Self driving buses, trains, trucks etc

05-06-2019 8:45pm

on autopilot

https://i.imgur.com/G0GZuN1.mp4

magicbastarder · 03-10-2019 8:11am

Collision course: pedestrian deaths are rising – and driverless cars aren’t likely to change that
For drivers, roads are safer than ever – but for people on foot, they are getting deadlier. Car companies and Silicon Valley claim that they have the solution. But is that too good to be true?
https://www.theguardian.com/technology/2019/oct/03/collision-course-pedestrian-deaths-rising-driverless-cars

cnocbui · 03-10-2019 10:45am

magicbastarder wrote: »

Collision course: pedestrian deaths are rising – and driverless cars aren’t likely to change that
For drivers, roads are safer than ever – but for people on foot, they are getting deadlier. Car companies and Silicon Valley claim that they have the solution. But is that too good to be true?
https://www.theguardian.com/technology/2019/oct/03/collision-course-pedestrian-deaths-rising-driverless-cars

I learned to drive in Australia. Pedestrians have right of way. They are obliged to use a pedestrian crossing if within 20m of it.

In the US, the car manufacturers lobbied the government to create the crime of J walking - not crossing a road except at a crosswalk or traffic lights - period. They also gave drivers some legal immunity regarding injuring/killing pedestrians:

Drivers who obey traffic laws and report an accident that includes hitting a pedestrian, will also likely receive an exception. Pedestrians who cross streets illegally will also allow you to avoid consequences for hitting them.

https://www.thetrafficticketattorneys.com/blog/what-to-expect-if-you-hit-a-pedestrian/

There are 3 times as many pedestrian deaths per head of population in the US than in Australia.

The US don't seem to actually have as much will to reduce road accident fatalities as other countries. Between 1990 and 2013, their road fatalities only declined by 36%. Switzerland and Germany, Slovenia and Austria managed 78% reductions.

If silicon valley has a plan, it's primary aim will be to line their pockets with money. The US should basically be ignored when it comes to what they say about road safety. They an appalling track record, an appalling attitude and are about the least expert country you will find when it comes to road safety.

cnocbui · 03-10-2019 10:24pm

We need automation to save lives because humans are bad drivers. The sooner we get more automation the sooner there will be no traffic accidents or deaths.

Or maybe not...

New safety features being rolled out by auto makers to keep drivers from hitting pedestrians don’t work at times in some of the most dangerous situations and frequently fail at night, according to a new study by AAA.

Testing performed by the association found that pedestrian-detection technology offered in four different models performed inconsistently and didn’t activate properly after dark, when many roadway deaths occur.

The uneven performance highlights the challenges the auto industry faces as it looks to automate more of the car’s driving functions and roll out new crash-avoidance technologies that rely on sensors and software to detect road hazards.

“Pedestrian fatalities are really becoming a crisis,” said Greg Brannon, AAA’s director of automotive engineering. While such pedestrian-detection systems have the potential to save lives, drivers shouldn’t become overly reliant on them to prevent accidents, Mr. Brannon said.

Like other advanced safety features becoming more widespread, pedestrian-detection technology uses cameras, radar and other sensors to identify people in the vehicle’s path and alert drivers to the danger ahead. If the driver doesn’t react quickly enough, the car can brake for them.

...

Nearly 6,000 pedestrians were killed in U.S. traffic accidents in 2017, the latest year data were available, according to the National Highway Traffic Safety Administration. That was up 35% from 2008, even though fatalities involving vehicle occupants were down 7% in the same period, NHTSA data show.

For the AAA study, testers picked four sedan models—the Chevrolet Malibu, Honda Accord, Toyota Camry and Tesla Model 3—and put the cars through scenarios meant to replicate some of the most dangerous situations for pedestrians. One test, for instance, simulates a child darting out from between parked cars, and another involves an adult crossing the road as the vehicle turns right.

At 20 miles an hour, the cars struggled with each test, AAA found. The child was struck 89% of the time, and all of the cars hit the pedestrian dummy after making a right turn. The systems were generally ineffective if the car was going 30 mph. The systems were also completely ineffective at night, Mr. Brannon said, the deadliest time for pedestrians. Three-quarters of all pedestrian fatalities occur after dark, according to AAA.

https://www.wsj.com/articles/new-cars-pedestrian-safety-features-fail-in-deadliest-situations-study-finds-11570075260

These semi autonomous technologies are dangerous. There are just too many gadgeteers who think technology is miraculous and near infallible and seemingly have utter faith in it, and so grossly overestimate its actual performance. It doesn't matter how much you warn people that it's not perfect and to be vigilant, a significant percentage will think the warnings are just ass-covering for legal reasons and that the real truth is that they are near perfect, and will act accordingly.

