CES-2018: the multimedia system of Mercedes is capable of self-learning

During the exhibition of consumer electronics CES-2018, Mercedes-Benz presented a new digital cockpit MBUX (Mercedes-Benz User Experience). The updated multimedia system and the digital instrument panel, which run on the Linux OS, will debut officially on the new A-Class. Zakmuflirovanny pre-serial model of the model has already been shown in Las Vegas, although the car itself will not be sold in the US.

Recall that in the A-class you can see three options for displays: two by seven inches (an assembly and a central one with a capacitive sensor), one for seven (assembly) and one for 10.25 inches (central) or two for 10.25 inches (with a resolution of 1920 x 720 pixels). And this is not counting the projection. Note that all the new "Mercedes" is gradually moving to a typical display layou.
The MBUX system is a complex with elements of artificial intelligence, updates "through the air" and self-learning. It is activated by the phrase "Hey Mercedes" and is controlled by touching in three places (central touch screen, touchpad on the tunnel and mini-touchpad on the spokes of the steering wheel). The computer with the Nvidia Reilly PX graphics processor (in the initial and medium versions of the complex) or the Nvidia Parker 128 (in the top versions) and the data processing rate of up to 500 billion operations per second, which is comparable to the powerful game consoles, is responsible for the system's recovery.
Note that the management of many functions of the multimedia system is based on the principle of tablets and smartphones with the movement of icons, scaling by the movement of fingers and so on.

The MBUX system will support near field communication standards (NFC) and USB Type C, 3D images with augmented reality, voice commands with self-learning function, and software update "by air". One of the components of the new multimedia will be the support of "smart" watches and built-in artificial intelligence to help the driver. The ability to learn means that the system generates individual offers based on the user's habits.
By the way, the developers promise to recognize natural speech, so instead of asking for weather, you can ask "Will I need sunglasses tomorrow?" And instead of verbal indication of the temperature in the cabin, you can simply say "I'm frozen".