Cars discussion, split from 'other news'

Seen something interesting in the news or on the intertubes? Discuss it here.

Moderators: Zamfir, Hawknc, Moderators General, Prelates

commodorejohn
Posts: 1179
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Wed Jul 01, 2015 5:54 am UTC

KrytenKoro wrote:The point was a bad point. People already do all of those things simply because they feel confident from muscle memory. Show me the data of a driver in an autopilot car actually doing these things more frequently than your average driver, or stop making up groundless horror stories.

Given that the whole point of the thing is for them to be able to do exactly that, I'm really not sure why you think it's so much of a stretch.

Also, complaining about "groundless horror stories" when half the reservations being expressed here are precisely that we have insufficient data on how these things perform in suboptimal, real-world conditions to know whether or not they're safe is just pig-headed.

If I ask you for directions to the emergency room, or Google, and you both get it wrong, I don't really give a fuck whose ""fault" it is. The person I was racing to get there is already dead. The fault occurred, people got hurt because of it.

But again, I'm not selling my services as a navigator.

There are, right now, people dying because they weren't able to figure out how to solve a driving problem, like black ice or deer, "RIGHT NOW". This is not a problem that the computers are introducing. Therefore, the only way it is relevant is if it occurs at a greater rate than the problems already occurring do.

And that's exactly the thing we don't know yet. Welcome to the point, enjoy your complimentary slice of cake.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

elasto
Posts: 3751
Joined: Mon May 10, 2010 1:53 am UTC

Re: Cars discussion, split from 'other news'

Postby elasto » Wed Jul 01, 2015 9:05 am UTC

commodorejohn wrote:
There are, right now, people dying because they weren't able to figure out how to solve a driving problem, like black ice or deer, "RIGHT NOW". This is not a problem that the computers are introducing. Therefore, the only way it is relevant is if it occurs at a greater rate than the problems already occurring do.

And that's exactly the thing we don't know yet. Welcome to the point, enjoy your complimentary slice of cake.

We here may not know it to the last decimal point but you can bet your bottom dollar the tech companies do.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Wed Jul 01, 2015 10:30 am UTC

commodorejohn wrote:But again, I'm not selling my services as a navigator.
Suggest a metric defining how good directions have to be to meet your standard. Mapping software and GPS are the best solution to navigation that has ever existed. Without exception. And as far as I know Google isn't selling Google Maps. You are using it because it's free. However assuming that you bought some type of GPS mapping solution then you would still run into the same type of problems. If you had been driving in 1960 you would have still had the same type of problem. If you were Hannibal crossing the Alps you still would have had this problem. Maps have never been 100 percent accurate.

I posted a link to a TED talk that hit a lot of these issues. Here are some excerpts from a transcript. On what drivers do when the system is active.
So this was music to our ears, but then we started to look at what the people inside the car were doing, and this was eye-opening. Now, my favorite story is this gentleman who looks down at his phone and realizes the battery is low, so he turns around like this in the car and digs around in his backpack, pulls out his laptop, puts it on the seat, goes in the back again, digs around, pulls out the charging cable for his phone, futzes around, puts it into the laptop, puts it on the phone. Sure enough, the phone is charging. All the time he's been doing 65 miles per hour down the freeway. Right? Unbelievable. So we thought about this and we said, it's kind of obvious, right? The better the technology gets, the less reliable the driver is going to get. So by just making the cars incrementally smarter, we're probably not going to see the wins we really need.


Here is a capture of the Google car identifying a policeman.Image

There is a amusing moment at around 11:04 of a woman in an electric wheelchair chasing a duck.

User avatar
HES
Posts: 4889
Joined: Fri May 10, 2013 7:13 pm UTC
Location: England

Re: Cars discussion, split from 'other news'

Postby HES » Wed Jul 01, 2015 12:05 pm UTC

ucim wrote:[No matter what the circumstance, something will happen that demands that the driver take over NOW.

Not exactly. All you need is a failsafe system that stops the car.

If both the main system and the failsafe catastrophically fail simultaneously, then sure, that's bad. But no worse than a human driver who suddenly falls alseep or gets distracted or crashes for the multitude of human failures that occur all the freaking time.
He/Him/His Image

User avatar
addams
Posts: 10226
Joined: Sun Sep 12, 2010 4:44 am UTC
Location: Oregon Coast: 97444

Re: Cars discussion, split from 'other news'

Postby addams » Wed Jul 01, 2015 12:07 pm UTC

What a fun clip. Thanks for that.
That thing is a better driver than any human I've ever known.
It can see better than any human I have ever known.

The 'hack' thing remains problematic.

I think it is kind'a like any other physical assault or murder.
When some persons contemplate that antisocial acton the thought of sever penalty might all that protects us.

We may have to live with the knowledge we can be 'hacked', then
the way we live with the knowledge we can be assaulted or murdered, now.
Life is, just, an exchange of electrons; It is up to us to give it meaning.

We are all in The Gutter.
Some of us see The Gutter.
Some of us see The Stars.
by mr. Oscar Wilde.

Those that want to Know; Know.
Those that do not Know; Don't tell them.
They do terrible things to people that Tell Them.

commodorejohn
Posts: 1179
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Wed Jul 01, 2015 2:03 pm UTC

morriswalters wrote:Suggest a metric defining how good directions have to be to meet your standard.

