3 ms·
Does anyone have good numbers for, say, deaths per mile driven in urban environments, and for the number of miles driven by driverless cars? The article says:
by mcguire 3y ago
Does anyone have good numbers for, say, deaths per mile driven in urban environments, and for the number of miles driven by driverless cars?
The article says:
"Both Cruise and Waymo say their vehicles are far safer than human drivers and compared to humans they've had relatively few incidents. They say they've driven millions of driverless miles without any human fatalities or life-threatening injuries."
which is about all the detail I've seen since Tesla's sketchy press release a few years back.
Here's the deal: California has 1.4 deaths per 100,000,000 vehicle miles traveled; "millions of miles" is a drop in the statistical bucket. But on the other hand, that's California overall while the autonomous vehicles are largely in urban areas, right?
The article claims, between Cruise and Waymo, there are roughly 300 AVs. If they average 30mph, that's 9000 miles per hour or 80M miles per year. That would be about 1 expected death per year, assuming the California statistics.
https://www.iihs.org/topics/fatality-statistics/detail/state-by-state https://www.iihs.org/topics/fatality-statistics/detail/state...
- throwawaylinux 3y ago30mph across the entire fleet, 24x7? I don't live in a city with them but that sounds optimistic. 30mph seems like a decent average in a big city when you're on the move. Add stops to recharge, pick and drop fares, maintenance, and down time between fares, I'd not be surprised if it was half that or less.
- culopatin 3y agoIn sf you rarely hit 30mph unless all you do is the highway and the bridges. Otherwise speed limit is often 25 and these cruise and way nos drive on the safe side, so their average is probably in the tens of mph
- oatmeal1 3y agoIntuitively I'd believe it. They are programmed to be very cautious around pedestrians and cyclists, and drive the speed limit.
- Legend2440 3y agoI don't think there's enough public data to make a good comparison on safety. Only the companies and presumably the NTSB have that data. As you mentioned, an apples-to-apples comparison is hard. The type of driving (urban vs freeway), the location, the weather conditions, even the times of day they drive could all bias accident rates.
- bastawhiz 3y agoI suspect if they wanted, Cruise and Waymo could easily have their fleet shuttle back and forth from SF to LA to just rack up miles, but exactly nobody would be impressed. A huge percentage of accidents occur around intersections, and intersections are rare on the longest stretches of road. I wish it was possible to measure safety as incidents per stops. That is, the number of accidents that occur per number of times the vehicle comes to a full stop. It's not perfect but it gives another good signal.
- mint2 3y agoEgads that would be something - cars slamming on their brakes going 25 in the city and then blocking traffic is one thing. now imagine it happens at 70 mph on i5 because a tumble weed blows by, of which I’ve seen plenty, and it takes the company 3 hours for the driver to get out there. Or similar for the usual construction temp lane closer. There’s a reason they don’t test on I-5 with their current tech.