Slashdot Mirror

Depends on how you measure the stats:
https://seriousaccidents.com/l...

Not too interested in death counts so I didn't look for that statistic. Many possible outcomes are worse than death.

FYI: I do not drink. Measures taken to combat the drunks haven't worked out and an expensive car add on to go wrong isn't going to fix that. Motivated people will get around it as easily as driving while revoked. Chinese hack devices from ebay will be out within a year... court-ordered add-on devices are really expensive and monitored but to mass produce a cheap alternative that is not monitored is not going to work.

I don't know what you're getting at. You can show otherwise? I'd love to see it... In the meantime...

https://seriousaccidents.com/l...

Yea, just like the F22 Raptor, [airforcetimes.com] right?

Not really: let's see how many differences there are:
1. The problem with the F-22 is with an oxygen system, not a computer system
2. The F-22 is still currently in it's testing phase. There will be extensive testing of autodrive systems before release.
3. Google probably has more self-driving cars than there are F-22s(IE there's already a larger base of self driving cars for testing purposes)

No, but it does blame any potential failure on the concept of improper road signage, not equipment malfunctions.

The SPECIFIC problem in question was 'driving into a busy farmer's market', a very specific potential failure, not a generic one. My response was that it would be a 'real possibility', IE high probability event, if they mismarked the street. It was more a statement of the sort of incidents I see auto-drive cars getting into. You could hang a white and black banner across a street above car level with 'STREET CLOSED' written on it and most humans would get the point, but unless an autodrive system is rather more AI than I'd expect, it'd miss it completely, resulting in a stream of cars going down the road. This decreases if it's marked per NTSB standards, because then in order to get onto the road you're going to have to maneuver around barriers at the least.

This is an example of an incident caused by 'failure to recognize strategic problems', as opposed to 'fast twitch' problems like somebody running into the road. It'd be dangerous even with 'fast twitch' accident avoidance responses to try to keep the car from hitting anybody/thing. Road closed improperly - Strategic problem. Somebody running into the road - fast twitch problem. Computers driven vehicles will likely be better at the latter than they are at the former.

What sorts of equipment malfunction could happen? Failure to recognize obstacle, unexpected acceleration, unexpected deceleration, veering into an obstacle as opposed to avoiding it, over-estimation of road surface resulting in loss of control, overcorrecting resulting in loss of control/rollover, etc...

What I have a problem with is all the Google-dick-sucking that goes on anytime someone (like me) posits any sort of doubt or question as to the infallible nature of such a system.

But that's not a good excuse when I lead off with proposals where the government specifically limits liability on the part of autodrive manufacturers in order to keep them from being financially slaughtered in liability lawsuits, indicating that I am neither sucking Google's cock or assuming that the system is infallible. Heck, I only posit cutting the accident rate [i]in half[/i]! Rain is #5, design defects is #10. Distracted driving is #1, Speeding is #2, and Drunk Driving is #3. Don't see percentages for #1&2, but #3 is around 30%. That's 90% of accidents right there that an autodrive system should readily solve almost completely.

Per this study, that would be:

#14 Tailgating
#11 Unsafe lane change
#1 Distracted driving of any sort
#20 Drowsy Driving
#3 drunk driving
#4 Reckless driving

#2 is speeding - It says your reactions slow, my correction would be that you need to react faster to avoid an accident at higher speed. The difference between 60 and 65 can be over 100ft in stopping distance, or the difference between just missing the bumper of the car in front of you and plowing through it.

I've heard that speeding isn't the leading, so some research:
Distracted Driving (speeding #2)
Distracted Driving (Speeding #4)
Not using turn signals 2x worse than distracted driving?
distracted driving
Disparities noticed:
Fatigue: #20 in the first list, #2 in the second

Anyways, I like being distracted; I don't particularly like driving. Bring on the self-driving vehicles! Or other way I can get to work/store without having to be behind the wheel.

So if about 10% of accidents are from distracted driving (actually sounds pretty low).

That does sound pretty low. I think, under most circumstances, you can avoid accidents if you're paying attention. Anyway, I did some searching the web, and here are some numbers that came up:

http://www.nationwide.com/newsroom/dwd-facts-figures.jsp: Driving while distracted is a factor in 25 percent of police reported crashes.

http://www.safetyresource.org/car/distracted_driving_statistics.html: Car safety experts recently announced that after a study of thousands of car accidents, it was determined that 40 percent, or almost half, are caused by distracted drivers.

http://seriousaccidents.com/legal-advice/top-causes-of-car-accidents/driver-distractions/: According to a study released by the National Highway Traffic Safety Administration (NHTSA) and the Virginia Tech Transportation Institute (VTTI), 80 percent of automobile accidents and 65 percent of near-accidents involve at least some form of driver distraction within three seconds of the crash or near-miss.

Funny how the numbers are all over the map.

sorry I can't offer more sources than this (heading into work shortly), but it's not unheard of for someone to miss a red light and cause an accident because of a cell phone call:

http://www.seriousaccidents.com/accident-blog/car-technology-accident-blog/cell-phones-car-accidents-part-5-call-of-death/

http://www.brentadams.com/blog/report-says-28-percent-of-accidents-caused-by-cell-phone-usage.cfmhttp://www.brentadams.com/blog/report-says-28-percent-of-accidents-caused-by-cell-phone-usage.cfm

I've also heard that some areas do increase the severity of the punishment if it's determined someone's on a cell phone when the accident occured. Driving a car is a privilage, and to be licensed you are agreeing to stay in full control of an exceptionally deadly object. It makes sense that if you are willfully ignoring your duty to stay vigilant behind the wheel, you should take more responsibility if something goes wrong.

Admittedly I drive on the cell phone a decent amount. I'd hate to get punished more for the cell phone in case of an accident, but the logic makes sense, and I can't deny it's distracting.