First self driving car fatality

Experts Break Down the Self-Driving Uber Crash

Second… A Tesla driver died last year.

3 Likes

And there was a recent Tesla crash and related death as well:

The goal for self-driving tech isn’t perfection, but rather a marked improvement over human capabilities.

This Uber situation suggests that full autonomy isn’t there yet. Tesla has surely had some failures as well. But it’s come a surprisingly long ways in just a decade or so since the concept was prototyped.

4 Likes

Just watched the two videos taken by the car. The outside and the inside view. Wonder how many people are aware that they are being video’s while inside the car? I understand why it is done, curious if anyone has ridden in one of these how well, if at all, this is made known.

I’m sure the passenger in this accident is enjoying himself on the web.

In fact, our roads are quite safe, really, and have been getting safer despite increases in population, # of drivers, and vehicle miles traveled (VMT):

There is (of course) a certain sensationalism around Tesla and Uber and driveler-less technology, and certain factions that find it beneficial to their interests to highlight and promote all negative information about them.

That said, it is concerning, a bit, that there have been several fatalities given the low number of real world VMT for driver-less cars (i.e. not not including development and testing). Deaths Per VMT (ex driver-less) for 2016 was ~1.18. I am trying to find a decent source for driver-less VMT, but so few mile have been driven by them in real world this is likely to be negligible at this point…hence the concern withe the already 3 (?) fatalities.

2 Likes

Pedestrian deaths going up.

Uber Crash Highlights Growing Safety Concern: Pedestrians

3 Likes

The article states that is pretty much true. Smart phones should be called Darwin Devices

What no song?

2 Likes

Here you go!

https://www.youtube.com/watch?v=Ldyx3KHOFXw

Part of my neighborhood -probably built in the 80s before anything else was around - doesn’t feature sidewalks and has largely eschewed streetlights. It’s always a game of “spot the pedestrian” at night. Curiously, it’s the Boomer demographic that seems to exercise the least sense - dark clothing, distracted by their phone, walking erratically, walking on the right side of the road - as opposed to those incorrigible X’ers and Millennials. Spotted one fella ~fifteen yards out only because I saw white socks leapfrogging along the road.

All of my experience with software tells me developers don’t have enough time or creativity to think of all the edge cases. It’s one thing when your edge case is a specific combination of PC hardware and driver releases, and the resulting error causes a video game to crash. It’s another thing entirely when it causes a self-driving car to crash and kill a pedestrian.

Do those two graphs prove that we are obsessed with outlawing the wrong cell phone behavior? Maybe cell phones in the car are not increasing fatalities while texting and waking is? Should police cite distracted walking?

It’s certainly a problem too. Several of my customers are large universities, they tell me students walk into their side by sides all the time. While they are not moving that is. They tell me they also walk into parked cars & poles.

What is interesting is that before cars the death rate of pedestrian was very high
from horse drwn behicles

https://www.npr.org/sections/thetwo-way/2017/10/25/559980080/honolulus-distracted-walking-law-takes-effect-targeting-phone-users

Perfect. At least that law is based, even if very weakly, on data science that that proves there is a problem to solve.

1 Like

They really need to do something about this distracted walking.

5 Likes

And Tesla’s claimed another life…

Inattentive driver that happened to be driving a Tesla claimed a life.

I’ve read a bit about this wreck. The highway design there is evidently not particularly good, with drivers often failing to realize the “lane” they’re driving in isn’t and they have to exit onto the ramp or merge back into traffic. Tesla drivers in particular report that “Autopilot” has trouble with lane differentiation there. And most damaging in this case, the crash attenuator preceding the Jersey barrier divider was largely absent because it had been crashed into previously; had it been present, odds are there would not have been a fatality.

Tesla’s adaptive cruise control is pretty good, but it’s best used like pretty much all other cruise control - proper highway driving with moderate-to-low traffic desnity, not urban highway driving where sh_t’s happening constantly and the system’s modest processing powers and sensory suite becomes a liability. As the ex-Uber CEO was found to have spoken during discovery in the litigation with Google - “laser is the sauce”, by which laser refers to LIDAR, which is orders of magnitude better than cameras assessing the local environment and making navigation decisions GPS can’t. Tesla doesn’t have LIDAR; it’s presently just too damned expensive with “spinny hat” sensors protruding on vehicle roofs like tumors. Until the sensors get cheaper and less obtrusive, expect level 4/5 vehicle autonomy to remain that elusive 5-10 years away.

Suspect the NHTSA got involved because there’s something to be learned about energetic EV wrecks (due to the small number of EVs on the road, they’re comparably rare) as well as the behavior of level 2 (or level 3 as Tesla drivers seem to feel it is) autonomy and traffic.