Autonomous cars? I think not

User avatar
Alex Simmons/RST
Expert
Posts: 4529
Joined: Fri May 02, 2008 3:51 pm
Contact:

Re: Autonomous cars? I think not

Postby Alex Simmons/RST » Wed Mar 21, 2018 7:31 am

Scintilla wrote:
march83 wrote:In the entire history of the world, 1 person has been killed by an autonomous car and it's a failure. There were probably hundreds killed by human drivers just in 1 day yesterday.


THAT ^^^ is not the point.

These vehicles are not even generally available for sale nor in public use as a general 'thing', yet they are currently failing to identify and avoid key things. Like people :roll:

They have technology that is fine for coping with a closed-circuit racetrack, or maybe restricted access freeways (even here I kinda doubt their safety with stray and unexpected hazards), but they simply are not, and are not likely to be in the foreseeable future, able to deal with public access roads, with people walking, changing conditions, roadworks, detours, bicycle riders, kids and dogs or other animals about. Scrap 'em!

They already are. These are cars on the roads in a regular US city.

Did you read the latest information on what happened? Police reviewed the video. Not even the superior reaction time of a car's systems could have avoided this, let alone a human driver.

That's not to say the systems can't be better. But an expectation of perfection, i.e. zero deaths, is unrealistic because that can only occur when there are no vehicles at all. But when the road toll is 1% of what it is with human drivers, is that a good outcome?

User avatar
bychosis
Posts: 5456
Joined: Sat Jan 14, 2012 1:10 pm
Location: Lake Macquarie

Re: Autonomous cars? I think not

Postby bychosis » Wed Mar 21, 2018 9:19 am

Alex Simmons/RST wrote:Did you read the latest information on what happened? Police reviewed the video. Not even the superior reaction time of a car's systems could have avoided this, let alone a human driver.


I wonder if the impact was reduced because of the superior reaction times.

I thought it ironic that I saw an ad for Volvo vehicles sprouting about semi autonomous driving. Not ideal timing!
bychosis (bahy-koh-sis): A mental disorder of delusions indicating impaired contact with a reality of no bicycles.

road_warrior
Posts: 9
Joined: Tue Feb 06, 2018 1:04 pm

Re: Autonomous cars? I think not

Postby road_warrior » Wed Mar 21, 2018 9:48 am

bychosis wrote:
Alex Simmons/RST wrote:Did you read the latest information on what happened? Police reviewed the video. Not even the superior reaction time of a car's systems could have avoided this, let alone a human driver.


I wonder if the impact was reduced because of the superior reaction times.

I thought it ironic that I saw an ad for Volvo vehicles sprouting about semi autonomous driving. Not ideal timing!


I think autonomous cars are inevitable. Here are some reasons: they can see in all directions simultaneously; they never get tired or distracted; eventually these cars will talk to each other to share information about road conditions up ahead.

Within 0.05 of a second they will detect a hazard, and plan a strategy... humans normally take 10-100 times longer than that, and often make mistakes.

These cars will do things like instantly detect tyre blow-outs, oil slicks, ice etc.. and safely stop the car. When they do lose traction, they will be able to do amazing things like avoid roll-overs, steer between objects (like trees and other cars), or even control an impact to avoid side impacts and target car crumple zones.

Will there be glitches on the way? Yes, but there is a chance to reduce out road toll remarkably over time.

User avatar
sogood
Posts: 17164
Joined: Thu Aug 31, 2006 7:31 am
Location: Sydney AU
Contact:

Re: Autonomous cars? I think not

Postby sogood » Wed Mar 21, 2018 10:28 am

bychosis wrote:I thought it ironic that I saw an ad for Volvo vehicles sprouting about semi autonomous driving. Not ideal timing!

Not at all. Given the latest evidences, the technology is being exhonerated. Further, I assume you have not experienced the current crop of sensor based driving assistance on new vehicles, from proximity to lane warning, including interventions (emergency stop), they are truly beneficial. And these are but tip of the iceberg of what these fully autonomous systems are capable of. Going further ahead, China is now building the world's first expressway with full integration of autonomous driving systems, permitting exchanging of traffic data via iOT, inter-vehicle communication, speed-spacing coordination, emergency management, in-lane charging of EVs amongst other features and is due for completion in 2022 and ready for the Asian Games in Hangzhou.

https://www.ndtv.com/world-news/china-t ... 22-1803770

Have to have an open mind on these tech issues and the possibility is only limited by imagination!
Bianchi, Ridley, Montague, GT, Garmin and All things Apple :)
RK wrote:And that is Wikipedia - I can write my own definition.

g-boaf
Posts: 9407
Joined: Mon Sep 26, 2011 6:11 pm

Re: Autonomous cars? I think not

Postby g-boaf » Wed Mar 21, 2018 10:47 am

bychosis wrote:
Alex Simmons/RST wrote:Did you read the latest information on what happened? Police reviewed the video. Not even the superior reaction time of a car's systems could have avoided this, let alone a human driver.


