Self driving buses, trains, trucks etc

devnull

magicbastarder wrote: »

and the funny thing is, in probably the majority of those programs, the plane crashes with multiple or total fatalities. and it's regularly human error to blame.

an aside; the episode i found most interesting of air crash investigation was one where a pilot who was trained in the eastern bloc, on planes where the artificial horizon works in the opposite way to western bloc planes. in short, human error.

I noticed that virtually all of the crashes that feature on there are normally the result of a chain of events, there normally isn't one single thing that can bring down a plane which happens in the air, normally it's a chain of events which come together which finally form an accident.

CFIT is by far the most common one, spatitual disorientation and loss of situational awareness normally come in shortly behind, then you often see other human error (maintenance, air traffic control, weather) and only after that generally are some kind of plane defect or malfunction and even a lot of them cannot bring a plane down on their own.

On another sidenote, did you see the episode about the Russian hockey team that let a kid in the cockpit?

magicbastarder

a friend of mine designs processes and UIs, and he cites a couple as being of interest to him - one was a flight where the weather was a major distraction during preflight, as well as an issue which caused the plane to have to return to the gate - and in all the fuss, the crew missed the part where they had to extend the flaps for low speed flight - so took off with the flaps fully retracted and couldn't gain lift. NASA were brought in to investigate and make recommendations, and designed a preflight check system which allowed the crew to track more easily what had or had not been done in case of distraction. simple stuff in reality, so simple no one had thought to make it robust before that.

prinzeugen

Ben D Bus wrote: »

If this your case then you need to show that increasing automation has reduced flight safety. Not by anectdote, but by cold hard statistics.

Honestly, most of the "it'll never happen" posts on this thread seem driven purely by fear of change and not by any real evidence.

Breakdown of air accidents by cause:
http://www.independent.co.uk/travel/news-and-advice/the-one-chart-that-shows-what-causes-fatal-plane-crashes-10494952.html

You are only looking at fatal accidents.

There are 1000s more where there was no loss of life because the pilots used knowledge and skill to get the plane on the ground.

Its not fear of change. Its fear for my life.

Plenty younger people that rely on tech to do everything for them wont have a problem, but I, and many others do.

magicbastarder

anyway, back to the human factor in the technology - one thing i was mulling over was, if this *does* lead to a drop in private car ownership and a greater takeup of car pooling (not in the 'let's all share a car together' sense, but the 'i'll use a car someone else finished using 30 minutes ago' sense); will this lead to a reduction in car sizes in urban areas?

most people buy five seaters for the 10% of times they might have people sitting in the back - but if it's just one person needing to get from stoneybatter to merrion square, why would a car pool company not maintain a fleet of single seaters for this purpose?

eeguy

prinzeugen wrote: »

You are only looking at fatal accidents.

There are 1000s more where there was no loss of life because the pilots used knowledge and skill to get the plane on the ground.

Its not fear of change. Its fear for my life.

Plenty younger people that rely on tech to do everything for them wont have a problem, but I, and many others do.

Ah yes, all the 'minor' accidents that are never reported so your point can't be backed up.
As I said before, the only reason we fit so many places in the sky without them all crashing is through advances in technology and taking decision out of the hands of pilots.

If 60% of crashes are due to pilot error then we could more than halve the chance of crashing with autonomy.
What will bringing less automation result in?

devnull

prinzeugen wrote: »

There are 1000s more where there was no loss of life because the pilots used knowledge and skill to get the plane on the ground.

There are a load of technological solutions and systems on board which have stopped pilots making mistakes as well and have made aviation safety and without them we'd have many more incidents.

A massive one is GPWS for example.

Its not fear of change. Its fear for my life. Plenty younger people that rely on tech to do everything for them wont have a problem, but I, and many others do.

