How do Google's robots "see" the world?

I remember that I was greatly impressed by the movie “Terminator-2”, this particular part. I especially liked the shots where the T800 model was analyzing the world (the same moment taking clothes and a motorcycle, for example). It turns out that in our time the situation is not much different in the case of real robots (though not as perfect as the T800, of course). The other day, a video with the presentation of the “vision of the world” by Google’s robots caught the eye.



These cars use for the analysis of the environment a huge number of sensors, including lasers, radar and so on. Separately, it is worth noting that the car collects about a gigabyte of data per second, of course, analyzing this information, and quite quickly. When an obstacle appears, the car reacts almost instantly, which indicates a high speed of the auto control system.

')

Here is the video itself with a demonstration of the system:







But a more detailed presentation:







Via Bill_Gross + Dvice