I wonder if the impact was reduced because of the superior reaction times.

I thought it ironic that I saw an ad for Volvo vehicles sprouting about semi autonomous driving. Not ideal timing!


Volvo stated very categorically that when they do go out with full autonomous driving features (which they've not done at the moment), they will take full responsibility if any accidents occur. I think that's a very bold thing for Volvo to do, but also I think it gives me a lot of confidence in them that they will try to do everything in their power to make sure it is right. Really, we expect that of Volvo anyway.

User avatar
sogood
Posts: 17164
Joined: Thu Aug 31, 2006 7:31 am
Location: Sydney AU
Contact:

Re: Autonomous cars? I think not

Postby sogood » Wed Mar 21, 2018 10:54 am

g-boaf wrote:Volvo stated very categorically that when they do go out with full autonomous driving features (which they've not done at the moment), they will take full responsibility if any accidents occur. I think that's a very bold thing for Volvo to do, but also I think it gives me a lot of confidence in them that they will try to do everything in their power to make sure it is right. Really, we expect that of Volvo anyway.

Indeed bold and a smart commercial move. But call me a cynic, I would bet the vehicle's retail price tag would have included multi-year insurance premium hidden within. :mrgreen:
Bianchi, Ridley, Montague, GT, Garmin and All things Apple :)
RK wrote:And that is Wikipedia - I can write my own definition.

AdelaidePeter
Posts: 212
Joined: Wed Jun 07, 2017 11:13 am

Re: Autonomous cars? I think not

Postby AdelaidePeter » Wed Mar 21, 2018 11:36 am

bychosis wrote:
Alex Simmons/RST wrote:Did you read the latest information on what happened? Police reviewed the video. Not even the superior reaction time of a car's systems could have avoided this, let alone a human driver.


I wonder if the impact was reduced because of the superior reaction times.


According to this LA times article, the car didn't brake before impact. http://www.latimes.com/business/la-fi-h ... story.html It also says "the woman was walking a bike at about 10 p.m. (so it was dark) "and moved into traffic from a dark center median".

My gut feeling is that this would have had the same outcome with the majority of human drivers. But I also think that the most defensive of human drivers would have spotted the pedestrian. So what I'm saying is that driverless cars are already good, but need to (and will) get better. I certainly think that in the long run they will dramatically reduce the road toll.

If you look at (what I think is) the crash site on Google maps, it's not pedestrian friendly and (if you turn and look behind to the left) there is even "no pedestrians" sign on the median strip. https://www.google.com.au/maps/@33.4367 ... 312!8i6656 (Edited to a closer location)

Jmuzz
Posts: 252
Joined: Mon Jan 29, 2018 1:42 pm

Re: Autonomous cars? I think not

Postby Jmuzz » Wed Mar 21, 2018 12:53 pm

sogood wrote:, I would bet the vehicle's retail price tag would have included multi-year insurance premium hidden within. :mrgreen:


I think autonomous cars (and other robots) will be "as a service" model where you have to stay current with the version and updates on a yearly subscription.
Otherwise the features will disable since it is unsafe to operate without the latest bugfixes.

So yeah the subscription would contain an insurance element.
May not have to get your own insurance at all since their package may be a per time/distance dynamic billing. These smart cars have that ability because they know exactly where they are and have back to base monitoring.

Jmuzz
Posts: 252
Joined: Mon Jan 29, 2018 1:42 pm

Re: Autonomous cars? I think not

Postby Jmuzz » Wed Mar 21, 2018 1:04 pm

AdelaidePeter wrote:According to this LA times article, the car didn't brake before impact.


It's all just being made up at this stage.
Remember USA police chiefs are low level elected politicians like Mayor's so they and potential candidates love to get in front of a camera and be seen being important even if they have no truth to report yet.

One thing is certain, a huge log of data more than ever seen before in an accident investigation is in that car.
It will all come out and be reviewed in engineering journals.

