Tesla autopilot will rely on NVIDIA Drive PX 2 platformOctober 31, 2016
Recently, the company Tesla Motors and NVIDIA reported that the auto driving system of electric Tesla will now rely on the hardware platform Drive PX 2.
On-board supercomputer Drive PX 2 was presented at the exhibition CES on January 2016. The system combines two Tegra processor and two discrete graphics chip based on a new generation of Pascal architecture. Computer provide up to 24 trillion deep learning operations per second – specialized commands, accelerating mathematical calculations in the deep learning network.
It is alleged that the Drive PX 2 module is capable of processing data from 12 cameras, and information from lidar, radar and ultrasonic sensors. The resulting indicators system combines for accurate recognition of objects, determine the location of the machine in the environment and the calculation of the optimal and safe trajectory.
Ability to be deeply training Drive PX 2 allow the system to cope with problems such as foreign objects on the road, distracted drivers and road works. In addition, the autopilot basis on Drive PX 2 will be able to work effectively in case of poor weather conditions such as rain, snow or fog, as well as in complex lighting conditions, such as sunrise and sunset and in complete darkness.
It is noted that the platform Drive PX 2 is applied in the current models Model S and Model X, and also in “national” electric car Model 3, which is expected to enter the market in 2017.