XPost: rec.arts.sf.written, rec.arts.sf.science
From: rexjohns@nospam.com
"Your Name" wrote in message
news:240120141310026807%YourName@YourISP.com...
> In article , rex
> wrote:
>
>> "Your Name" wrote in message
>> news:240120140855359808%YourName@YourISP.com...
>> > J. Clarke wrote:
>> >> In article <52e0c728$0$52806$742ec2ed@news.sonic.net>,
>> >> dtravel@sonic.net
>> >> says...
>> >> > On 1/21/2014 9:49 PM, Your Name wrote:
>> >> > > In article <52df5822$0$52798$742ec2ed@news.sonic.net>, Dimensional
>> >> > > Traveler wrote:
>> >> > >> On 1/21/2014 7:46 PM, Greg Goss wrote:
>> >> > >>> "John F. Eldredge" wrote:
>> >> > >>>>
>> >> > >>>> Several years back, something got royally screwed up in the
>> >> > >>>> Google
>> >> > >>>> Maps
>> >> > >>>> database. If I recall correctly, going from one particular
>> >> > >>>> town
>> >> > >>>> to
>> >> > >>>> another in Scotland showed, correctly, a land trip of about 20
>> >> > >>>> miles.
>> >> > >>>> Going in the reverse direction showed a trip of several hundred
>> >> > >>>> miles,
>> >> > >>>> passing through three different countries and involving two
>> >> > >>>> trips
>> >> > >>>> across
>> >> > >>>> the North Sea.
>> >> > >>>
>> >> > >>> I forget whether it was Google Maps, or some predecessor mapping
>> >> > >>> internet program, but I was once told to turn left from one car
>> >> > >>> ferry
>> >> > >>> onto another near Nanaimo.
>> >> > >>>
>> >> > >> I once had map software on my laptop while I was entering St.
>> >> > >> Louis
>> >> > >> via
>> >> > >> a bridge across the Mississippi try to tell me to take a left
>> >> > >> turn
>> >> > >> from
>> >> > >> the divided interstate highway bridge a hundred feet up in the
>> >> > >> air
>> >> > >> onto
>> >> > >> the riverside jogging/bike path below.
>> >> > >
>> >> > > One example of NUMEROUS that prove self-driving cars simply aren't
>> >> > > going to happen any time soon.
>> >>
>> >> If a self-driving car took its _only_ navigational information from a
>> >> GPS this would be an issue. But the Google cars take their
>> >> information
>> >> from a variety of sources and should not drive off of the road simply
>> >> because a GPS tells them to.
>> >
>> > Without massive amount of expensive changes to the roading system and /
>> > or a sudden, and currently immpssible, increase in the functionality of
>> > "artificial intelligence", there's no way to stop the cars going the
>> > wrong way down a one-way street (as one example) simply because Google
>> > Maps told them to.
>>
>> It wouldn't be hard to ensure that google maps is always up to date
>> with any road change like that,
>
> It would be a MASSIVE task to try and keep it up to date with all the
> local changes worldwide,
No, it's done right now.
> especially temporary ones.
While those aren't don't currently, they don't increase the volume much.
>> and to even have the self driving car check the signs on the road that
>> the human drivers use too.
>
> That was the point of the example - the sign may well be in different
> places on different road, partially or fully obscured, stolen, graffiti
> altered, etc.etc. A human driver can tell in most cases (fully obscured
> being the exception), but a computer simply can't.
The computer can in fact do a much better job of doing
that because it can have part of its system dedicated to
observing the street signs and no human driver can do that.
> And that was only one example of numerous possibilities where a
> computer simply can't perform the same way as a human driver.
>
> The current and soon-to-be released self-driving cars all perform
> extremely simple tasks within set boundaries, but a self-parking car
> cannot tell whether or not the "parking space" is actually a real or
> legal one,
Yes they can when there is a proper database of that.
and the amount of information needed to make that decision
> is huge, let alone actually drive a car fully.
And the google cars do that already.
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
