Toyota has suspended trials in America following Monday's tragedy. A self driving Range Rover drove through Milton Keynes yesterday,parked and then returned to its starting point in a government funded trial.
Ive seen the dash cam footage and it was the fault of the pedestrian.
Yeah i have seen the footage, its an immensely stupid place to cross in the dark! RIP to the pedestrian though, very sad
Very sad. You wonder what percentage of drivers would have stopped in time. I assume the driver in the car had a manual override? Though she probably wasn't 100% concentrating
Whilst the pedestrian was in the wrong, it's a well lit, flat, straight street. The video is poor quality and doesn't accurately represent what a human eye would see. There's no way a human driver, paying attention, wouldn't have stopped in time.
That's all moot anyway, the car is fitted with LIDAR, it should be able to spot a pedestrian even if they were wearing black, there was no lighting and it was foggy. This was a systems failure as the car doesn't appear to have even slowed, let alone stopped or swerved.
Those who like a conspiracy theory are suggested the video was doctored to make it look darker and make the pedestrian harder to spot, whilst the original police statement blaming the victim for dashing out from behind bushes was nothing more than a disgusting lie.
Let me guess, judging by the title of this thread you don't like Uber?
Or maybe, the OP doesn't like people dying on the road, the thread title just summarises what happened, just a thought.
In that case it should read "Self driving car involved in fatal accident with pedestrian". Just a thought.
But it was a self driving uber vehicle whether you like them or not?
Anyway forgetting all that. She was crossing four lanes. The difference with a human driving vehicle would be for example.
You are in the fourth lane a few hundred yards away from someone crossing. You are getting closer and you spot someone crossing further ahead so you slow down. The pedestrian is then into the 2nd lane so you have most probably spotted the pedestrian. The pedestrian is then in the 3rd lane so you are more than likely to stop/maybe toot your horn.
A self driving vehicle is not going to spot someone crossing a few hundred yards ahead and slow down or stop unless they are directly in front of the self driving vehicle.
Let me guess, judging by the title of this thread you don't like Uber?
Or maybe, the OP doesn't like people dying on the road, the thread title just summarises what happened, just a thought.
In that case it should read "Self driving car involved in fatal accident with pedestrian". Just a thought.
But it was a self driving uber vehicle whether you like them or not?
Anyway forgetting all that. She was crossing four lanes. The difference with a human driving vehicle would be for example.
You are in the fourth lane a few hundred yards away from someone crossing. You are getting closer and you spot someone crossing further ahead so you slow down. The pedestrian is then into the 2nd lane so you have most probably spotted the pedestrian. The pedestrian is then in the 3rd lane so you are more than likely to stop/maybe toot your horn.
A self driving vehicle is not going to spot someone crossing a few hundred yards ahead and slow down or stop unless they are directly in front of the self driving vehicle.
Makes sense to me!
I guess the only way to make roads reasonably safe for pedestrians - particularly from self driving cars - is to make sure they are fenced off like railway lines. This will probably be the future for our towns and cities with cars and people always separated.
The 'driver' should have been paying more attention and the car's radar should have picked up the cyclist in the dark but no human driver would have been able to avoid the crash imho.
That's not how self-driving cars (are supposed to) work. They scan all around them and do real time risk assessment against anything that could move into their path. It could be possible the pedestrian paused whilst crossing and was wrongly identified as a stationary object. We'll know more if/when the logs are released. If reports are to be believed the car never slowed, which would suggest a fault in either the software or hardware.
Seen the footage, beyond tragic but the woman was crossing a fast multi lane road wearing mostly black without looking in the pitch black. At the very best a human driver would have swerved and potentially done a huge amount of damage to themselves in the process.
Let me guess, judging by the title of this thread you don't like Uber?
Or maybe, the OP doesn't like people dying on the road, the thread title just summarises what happened, just a thought.
In that case it should read "Self driving car involved in fatal accident with pedestrian". Just a thought.
Your response needs more thought. As I said it doesn't give all the facts does it. Or are you an Uber driver?
The main story here is about the driverless car, not who owns it.
