Here's the beginnings of a rational approach to the statistics of murder-cars. I haven't been through it in detail yet and it is not definitive
Abstract
The safety of increasingly automated vehicles is of great concern to regulators, yet crash rates are generally reported by manufacturers with proprietary metrics. Without consistent definitions of crashes and exposure, comparing automated vehicle crash rates with baseline datasets becomes challenging. This study investigates the reported on-road crash rates of one manufacturer’s partially automated driving system. Their reported crash rates are adjusted based on roadway classification and driver demographics to allow for direct comparison with the manufacturer’s own advanced driver assistance systems. Recommendations for uniform crash reporting standards are provided.
5 Conclusions
By correcting for roadway usage differences between the Autopilot and active safety only data, much of the crash reduction seen by vehicles using Autopilot appears to be explained by lower crash rates experienced on freeways. Correcting for age demographics likewise produced a 10% increase in the estimated crash rate, although it remains well below the rate of severe crashes in the SHRP 2 NDS.
Several other factors may explain difference in safety rates of new vehicle technologies based on who is using them, where they are being used, and when they are being used. Some safety features cannot be used in rain or snow, for example, which may bias the data towards clear weather and lower crash rates generally.
I think we must consider the parallel problems. The baseline of collisions is way up right now. Locally, we have a bunch of hoopties driving without license plates and crashing in to everybody. We have more stop and go traffic vs gridlock. Coupled with increased cell phone use were seeing a large uptick in serious rear end collusions. Throw the homeless pandemic on top with pedestrians wandering in high speed roadways... There are so many confounding factors going on at the moment that are causing serious upticks in the number of and severity of vehicle incidents. Comparison to data from the last two years is going to be very challenging.
Just this week Tesla recalled a full self-driving software package installed in 54,000 cars. Cars will get a software update. Link
Seems the software would allow for a rolling-stop at stop signs which is illegal in many states. Also it's a bad practice for a human driver to do and can confuse other drivers as to your intent.
I assume that this was NOT an autopilot failure, but I didn't want to start a new thread. This is just too stupid to even contemplate yet there are idiots out there who'll try this kind of stuff, and of course they invited all their friends to film it:
Dangerous Tesla stunt caught on video ends in crash in Echo Park neighborhood, driver on the loose
Possibly, they've made great strides in the last 15 years or so; it's unclear what new problems have been introduced, and whether all the old problems have been eliminated. Waymo's fatal collision in Arizona was only about 4 years ago, and the software clearly had a serious lack of common sense at that time.
TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!
![[pipe]](/data/assets/smilies/pipe.gif "[pipe] [pipe]")
![[cry]](/data/assets/smilies/cry.gif "[cry] [cry]")