Easy: it doesn't take me to where the place I want to go isn't.

And as far as I know Google isn't selling Google Maps. You are using it because it's free.

Right. But when Google starts selling cars that drive themselves, they will be selling their navigational services - which, unless something changes between now and then, are mediocre to merely-okay.

HES wrote:Not exactly. All you need is a failsafe system that stops the car.

I've got news for you: a car going 60 MPH does not stop on a dime, and coming to a complete halt as fast as possible has a non-negligible influence on surrounding traffic.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

KrytenKoro
Posts: 1487
Joined: Tue Apr 05, 2011 2:58 pm UTC

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Wed Jul 01, 2015 2:13 pm UTC

Given that the whole point of the thing is for them to be able to do exactly that, I'm really not sure why you think it's so much of a stretch.

Pretty sure the whole point is to improve upon the risk of driving. That will likely lead to less diligence given by humans, but that has always been true, whether the safety improvements involved taking control from the human or not.

People text, and read, and eat, and do make-up while driving cars now, so it's ludicrous to be saying "but if you have autopilot they may not always be paying attention to the road!" unless you can provide data. And data on road behavior in google cars does exist.

But again, I'm not selling my services as a navigator.

For the purposes of all these examples, replace your name with any human's name and the point is the same. Who to blame when an accident occurs is not relevant here, how to reduce the accidents across the board is.

Re:your other responses: If you're clarifying that you're not actually claiming that it's unlikely computers will be able to handle adverse conditions, and that you're just uncertain whether they have run enough tests at present to justify yourself buying the car now, then fine, I guess no one's actually disputing that, neh?
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

User avatar
PeteP
What the peck?
Posts: 1451
Joined: Tue Aug 23, 2011 4:51 pm UTC

Re: Cars discussion, split from 'other news'

Postby PeteP » Wed Jul 01, 2015 2:37 pm UTC

KrytenKoro wrote:
Given that the whole point of the thing is for them to be able to do exactly that, I'm really not sure why you think it's so much of a stretch.

Pretty sure the whole point is to improve upon the risk of driving. That will likely lead to less diligence given by humans, but that has always been true, whether the safety improvements involved taking control from the human or not.

People text, and read, and eat, and do make-up while driving cars now, so it's ludicrous to be saying "but if you have autopilot they may not always be paying attention to the road!" unless you can provide data. And data on road behavior in google cars does exist.


Whether the data of professionals test driving not completely finished self-driving cars is representative of normal people driving a finished product is questionable. That they aren't yet finished already changes the behaviour quite a bit. If you know that it can't deal with everything yet you have much more reason to pay attention. If your car however normally drives without ever needing your interference and it's only needed in unusual circumstances you are far more likely to not pay any attention. Though there are some technical things you could do if you need people to pay attention, first sense whether they take their hands of the wheel, perhaps also use some eye tracking. Though that would make self-driving less useful.

But honestly nothing in a self-driving car should be designed to rely on fast reflexes for the driver. Though if there isn't a sudden system failure in the worst moment the driver probably has a moment to react and you need something to get their attention.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Wed Jul 01, 2015 2:40 pm UTC

commodorejohn wrote:Right. But when Google starts selling cars that drive themselves, they will be selling their navigational services - which, unless something changes between now and then, are mediocre to merely-okay.
As compared to what? I'm trying to decide if you are trolling. Circa 1990 this none of this existed and you had to do this without the benefit GPS at all, and maps were outdated at publication. Locations for buildings weren't available at all. You got street numbers and if you were lucky and the street hadn't been built in the last couple of years the street might show on the map. And then you had to do it the hard way. Call and ask for directions.

I just planned a trip using Google Maps from my door to a campground in a National park. Not only did it get me door to door instructions with the accompanying maps, I was able to use satellite imaging to verify that the route exists. All for free. Gee! Welcome to the world. Google misplaced your building, and you were pissed. If you don't like Google's products, don't buy them. Buy one from someone else or don't buy one at all. No one is going to hold a gun to your head and make you do anything. On a more practical level, don't use Google Maps. You don't seem to understand the limitations that any map is going to have. Try navigating with paper maps for a while.

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Wed Jul 01, 2015 2:40 pm UTC

commodorejohn wrote:
If I ask you for directions to the emergency room, or Google, and you both get it wrong, I don't really give a fuck whose ""fault" it is. The person I was racing to get there is already dead. The fault occurred, people got hurt because of it.

But again, I'm not selling my services as a navigator.


Who gives a crap if selling is involved? Is that relevant?

If solution A is better than previously existing solution B, are you gonna freak out because the person who came up with solution A wants to make a buck off it?

Without GPS, google maps, etc, people STILL had to navigate. If the tech is better than humans, great. That has value. I buy a GPS device for my car because of that value. If it was actually way worse than human navigation, GPS wouldn't have been so popular. Nobody is forcing GPS and google maps on people. We just use it because we like it. The end.

commodorejohn
Posts: 1179
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Wed Jul 01, 2015 3:02 pm UTC

KrytenKoro wrote:People text, and read, and eat, and do make-up while driving cars now, so it's ludicrous to be saying "but if you have autopilot they may not always be paying attention to the road!" unless you can provide data.