Unfortunately the stats are not on your side, technology has made the skies safer, not more dangerous.

magicbastarder

anyway, the topic of autonomous planes and autonomous ground based vehicles are two very different things. concerns about one do not necessarily reflect on concerns about the other.
i was once talking to someone who had an 'interesting' flight on a cargo plane in the 70s, and when i asked him if it made him more nervous of flying, he said, no, it made him more nervous of driving. because at least in a jet plane, you know the crew are trained professionals.

prinzeugen

magicbastarder wrote: »

anyway, back to the human factor in the technology - one thing i was mulling over was, if this *does* lead to a drop in private car ownership and a greater takeup of car pooling (not in the 'let's all share a car together' sense, but the 'i'll use a car someone else finished using 30 minutes ago' sense); will this lead to a reduction in car sizes in urban areas?

most people buy five seaters for the 10% of times they might have people sitting in the back - but if it's just one person needing to get from stoneybatter to merrion square, why would a car pool company not maintain a fleet of single seaters for this purpose?

Bring back this.. No need for AI or Google cars.

https://www.lanemotormuseum.org/collection/cars/item/peel-p50-1964

This is the thing about the whole AI think that does not make sense.

You will still get phantom traffic jams on motorways, trucks will still need to be diesel powered, and at the end of the day they would be banned from college green anyway!

Maybe in a 100 years.

prinzeugen

eeguy wrote: »

Ah yes, all the 'minor' accidents that are never reported so your point can't be backed up.

'Minor' accidents??


BA009

http://www.dailymail.co.uk/news/article-431802/The-story-BA-flight-009-words-passenger-dreads-.html

Canada Air 624

https://en.wikipedia.org/wiki/Air_Canada_Flight_624

Flight 243

https://en.wikipedia.org/wiki/Aloha_Airlines_Flight_243

https://en.wikipedia.org/wiki/TACA_Flight_110

Is that enough for you?? How could AI have ever got these planes down with all crew/passengers surviving?

magicbastarder

the first link i clicked:
"Moody took drastic action: to prevent his passengers dying of oxygen starvation, he went into a nosedive, dropping 6,000ft in one minute, to an altitude where there was enough oxygen in the outside atmosphere to fill the cabin once more.

And quite unexpectedly, this action almost certainly saved the lives of every person on board."
so it was not skill of the pilots, but luck.

magicbastarder

second link:
"The final report was released in May 2017, finding Air Canada crew procedures to be the primary cause of the accident."

magicbastarder

prinzeugen wrote: »

There are 1000s more where there was no loss of life because the pilots used knowledge and skill to get the plane on the ground.

god be with the days when real men flew airplanes, not computers, and planes never crashed.

eeguy

magicbastarder wrote: »

the first link i clicked:
"Moody took drastic action: to prevent his passengers dying of oxygen starvation, he went into a nosedive, dropping 6,000ft in one minute, to an altitude where there was enough oxygen in the outside atmosphere to fill the cabin once more.

And quite unexpectedly, this action almost certainly saved the lives of every person on board."
so it was not skill of the pilots, but luck.

3 stories from the 1980's and plane in the 2015 story was 24 years old.

I think technology in general has moved on to stop these incidents happening.

Ben D Bus

prinzeugen wrote: »

Plenty younger people that rely on tech to do everything for them wont have a problem, but I, and many others do.

That's a telling comment tbh. Most people are comfortable with technology that arrived when they were younger but have a fear of technology that came later.

Change is the issue, not technology. I say this as someone approaching 50. I expect the fear to set in soon

devnull

The Daily Mail is hardly a great source! )

Nobody says that there are not incidents where there are not faults which happen with aircraft but these causing accidents alone are much lower than those which are caused by human error as shown by the various statistics on what causes fatal accidents.

GPWS alone has probably stopped huge amounts of accidents because it gives warnings of when a plane is close to terrain even if the pilot cannot see, we have no way of knowing exactly how many lives it has saved.

The Air Canada flight link you give says: "The TSB stated that there were no mechanical or maintenance faults uncovered with the aircraft and "Air Canada crew procedures to be the primary cause of the accident"

The Aloha flight that you mention is not a mechanical failure due to plane malfunction as such, it's the result of shoddy and improper maintenance according to the final repost by the NTSB.

prinzeugen

eeguy wrote: »

3 stories from the 1980's and plane in the 2015 story was 24 years old.

I think technology in general has moved on to stop these incidents happening.

How would AI handle all 4 engines going on a 747?? Reboot while the plane plunges?

