaging, crash, weight, gee-whiz factor, ability to do some seriously weird stuff. For instance by steering each wheel independently, well I don't know, but you could do some seriously odd things. But the main advanatage is that you no longer need a steering colum or rack, so the engine bay can suddeny be full of engine and batteries etc, instead of needing to leave long inconvenient packaging volumes for the intermediate shaft and rack. Also your steering wheel becomes part of your crash solution rather than being the main problem.



#271 carlt

carlt
  • Member

  • 1,053 posts
  • Joined: June 09

Posted 04 November 2012 - 10:39

Not sure if anybody posted this yet:

race between self-driving car and pro driver



reading the last part of your link brought one immediate feeling - 'would I be comfortable giving over control of a potential lethal machine and possibly also my life' ,
daily transport is so much part of our lives, not being in control and ultimately taking responsibility would be very difficult .

I can't even do fairground rides , but am quite happy being on the edge of the performance envelope and a potential major accident during a speed hillclimb - its about being in charge/control of ones life

#272 Bloggsworth

Bloggsworth
  • Member

  • 7,470 posts
  • Joined: April 07

Posted 04 January 2013 - 21:26

Read this then ask yourself "Do I really want a driverless car..."

#273 Lee Nicolle

Lee Nicolle
  • Member

  • 5,853 posts
  • Joined: July 08

Posted 04 January 2013 - 22:55

reading the last part of your link brought one immediate feeling - 'would I be comfortable giving over control of a potential lethal machine and possibly also my life' ,
daily transport is so much part of our lives, not being in control and ultimately taking responsibility would be very difficult .

I can't even do fairground rides , but am quite happy being on the edge of the performance envelope and a potential major accident during a speed hillclimb - its about being in charge/control of ones life

'But only by a few measly seconds' How many seconds? Probably about 10, maybe more. Which says it all.Meaning the car is about as fast as any idiot driver,, and in many cases an idiot can be trained to drive far better. Sometimes!!

#274 mariner

mariner
  • Member

  • 1,374 posts
  • Joined: January 07

Posted 05 January 2013 - 16:37

I have terrible habit of skim reading books I buy as presents so when I bought my son ( an avid " air rash investigators" fan) a copy of the story of the Quantas A380 flight engine failure by the Captain of the aircaft I skim read it.

Althought the events whereby the plane had an uncontained engine explosion are well known I had not realized just how badly damaged the plane was.

Many critical sensors and actuators were lost. So the Fly by wire computers were simply unable to operate properly even in back up to back up mode as they lacked enough sensor data .

The plane had four highly experienced pilots on the flight deck and it took all of them to understand what was wrong , and critically, what bad info. the computers were telling them.

At one point the systems clearly told them to pump fuel into the damaged wing. The captain called a flight crew "sense chack " and all four voted " NO WAY"

Obviously a car can stop easier than plane etc. but the next time anybody tells me computers can always fly a plane , or car, safer than an experienced pilot I'll point them at that book.

#275 Lee Nicolle

Lee Nicolle
  • Member

  • 5,853 posts
  • Joined: July 08

Posted 06 January 2013 - 01:02

I have terrible habit of skim reading books I buy as presents so when I bought my son ( an avid " air rash investigators" fan) a copy of the story of the Quantas A380 flight engine failure by the Captain of the aircaft I skim read it.

Althought the events whereby the plane had an uncontained engine explosion are well known I had not realized just how badly damaged the plane was.

Many critical sensors and actuators were lost. So the Fly by wire computers were simply unable to operate properly even in back up to back up mode as they lacked enough sensor data .

The plane had four highly experienced pilots on the flight deck and it took all of them to understand what was wrong , and critically, what bad info. the computers were telling them.

At one point the systems clearly told them to pump fuel into the damaged wing. The captain called a flight crew "sense chack " and all four voted " NO WAY"

Obviously a car can stop easier than plane etc. but the next time anybody tells me computers can always fly a plane , or car, safer than an experienced pilot I'll point them at that book.

Cannot compute, cannot compute, need more information. Mindless moronic computers. And often there inventors!

