Human driver vs Autonomous solution

93Regina · Post by **93Regina** » 10 Dec 2018, 18:20

12/10/2018 - "A study from the RAND Corporation in 2016 demonstrated
that a vehicle would need to be driven for 275 million failure-free
miles to demonstrate a fatality rate equal to that of a human driver"
==========

10/15/2015 - "...Volvo. Its CEO, Håkan Samuelsson, has set the company
a goal: by 2020, no one should be killed or seriously injured in one
of its new cars. Samuelsson also had some bold advice for other car
makers on the contentious topic of liability. "When you drive
manually, the driver is responsible. When it's automatic, we as the
manufacturer are liable. If you're not ready to make such a statement,
you're not ready to develop autonomous solutions," he said.

Post by **velorider** » 21 Dec 2018, 00:07

The RAND study is misleading. The methodology they are using are used to prove experiment results are true or that defect rates are within normal range. That is not the method to be used to prove the safety of atonomous cars.

UL labs for example, they do not measure failure rates to determine if a device is safe. They test the safety of the device. They perform tests to examine how the device handles extreme situations. Validating the safety of atonomous cars is definitely going to be difficult but the RAND report is misleading. It's definitely not the method that will be used to validate the safety.

93Regina · Post by **93Regina** » 21 Dec 2018, 09:09

velorider wrote: ↑21 Dec 2018, 00:07The RAND study is misleading.

Here's their write-up...RAND_RR1478....

I might look at it later...busy now!

Post by **oragex** » 21 Dec 2018, 09:38

Notice how many studies focus on fatalities, rather than on all the accidents with or without wounds. So it's a limitation that doesn't represent the whole scenario - same with the statistics of car accidents by country

I am a firm believer that autonomous driving will reduce the number of accidents all together by a very large margin. At this point, these systems are not perfect - there were several Tesla fatalities and even more Tesla accidents while driving autonomously.

I like this debate because autonomous driving does not mean 'blind driver' driving because autonomous systems have their clear limitations. What is interesting is that drivers using the autonomous systems will tend to think they have no longer to pay attention. I think it's a psychological issue here. I drive Sonata with the adaptive cruise control that breaks and accelerates to follow the car driving ahead, but the system is useless if the road takes a sharp turn and the car needs to break: if there is no car driving ahead, the cruise control will drive full speed into the turn. I have to say that I had the tendency to forget that the car won't brake for me in such situation. This is to say that psychologically we tend to think these systems will do everything - which of course is very wrong.

By the way, maybe one day in the future when these systems will reach near perfection, that would be a first for actual blind persons 'driving' a car, which was still quite a dream years ago - just a fun fact

93Regina · Post by **93Regina** » 21 Dec 2018, 12:49

oragex wrote: ↑21 Dec 2018, 09:38Notice how many studies focus on fatalities

Quick Scan - "Given that fatalities and injuries are rare events, we will show that fully autonomous vehicles5 would have to be driven hundreds of millions of miles and sometimes hundreds of billions of miles to demonstrate their reliability in terms of fatalities and injuries."

So, crashes are eliminated unless fatalities and injuries are involved.

Non injury crashes do exist on both sides, but governments tend to focus on citizens "well being."

>autonomous driving will reduce the number of accident

I'm sure it could....but implementing cheaper tech that reduces collisions with animals, humans, and vehicles could greatly reduce accidents also.

I'm all in favor of this cheaper tech, that will not depend upon 5G/satellites....and where, the driver is behind the wheel.

93Regina · Post by **93Regina** » 21 Dec 2018, 12:56

oragex wrote: ↑21 Dec 2018, 09:38By the way, maybe one day in the future when these systems will reach near perfection, that would be a first for actual blind persons 'driving' a car, which was still quite a dream years ago -

Google has already done this....for show and tell...

Dec 13, 2016 Google is getting serious about self-driving cars. So serious that it put a legally blind man in one that drove him around safely on his own. The successful trip means that the tech giant can now launch its own self-driving car company, which it's calling Waymo.

: Capturevvv.PNG (526.11 KiB) Viewed 483 times

[youtube][/youtube]