It's not "they may not always be paying attention to the road," it's "they may not ever be paying attention to the road." Are a lot of people now not as careful as they should be? Yep. Some of them are even pretty terrible about it. But as it stands they still have an immediate, concrete incentive to be at least partially engaged, or they don't stay driving for very long. Once they can rely on their car to do it all for them (or once they think they can,) that will go right out the window. And that's a huge, huge problem if we design with the assumption that the "driver" is going to remain available as a failsafe for what the computer can't handle.

For the purposes of all these examples, replace your name with any human's name and the point is the same. Who to blame when an accident occurs is not relevant here, how to reduce the accidents across the board is.

Except I'm not talking about who gets blamed for accidents, I'm talking about the quality of automated navigation. The difference between J. Random Human and the manufacturer of an automated car is that J. Random Human is not charging money for their services as a navigator; the manufacturer is, therefore they have a responsibility to be reliable as a navigator. Hopefully automated cars will at least be able to avoid turning onto streets that aren't there in the event of faulty map data, but if you're paying robocar prices for navigation service, you have every right to expect that not to happen in the first place, and no amount of "well, maps aren't perfect!" changes that.

Re:your other responses: If you're clarifying that you're not actually claiming that it's unlikely computers will be able to handle adverse conditions, and that you're just uncertain whether they have run enough tests at present to justify yourself buying the car now, then fine, I guess no one's actually disputing that, neh?

Both, actually. There isn't enough data outside of essentially optimal conditions, and I'm skeptical about the ability of computers to cope with the wide, wide range of possible issues that crop up in real driving circumstances. People's repetitions of "well, all we have to do to solve this complex problem is solve this complex problem, I'm sure that'll happen inevitably because it's inevitable that it'll happen!" do little to reassure me.

PeteP wrote:But honestly nothing in a self-driving car should be designed to rely on fast reflexes for the driver. Though if there isn't a sudden system failure in the worst moment the driver probably has a moment to react and you need something to get their attention.

Again, though, the problem is that pretty much any situation serious enough to require intervention is not a situation in which "a moment to react" is an available luxury.

Tyndmyr wrote:If solution A is better than previously existing solution B, are you gonna freak out because the person who came up with solution A wants to make a buck off it?

No, I'm going to freak out because I'm still not past the if in that sentence and yet people are proposing putting solution A in charge of 1500+ pounds of machine hurtling down the road at 60 MPH.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

icanus
Posts: 478
Joined: Mon Aug 13, 2007 1:19 pm UTC
Location: in England now abed

Re: Cars discussion, split from 'other news'

Postby icanus » Wed Jul 01, 2015 3:08 pm UTC

commodorejohn wrote:
HES wrote:Not exactly. All you need is a failsafe system that stops the car.

I've got news for you: a car going 60 MPH does not stop on a dime, and coming to a complete halt as fast as possible has a non-negligible influence on surrounding traffic.
It's still safer than what meat-piloted cars do if their driver is incapacitated, or swallows a bug, or spills coffe in it's lap, which is even less predicatble than a sudden stop.

It'd also be a lot safer once surrounding traffic wasn't limited human reaction times, and consistently maintained a safe stopping distance between vehicles, which humans are awful at.
But again, I'm not selling my services as a navigator.
You do realise that navigation and driving are two seperate tasks? Outside of emergency services getting where they need to be on time, navigation isn't a safety feature.

Of course there will be occasional navigational errors and people will be driven to the wrong place, but blindly following sat-nav off non-existent bridges is a human failure caused by not being able to deal with navigation instructions and sensory data at the same time.

commodorejohn
Posts: 1179
Joined: Thu Dec 10, 2009 6:21 pm UTC
Location: Placerville, CA
Contact:

Re: Cars discussion, split from 'other news'

Postby commodorejohn » Wed Jul 01, 2015 3:15 pm UTC

icanus wrote:It's still safer than what meat-piloted cars do if their driver is incapacitated, or swallows a bug, or spills coffe in it's lap, which is even less predicatble than a sudden stop.

I'm sorry, but "it's probably safer than human drivers are in the worst possible circumstances for human drivers" would not be a glowing recommendation even if I were convinced that was the case.
"'Legacy code' often differs from its suggested alternative by actually working and scaling."
- Bjarne Stroustrup
www.commodorejohn.com - in case you were wondering, which you probably weren't.

User avatar
HES
Posts: 4889
Joined: Fri May 10, 2013 7:13 pm UTC
Location: England

Re: Cars discussion, split from 'other news'

Postby HES » Wed Jul 01, 2015 3:34 pm UTC

commodorejohn wrote:I'm sorry, but "it's probably safer than human drivers are in the worst possible circumstances for human drivers" would not be a glowing recommendation even if I were convinced that was the case.

You missed an important part:
"In the worst possible circumstances for a machine, it's probably safer than human drivers are in the worst possible circumstances for human drivers"

commodorejohn wrote:And that's a huge, huge problem if we design with the assumption that the "driver" is going to remain available as a failsafe for what the computer can't handle.

Who is doing that?

ETA:
commodorejohn wrote:I'm skeptical about the ability of computers to cope with the wide, wide range of possible issues that crop up in real driving circumstances.