#276 Ross Stonefeld

Ross Stonefeld
  • Member

  • 57,301 posts
  • Joined: August 99

Posted 06 January 2013 - 14:07

I bet the computer would have picked up the typo, though.

#277 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 07 January 2013 - 00:16



Once again "Four Corners" provides the definitive piece of investigative journalism.

Yes the plane would have crashed but for the skill of the humans in control. OTOH computers have made aviation much safer by reducing opportunities for human error. The same is already true in the automotive arena and will become increasing so - yes even to the point of autonomous control when it hits the market.

#278 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 07 January 2013 - 00:49

Obviously a car can stop easier than plane etc. but the next time anybody tells me computers can always fly a plane , or car, safer than an experienced pilot I'll point them at that book.

for cars it doesn't have to be /always better than an experienced driver/, it just has to be rather better than the average driver paying the average level of attention to his surroundings.

That is a rather simpler task, to say the least.

#279 desmo

desmo
  • Tech Forum Host

  • 12,988 posts
  • Joined: January 00

Posted 07 January 2013 - 04:47

It's not a safety issue as much as a sales issue, everyone thinks they are an elite driver so people will tell themselves "it may be better than the average driver but not than I" and invent all sorts of fanciful rationalizations such as contrived highly unlikely hypotheticals they believe an "elite driver"-- e.g. themselves-- could do better than the automated driver. So even if the systems are significantly safer overall than real drivers, those drivers' hubris will prevent them accepting them as such.

Advertisement

#280 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 07 January 2013 - 05:29

Agreed. I think the bar will be set quite high because litigation follows the big money, and in this case rather than bankrupting some average Joe, they'll be after the car company.

http://www.prnewswir...-126563103.html

#281 404KF2

404KF2
  • Member

  • 2,264 posts
  • Joined: October 99

Posted 07 January 2013 - 06:18

It's got to be a load of fun to sit in a car that some chips are driving for you. Imagine the exhilaration! Boo yeah!

#282 Youichi

Youichi
  • Member

  • 2,999 posts
  • Joined: January 04

Posted 07 January 2013 - 17:29

It's not a safety issue as much as a sales issue, everyone thinks they are an elite driver so people will tell themselves "it may be better than the average driver but not than I" and invent all sorts of fanciful rationalizations such as contrived highly unlikely hypotheticals they believe an "elite driver"-- e.g. themselves-- could do better than the automated driver. So even if the systems are significantly safer overall than real drivers, those drivers' hubris will prevent them accepting them as such.


But everyone's wife doesn't think she's an 'elite' driver, so will be quite happy to be chauffeured around.........


#283 Tony Matthews

Tony Matthews
  • Member

  • 17,499 posts
  • Joined: September 08

Posted 07 January 2013 - 19:15

Yep, and I think I would prefer to be driven by a computer than my wife...

#284 BRG

BRG
  • Member

  • 11,588 posts
  • Joined: September 99

Posted 07 January 2013 - 19:19

It's got to be a load of fun to sit in a car that some chips are driving for you. Imagine the exhilaration! Boo yeah!

Yes, but safe & incident-free driving isn't really meant to be exhilarating. That should be saved for the race track or the rally stage. Sorry. boring, I know, but out there in the real world, that's how people think.

#285 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 07 January 2013 - 21:16

Yes, but safe & incident-free driving isn't really meant to be exhilarating. That should be saved for the race track or the rally stage. Sorry. boring, I know, but out there in the real world, that's how people think.


And its how they drive. When CANBUS was introducued we built some real time data loggers and recorded 7 million seconds of real driver usage. Apart from two seconds of data from Sydney at 0.8g braking there was no sign that anybody used more than 70% of the maximum capability of the vehicle in that time.

If you have one of those g analyst things, try and drive for 0.5g through 3 connected corners on public roads. I won't post bail for you, but I might send flowers if you end up in hospital.

#286 Ross Stonefeld

Ross Stonefeld
  • Member

  • 57,301 posts
  • Joined: August 99

Posted 07 January 2013 - 21:17

I'm trying to remember the title of the thread so I can look up phantom II's g-claims in his Vette...

#287 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 08 January 2013 - 00:04

I'm trying to remember the title of the thread so I can look up phantom II's g-claims in his Vette...

