GPU developer Nvidia launched today its HD Mapping program, describing it as a highly-efficient, high-definition mapping program for automakers, Tier One suppliers, and startups that are developing autonomous vehicle computing systems.
Co-Founder and CEO Jen-Hsun Huang announced the new system during his keynote address at the Nvidia GPU Technology Conference in San Jose, Calif. The conference began April 4 and runs to April 7 at the McEnery Convention Center.
The new HD Mapping program moves the data processing of data into the vehicle itself, reducing communication with the cloud. That means map creation and updating can occur in “near real-time,” according to Nvidia.
The GPU Conference has grown as the GPU evolved from a graphics chip for gaming to an essential component of driverless technology. The top GPUs can process simultaneously massive amounts of data and graphics. Nvidia says more than 110 companies are represented at the event, along with 95 startups.
Nvidia’s Graphic Processing Units are used by such auto OEMs as Audi, BMW, Honda, Tesla, and Volvo.
It launched its Drive PX 2 chip for driverless cars at the CES 2016 in Las Vegas, claiming it could run 24 trillion operations per second, and power 12 video cameras, LiDAR, radar, and other sensors. DriveWorks software tools and modules were also introduced.
In the Netherlands in January Nvidia’s system on a chip helped to control the steering and braking of the six-passenger shuttles in a live test of the autonomous WEPod. The vehicles max out at 25 kilometers per hour.
The shuttle can be called with a smartphone app. A human operator monitors the WEPod from a control room and can take control if necessary. (See “WEPod Autonomous Shuttle Takes 6 Passengers for a Successful Ride on Dutch Public Roads”).
Nvidia (NVDA) is a component stock of the Driverless Transportation (D20) Stock Index.