Continental develops auto-driving computing platform with Nvidia
The close engineering partnership both companies have agreed upon will enable the production of AI computer systems that scale from automated Level 2 features through full Level 5 self-driving capabilities, where the vehicle has no steering wheel or pedals.
Dedicated engineering teams from each company will work together to develop self-driving solutions based on the Nvidia DRIVE platform — which includes the Nvidia DRIVE Xavier high-performance system-on-a-chip, as well as the Nvidia DRIVE OS operating system and DRIVE AV (autonomous vehicle) software stacks.
The solutions will utilize Continental’s experience in system and software engineering for ASIL-D rated safety — the highest rating level — and integrate a range of Continental sensor technologies, including radar, camera and high-resolution 3D lidar.
See also: AI computing platform targets level 5 autonomy
“We now have all the key elements in place to take AI self-driving cars from development to mass production,” said Jensen Huang, founder and CEO of Nvidia. “Our newly arrived DRIVE Xavier processor, extensive Nvidia DRIVE software, and cloud-to-car approach for testing, validation and functional safety, combined with Continental’s expertise and global reach, will bring autonomous cars to the world.”
As the brain of the Continental system, Nvidia DRIVE Xavier can deliver 30 TOPS (trillion operations per second) for deep learning, while consuming only 30 watts of energy. This level of performance is necessary to handle the massive amount of data processing that self-driving vehicles must perform. These include running deep neural networks to sense surroundings, understand the environment, localize the vehicle on an HD map, predict the behavior and position of other objects, compute vehicle dynamics, and plan a safe path forward.
Continental and Nvidia will initially develop highly automated driving features, including 360-degree perception and automatic lane changing on highways, plus the ability to merge in traffic. In addition, the system will integrate HD maps, enabling vehicles to localize themselves and provide mapping updates.
Continental’s expertise in advanced driver assistance systems integrates multi-function cameras, fish-eye cameras with surround view, short- and long-range radar sensors, high-resolution 3D lidar technologies, as well as central control units for assisted and automated driving. In 2016, Continental’s sales of advanced driver assistance systems exceeded €1.2 billion and the company expects it to grow to €2.5 billion by 2020.
See also: Korean researchers devise single-chip brains for autonomous cars
Before the current deal with Continental, Nvidia already stroke similar deals for AI-enabled platforms based on Nvidia’s chips, with Bosch, Daimler and Volkswagen being only the most recent ones.
Mercedes HMI makes massive use of Nvidia architecture
Volkswagen and Nvidia plan to make cars smarter
Voice assistant will clean up the car cockpit