That's the one where FatBoy went ballistic at him for driving like a neejit.

#288 Ross Stonefeld

Ross Stonefeld
  • Member

  • 57,301 posts
  • Joined: August 99

Posted 08 January 2013 - 01:24

We don't have enough threads with annotated satellite images.

#289 Dmitriy_Guller

Dmitriy_Guller
  • Member

  • 4,044 posts
  • Joined: July 01

Posted 09 January 2013 - 02:38

I have terrible habit of skim reading books I buy as presents so when I bought my son ( an avid " air rash investigators" fan) a copy of the story of the Quantas A380 flight engine failure by the Captain of the aircaft I skim read it.

Althought the events whereby the plane had an uncontained engine explosion are well known I had not realized just how badly damaged the plane was.

Many critical sensors and actuators were lost. So the Fly by wire computers were simply unable to operate properly even in back up to back up mode as they lacked enough sensor data .

The plane had four highly experienced pilots on the flight deck and it took all of them to understand what was wrong , and critically, what bad info. the computers were telling them.

At one point the systems clearly told them to pump fuel into the damaged wing. The captain called a flight crew "sense chack " and all four voted " NO WAY"

Obviously a car can stop easier than plane etc. but the next time anybody tells me computers can always fly a plane , or car, safer than an experienced pilot I'll point them at that book.

"Always" is way too high of a standard. Any technological advancement will bring about new modes of failure that weren't present before. The decider is whether the new modes of failure bite less often than the old modes of failure they minimize.

#290 V8 Fireworks

V8 Fireworks
  • Member

  • 5,468 posts
  • Joined: June 06

Posted 20 January 2013 - 07:01

It takes my shiny new Sony television 3 seconds to change channels,

I think Sony televisions are made cheaply as they seem to take longer to decode the signal than other televisions.

In the meantime to fit that emergency brake system with radar to all cars would be a modest compromise I think, wouldn't that prevent most rear ending crashes?

#291 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 21 January 2013 - 00:37

In the meantime to fit that emergency brake system with radar to all cars would be a modest compromise I think, wouldn't that prevent most rear ending crashes?

Or create some? Perhaps drivers of such cars will become complacent resulting in a higher number of cars performing emergency stops . . . . . and the car behind may or may not be fitted with the system . . . . :) . . . just saying . . .

#292 Greg Locock

Greg Locock
  • Member

  • 4,514 posts
  • Joined: March 03

Posted 21 January 2013 - 02:59

Or create some? Perhaps drivers of such cars will become complacent resulting in a higher number of cars performing emergency stops . . . . . and the car behind may or may not be fitted with the system . . . . :) . . . just saying . . .

That's a great argument for not introducing disc brakes and radial tires, both of which gave a step change in performance compared with the preceeding system.

I suggest you game out the variosu combinations of Radar assisted brakes and drivers too close to car in front, RAB still results in a net reduction in accidents.
1)
car 1 stops suddenly

car2 inattentive driver+normal brakes crash
Car 3 inattentive driver+normal brakes crash

3 cars damage
2)
car 1 stops suddenly

car2 inattentive driver+RAB no crash
Car 3 inattentive driver+normal brakes crash

2 cars damaged

3)
car 1 stops suddenly

car2 inattentive driver+normal brakes crash
Car 3 inattentive driver+RAB no crash

2 cars damaged

4)

car 1 stops suddenly

car2 inattentive driver+RAB no crash
Car 3 inattentive driver+RAB no crash

0 cars damaged

I think of the various technologies RAB is the one that has the most immediate payoff and has a relatively small effect on normal driving behaviour.

Edited by Greg Locock, 21 January 2013 - 03:07.


#293 johnny yuma

johnny yuma
  • Member

  • 928 posts
  • Joined: May 07

Posted 21 January 2013 - 03:55

I also with magoo welcome our new masters,the self driving cars of the future.I bow down before them and trust in their peerless wisdom.
I'll be buying the one which fills in the insurance form for you.In fact we won't need insurance will we,in
this brave new world of perfectly driven cars ?

Edited by johnny yuma, 21 January 2013 - 05:26.


