'Self-driving' cars, aircraft, and GPS apps

[Socal Fan]

Mar 21, 2018 5:20:49 GMT -5 BOGC said:

Because even ordinary driving is like the Blue Angels or Thunderbirds flying in tight formation.  Rest assured, they aren't on autopilot!

Interesting issue. I'd bet that computers with sensors and communications could fly Blue Angels or Thunderbirds better than humans.

Edit add: but ironically, they would never bother to build a computer system to do that. Flying Blue Angels or Thunderbirds is hard for humans but easy for computers. Nobody would be interested in watching easy things. That's why Jackie got so much attention. Singing opera is hard for a child but relatively easier for an adult.

[Socal Fan]

www.forbes.com/sites/samabuelsamid/2018/03/21/uber-crash-tape-tells-very-different-story-from-police-report-time-for-some-regulations/

My thoughts:

1. The victim was crossing at night in an unlit part of the road. She was invisible until 3 secs into the video and was hit at 4 secs. A human driver would very likely have hit her too.

2. The victim should clearly have seen the car coming and crossed later. What happened to "look both ways before crossing the road"? Was the victim impaired?

3. The Uber self driving car had RADAR and LIDAR which would easily have detected the victim even in complete darkness. So clearly something was wrong with the car.

4. The Uber operator was looking down just before the accident. She should have been looking at the road, but I don't think it would have made much of a difference.

5. I was reading that Arizona has one of the highest pedestrian accident rates in the country. Arizonans apparently need to learn how to cross roads.

6. All cars are now required by law to have backing up cameras. I think they should also have forward cameras with 10 secs of recording. I think that people would drive more carefully if they knew their driving was being recorded and the video and car computer information would be used against them if they ever got into an accident. Of course, this might also be used to exonerate them and incriminate others.

[Deleted]

Thanks for the video, socalfan. If you are correct about your assessment (number 3), this may indeed be the car's fault.
Have you learned anything about the car's speed?

[Deleted]

It also appears that the driver was not really paying attention. At night it is especially important to do so. Nevertheless, I agree that this accident would probably be unavoidable by the 'safety' driver. 

I feel sorry for the driver....I bet she's feeling very badly about all of this.

[Socal Fan]

Mar 22, 2018 11:11:14 GMT -5 @ben said:

Thanks for the video, socalfan. If you are correct about your assessment (number 3), this may indeed be the car's fault.
Have you learned anything about the car's speed?

It was primarily the pedestrian's fault for not looking before crossing. It was secondarily the car's fault for not making a better effort to avoid the pedestrian. It is important to note that the car's fault was not that it underperformed a human driver; rather, its fault was that it was designed to outperform a human driver but, for whatever reason, did not do so.

Reports are that the car was going at 38 or 40; the speed limit was reported to be either 35 or 45. It is likely that a human driver would be going at the same speed or faster.

[Deleted]

Mar 22, 2018 11:31:45 GMT -5 Socal Fan said:

Mar 22, 2018 11:11:14 GMT -5 @ben said:

Thanks for the video, socalfan. If you are correct about your assessment (number 3), this may indeed be the car's fault.
Have you learned anything about the car's speed?

It was primarily the pedestrian's fault for not looking before crossing. It was secondarily the car's fault for not making a better effort to avoid the pedestrian. It is important to note that the car's fault was not that it underperformed a human driver; rather, its fault was that it was designed to outperform a human driver but, for whatever reason, did not do so.

Reports are that the car was going at 38 or 40; the speed limit was reported to be either 35 or 45. It is likely that a human driver would be going at the same speed or faster.

   Agreed, especially if there was no crosswalk provided: 

   Arizona law says that....Motorists must yield the right-of-way to pedestrians who are crossing the street. This applies to pedestrians who are in marked and unmarked crosswalks.

Additionally, motorists must yield to pedestrians walking through intersections when the light turns green.

Furthermore, a motorist must come to a complete stop at any school crossing when anyone is in the crosswalk. A motorist cannot pass another vehicle that is allowing for pedestrians to cross.

