You don’t understand what Google is using LIDAR for. Think in terms of facial recognition. It uses a cloud of points to identify “traffic cone”.
So no, it is not using the LIDAR to try and gain any sort of high precision. It is using it to map the world around the car, and simplify it - again into ‘blocks’. And in a GIS type way, they are assigning ‘attributes’ to the block, which have a ‘library’ of possibly actions that the block might take.
Since you have stated that GPS could handle half of the examples from my morning commute, I have a simple question:
Name one.
Name one of the events from my morning commute that GPS can handle. And explain in detail how GPS would work through the situation.
We’re in circles here. So here’s my final comment on this one.
First off understand that Google has ALREADY done thousand of miles self driven, and is planning on 100 more cars by the end of the year:
http://www.freerepublic.com/focus/f-chat/3160840/posts
Then remember these great declarations throughout history:
It will take 2000 years to explore America
all the major scientific discoveries have been made
man will never run a mile in under 4 minutes
man will never fly under control
man will never control a vehicle over 100MPH
man will never break the speed of sound under control
man will never land on the moon and come back safely
You are, quite simply, on the wrong side of history. This problem you insist is insurmountable has been surmounted, now it’s just a matter of doing it cheaply enough for a production car.