NVIDIA is building a complete self-driving car platform with Continental
NVIDIA is building a complete self-driving car platform with Continental
German automotive components maker Continental has announced its intention to build a full-scale, self-driving system with NVIDIA. Continental is best known for its tires, but the company also produces a wide range of automotive parts and systems. It is partnering with NVIDIA to use the latter’s DRIVE autonomous vehicle platform for its system, with an eye on launching by 2021.
This could make it easier for automakers to build their own self-driving cars, as they could simply use Continental’s system instead of designing their own.
The platform will use NVIDIA’s Xavier chips, operating system, and software, all of which are part of its DRIVE suite, while Continental will provide its security certification expertise, and LiDAR, camera, and radar sensors.
The Xavier SoCs were recently unveiled at CES 2018, and they feature a custom 8-core CPU, a 512-core Volta GPU, an 8K HDR video processor, a deep learning accelerator and even new computer vision accelerators.
It is also 15 times more efficient than before, according to NVIDIA, and capable of performing 30 trillion operations per second while drawing only 30W of power. This is good news for electric vehicles that require every drop of juice they can make use of, and autonomous vehicles that need to handle huge amounts of sensor and vehicle data.
The goal is to offer self-driving autonomy from levels 2 through 5, which includes capabilities ranging from advanced cruise controls to true self-driving without needing a safety driver or even a steering wheel.
This tie-up is just the latest in NVIDIA’s extensive web of partnerships, which already includes many leading manufacturers, suppliers, and internet giants. Continental isn’t a new player in the field of autonomous driving – last year, it carried out a test with Canadian automotive component maker Magna to attempt the first self-driving car trip across the US-Canada border.
Source: NVIDIA