Waymo robotaxis have now driven 1 million miles autonomously with no recorded injuries:
[...] Robotaxi company Waymo has released a trove of data that could help convince pessimists that fully autonomous vehicles are safer than those controlled by humans. The Alphabet subsidiary announced that its taxis reached the milestone of driving one million miles in January without anyone behind the wheel, and there were no reported injuries or deaths throughout that time.
In addition to nobody being hurt as Waymo racked up one million miles of autonomous driving, there were only two incidents that meet the criteria for inclusion on the National Highway Safety Transportation Administration's database for car crashes, called the Crash Investigation Sampling System (CISS). This criteria includes the need to file a police report or one or more vehicles being towed away.
Those incidents weren't the fault of the self-driving systems. The more serious of the two involved a Waymo taxi being hit from behind by a vehicle whose driver was looking at their cell phone while approaching a red light.
The other incident happened when a car pulled into the same lane as a Waymo taxi and suddenly hit its brakes, causing the self-driving car to go into the back of said vehicle as the taxi didn't have enough time to slow down.
There were an additional 18 minor contact events that did not meet the criteria for the database, including a car backing out of a parking space into a stationary Waymo waiting to pick up a rider, and a portable plastic sign stand getting blown by the wind and making contact with a vehicle. Waymo said that more than half of all these contact events were the result of a human driver hitting one of its stationary taxis.
"Despite 24/7 driving across major U.S. cities, Waymo experienced no collisions at all of the types that are responsible for 94% of fatal collisions in NHTSA's crash investigation database," the company writes.
(Score: 3, Interesting) by Runaway1956 on Friday March 03, @11:20PM (4 children)
Tens of thousands, maybe hundreds of thousands, American truck drivers can make the same claim to fame. In fact, I can claim almost 5 million miles, with no injuries. One million safe miles doesn't cut it. A hundred million will start to convince me how safe they are.
Abortion is the number one killed of children in the United States.
(Score: 3, Insightful) by SomeRandomGeek on Friday March 03, @11:53PM (3 children)
Depends on what you want from the numbers. To know how Waymo compares to a typical human driver, I agree that you would need at least a hundred million miles of data. But the million miles that they already have is enough to say that Waymo isn't considerably more dangerous than a typical human. If the Waymo driver has serious accidents at ten times the human rate, there would have been multiple accidents that were Waymo's fault by now. To show that the Waymo driver is a lot safer than a typical human driver (which appears to be the case based on simulation data) will take a lot more road miles.
(Score: 4, Insightful) by Runaway1956 on Saturday March 04, @12:05AM (2 children)
We apparently read that title quite differently. Waymo has driven a million miles without injuring anyone. That doesn't necessarily mean that it has had no accidents. I had a couple accidents during my career, but I never injured anyone. Food for thought there, huh?
Abortion is the number one killed of children in the United States.
(Score: 5, Informative) by SomeRandomGeek on Saturday March 04, @12:23AM
Waymo is extremely transparent with their safety data. https://blog.waymo.com/ [waymo.com]
They have had a total of 20 accidents, which seem to generally fall under the category of not very serious and also someone else's fault.
I found this post in particular to be very interesting: https://blog.waymo.com/2022/09/benchmarking-av-safety.html [waymo.com]
(Score: 0) by Anonymous Coward on Sunday March 05, @04:51AM
Maybe not a million miles but my numbers can be added to other human drivers like us who haven't injured anyone.
And I'm far from the best or safest driver.
(Score: 5, Informative) by Anonymous Coward on Friday March 03, @11:35PM
Google tells me that in the USA there are something around 3.2 trillion miles driven per year.
It also says there are 4.8 million medically consulted injuries (this includes deaths, I believe).
Doing the division gives the delightful result of 666,666 miles between injury accidents. It's a really big system, and looked at this way, quite safe.
Best I can tell, Waymo started AV taxi service in 2020, so let's say that million miles was driven over three years. That's 333,333 miles/year, so if their driving is "average" then they are about halfway to their first injury accident (statistically).
I swear that I didn't jigger the numbers to come out so neatly, I really did just grab data from searches and round to a couple of places.
Here's the punch line: My demographic is likely about 10X "safer" than average -- don't drive impaired, don't use a cell phone, not a teen, and not decrepit yet.
I'll start thinking about a Waymo ride after they complete a few years with 6,666,660 miles, with no injury accidents. That's what I estimate it takes to match my current odds.
(Score: 5, Interesting) by edinlinux on Saturday March 04, @12:04AM (2 children)
Are these miles random all over random roads (in varying conditions and weather), or just driving over and over on the same 'curated' list of roads that it is already known the autonomous driving software can handle correctly?
(Score: 3, Insightful) by Anonymous Coward on Saturday March 04, @12:50AM
Pretty sure it's the latter. Their demonstration projects are confined to specific parts of a city.
(Score: 2) by acid andy on Saturday March 04, @02:48PM
Kinda like an evolution of the tram...
Master of the science of the art of the science of art.
(Score: 2) by PiMuNu on Saturday March 04, @10:40AM (3 children)
Highways are considerably safer than back roads, for example.
But it is a good statistic
(Score: 0) by Anonymous Coward on Saturday March 04, @02:10PM (2 children)
> But it is a good statistic
Um, no. Given the size of the system, a million miles is a drop in the bucket. Waymo needs many times that mileage with no injury accidents to convince me. See several previous posts that put this in perspective.
(Score: 2) by Freeman on Monday March 06, @05:30PM (1 child)
Also, unlike Tesla's "full" sellf-driving, Waymo is going over a specific set of roads. It's not likely that Waymo's current data will 100% translate to "drive anywhere".
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(Score: 0) by Anonymous Coward on Monday March 06, @10:01PM
Speaking of this, did Waymo keep operating last week(??) when most of California had some snow? In general, I believe Waymo has restricted their demonstrations to areas without winter weather conditions.