So you’re saying you’ve never driven in a snow storm at night? Some local you turned out to be.
Printable View
Your sarcasm meter needs calibration.
it would take around a week before people figured out that if they saw a self driven car driving, they could cut them off and the computer would automatically jam on the brakes. For lane changes, and just for fun (because that's how most folks are, right?)
Jaywalking around self-driving cars would be a hoot (once they get shit figured out of, this lady getting killed proves it's not that time quite yet).
Yep the really tricky part of the self driving Uber software will be be dialing in the correct level of having learnt to drive in Somalia to it.
Random stops, turns made from wrong lanes, veering wildly across street to pick up some cell phone waving douche bag at opposite curb... that kind of shit.
He just flipped.
Again.
I guess that's the way he rolls.
I know I'd vote for a politician with the emotional fortitude of a middle school girl.
if it was a self driving gun there would be no issues...
Breaking news
https://jalopnik.com/video-shows-dri...s-l-1823970417Quote:
Police in Tempe, Arizona, have released video of a fatal crash involving an Uber-owned self-driving car that fatally struck a woman on Sunday. The clip includes footage from two cameras attached to Uber’s Volvo, and shows the car’s safety driver looking away moments before the collision. The victim, contrary to earlier reports, isn’t seen darting in front of the car.
Includes video of fat guy surfing internets before impact.
Safe to view Danno, cuts out before impact with no blood. Surprised "fat guy" reacts in shock just before their car hits woman.
You naysayers are no different than those who said the horse will always be the dominant form of transportation. Driving a car will become a leisure only activity on select roads and tracks. Like horseback riding.
http://www.saturdayeveningpost.com/2...tomobiles.html
The way I see it the change will be gradual. At first driverless cars will coexist with drivers and bad weather will require human drivers. But then certain expressways with terrible traffic will say no more drivers and only driverless cars that move far more efficiently than human drivers are allowed in certain lanes or during rush hour. Then a city will go to all driverless cars. Once that is successful every major city will follow suit and then the suburbs.
Likewise cars and roads will get better and better at driving in rain, snow, fog and weather. They will use a host of different systems; visual, lidar, GPS, radar, etc. Roads will have sensors that identify lanes, road friction, temperature, etc.
Sent from my iPhone using TGR Forums
So is it propaganda or real? Some big money at stake here.
This is a situation--pedestrian in dark clothes on unlighted road--where the car should have detected the woman sooner than a human driver. People keep saying autonomous cars will save lives like it's a fact--because people are so shitty at it. It remains to be seen whether autonomous cars are better.
Bigger obstacle--we were talking about a new roundabout in Truckee and people were talking about how while cars entering the roundabout normally have to yield to cars in the roundabout, when traffic is wall to wall the de facto mode is zipper merging. That's condoned but technically illegal. How does a car with no manual override or even any human in the vehicle (Uber on the way to a pickup) handle that? Sit waiting to enter the roundabout for hours until traffic finally eases?
I'm sure every technical obstacle can be overcome, in theory. The obstacle, as I said before, is money. Lot's of countries have high speed rail, we can't afford even old fashioned slow rail. Now they're talking about dropping the California Zephyr between SF and Chicago, through Truckee--adding a few more cars to the Tahoe traffic mess.
Why anyone thinks this tech needs to be "tested" by unleashing it around unsuspecting pedestrians is beyond me. I'm fully onboard that technical challenges can be overcome, but can you name any other potentially lethal tech that is prototyped and tested on the general public like this? The level of testing that other vehicles have to go through before being allowed into service is dramatically higher, even when they're basically copies of existing designs. I always assumed the regulatory hurdles would be too high to make autonomous profitable, but lowering the bar to make it easier was an even worse idea.
Basic stuff like making sure the cameras and sensors can "see" in all the conditions that human eyes see (Tesla) is a huge challenge and not something that needs an actual autonomous car on public roads to find out. How many other corners have been cut? It shouldn't be that hard to create regimes to test the sensors and the logic in a safe environment. Instead there's a rush to rack up massive numbers of highway miles (while making sure the driver is alert) to reinforce the "success" of a system. Is that by comparing it to all miles driven by humans, or to highway miles?
Jesus, massive fail for everyone concerned (including whoever was responsible for there being a median path that just plain ends far short of the intersection, by the way).
The ped: She’s crossing a boulevard, in the dark, completely oblivious to at least one vehicle that was headed her way. I don’t have a beef with jaywalking; I do it all the time. But I don’t even trust car traffic to stop for me when I’m in a crosswalk and the sign says walk. When I don’t have right-of-way, I wait for a clean gap and don’t assume anyone’s going to slow down even a little. Walking out there totally oblivious like she’s at a city park, IN THE DARK, come on, WTF.
On the other hand, dumb pedestrians are nothing new; driverless cars are. Uber’s the one with something to prove here, and they failed. Even though the car had right-of-way and the ped was an idiot.
She didn’t just dart out of nowhere; she’s taking a casual stroll, and is now crossing the third lane of her slow ramble. First, the video camera’s dynamic range surely understates what was actually visible, and second, if not, a slowly walking woman with bike, broadside direction, would have been visible with high beams. Which is beside the point (but not quite, because part of what I’m saying is that a human probably would have had time to do something to try to avoid hitting her), because a night driving computer system shouldn’t be relying only on visible information. A human driver would have an excuse, but an automated system has no excuse for not sensing something that’s steadily approaching the car’s path within a period of time adequate for at least swerving in time (apparently the car did nothing before impact).
And yeah, that “driver” or chaperone or whatever was kind of useless. Way to hire the best there, Uber.
This should have been a relatively easy accident for software to avoid. How are these guys going to do something actually challenging, like dropping off a fare in front of Costco?
Noticed something else in the video: There are streetlights along that section of road. Granted, they're just dots of light amidst a pitch-black surround that envelops everything except about 50 feet or so in front of the car's low-beams, but there are functioning streetlights. My hunch is that the streetlights illuminate things better than zero amount that the video suggests.
After watching the video it makes me wonder if any current avoidance technology would have done better?
Seems like the Uber car should have performed better in detecting that woman. She didn't seem to be moving very fast.
Ill just keep driving my car. I've never run anyone over.
Howany people got killed by cars with drivers ?
I think that was a deer the other night.......
Not really. He came in to let everyone know he was smarter than them and when no one agreed, he realized that he couldn't project his mind's echo chamber onto TGR as a wider audience. It also didn't work in his last election but people with long egos have to have short memories so I bet he's off to see if the real world will buy his brand again. As brit said, he will be back here again too to tell us why the real world is dumb for not having done so. Hell, he's reading right now.
The current rate for ped and cyclist fatalities, by car, in the US is 1.25 per 100 million miles driven. One incident doesn't establish a statistical record, but let's just say that this incident gives self-driving cars (and Uber in particular) a pretty bad head start on whatever statistical record they end up establishing.
WSJ reporting Uber "safety driver" involved in fatal accident had criminal record.
For a company that has already had issues with it's background check policy for drivers.. that's going to hurt.
Homeless drug addict dies from walking into the path of robot car carrying transgender armed robber… welcome to the future, Earthlings.