Do you realise how much they already do?
Last edited by HES on Wed Jul 01, 2015 3:37 pm UTC, edited 1 time in total.
He/Him/His Image

icanus
Posts: 478
Joined: Mon Aug 13, 2007 1:19 pm UTC
Location: in England now abed

Re: Cars discussion, split from 'other news'

Postby icanus » Wed Jul 01, 2015 3:36 pm UTC

commodorejohn wrote:
icanus wrote:It's still safer than what meat-piloted cars do if their driver is incapacitated, or swallows a bug, or spills coffe in it's lap, which is even less predicatble than a sudden stop.

I'm sorry, but "it's probably safer than human drivers are in the worst possible circumstances for human drivers" would not be a glowing recommendation even if I were convinced that was the case.

We're talking about failsafes for a total system failure in an electronic "driver", so it seems like we should compare to a total system failure of a human driver. I appologise for muddying the issue by introducing less serious failures like swallowing a bug or spilling a hot drink.

Please feel free to outline the failsafes that come into play when a human driver dies or becomes unconscious at the wheel, and how these are superior to stopping the vehicle.

User avatar
Dauric
Posts: 3989
Joined: Wed Aug 05, 2009 6:58 pm UTC
Location: In midair, traversing laterally over a container of sharks. No water, just sharks, with lasers.

Re: Cars discussion, split from 'other news'

Postby Dauric » Wed Jul 01, 2015 3:37 pm UTC

HES wrote:
commodorejohn wrote:And that's a huge, huge problem if we design with the assumption that the "driver" is going to remain available as a failsafe for what the computer can't handle.

Who is doing that?

Any system that has a manual override for emergencies.
We're in the traffic-chopper over the XKCD boards where there's been a thread-derailment. A Liquified Godwin spill has evacuated threads in a fourty-post radius of the accident, Lolcats and TVTropes have broken free of their containers. It is believed that the Point has perished.

User avatar
HES
Posts: 4889
Joined: Fri May 10, 2013 7:13 pm UTC
Location: England

Re: Cars discussion, split from 'other news'

Postby HES » Wed Jul 01, 2015 3:48 pm UTC

I don't foresee automated driving systems expecting humans to suddenly take over. Barring an emergency stop button, any manual control will be for convenience rather than emergency control and switching would be when the driver wants, not when the car needs.
He/Him/His Image

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Wed Jul 01, 2015 3:52 pm UTC

Dauric wrote:
HES wrote:
commodorejohn wrote:And that's a huge, huge problem if we design with the assumption that the "driver" is going to remain available as a failsafe for what the computer can't handle.

Who is doing that?

Any system that has a manual override for emergencies.
The term fail safe and manual over ride have to very different meanings. A fail safe doesn't need anyone to take control. It means that it fails safely, no more no less. it might be nothing more than a smooth application of the brakes to bring the vehicle to a halt. A manual override makes sure that you don't have to fight for control, it takes the computer out of the loop. Testing vehicles currently have overrides, and it is why they are called test systems. Some combination of the two might be used but, long term it would probably be better to let the car fail safe.

User avatar
Whizbang
The Best Reporter
Posts: 2238
Joined: Fri Apr 06, 2012 7:50 pm UTC
Location: New Hampshire, USA

Re: Cars discussion, split from 'other news'

Postby Whizbang » Wed Jul 01, 2015 3:53 pm UTC

After a few years, a decade at most, I wouldn't want a human to take over driving in an emergency anyway. Could you imagine a situation in which a driver's ed student taking over during an emergency (which a computer is unable to handle safely) would be a good thing?

User avatar
ivnja
The spirit of things can bugger right off.
Posts: 823
Joined: Mon Dec 04, 2006 5:30 am UTC
Location: 19T526268 4971339 (NAD 83)

Re: Cars discussion, split from 'other news'

Postby ivnja » Wed Jul 01, 2015 4:35 pm UTC

BattleMoose wrote:As a cyclist I eagerly await the advent of autonomous cars. The, complete disregard for the life of cyclists in Australia is boarder line and sometimes is actually criminal. A computer will value my safety more than most humans.

As an occasional bicycle commuter myself, I'm curious how well these systems will interact with vehicular cyclists. Especially those whose interpretation of traffic laws may...differ somewhat from what is actually written. Running stop signs/lights suddenly becomes a lot less hazardous if you know the cross-traffic will automatically stop for you. With enough cyclists (I'm picturing New York City couriers, for example), could you "hack" an intersection and get traffic to just sit there?

Not that that sort of thing matters too much to me personally. Most of my riding is in a little college town, not a city, and so like you, I'd love to just have the cars always give space when passing and not pull in/out in front of me. Even if the human drivers are for the most part pretty good where I am.
Hi you.
she/her

Trebla
Posts: 387
Joined: Fri Apr 02, 2010 1:51 pm UTC

Re: Cars discussion, split from 'other news'

Postby Trebla » Wed Jul 01, 2015 4:36 pm UTC

commodorejohn wrote:
morriswalters wrote:Suggest a metric defining how good directions have to be to meet your standard.

Easy: it doesn't take me to where the place I want to go isn't.


What single source gives you more correct and accurate directions more consistently than does any automated navigation software?

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: Cars discussion, split from 'other news'

Postby Tyndmyr » Wed Jul 01, 2015 4:45 pm UTC

commodorejohn wrote:
Tyndmyr wrote:If solution A is better than previously existing solution B, are you gonna freak out because the person who came up with solution A wants to make a buck off it?