Uber don't just own it, they helped build it and develop the technology, as well as running the testing program and hiring the "driver" in the car at the time. This is very much about Uber as much as one of Google's autonomous cars crashing would be about Google.
Presumably driverless cars don't need lights. Wonder if the lights were unusually low and surprised the pedestrian who was used to seeing cars out of the corner of her eye?
1 person dies from a driverless car. 400 probably died from human-driven cars on the same day.
Except that driverless cars (in total) have not yet even driven a million miles yet have registered a death, whilst the typical fatality rate in the West is one death per 100 million miles driven.
Moreover driverless cars are most likely to be rolled out in a meaningful way in developed market city centres (where the fatality rate is extremely low compared to rural areas or the developing world).
1 person dies from a driverless car. 400 probably died from human-driven cars on the same day.
Except that driverless cars (in total) have not yet even driven a million miles yet have registered a death, whilst the typical fatality rate in the West is one death per 100 million miles driven.
Moreover driverless cars are most likely to be rolled out in a meaningful way in developed market city centres (where the fatality rate is extremely low compared to rural areas or the developing world).
Google’s driverless cars alone have driven more than 2 million miles.
1 person dies from a driverless car. 400 probably died from human-driven cars on the same day.
Except that driverless cars (in total) have not yet even driven a million miles yet have registered a death, whilst the typical fatality rate in the West is one death per 100 million miles driven.
Moreover driverless cars are most likely to be rolled out in a meaningful way in developed market city centres (where the fatality rate is extremely low compared to rural areas or the developing world).
Google’s driverless cars alone have driven more than 2 million miles.
...mainly on city streets, precisely where they're not needed from a safety standpoint (though there are other potential benefits of course).
I've said this for sometime that driverless cars (in the form cars are today) won't happen (for a long time) its not just the technological issues, its the litigation issues and responsibility for a deadly machine which a car potentially is, the state of the roads, etc, that will prevent it. Some car companies cant even get sat navs to work properly.
Driverless cars will learn from their mistakes. Like the aviation industry, each accident is a lesson learned.
In aviation there are lessons from every incident which are shared and become required practice by every airline subsequently. Matthew Syed's book 'Black Box Thinking' is a good read on this topic.
Meanwhile the companies developing driverless cars are involved in an arms race where any useful nugget of information (relating to safety or otherwise) is deemed proprietary.
Comments
A self driving Range Rover drove through Milton Keynes yesterday,parked and then returned to its starting point in a government funded trial.
Whilst the assurances out there would make it seem this shouldn't happen - There is no way a human wouldn't have killed her.
That's all moot anyway, the car is fitted with LIDAR, it should be able to spot a pedestrian even if they were wearing black, there was no lighting and it was foggy. This was a systems failure as the car doesn't appear to have even slowed, let alone stopped or swerved.
Those who like a conspiracy theory are suggested the video was doctored to make it look darker and make the pedestrian harder to spot, whilst the original police statement blaming the victim for dashing out from behind bushes was nothing more than a disgusting lie.
Anyway forgetting all that. She was crossing four lanes. The difference with a human driving vehicle would be for example.
You are in the fourth lane a few hundred yards away from someone crossing. You are getting closer and you spot someone crossing further ahead so you slow down. The pedestrian is then into the 2nd lane so you have most probably spotted the pedestrian. The pedestrian is then in the 3rd lane so you are more than likely to stop/maybe toot your horn.
A self driving vehicle is not going to spot someone crossing a few hundred yards ahead and slow down or stop unless they are directly in front of the self driving vehicle.
I guess the only way to make roads reasonably safe for pedestrians - particularly from self driving cars - is to make sure they are fenced off like railway lines. This will probably be the future for our towns and cities with cars and people always separated.
Presumably driverless cars don't need lights. Wonder if the lights were unusually low and surprised the pedestrian who was used to seeing cars out of the corner of her eye?
Moreover driverless cars are most likely to be rolled out in a meaningful way in developed market city centres (where the fatality rate is extremely low compared to rural areas or the developing world).
Meanwhile the companies developing driverless cars are involved in an arms race where any useful nugget of information (relating to safety or otherwise) is deemed proprietary.