03-10-2019 10:51pm

cnocbui wrote: »

These semi autonomous technologies are dangerous. There are just too many gadgeteers who think technology is miraculous and near infallible and seemingly have utter faith in it, and so grossly overestimate its actual performance. It doesn't matter how much you warn people that it's not perfect and to be vigilant, a significant percentage will think the warnings are just ass-covering for legal reasons and that the real truth is that they are near perfect, and will act accordingly.

There's no accounting for stupid

Still not a good enough reason to stop progression of this tech, in fact it reinforces the fact that development should be continued

magicbastarder · 03-10-2019 10:54pm

oh, i think the technology will keep getting better and better, there's no doubt in my mind about that.
what i find funny is the belief that this will somehow lead to a lessening of the car's impact upon society. we should be using technology to get rid of cars, not to enhance them.

cnocbui · 25-10-2019 12:35pm

This is absolutely wonderful. Confirmation that all those who paid €6,000 for the Full Self Driving option for their Tesla on the promise it could be used as a driverless robo taxi, and thus earn back some or all of the cost of the vehicle, have been well and truly Musked.

So no fully autonomous within the lifetime of any existing Teslas, so no paying off the car by pimping it out for robo rides.

Musk:
Yeah, feature-complete, I mean, it’s the car able to drive from one’s house to work, most likely without interventions. So it will still be supervised, but it will be able to drive — it will fill in the gap from low-speed autonomy with Summon. You’ve got high-speed autonomy on the highway, and intermediate speed autonomy, which really just means traffic lights and stop signs.
...
So feature-complete means it’s most likely able to do that without intervention, without human intervention, but it would still be supervised. And I’ve gone through this timeline before several times, but it is often misconstrued that there’s three major levels to autonomy. There’s the car being able to be autonomous, but requiring supervision and intervention at times. That’s feature complete. Then it doesn’t mean like every scenario, everywhere on earth, including ever corner case, it just means most of the time.
...
While it’s going to be tight, it still does appear that we will be at least in limited early access release of a feature-complete Full Self-Driving feature this year. So, it’s not for sure, but it appears to be on track for at least an early access release of a fully functional Full Self-Driving by the end of this year.
...
And then, there’s another level which from a Tesla standpoint, we think the car is safe enough to be driven without supervision. Then the third level would be that regulators are also convinced that the car can be driven autonomously without supervision. Those are three different levels.

cnocbui · 31-10-2019 8:08pm

Do you remember those boardsies saying driverless cars would be here any minute and after they had taken over there would be no more road fatalities?

I said they were wrong, and that general AI was likely needed:

Jim Hackett, the boss of Ford, acknowledges that the industry “overestimated the arrival of autonomous vehicles”. Chris Urmson, a linchpin in Alphabet’s self-driving efforts (he left in 2016), used to hope his young son would never need a driving licence. Mr Urmson now talks of self-driving cars appearing gradually over the next 30 to 50 years. Firms are increasingly switching to a more incremental approach, building on technologies such as lane-keeping or automatic parking. A string of fatalities involving self-driving cars have scotched the idea that a zero-crash world is anywhere close. Markets are starting to catch on. In September Morgan Stanley, a bank, cut its valuation of Waymo by 40%, to $105bn, citing delays in its technology.

The future, in other words, is stuck in traffic. Partly that reflects the tech industry’s predilection for grandiose promises. But self-driving cars were also meant to be a flagship for the power of AI. Their struggles offer valuable lessons in the limits of the world’s trendiest technology.

https://www.economist.com/leaders/2019/10/10/driverless-cars-are-stuck-in-a-jam

Del.Monte · 31-10-2019 8:13pm

But, but, Ireland's different to everywhere else, we will have self-driving, free powered electric cars everywhere in the New Year.

magicbastarder · 18-11-2019 1:16pm

a researcher in the US provided chauffeurs to 13 test subjects (to see how their car use would change if they had access to cars which were 'self driving') and the results were a clear increase in traffic on the roads:

https://jalopnik.com/zombie-miles-and-napa-weekends-how-a-week-with-chauffe-1839648416