#294 gruntguru

gruntguru
  • Member

  • 5,290 posts
  • Joined: January 09

Posted 22 January 2013 - 07:40

I think of the various technologies RAB is the one that has the most immediate payoff and has a relatively small effect on normal driving behaviour.

. . . . :) . . . just saying . . .

Couldn't find a "tongue in cheek" smiley. I'm sure you know I am actually a staunch advocate of technology over human frailty.

#295 mariner

mariner
  • Member

  • 1,374 posts
  • Joined: January 07

Posted 12 February 2013 - 09:08

I missed this last year, seems quite well balanced article.

http://news.bbc.co.u...ine/9755210.stm

I don't now where the "google must do 300 million miles without a fatality "comes from , but it does illustrate the problems with certification that driverless cars will face before becoming proven to be safe.

Lets just suppose 300 M mile is right (?). Its pretty hard to do more than 300 miles per day in representative trafific ( ten hrs at 30 mph suburban average ) because driving it round at night isnt a real test. So 360*300 = 108K miles per year. So you for our 300M miles validation you need about 500 cars for 6 years.

Thats not imposible or unaffordable but it illustrates that what Google are doing is really just another "proof of concept", not an FAA type certification process.

#296 desmo

desmo
  • Tech Forum Host

  • 12,988 posts
  • Joined: January 00

Posted 12 February 2013 - 15:10

That's a ridiculous standard. Particularly given that it is altogether likely a fatality will be the fault of another driver.

#297 CSquared

CSquared
  • Member

  • 621 posts
  • Joined: December 09

Posted 12 February 2013 - 17:28

The way I read the article, someone's just saying that with the stat of 1.11 fatalities per 100 million miles, 300 million miles is the sample size needed to be able to make a claim about that stat. It says nothing about certification or Google's intentions.

#298 mariner

mariner
  • Member

  • 1,374 posts
  • Joined: January 07

Posted 12 February 2013 - 17:43

I don't think a target of one fatality per 300 M miles for a driverless car is at all unreasonable.

The Uk is reported at 5.9 fatalities per Billion vehicle Km versus 8.5 per Billion for the USA. There are lots of reasons for the difference but the Uk number is not a bad US goal . so the current US 1.3 per 100 M miles would be about one fatality per 110 milion miles.

It seems hardly worth doing driverless cars unless a significant reduction in road deaths is a result ( and that is Googles awoved reason for persuing driverless cars).

So if you say other measures could get the US to the Uk level then one death per 110M miles is acheivable without driverless cars so a goal of one death per 300M seems no more than a modest target to me.

Also current actual death rates have, sadly, a vast actual database. The driverless rate wil be hypothetical for a long time as milage builds up so aiming for less than one per 300 M would be risking driverless cars being LESS safe in practice.

Regardles of the sort of goals the enginers would set any outcome below 300M is likely to be legally and socially un-acceptable in today's blame society in the west.

The 300M was calculated by a Californian lawyer BTW.

#299 CSquared

CSquared
  • Member

  • 621 posts
  • Joined: December 09

Posted 12 February 2013 - 20:05

There are lots of benefits to autonomous cars besides safety:
-Increased access for the disabled
-More possibility for car-sharing, which means fewer cars, more parking, and money saved
-Increased productivity

All they have to do is show that the accident, injury, and fatality rates are no worse than human drivers and it's a win. As you point out, even if a car is perfectly driven other drivers will still run into it.

Advertisement

#300 pugfan

pugfan
  • Member

  • 174 posts
  • Joined: August 09

Posted 13 February 2013 - 00:25

Regardles of the sort of goals the enginers would set any outcome below 300M is likely to be legally and socially un-acceptable in today's blame society in the west.


A reduction in fatalities isn't the only benefit though. How about the benefit through a (presumably) massive reduction in accidents that don't result in fatalities. Medical fees, crash repair fees, insurance fees, reduction in sick leave, etc, etc, all that broken window economic activity that could be redirected elsewhere.

Presumably the fatal accident rate could actually be worse and still result in a net benefit to society but then you'd be putting a price on someone's life, which for some reason people feel uncomfortable about.