It will have impacts on accident investigation, what accident forensics determined by "traditional" investigation vs the undeniable record of exactly what the robot saw and did and thought (unless it introduces randomisation in decision making) down to the millisecond.

But it could be 6 months plus before the full details are out.
Video/stills or renders could be soon, could be weeks.

Nobody knows until more data comes out.
Right now too many people with agendas trying to manipulate impressions.

User avatar
sogood
Posts: 17164
Joined: Thu Aug 31, 2006 7:31 am
Location: Sydney AU
Contact:

Re: Autonomous cars? I think not

Postby sogood » Wed Mar 21, 2018 1:16 pm

Jmuzz wrote:I think autonomous cars (and other robots) will be "as a service" model where you have to stay current with the version and updates on a yearly subscription.

I agree that'll likely to be the model. At the same time, Tesla is not officially labelling such a subscription for their current sales. Yet again, given the annual service of our vehicles from decades back, they are effectively a form of subscription for continued vehicle ownership.
Bianchi, Ridley, Montague, GT, Garmin and All things Apple :)
RK wrote:And that is Wikipedia - I can write my own definition.

g-boaf
Posts: 9407
Joined: Mon Sep 26, 2011 6:11 pm

Re: Autonomous cars? I think not

Postby g-boaf » Wed Mar 21, 2018 1:29 pm

sogood wrote:
g-boaf wrote:Volvo stated very categorically that when they do go out with full autonomous driving features (which they've not done at the moment), they will take full responsibility if any accidents occur. I think that's a very bold thing for Volvo to do, but also I think it gives me a lot of confidence in them that they will try to do everything in their power to make sure it is right. Really, we expect that of Volvo anyway.

Indeed bold and a smart commercial move. But call me a cynic, I would bet the vehicle's retail price tag would have included multi-year insurance premium hidden within. :mrgreen:


They are doing subscription model pricing anyway with XC40, others later on. Probably about USD$600 per month. Porsche is also going that way. These pricing models are all inclusive.

Any of these modern cars generate an enormous amount of logged data - so it will be possible to track whatever happens. Autonomous vehicles are the way to go. It must happen.

User avatar
sogood
Posts: 17164
Joined: Thu Aug 31, 2006 7:31 am
Location: Sydney AU
Contact:

Re: Autonomous cars? I think not

Postby sogood » Wed Mar 21, 2018 1:59 pm

g-boaf wrote:They are doing subscription model pricing anyway with XC40...

Indeed. Or 'factory bonus' free servicing for X years... Effectively the insurance and other costs have already been built into their financial spreadsheets on the pricing. Bingo! If there's any additional subscription collected from the following years, it'll be a bonus to the bottomline.

Yes, autonomous and AI technologies are here to stay. No questions about it.
Bianchi, Ridley, Montague, GT, Garmin and All things Apple :)
RK wrote:And that is Wikipedia - I can write my own definition.

User avatar
Howzat
Posts: 758
Joined: Wed Aug 15, 2012 7:08 pm

Re: Autonomous cars? I think not

Postby Howzat » Wed Mar 21, 2018 2:05 pm

What this shows is that this autonomous car can’t see well enough to avoid hitting a pedestrian with a bicycle in less than ideal conditions.

New version of “victim cyclist wasn’t wearing helmet” will be “victim was riding on the roads in less than ideal conditions”.

The other question raised is this: why does anyone think a taxi-company-with-an-app has any hope of mustering the engineering or computational wherewithal to develop driverless cars?

fat and old
Posts: 3209
Joined: Sat Aug 30, 2014 12:06 pm
Location: Mill Park

Re: Autonomous cars? I think not

Postby fat and old » Wed Mar 21, 2018 6:10 pm

Howzat wrote:
The other question raised is this: why does anyone think a taxi-company-with-an-app has any hope of mustering the engineering or computational wherewithal to develop driverless cars?


I'd have thought they were merely the test bed. Makes sense.... No cost to them, forefront of developments in this area, social responsibility marketing advantages....list goes on.

I think autonomous cars (and other robots) will be "as a service" model where you have to stay current with the version and updates on a yearly subscription.
Otherwise the features will disable since it is unsafe to operate without the latest bugfixes.


This is why i'm skeptical at best. Suffering the latest windows updates for 3 weeks caused a lot of late nights. Better that than deaths. Thanks, but I'll pass until I see a 100% unhackable safety record after 10 years in service.....by which time I'll need it as I'll be too old and decrepit to drive anyway and ready to accept a dodgy system.