Would AI know about a closed airfield?

Would AI be able to glide a plane to flattish strip and land?

It was human actions that saved these planes.

AI could hot have done it.

The Air Canada was human error in fuel loads, however if it was not for human knowledge, ie the pilot knowing that was there, the the outcome could have been very different.

magicbastarder

and this is relevant to landing a self-driving car in a disused airfield?

prinzeugen

devnull wrote: »

The Daily Mail is hardly a great source! )

Nobody says that there are not incidents where there are not faults which happen with aircraft but these causing accidents alone are much lower than those which are caused by human error as shown by the various statistics on what causes fatal accidents.

GPWS alone has probably stopped huge amounts of accidents because it gives warnings of when a plane is close to terrain even if the pilot cannot see, we have no way of knowing exactly how many lives it has saved.

The Air Canada flight link you give says: "The TSB stated that there were no mechanical or maintenance faults uncovered with the aircraft and "Air Canada crew procedures to be the primary cause of the accident"

The Aloha flight that you mention is not a mechanical failure due to plane malfunction as such, it's the result of shoddy and improper maintenance according to the final repost by the NTSB.

You are missing the point of my posts.

There have been 1000's of incidents with aircraft in which human knowledge and skill has prevented loss of life.

AI would not be able to understand what has gone on.

Would AI know if there was a hydraulic leak from xx part? Can AI look out the window and look?

Or does the plane need a person. A person that knows how the plane works, so that when the AI stops working, could step in?

Oh wait. We have them already.

prinzeugen

magicbastarder wrote: »

and this is relevant to landing a self-driving car in a disused airfield?

No. Just proving self driving planes/cars etc have issues.

With the way google and other sat navs work there is every chance the will end up on a disused runway.

Unreliable. My street, been here since 1896, does not show on sat navs.. South Dublin.

So how are these cars going to work??

magicbastarder

prinzeugen wrote: »

You are missing the point of my posts.

There have been 1000's of incidents with aircraft in which human knowledge and skill has prevented loss of life.

there have been thousands of incidents caused by human failure. you referenced 'air crash investigation' earlier. knock yourself out reading how many people have died as a result of human error:

https://en.wikipedia.org/wiki/List_of_Mayday_episodes

devnull

prinzeugen wrote: »

You are missing the point of my posts.

There have been 1000's of incidents with aircraft in which human knowledge and skill has prevented loss of life.

There have been thousands more incidents where automation and tech systems have warned humans of an upcoming disaster which they otherwise would have not known was about to happen without it and the disaster would have happened and everyone would be dead without any doubt.

If a pilot cannot see but the plane has detected that there is terrain approaching and warns the pilots without that warning there would be a crash. If a plane did not have anti stall systems and warnings pilots would stall their plane more often which would also lead to more crashes.

There's also systems that do not allow certain parameters to be put in to flight computers or prevent pilots from taking unsafe actions by both issuing a warning and preventing something from happening by simply refusing to do it in the interests of safety of passengers and those people who are on board.

If you go and watch Air Crash Investigation and a lot of episodes you will see lots of experts on aviation on there and pretty much every time they all agree that automation and systems like these which are put in place have vastly improved standards of safety in the skies.

prinzeugen

magicbastarder wrote: »

there have been thousands of incidents caused by human failure. you referenced 'air crash investigation' earlier. knock yourself out reading how many people have died as a result of human error:

https://en.wikipedia.org/wiki/List_of_Mayday_episodes

Take the blinkers off. Stop looking at fatal accidents only!

Air crashes only make the news when people die.

Things like this get little attention. Undercarriage failure.

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

Would AI have been able to tell all the passengers to move to the left of the plane?

devnull

prinzeugen wrote: »

Would AI have been able to tell all the passengers to move to the left of the plane?

Would a human be able to tell you terrain was below them or that they are about to crash into a bunch of trees or a mountain if they cannot see anything from the window because their vision is blocked by weather?

There are so many warnings and restrictive systems now in the cockpit that the automation means that if a pilot does something wrong the automation will tell them and/or try and stop and correct the actions that they are taking.

