[PLUG-TALK] Google Self-Driving Car

[Bill Barry]

On Wed, May 28, 2014 at 1:13 PM, Rich Shepard <rshepard at appl-ecosys.com> wrote:
> On Wed, 28 May 2014, John Jason Jordan wrote:
>
>> That is precisely the problem. Self-driving cars could be here now, but
>> they must co-exist with manually driven cars. If all cars were
>> self-driving they could communicate with each other and negotiate safe
>> passage for all. But throw one manually driven car into the mix and
>> everything becomes impossible to predict.
>
>    Notice that the vehicle is the size of the tiny, urban "smart" car. It
> makes sense as an urban commuting vehicle but I would not want to be in one
> of those on a county road in Harney County (will it recognize a cow standing
> in the middle of the road) or on a forest road under heavy vegetation or at
> the bottom of a draw where satellite visibility is problamatic. And, how
> good are they on black ice or a blizzard?
>

They are not strongly dependent on having a strong satellite signal.
Most of the detailed location is done by matching the surrounding
environment to a recording of the surrounding environment done by a
recording vehicle.  This definitely causes trouble in black ice and
blizzard because all of those landmarks disappear. It would probably
see a cow in the middle of the road faster and better than me because
it uses a wider spectrum for detection and it's always paying
attention.  It really does not have to be perfect. If it hits just a
few less people than we currently do, the insurance company will
charge us a larger premium  to actually drive our own car.

Bill