18-11-2019 2:02pm

magicbastarder wrote: »

a researcher in the US provided chauffeurs to 13 test subjects (to see how their car use would change if they had access to cars which were 'self driving') and the results were a clear increase in traffic on the roads:

https://jalopnik.com/zombie-miles-and-napa-weekends-how-a-week-with-chauffe-1839648416

Ní surprises there, sure if I had one, I would go to the pub in town and it would be sent to bring the adult children home after a night out.

magicbastarder · 22-11-2019 12:23pm

i know it's not necessarily a self-driving car - but has elon musk finally jumped the shark?

https://twitter.com/etienneshrdlu/status/1197804644224950278

markodaly · 23-11-2019 1:44am

cnocbui wrote: »

Do you remember those boardsies saying driverless cars would be here any minute and after they had taken over there would be no more road fatalities?

Care to link us the posts which stated exactly that?

LoughNeagh2017 · 24-11-2019 1:49am

It couldn't handle the roads of Ulster, you need to slow down frequently when you see a big crater hole in the road as well as the giant puddles. The tyres would be burst after every drive

30-12-2019 3:19pm

Another reason to be aware of the flaws in self driving technology.

https://www.facebook.com/callumcraddy/videos/10156493314356857

EDIT: After hearing from a lot of VAG owners, apparently their anti collision system is known to malfunction and stop a car for absolutely no reason in light changing conditions? WHY HAVE THEY BEEN ALLOWED TO STAY ON THE ROAD IF THIS IS THE CASE? And if this happened to you.. Would you have continued on with your journey like these people did seconds later?

This was a fun one a few hours ago passing Wetherby... NOT. What an absolute asshat. Bloody good thing there was a hard shoulder as I wasn't stopping 40ton in 150 yards on a motorway! (Cam lense makes it look further). Also had a wagon overtaking me on the middle lane! They overtook 2 mins later again as if nothing had happened. No acknowledgement or anything. Female in the passenger side still with her legs up on the dashboard playing on her phone ��*♂️ also a good thing I had a powder tank rather than an un-baffled liquid else it could have been worse for me avoiding them!

Unfortunately can't quite get the reg. It was reported over the phone whilst I remembered it at the time but I can't remember it exactly now I've parked up. Share the **** out of this. Learn to ****ing drive people! **** like this, and worse, is happening far too often nowadays. The driving standard across the UK is worsening rapidly.

Sounds like a collision creating application!

magicbastarder · 27-06-2020 12:29pm

interesting video about the issue of camera frame rates and use of LED lights, especially if the camera is used by an autonomous vehicle:

https://www.youtube.com/watch?v=c6NFuJMfmYU

magicbastarder · 03-01-2021 10:14pm

'Peak hype': why the driverless car revolution has stalled
As Uber parks its plans for robotaxis, experts admit the autonomous vehicle challenge is bigger than anticipated
https://www.theguardian.com/technology/2021/jan/03/peak-hype-driverless-car-revolution-uber-robotaxis-autonomous-vehicle

03-01-2021 11:47pm

Surprised it has taken so long for that penny to drop, AV's will succeed on roads where many of the variables can be controlled and have extensive telemetry to provide the vehicles with precise information on road layout and the like (like a railway route finder).
Plus of course no other vehicular traffic (or pedestrians) that tends to violate traffic rules.
Motorways, should be safe for AV's, with the exception that the AI may not be capable of predicting the habits of certain types of moronic drivers, those who switch four lanes to leave at the exit they've almost just passed for example.

cnocbui · 04-01-2021 12:21am

Assuming rationality, about the only possible reason Tesla's market cap could be more than that of the seven largest car makers put together, particularly when they only make a profit by selling green credits, is the belief that Musk was being near enough truthful when he said Teslas would all soon be fully autonomous and robo-taxi capable by the end of 2020.

Elon Musk expects Tesla to have produced a fleet of “robo-taxis” that don’t need a driver by the end of next year.

https://www.thetimes.co.uk/article/get-ready-to-turn-your-tesla-into-a-robo-taxi-says-musk-drxm9lt9b

That is the only thing that could possibly deliver a revenue stream sufficient to justify the valuation. And they would have to have exclusivity on the tech, also.

I think fully autonomous vehicles are not going to happen, because the task is too complex for the not-AI that is currently touted as AI. I think driving requires general intelligence, something which if ever achievable artificially, is far further out than viable commercial scale fusion energy generation.

