Google Cars heading down the road to cover more of California
How far Google has gotten down the road toward autonomous cars can be measured by the fact that members of the media are now invited to be chauffeured about Google’s headquarters hometown of Mountain View, California in its latest models of self-driven automobiles. Included in this tour is plenty of positive data concerning the Google car project, where it’s been, how it got there, and where it is headed.
Sebatian Thrun is the Google scientist who lead the charge in 2009 to design and manufacture hardware and software to create a vehicle that could drive autonomously in a standard freeway situation. This accomplishment, while fantastic in and of itself, pales in comparison to the task of mapping out and devising a car to interact and drive safely in a random, city-street environment.
The process has moved at what seems a snail’s pace, as the method has been rather bottom-up, meaning each area the Google car can operate in must first be mapped and integrated into the storage of the onboard computer. A new street are added at the rate of about one a week. Clearly, this method will need to be augmented by satellites and a more rigorous algorithm to be able to negotiate driving on any street, any where.
In addition, as long as Google car’s share the roads with cars with human drivers, as well as bicyclists, pedestrians, and other random events, their very impressive sensory abilities, (360 vision all the time, radar, road sign reading and recognition), will need to be controlled by a very sophisticated, near-AI level of cognition and decision-making. The current models are specially modified Lexuses with Google’s very special imaging, processing and sensory hardware and software installed.
While progress has been slow and steady by design, the results have been impressive. Google’s autonomous vehicles, have been tested almost constantly in a growing area surrounding Mountain View that now adds up to more then 700,000 miles of driving. They have never been involved in even the most minor of accidents. 100% accuracy in analyzing the environment, and absolutely zero accidents are the stated goals of the project. This, in contrast to the fact that over 95% of all auto accidents are caused by human error, not mechanical failure.