No, I'm going to freak out because I'm still not past the if in that sentence and yet people are proposing putting solution A in charge of 1500+ pounds of machine hurtling down the road at 60 MPH.


So? Horrible, horrible drivers are hurtling down the road in such vehicles all the time. Even if they are worse than average, the appreciable change in risk to you should be neglibile.

Do you also run for a bunker whenever a car drives by because it could be driven by a terrible human driver?

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26740
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Wed Jul 01, 2015 5:08 pm UTC

commodorejohn wrote:
morriswalters wrote:Suggest a metric defining how good directions have to be to meet your standard.

Easy: it doesn't take me to where the place I want to go isn't.
That's a metric, but what's your standard? How frequently can it do this and still meet your standard? (And if you say "never", then you have an impossible standard that no navigational system has ever achieved or will ever achieve.)

commodorejohn wrote:The difference between J. Random Human and the manufacturer of an automated car is that J. Random Human is not charging money for their services as a navigator; the manufacturer is, therefore they have a responsibility to be reliable as a navigator. Hopefully automated cars will at least be able to avoid turning onto streets that aren't there in the event of faulty map data, but if you're paying robocar prices for navigation service, you have every right to expect that not to happen in the first place, and no amount of "well, maps aren't perfect!" changes that.
No, you'd be paying robocar prices for a robocar, one aspect of which is navigation.

But before free online sources, mapmakers charged money for its services, despite the fact that everything they ever published was outdated even before the first sale was ever made.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

KrytenKoro
Posts: 1487
Joined: Tue Apr 05, 2011 2:58 pm UTC

Re: Cars discussion, split from 'other news'

Postby KrytenKoro » Wed Jul 01, 2015 7:03 pm UTC

commodorejohn wrote:Except I'm not talking about who gets blamed for accidents, I'm talking about the quality of automated navigation. The difference between J. Random Human and the manufacturer of an automated car is that J. Random Human is not charging money for their services as a navigator; the manufacturer is, therefore they have a responsibility to be reliable as a navigator. Hopefully automated cars will at least be able to avoid turning onto streets that aren't there in the event of faulty map data, but if you're paying robocar prices for navigation service, you have every right to expect that not to happen in the first place, and no amount of "well, maps aren't perfect!" changes that.

"I'm not talking about who's responsible when things don't work, but I'm talking about who's responsible when things don't work."

There is no perfect system, human or machine. No one has all the information. No one is disputing that Google Maps is fallible, they're just pointing out that it's much less fallible than any given human. Human-driven taxis can and do have the -exact same problems- that you're hypothesizing will plague automated cars, and no one has said to the concept of taxis "well, the technology's just not there yet."

This entire line of argument you've been spouting is absolutely irrelevant to any kind of meaningful decision regarding automated cars.

Both, actually. There isn't enough data outside of essentially optimal conditions, and I'm skeptical about the ability of computers to cope with the wide, wide range of possible issues that crop up in real driving circumstances.

You need to commit yourself to what your argument is, because you keep on conveniently forgetting that you're arguing that point whenever people respond to it.

People's repetitions of "well, all we have to do to solve this complex problem is solve this complex problem, I'm sure that'll happen inevitably because it's inevitable that it'll happen!" do little to reassure me.

It's a good thing no one's saying that. Instead, they're pointing out the actual scope of the problem, and giving examples of how it could be conceivably solved. In fairness, I don't think anyone has actually produced a report demonstrating that testing plans have already been drafted. In unfairness, you haven't demonstrated that it would be impossible beyond repeated "I don't think computers would necessarily be able to handle [this thing that humans already fail at handling with hilarious frequency]."

ivnja wrote:
BattleMoose wrote:As a cyclist I eagerly await the advent of autonomous cars. The, complete disregard for the life of cyclists in Australia is boarder line and sometimes is actually criminal. A computer will value my safety more than most humans.

As an occasional bicycle commuter myself, I'm curious how well these systems will interact with vehicular cyclists. Especially those whose interpretation of traffic laws may...differ somewhat from what is actually written. Running stop signs/lights suddenly becomes a lot less hazardous if you know the cross-traffic will automatically stop for you. With enough cyclists (I'm picturing New York City couriers, for example), could you "hack" an intersection and get traffic to just sit there?

Not that that sort of thing matters too much to me personally. Most of my riding is in a little college town, not a city, and so like you, I'd love to just have the cars always give space when passing and not pull in/out in front of me. Even if the human drivers are for the most part pretty good where I am.

This is a really good point and I'd be interested to learn whether google has done any testing in regards to cyclist traffic. It would definitely be a killing blow if they are unable to solve it.
From the elegant yelling of this compelling dispute comes the ghastly suspicion my opposition's a fruit.

User avatar
ivnja
The spirit of things can bugger right off.
Posts: 823
Joined: Mon Dec 04, 2006 5:30 am UTC
Location: 19T526268 4971339 (NAD 83)

Re: Cars discussion, split from 'other news'

Postby ivnja » Wed Jul 01, 2015 7:16 pm UTC

Trebla wrote:
commodorejohn wrote:
morriswalters wrote:Suggest a metric defining how good directions have to be to meet your standard.
Easy: it doesn't take me to where the place I want to go isn't.
What single source gives you more correct and accurate directions more consistently than does any automated navigation software?