The schilling of autonomous driving for superior safety is noxious, It's stated as factual that AD will be safer than humans. It's a complete nonsense. Most of the drivers who are, or who have ever been, never cause the death of someone because of an error in driving. Their performance is perfect, something which can not be improved upon. The same can not be said for complex software. Even the comparatively simple software that caused the Toyata such problems with their braking system was too complex to ever enable the fault to be found, even after hiring some of the best people possible and spending millions.

The non-AI AD systems that learn and are trained are essentially black-boxes. They develop a complex algorithm, but it can't be comprehended by humans. If mass deployment of not-AI AD systems ever happens, and there are fatalities, it's almost certain that the cause will be beyond the comprehension of the system creators and be unfixable.

04-01-2021 2:32am

cnocbui wrote: »

Assuming rationality, about the only possible reason Tesla's market cap could be more than that of the seven largest car makers put together, particularly when they only make a profit by selling green credits, is the belief that Musk was being near enough truthful when he said Teslas would all soon be fully autonomous and robo-taxi capable by the end of 2020.

https://www.thetimes.co.uk/article/get-ready-to-turn-your-tesla-into-a-robo-taxi-says-musk-drxm9lt9b

That is the only thing that could possibly deliver a revenue stream sufficient to justify the valuation. And they would have to have exclusivity on the tech, also.

I think fully autonomous vehicles are not going to happen, because the task is too complex for the not-AI that is currently touted as AI. I think driving requires general intelligence, something which if ever achievable artificially, is far further out than viable commercial scale fusion energy generation.

The schilling of autonomous driving for superior safety is noxious, It's stated as factual that AD will be safer than humans. It's a complete nonsense. Most of the drivers who are, or who have ever been, never cause the death of someone because of an error in driving. Their performance is perfect, something which can not be improved upon. The same can not be said for complex software. Even the comparatively simple software that caused the Toyata such problems with their braking system was too complex to ever enable the fault to be found, even after hiring some of the best people possible and spending millions.

The non-AI AD systems that learn and are trained are essentially black-boxes. They develop a complex algorithm, but it can't be comprehended by humans. If mass deployment of not-AI AD systems ever happens, and there are fatalities, it's almost certain that the cause will be beyond the comprehension of the system creators and be unfixable.

That's a lot of nonsense for 1 post, too much to pick apart at this time of night, might come back to it tomorrow

magicbastarder · 04-01-2021 3:34am

cnocbui wrote: »

Most of the drivers who are, or who have ever been, never cause the death of someone because of an error in driving. Their performance is perfect, something which can not be improved upon.

You have a definition of perfect waaaay removed from the one I'd use.

cnocbui · 04-01-2021 2:14pm

I think it's wrong to lump all drivers together and to say something along the lines of 'humans are bad at driving' not-AI ones will be better. It's like saying all humans have a drinking problem so not-AI will solve drink driving.

Another problem is so much of this is US centric in thinking with US drivers being the metric. They are not great drivers, particularly when their twisted political system allowed road safety to be sidelined by corporate interests wherin pedestrians were made almost legitimate targets for motorists if they were to set foot on a road. Where I am from, pedestrians always had a right of way, but were expected to use common sense and cross-walks, when near them.

Off the top of my head, and without re-checking, I think the US death rates is around 1 per 150 million km of distance driven, Ireland and Australia are about half that at 1 per 300 million and the show-off Norwegians are at 1 per 600 million km driven. That is pretty epic, IMO, not a sign of terrible human drivers. Ignoring Tesla's usual BS, the total distance driven by not-AI systems operated by more rational developers like Waymo and GM, was a tiny fraction of even 150 M km, yet idiots spout off about definitive proof that not-AI's are safer than humans. As I said, that should always be a sub-set of humans, as most are indistinguishable from perfect. My parents were still driving in their mid 80's and drove for more than a combined 120 years without causing or being in a serious accident. You may disagree with my idea of perfection, but objectively you can't measure their failure rate as they never failed. And I think we shouldn't be using the woeful US driving stats as a general metric of human driving performance, for a start.

AndrewJRenko · 04-01-2021 5:55pm

cnocbui wrote: »

I think it's wrong to lump all drivers together and to say something along the lines of 'humans are bad at driving' not-AI ones will be better. It's like saying all humans have a drinking problem so not-AI will solve drink driving.