My childhood books reckoned I'd be living on Mars with silver knee highs by now :lol:

User avatar
ldrcycles
Posts: 8461
Joined: Thu Oct 28, 2010 3:19 pm
Location: Kin Kin, Queensland

Re: Autonomous cars? I think not

Postby ldrcycles » Wed Mar 21, 2018 6:35 pm

After working with robots for the last 11 years, I don't have the faith in technology that a lot of people seem to have. Short of aerospace/military grade stuff (which would make for cars that 99% of people couldn't afford), stuff breaks or just glitches quite a lot. When the worst that can happen is a bit of inconvenience that's not so bad, when 1.5 tonnes of metal is going down the road out of control, it's a problem.
When man invented the bicycle he reached the peak of his attainments- Elizabeth West.

User avatar
Strawburger
Posts: 1700
Joined: Tue Aug 26, 2008 12:19 pm
Location: Dulwich Hill, Sydney

Re: Autonomous cars? I think not

Postby Strawburger » Wed Mar 21, 2018 8:17 pm

Howzat wrote:
The other question raised is this: why does anyone think a taxi-company-with-an-app has any hope of mustering the engineering or computational wherewithal to develop driverless cars?


Uber is not a taxi service company. They are a technology company.

Stage 1 of their strategy is to collect data by using the public to travel around the road network. Their ultimate goal is to service the transport industry with their own connected autonomous vehicles.
n=10 (2013 & 2004 roads,2010 track,2x 2009 foldups,1990 hybrid,1992 trainer,2007 rental,1970's step through,1980's zeus)

fat and old
Posts: 3209
Joined: Sat Aug 30, 2014 12:06 pm
Location: Mill Park

Re: Autonomous cars? I think not

Postby fat and old » Wed Mar 21, 2018 8:38 pm

Strawburger wrote:
Howzat wrote:
The other question raised is this: why does anyone think a taxi-company-with-an-app has any hope of mustering the engineering or computational wherewithal to develop driverless cars?


Uber is not a taxi service company. They are a technology company.


Uber pays for this themselves? I stand corrected. :)

User avatar
WyvernRH
Posts: 2322
Joined: Tue Nov 06, 2007 6:41 pm
Location: Newcastle NSW

Re: Autonomous cars? I think not

Postby WyvernRH » Wed Mar 21, 2018 8:43 pm

ldrcycles wrote:After working with robots for the last 11 years, I don't have the faith in technology that a lot of people seem to have. Short of aerospace/military grade stuff (which would make for cars that 99% of people couldn't afford), stuff breaks or just glitches quite a lot. When the worst that can happen is a bit of inconvenience that's not so bad, when 1.5 tonnes of metal is going down the road out of control, it's a problem.


Hmm afraid that last sentence describes an awful lot of human drivers on a daily basis that I encounter.
Software is at least consistent, doesn't drink, get annoyed, have 'bad times of the month' or get distracted texting the sig other.
I have no idea what sort of level autonomous cars software has got to these days but I was around for some of the higher grade aero stuff many (many) years ago and I would hope that they are at least as good today as our stuff was back then at avoiding stuff (like trees, pylons and rocks) at significantly high speeds and finding other stuff to do nasty things to (avoid in this case). I don't think I would offend many military pilots if I said they sometimes sat in the cockpit thinking '!! BAN ME NOW FOR SWEARING !! I hope this stuff really works!' - and it did...

User avatar
Howzat
Posts: 758
Joined: Wed Aug 15, 2012 7:08 pm

Re: Autonomous cars? I think not

Postby Howzat » Wed Mar 21, 2018 9:05 pm

The other thing that this demonstrates is that autonomous cars - today - are not safer than human drivers. They are worse. Much worse.

At the current rate, they are fifty times more dangerous. Uber has run up 2 million vehicle miles in autonomous car testing, and killed one pedestrian. Human drivers in the US do 100 million vehicle miles per death.

Until now, there has been a lot of "imagineering" speculation about our brave new automative future. "Autonomous cars will be safer because (I imagine) they can see in all directions and (I imagine) they have fast reaction times and (I imagine) they will so much better than humans at everything". Well, possibly, maybe one day.

But not this day, plainly.

User avatar
Strawburger
Posts: 1700
Joined: Tue Aug 26, 2008 12:19 pm
Location: Dulwich Hill, Sydney

Re: Autonomous cars? I think not

Postby Strawburger » Wed Mar 21, 2018 9:24 pm

fat and old wrote:
Strawburger wrote:
Howzat wrote:
The other question raised is this: why does anyone think a taxi-company-with-an-app has any hope of mustering the engineering or computational wherewithal to develop driverless cars?