It should probably be noted that this is a two-part problem. The issue commodorejohn had, the FedEx center actually being located at 123 Outer Rd rather than 456 Downtown St, is simple enough to fix - just update the address lookup for the FedEx center in the GIS database. And for the driver who finds himself in that situation of being brought to the wrong place by inputting the business name into the nav system, pull up the internet, search for the FedEx webpage, see what the address of the center is, and tell the car to go to the address. Or ask somebody like commodorejohn did and tell the car to go to the correct address.

The more annoying problem is when the mapping software has the physical location of street addresses wrong. Generally, the addresses are built right in to the line segments that make up the roads in the map. There will be an attribute for numbers on the left side of the street, and an attribute for numbers on the right side of the street (as it is drawn). The mapmaker will put the range of addresses into those attribute fields (so for example, one line segment of Smith St. will have numbers 1-15 on one side and 2-16 on the other), and the software will subdivide the length of the line segment into the appropriate numbers. Then the end user can query "6 Smith St." and it will bring you to a point 3/8 of the way down the right side of the street. But as anyone who has driven in rural or even suburban areas knows, the nicely divided segments on the GPS screen or in Google Maps don't always line up with where those building numbers actually fall on the ground. Sometimes they're quite some distance off, especially on long roads. So then when you put in your correct address, the car may still bring you to the wrong place, and there's no immediate recourse in that you can't give the car any more address data than you already have. Fortunately, you will likely end up close enough that you'll be able to tell the car "go back three houses" or "go forward two more parking lot entrances" or whatever, but it's still a problem with a less simple solution, from a GIS standpoint.
Hi you.
she/her

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Wed Jul 01, 2015 7:20 pm UTC

@KrytenKoro

From the same TED talk I linked to earlier. To understand the complete context you need to see the video.
Now, I just fire-hosed you with a lot of stuff there, so I'm going to break one of these down pretty quickly. So what we're looking at is the scene with the cyclist again, and you might notice in the bottom, we can't actually see the cyclist yet, but the car can: it's that little blue box up there, and that comes from the laser data. And that's not actually really easy to understand, so what I'm going to do is I'm going to turn that laser data and look at it, and if you're really good at looking at laser data, you can see a few dots on the curve there, right there, and that blue box is that cyclist. Now as our light is red, the cyclist's light has turned yellow already, and if you squint, you can see that in the imagery. But the cyclist, we see, is going to proceed through the intersection. Our light has now turned green, his is solidly red, and we now anticipate that this bike is going to come all the way across. Unfortunately the other drivers next to us were not paying as much attention. They started to pull forward, and fortunately for everyone, this cyclists reacts, avoids, and makes it through the intersection. And off we go.

User avatar
Whizbang
The Best Reporter
Posts: 2238
Joined: Fri Apr 06, 2012 7:50 pm UTC
Location: New Hampshire, USA

Re: Cars discussion, split from 'other news'

Postby Whizbang » Wed Jul 01, 2015 7:26 pm UTC

Question:

What is the current standing with regards to automated parking space location? Like, say your auto-auto drives you to a location. Fine. But now you need to park. In many situations (parking lots and driveways) this is fairly straightforward, but for some locations it is extremely difficult even for an experienced driver knowledgeable about the area.

[Edit]

Although, there is no reason your car has to be parked particularly near your destination. You can get out and your auto-auto can drive to the nearest public parking lot and park there. Then, when you are ready to go, you simply call your car from your phone and it comes to pick you up. No more tight roadways with cars parked on either side. No more people fighting over parking locations and leaving lawn chairs out after snow storms and the like. This would probably lead to people having to rent parking spaces somewhere, however.

Trebla
Posts: 387
Joined: Fri Apr 02, 2010 1:51 pm UTC

Re: Cars discussion, split from 'other news'

Postby Trebla » Wed Jul 01, 2015 8:03 pm UTC

Whizbang wrote:Although, there is no reason your car has to be parked particularly near your destination. You can get out and your auto-auto can drive to the nearest public parking lot and park there. Then, when you are ready to go, you simply call your car from your phone and it comes to pick you up. No more tight roadways with cars parked on either side. No more people fighting over parking locations and leaving lawn chairs out after snow storms and the like. This would probably lead to people having to rent parking spaces somewhere, however.


This'll make metered parking fun if you can instruct your car to drive off and find another space if it "sees" a meter attendant.

User avatar
Whizbang
The Best Reporter
Posts: 2238
Joined: Fri Apr 06, 2012 7:50 pm UTC
Location: New Hampshire, USA

Re: Cars discussion, split from 'other news'

Postby Whizbang » Wed Jul 01, 2015 8:05 pm UTC

I would imagine metered parking would be a thing of the past. After all, the parking space is only there because people don't want to walk a long distance between their car and the enterance to the building. Instead I see these places becoming pick up and drop off zones.

Trebla
Posts: 387
Joined: Fri Apr 02, 2010 1:51 pm UTC

Re: Cars discussion, split from 'other news'

Postby Trebla » Wed Jul 01, 2015 8:11 pm UTC

Whizbang wrote:I would imagine metered parking would be a thing of the past. After all, the parking space is only there because people don't want to walk a long distance between their car and the enterance to the building. Instead I see these places becoming pick up and drop off zones.


Oh yeah, but in the meantime (while adoption is still only partial)...