Another problem is so much of this is US centric in thinking with US drivers being the metric. They are not great drivers, particularly when their twisted political system allowed road safety to be sidelined by corporate interests wherin pedestrians were made almost legitimate targets for motorists if they were to set foot on a road. Where I am from, pedestrians always had a right of way, but were expected to use common sense and cross-walks, when near them.

Off the top of my head, and without re-checking, I think the US death rates is around 1 per 150 million km of distance driven, Ireland and Australia are about half that at 1 per 300 million and the show-off Norwegians are at 1 per 600 million km driven. That is pretty epic, IMO, not a sign of terrible human drivers. Ignoring Tesla's usual BS, the total distance driven by not-AI systems operated by more rational developers like Waymo and GM, was a tiny fraction of even 150 M km, yet idiots spout off about definitive proof that not-AI's are safer than humans. As I said, that should always be a sub-set of humans, as most are indistinguishable from perfect. My parents were still driving in their mid 80's and drove for more than a combined 120 years without causing or being in a serious accident. You may disagree with my idea of perfection, but objectively you can't measure their failure rate as they never failed. And I think we shouldn't be using the woeful US driving stats as a general metric of human driving performance, for a start.

Maybe we should start with metrics like the 98% of drivers that break urban speed limits here, rather than getting confused between 'perfect' and 'lucky'?

daniel_t1409 · 08-01-2021 6:17pm

Deleted User wrote: »

Don't know about the rest of you, but I am looking forward to seeing driverless buses, trains, taxis and luas

It's really not as far away as people think

Scratch that, they will never happen.

GT89 · 08-01-2021 6:39pm

daniel_t1409 wrote: »

Scratch that, they will never happen.

I think the OP in question has an unhealthy dislike of public transport workers. Public transport vehicles will be the last to be fully automated mark my word.

08-01-2021 7:09pm

GT89 wrote: »

I think the OP in question has an unhealthy dislike of public transport workers. Public transport vehicles will be the last to be fully automated mark my word.

Dislike of unions and overpaid workers who like to strike or threaten to at the drop of a hat....just for clarification

Public transport i.e. buses & trains, are perfect for automation. Static routes, easily mapped. Perfect for automation. Its why there are towns and cities all over the world doing small scale trials.

I think this particular landscape is going to look very, VERY different by 2025

cnocbui · 08-01-2021 7:10pm

AndrewJRenko wrote: »

Maybe we should start with metrics like the 98% of drivers that break urban speed limits here, rather than getting confused between 'perfect' and 'lucky'?

Yay, and then we can talk about the 86% of fatal accidents that are not caused by speeding. But yes, speeding is most definetely the most important issue in road safety, even though it is only the direct cause of 21.7 deaths in Ireland 2019, (by abstraction) which in turn is a whopping 0.29% of all deaths that year. Definitely something to be irate about. Someone should do something.

GT89 · 08-01-2021 7:21pm

[Deleted User] wrote: »

Dislike of unions and overpaid workers who like to strike or threaten to at the drop of a hat....just for clarification

Public transport i.e. buses & trains, are perfect for automation. Static routes, easily mapped. Perfect for automation. Its why there are towns and cities all over the world doing small scale trials.

I think this particular landscape is going to look very, VERY different by 2025

Public transport workers are not overpaid and they do not threathen to strike at the drop of a hat.

You do not understand the responsibility of a bus or particulalry a train driver. To automate railway lines there would need to be complete grade separation like the DLR or Copenhagen Metro which not be possible on 90% of railway lines in Ireland which were built in the Victorian. Anyone familar with the public transport system in Ireland would know that it takes an unhealthy amount of time to make changes.

Small scale trials does not fully automated nuses and trains are imminent. It's likely there will be automated buses in the next 10-15 years but with still with a steering wheel and a driver. Also even when these are brought in there would be reducndancy packages for drivers.

Oscar Madison · 09-01-2021 7:34pm

I can't believe that all this talk about AV's is still ongoing!

It will never ever happen for so many reasons to list here.

The biggest reasons why it won't work are humanity and the individual!

donvito99 · 09-01-2021 11:33pm

GT89 wrote: »

Anyone familar with the public transport system in Ireland would know that it takes an unhealthy amount of time to make changes.

Thanks to drivers' unions.

Automated trains have been around for years now. Roll on MetroLink and 24 hour services between the CC, the Airport and Swords!

Self driving buses, trains, trucks etc

Comments