Source:  www.phillipslaw.com/blog/arizona-right-of-way-laws 


It is my understanding that speed limits can/should be exceeded (or lessened) depending upon the
road conditions, time of day, and average speed of surrounding traffic.  So, in this case, I believe
that the car was going too fast, no matter what the posted limit was.    

[Socal Fan]

Here is some food for thought.

Let us assume the following:
1. Auto deaths are 35K/year (FYI in 2016 it was 37K).
2. Self driving cars will reduce fatalities by 5%, ie 1750 per year.
3. Congress can pass laws allowing the fast testing and implementation of self driving cars or pass laws requiring a slow and cautious process.
4. Going fast will result in 250 people getting killed while ironing out the bugs.
5. The slow approach will delay implementation by 1 year but kill nobody.

Given the above assumptions, the fast approach will kill 250 but save 1750 with a net saving of 1500 lives. But Congress would never take the fast approach because it would then be held responsible for 250 deaths. The public has already deemed the 35K deaths per year as acceptable. What is unacceptable is the 250 additional deaths, regardless of the net decrease in deaths.

Unfortunately, the public is not very good at evaluating and understanding risks.

[BOGC]

Mar 23, 2018 14:16:49 GMT -5 Socal Fan said:

Here is some food for thought.

Let us assume the following:
1. Auto deaths are 35K/year (FYI in 2016 it was 37K).
2. Self driving cars will reduce fatalities by 5%, ie 1750 per year.
3. Congress can pass laws allowing the fast testing and implementation of self driving cars or pass laws requiring a slow and cautious process.
4. Going fast will result in 250 people getting killed while ironing out the bugs.
5. The slow approach will delay implementation by 1 year but kill nobody.

Given the above assumptions, the fast approach will kill 250 but save 1750 with a net saving of 1500 lives. But Congress would never take the fast approach because it would then be held responsible for 250 deaths. The public has already deemed the 35K deaths per year as acceptable. What is unacceptable is the 250 additional deaths, regardless of the net decrease in deaths.

Unfortunately, the public is not very good at evaluating and understanding risks.

Maybe they're not as bad as it sounds.  Statistics can make precise predictions about very large numbers.  But for the individual, the difference in risk is roughly 0.00054%.  Even for someone who knows 1,000 people, it's still unlikely they'll know anyone saved by a self-driving car; and of course, since one can never be sure what would have happened differently, even if they do, they may never be aware of it.

[Deleted]

Whoa, navigator! 


www.foxnews.com/us/2018/03/27/uber-driver-takes-stairs-blames-apps-navigation.html

[Deleted]

The magnificent obsession for high technology has consequences we may not like one day. 

Are we becoming so lazy that we will depend upon computers for conventional wisdom, and, hence, matters of life and death?   This may be yet another condition of amorality we have accepted as the norm. 
And the concept of 'responsible driving' has just evaporated.   

Driving with eyes wide shut.  


                             

[Deleted]

www.ksl.com/?sid=46288489&nid=157&title=new-cars-are-quickly-getting-self-driving-safety-features





                                                         

[Socal Fan]

Mar 28, 2018 19:56:02 GMT -5 @ben said:

www.ksl.com/?sid=46288489&nid=157&title=new-cars-are-quickly-getting-self-driving-safety-features

Thanks. One line was particularly interesting: "The government says 94 percent of crashes are caused by human error."

That's why I'm so optimistic about self driving cars. They have a very low bar to clear.

[Deleted]

Mar 28, 2018 20:20:18 GMT -5 Socal Fan said:

Mar 28, 2018 19:56:02 GMT -5 @ben said:

www.ksl.com/?sid=46288489&nid=157&title=new-cars-are-quickly-getting-self-driving-safety-features

Thanks. One line was particularly interesting: "The government says 94 percent of crashes are caused by human error."

That's why I'm so optimistic about self driving cars. They have a very low bar to clear.

   Yes, I noticed that percentage too.  And I encourage optimism, but when there is only one driver,
who happens to be human, of course that would be true.  Who else could be at fault?    

   Blame the remaining 6 percent on bad brakes. 
     

   

[Deleted]

100 percent human error. (Blame it on Louise.)  


                  

[BOGC]

Mar 29, 2018 0:21:18 GMT -5 @ben said:

100 percent human error. (Blame it on Louise.)