User avatar
Dauric
Posts: 3989
Joined: Wed Aug 05, 2009 6:58 pm UTC
Location: In midair, traversing laterally over a container of sharks. No water, just sharks, with lasers.

Re: Cars discussion, split from 'other news'

Postby Dauric » Wed Jul 01, 2015 8:26 pm UTC

Trebla wrote:
Whizbang wrote:I would imagine metered parking would be a thing of the past. After all, the parking space is only there because people don't want to walk a long distance between their car and the enterance to the building. Instead I see these places becoming pick up and drop off zones.


Oh yeah, but in the meantime (while adoption is still only partial)...


I'm pretty sure you threaten a municipal revenue stream like that and countermeasures will be developed quickly.
We're in the traffic-chopper over the XKCD boards where there's been a thread-derailment. A Liquified Godwin spill has evacuated threads in a fourty-post radius of the accident, Lolcats and TVTropes have broken free of their containers. It is believed that the Point has perished.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Cars discussion, split from 'other news'

Postby morriswalters » Wed Jul 01, 2015 8:36 pm UTC

Or you simply make it cheap enough that you don't take a car a park it anywhere, you call for one with your smartphone. It picks you up and takes you to your destination and departs. And when you're ready to go, another one picks you up and returns you home. Parking meters go extinct. Uber is certainly thinking along this line.

cphite
Posts: 1360
Joined: Wed Mar 30, 2011 5:27 pm UTC

Re: Cars discussion, split from 'other news'

Postby cphite » Wed Jul 01, 2015 8:42 pm UTC

Whizbang wrote:I would imagine metered parking would be a thing of the past. After all, the parking space is only there because people don't want to walk a long distance between their car and the enterance to the building. Instead I see these places becoming pick up and drop off zones.


I don't see any municipal area giving up the revenue stream based on parking. No way. You might see parking lots moved further and further away from the places where folks want to be; but they're not going to be free, or even necessarily cheaper.

User avatar
Whizbang
The Best Reporter
Posts: 2238
Joined: Fri Apr 06, 2012 7:50 pm UTC
Location: New Hampshire, USA

Re: Cars discussion, split from 'other news'

Postby Whizbang » Wed Jul 01, 2015 8:46 pm UTC

cphite wrote:
Whizbang wrote:I would imagine metered parking would be a thing of the past. After all, the parking space is only there because people don't want to walk a long distance between their car and the enterance to the building. Instead I see these places becoming pick up and drop off zones.


I don't see any municipal area giving up the revenue stream based on parking. No way. You might see parking lots moved further and further away from the places where folks want to be; but they're not going to be free, or even necessarily cheaper.


I did not say they would be free or cheaper, merely that street-side metered parking as we use today will go away.

[Edit]
In fact, I see parking lots becoming booming commercial ventures.
Last edited by Whizbang on Wed Jul 01, 2015 8:54 pm UTC, edited 2 times in total.

User avatar
Dauric
Posts: 3989
Joined: Wed Aug 05, 2009 6:58 pm UTC
Location: In midair, traversing laterally over a container of sharks. No water, just sharks, with lasers.

Re: Cars discussion, split from 'other news'

Postby Dauric » Wed Jul 01, 2015 8:46 pm UTC

cphite wrote:
Whizbang wrote:I would imagine metered parking would be a thing of the past. After all, the parking space is only there because people don't want to walk a long distance between their car and the enterance to the building. Instead I see these places becoming pick up and drop off zones.


I don't see any municipal area giving up the revenue stream based on parking. No way. You might see parking lots moved further and further away from the places where folks want to be; but they're not going to be free, or even necessarily cheaper.


The ultimate model may be something along the lines of: municipal parking garages replace curbside parking, an auto-car that enters the garage starts it's meter, the meter stops when the car is recorded as having left the garage, and the owner of the car is charged for the parking. If it's a rental-share car the cost of the parking gets folded in to the cost of the service.
We're in the traffic-chopper over the XKCD boards where there's been a thread-derailment. A Liquified Godwin spill has evacuated threads in a fourty-post radius of the accident, Lolcats and TVTropes have broken free of their containers. It is believed that the Point has perished.

User avatar
HES
Posts: 4889
Joined: Fri May 10, 2013 7:13 pm UTC
Location: England

Re: Cars discussion, split from 'other news'

Postby HES » Wed Jul 01, 2015 8:52 pm UTC

This removal of kerbside parking answers the cyclist question: The additional space allows for fully segregated bike lanes.
He/Him/His Image

User avatar
Whizbang
The Best Reporter
Posts: 2238
Joined: Fri Apr 06, 2012 7:50 pm UTC
Location: New Hampshire, USA

Re: Cars discussion, split from 'other news'

Postby Whizbang » Wed Jul 01, 2015 8:58 pm UTC

Cycling will be a thing of the past too, as auto-cycling becomes a thing.

cphite
Posts: 1360
Joined: Wed Mar 30, 2011 5:27 pm UTC

Re: Cars discussion, split from 'other news'

Postby cphite » Wed Jul 01, 2015 8:59 pm UTC

KrytenKoro wrote:
People's repetitions of "well, all we have to do to solve this complex problem is solve this complex problem, I'm sure that'll happen inevitably because it's inevitable that it'll happen!" do little to reassure me.


