Daimler, the parent company of Mercedes-Benz, and major auto part manufacturer Bosch made a surprise announcement earlier this week that they would collaborate on future autonomous driving taxis and that the service will start in the heart of Silicon Valley.
To power the fleet of Super Ubers will be Nvidia’s DRIVE Pegasus AI supercomputer designed specifically for autonomous vehicles that can deliver 320 TOPS (trillions of operations per second) to help them navigate the mean streets of Palo Alto, California.
A far cry from the room-sized supercomputers seen in sci-fi films, Nvidia’s hardware is the size of a license plate, but packs in the performance equivalent of six desktop workstations.
The goal of the partnership is to put self-driving cars on the road by 2021, at which point the companies will begin offering rides for free on "selected routes" to a limited number of customers.
Data for days
What Nvidia adds to the mix is obviously the brains behind the brawn - autonomous vehicles, especially those in Level 4 and 5 of autonomy, require a vast amount of data to function.
According to Bosch, a single video camera generates 100 gigabytes of data per kilometer, which then needs to be processed and properly converted into driving commands.
The Bosch, Daimler and Nvidia partnership isn't the only name in the autonomous vehicle game at the moment, and it faces competition from BMW, Intel and Delphi who have a similar arrangement, as well as Waymo, GM and Uber itself.
Who's going to win the race to build the first mass-market autonomous vehicle? We'll just have to wait and see.
Source: Nvidia Blog
"Knowledge and community" - IBM and the benefits of the AI Alliance one year on
Analog AI chip startup wants to achieve 1 POPS per Watt: 1000 TOPS/W is 'within reach' says firm co-founder as he looks to analog-first Gigabit-sized SLMs
My dream Christmas gift is Samsung's The Frame TV and it's down to a record-low price