You won't read about that online because of the fact the automation has saved the pilots skin because it realised what a mistake they made. A lot of the pilot error incidents are in relation to loss of situational awareness and spatial disorientation where the automation and systems have screamed at them they are doing things wrong but they just ignore them time and time again until it's too late. =

Ben D Bus

So let's say there are some accidents prevented by human intervention, and some prevented by total automation.

Which is likely to result in less injury to people, whether in the air or on the roads? Statistics and the history of technology suggest that sooner or later (probably sooner), automation will, on balance, be the safer option.

magicbastarder

this is fundamental - your examples are all examples where something has already gone wrong. you are using examples where *the failure has already happened*.
you're refusing to take into consideration the examples where the same automation you hate so much has prevented accidents, simply because *nothing happened* that was noteworthy.

it's like the issue with changing to permanent DST that the UK has mulled over - the tabloids will jump on the people who died during the darker mornings, and not the `(greater number of) people whose lives were saved during the brighter evenings. precisely because the latter have stories that are simply mundane - 'i could have died but came out without a scratch' doesn't sell newspapers, but people dying does.

devnull

prinzeugen wrote: »

Things like this get little attention. Undercarriage failure.

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

Would AI have been able to tell all the passengers to move to the left of the plane?

Human (maintenance) error again:
http://avherald.com/h?article=47f7b9f9&opt=0

On Oct 8th 2015 the AAIB released their bulletin into the serious incident reporting that it turned out the right hand main gear hydraulic retract actuator had been incorrectly installed.

magicbastarder

again, trying to get this back on track - would be interested to know how many trucks are delivering a complete (20KG? is a 50 ton truck full of gravel the archetypal maximum?) load from A to B, and the full load being onloaded at A, and fully offloaded at B. loads where there's a string of deliveries from a loaded truck would imply someone to manage the load, i guess? e.g. last time i had a truck come to the house, it was a mattress delivery, which required two people - not something which could be skimped on, even if the truck drove itself.

in short - how many truck journeys don't require someone to oversee what's taken off the truck?

Stephen15

magicbastarder wrote: »

there have been thousands of incidents caused by human failure. you referenced 'air crash investigation' earlier. knock yourself out reading how many people have died as a result of human error:

https://en.wikipedia.org/wiki/List_of_Mayday_episodes

And who are controlling the technology behind these autonomous humans. Technology can only be as itelligent as the humans who create it all humans make errors no matter how inteligent they are. There is no such thing as AI as humans are behind it so it is not artificial intelligence, electronic itelligence would be a more accurate name for it.

Why is everyone here so pro driverless vehicles at the end of the day thousands of people are going to be left unemployed and mostly working class people who have no other formal qualifications other their trade or indeed their driving licence. These people are our truck drivers, taxi drivers, train drivers, chauffeurs, bus drivers and our mehanics all these will more than likely join the dole queue why is everyone so anti the people I mentioned are they not entitled to make a living for what they do and do well.

The way technology is going or the way some people here think it is will just replace all mehanical jobs. This will only end up leading to a working class revolt.

prinzeugen

devnull wrote: »

Human (maintenance) error again:
http://avherald.com/h?article=47f7b9f9&opt=0

On Oct 8th 2015 the AAIB released their bulletin into the serious incident reporting that it turned out the right hand main gear hydraulic retract actuator had been incorrectly installed.

So is AI going to fit it right?? Do you think a robot could have avoided that situation?

Is AI going to design, build and fly/drive stuff for us?

https://www.youtube.com/watch?v=EsUndkXAQ_8

Google cars have privacy issues as well. This "recommended" stuff will happen.

"google car take me to a place that serves ice cream"

"taking you to Iceland. journey time 123 hours. here are your flights"

magicbastarder

Stephen15 wrote: »

Why is everyone here so pro driverless vehicles at the end of the day thousands of people are going to be left unemployed and mostly working class people who have no other formal qualifications other their trade or indeed their driving licence.

this is the problem with this thread. there's a genuine discussion to be had about the above issues, but the people who want to have the discussion keep bringing it back to a technological argument (one which cannot be won, imho) rather than a human one. they're picking the wrong battle.