Uber is not a taxi service company. They are a technology company.


Uber pays for this themselves? I stand corrected. :)


That I can't answer, who knows where they are finding it from. I would guess they are partnering with others and part finding from their delivery services
n=10 (2013 & 2004 roads,2010 track,2x 2009 foldups,1990 hybrid,1992 trainer,2007 rental,1970's step through,1980's zeus)

User avatar
sogood
Posts: 17164
Joined: Thu Aug 31, 2006 7:31 am
Location: Sydney AU
Contact:

Re: Autonomous cars? I think not

Postby sogood » Wed Mar 21, 2018 11:07 pm

Howzat wrote:The other thing that this demonstrates is that autonomous cars - today - are not safer than human drivers. They are worse. Much worse.

Huh? Where's the evidence?

Ok, if that view makes you sleep better at night... :mrgreen: :mrgreen: :mrgreen:
Bianchi, Ridley, Montague, GT, Garmin and All things Apple :)
RK wrote:And that is Wikipedia - I can write my own definition.

road_warrior
Posts: 9
Joined: Tue Feb 06, 2018 1:04 pm

Re: Autonomous cars? I think not

Postby road_warrior » Thu Mar 22, 2018 1:44 am

Howzat wrote:The other thing that this demonstrates is that autonomous cars - today - are not safer than human drivers. They are worse. Much worse.

At the current rate, they are fifty times more dangerous. Uber has run up 2 million vehicle miles in autonomous car testing, and killed one pedestrian. Human drivers in the US do 100 million vehicle miles per death.

Until now, there has been a lot of "imagineering" speculation about our brave new automative future. "Autonomous cars will be safer because (I imagine) they can see in all directions and (I imagine) they have fast reaction times and (I imagine) they will so much better than humans at everything". Well, possibly, maybe one day.

But not this day, plainly.


You cannot say that based on one incident. Unfortunately you need much more data to determine this empirically.

Without enough empirical data, the standard way to do this is to create a statistical model in order to estimate failure rate. Then using testing you can verify elements of the model to ensure its integrity. All of this happens within the design teams and we don't have access to this data.

Jmuzz
Posts: 252
Joined: Mon Jan 29, 2018 1:42 pm

Re: Autonomous cars? I think not

Postby Jmuzz » Thu Mar 22, 2018 7:36 am

Strawburger wrote:That I can't answer, who knows where they are finding it from. I would guess they are partnering with others and part finding from their delivery services


Uber is Google, they have near unlimited funds.
It is a loss making operation in the current human driver mode, they are only doing it to establish control of the market so that they dominate when driverless is ready.

User avatar
Howzat
Posts: 758
Joined: Wed Aug 15, 2012 7:08 pm

Re: Autonomous cars? I think not

Postby Howzat » Thu Mar 22, 2018 8:45 am

road_warrior wrote:You cannot say that based on one incident. Unfortunately you need much more data to determine this empirically.


Agreed - but obtaining a more statistically confident sample size of dead pedestrians per million vehicle-miles is not really on the cards. This is the first empirical data we have - and we have to pay attention to it. It's a bit of a fell-at-the-first-hurdle situation. It contradicts the utopian press-release vision of autonomous cars being by definition "safer than human drivers", and it should also indicate that this problem space is harder to solve than your average popular-science prognosticator might think.

The letter to the editors pages are full of people asking why we're spending money trams and trains when self driving pod cars are just around the corner. Well, that ain't necessarily so.

User avatar
mikesbytes
Super Mod
Super Mod
Posts: 18734
Joined: Thu Dec 01, 2005 11:42 pm
Location: Tempe, Sydney
Contact:

Re: Autonomous cars? I think not

Postby mikesbytes » Thu Mar 22, 2018 9:13 am

While not specifically commenting on this incident, I can see another "not my fault" excuse coming to the fray;
1. The Autonomous mode drives conservatively complying with all traffic laws and applying safe practices
2. The occupant wants to get there quicker, not leave a gap to the car in front to prevent others cutting in etc etc
3. To achieve 2. the occupant drives manually
4. An incident occurs
5. The occupant says the vehicle was in autonomous mode
If the R-1 rule is broken, what happens to N+1?

Who is online

Users browsing this forum: Scintilla