autonomous vehicle lane

Car_54

reddwarf2002 wrote: »

If that stat can be relied upon then I can see in the not too distant future where it would be near enough unpractical to drive manually because A) insurance price would be prohibitive or government would introduce so many legislations to deter people from manual driving.

Of course it can't be relied upon. It's a prediction, made four years ago, and based on no experience whatever. More accurately described as a guess.

forgotmyname

Johno100 wrote: »

But my question was would a human controlled driver have detected the hazard and been able to respond in time?

With the cyclist detected at some 6 seconds before impact, I would have expected a driver who was looking at the road to have seen them. If it was as dark as the dashcam i would have said OK a tought one. But it was nowhere near that dark in reality. Proven many times since with true video of the actual light levels. I think it was only doing 40mph? So over 350ft travelled between detection and to the impact.

But that was not an autonomous vehicle, as none are currently commercially available, it was a driver assist function that was being utilised by the driver. I've no idea if that particular system incorporates steering wheel detection. From some of the videos I've seen posted of individuals driving such equipped vehicles and taking their hands off the wheel I'd say many don't.

It was a self drive Tesla, call it driver assist if you want. So why did it not pull over safely when it detected there was no driver??
Yes they know how long the driver had his hands off the wheel.
The system knows that. When the first fatal accident occured in a Tesla, the one where it hit a trailer. They said the driver had his hands off the wheel for x time / x distance.

Why didnt they pickup on the fact that as an assistance device why is it not reacting to irresponsible drivers?

Beenie

Johno, there is evidence that the autonomous car was in the wrong.

The code was written by the Volvo engineers as a 'fail safe' that is it slowed or stopped the vehicle if in doubt. That caused the ride to be slow and frustrating, so it was over-ridden by Uber. That would be fine in 999 cases but in the 1000 it ran over a pedestrian dressed in black crossing the road at night.

Car_54

Beenie wrote: »

Johno, there is evidence that the autonomous car was in the wrong.

The code was written by the Volvo engineers as a 'fail safe' that is it slowed or stopped the vehicle if in doubt. That caused the ride to be slow and frustrating, so it was over-ridden by Uber. That would be fine in 999 cases but in the 1000 it ran over a pedestrian dressed in black crossing the road at night.

But the pedestrian was also in the wrong. Highway Code Rules 3 and 7.

forgotmyname

Highway code? The accident was in America.

You think it's OK to kill cyclists crossing the road? Have you seen the reports? Dont watch the dark dashcam video look at the reports which show the true light levels to be far brighter and the data where the system spotted the cyclist and knew it needed to brake. But it had been disabled and gave the driver no warning to stop.

The driver was looking down at the screen in the dash at the data from the self drive system. Not looking at the road at all.

If you want a guide on how not to do it, I think that incident covers most of it.

Car_54

forgotmyname wrote: »

Highway code? The accident was in America.

You think it's OK to kill cyclists crossing the road?

Fair point, rules may be different in the USA.



But I was talking about killing pedestrians, not cyclists.

forgotmyname

[quote=[Deleted User];74389416]Fair point, rules may be different in the USA.

But I was talking about killing pedestrians, not cyclists.[/QUOTE]


It was a pedestrian pushing a bicycle. 6 seconds from spotting them initially and the point of impact. Took the driver 2 seconds from impact to hitting the brakes. Driver was looking down at the screen.

I wonder what was on the screen though? Self drive system clearly detected them, a good 6 seconds before. Why did the driver not look up or take action in all that time?

AnotherJoe

forgotmyname wrote: »

It was a pedestrian pushing a bicycle. 6 seconds from spotting them initially and the point of impact. Took the driver 2 seconds from impact to hitting the brakes. Driver was looking down at the screen.

I wonder what was on the screen though? Self drive system clearly detected them, a good 6 seconds before. Why did the driver not look up or take action in all that time?

From what I've read, the "driver " was looking at a log of stuff going on. It most likely didn't include the pedestrian witha bicycle since it hadn't worked out what it was yet, but tellingly, whoever created this autonomous system didn't think it worthwhile programming it to avoid (or to learn to avoid) obstacles in the road ! Imagine a learner driver using an excuse "well I didn't know what it was so I thought I'd just carry on and pole right into it"

It's laughable that a few posts up there's someone repeating utter guff that autonomous systems will have an error rate one ten thousandth that of a human when currentiy they can't even "understand" that if there's an obstacle in the road, be that someone pushing a bike, or a bloody great big red stationary fire engine, avoid it !

forgotmyname

It detected the object 6 seconds before the impact. Approx 360feet away.

Not seen the timings for how quickly it went from object to bicycle though, but i imagine it would have been fairly quick. The software knew the object was moving and in a collision path.

Would the display the driver was looking at not have shown the object detected and then the path it was taking?

prowla

I know!


We could give them their own lane, maybe fence it off, maybe we could even put rails along it, so they'd never need to steer...


...then, instead of calling it a motorway, we could call it a railway.


Nah - sounds like a silly idea!