It's a good thing no one's saying that. Instead, they're pointing out the actual scope of the problem, and giving examples of how it could be conceivably solved. In fairness, I don't think anyone has actually produced a report demonstrating that testing plans have already been drafted. In unfairness, you haven't demonstrated that it would be impossible beyond repeated "I don't think computers would necessarily be able to handle [this thing that humans already fail at handling with hilarious frequency]."


Actually, several people have pointed to specific problems that still have folks in the industry scratching their heads, and so far most of the solutions being offered fall into the "Well, surely someone will get around to fixing that..." category. People assume that, since technology tends to improve "fast" that it's basically a given that it'll happen when it needs to happen. According to Google, they're a year or so away from wanting to sell these things, despite not being a year or so away from fixing some of the issues.

Do I think they'll solve them eventually? Sure. Do I think it'll be next year... not so sure.

Security is the big one. These vehicles will be accepting input from the outside world, and there is no industry standard (that is, agreed upon) way to make that secure. And rest assured, there is a LOT of damage someone with ill-intent could do with a car, or multiple cars, if that problem isn't nailed down tight. And to be clear, if it communicates it can be hacked; period. And if hacking is too hard, it can be interfered with. Even if we set aside the possibility of some evil person deliberately crashing cars... imagine the results if you could simply jam or shut down the communication during rush hour? All you'd need is a transmitter with enough juice on the right frequency, and you could gridlock everybody.

The second one is data processing. The systems that control these cars will have to depend on outside information; GPS, sensor arrays, and so forth. And, they have to process that information reliably. There have to be measures in place for bad data, missing data, delayed data. There have to be measures in place for when the computer freezes up, experiences an error, etc. So far the most specific solution we've gotten for that in this thread has been that the driver will be able to immediately take over, regardless of what he or she happens to be doing.

elasto
Posts: 3751
Joined: Mon May 10, 2010 1:53 am UTC

Re: Cars discussion, split from 'other news'

Postby elasto » Wed Jul 01, 2015 9:23 pm UTC

cphite wrote:Actually, several people have pointed to specific problems that still have folks in the industry scratching their heads, and so far most of the solutions being offered fall into the "Well, surely someone will get around to fixing that..." category. People assume that, since technology tends to improve "fast" that it's basically a given that it'll happen when it needs to happen. According to Google, they're a year or so away from wanting to sell these things, despite not being a year or so away from fixing some of the issues.

They are? Can you link to that? The article I saw said they were 15 years away from full autonomy.

Are you referring to just highway driving or something? Cos Google has driven millions of miles of public highways without a single accident so far, so I could see that being a year away from production for sure.

Security is the big one. These vehicles will be accepting input from the outside world, and there is no industry standard (that is, agreed upon) way to make that secure.

Do you have a link for that also? My understanding is that the driving itself is autonomous - which it kinda has to be for speed of reaction purposes.

And rest assured, there is a LOT of damage someone with ill-intent could do with a car, or multiple cars, if that problem isn't nailed down tight. And to be clear, if it communicates it can be hacked; period. And if hacking is too hard, it can be interfered with. Even if we set aside the possibility of some evil person deliberately crashing cars...


Well of course. And someone with a home-made pipe bomb or an automatic weapon or a biological agent could kill hundreds at a public gathering. It's extremely easy to carry out a mass-killing if the will exists.

imagine the results if you could simply jam or shut down the communication during rush hour? All you'd need is a transmitter with enough juice on the right frequency, and you could gridlock everybody.


I'm not sure that's the case, but you could hack traffic light systems right now and cause gridlock. Or you could cause it by dropping a load of nails from the back of your car. Or, if you were particularly inventive, you could build yourself a laser, put it at the side of the road on a timer, and blind every driver that drives past, and cause gridlock that way...

The second one is data processing. The systems that control these cars will have to depend on outside information; GPS, sensor arrays, and so forth. And, they have to process that information reliably. There have to be measures in place for bad data, missing data, delayed data. There have to be measures in place for when the computer freezes up, experiences an error, etc.

Of course. But Google has reported that for the millions of miles driven so far, that hasn't happened. So, while it remains a risk, it still seems a lower level of risk than allowing humans to drive.

Bottom line is, however nervous and skeptical you are, Google is going to be ten times so. The US is a hugely litigious society, and class-action lawsuits are a powerful weapon. Google is NOT going to go public with this until they are as certain as is conceivably possible that this will all work. The bad publicity and reputational damage itself would be nightmare enough, let alone the settlement costs.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26740
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Cars discussion, split from 'other news'

Postby gmalivuk » Wed Jul 01, 2015 10:05 pm UTC

cphite wrote:Even if we set aside the possibility of some evil person deliberately crashing cars... imagine the results if you could simply jam or shut down the communication during rush hour? All you'd need is a transmitter with enough juice on the right frequency, and you could gridlock everybody.
So don't make the car come to a complete stop every time external data is interrupted?

While being aware of the positions of all other cars in the area could make everything much smoother, I don't see any reason why a car couldn't still do basic automatic driving even without that data, using internal sensors.

(And traffic lights can already accept external signals from emergency vehicles, yet I haven't heard of anyone causing city-wide gridlocks by hacking those...)
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)


Return to “News & Articles”

Who is online

Users browsing this forum: No registered users and 19 guests