Last year, Daimler and electronics company Bosch announced that they will join forces to develop a fully automated system for autonomous driving as well as for other things.
However, the two companies have gone a step further on this issue and announced they have chosen Nvidia’s Drive Pegasus as a computer intelligence platform (AI) and will start testing their stand-alone vehicles in California in the second half of 2019.
“Automated vehicles are in fact complex computers,” the two companies said in a press release “and they need even more computing power when they are in adverse conditions, such as traffic jams in the city, which need to work at the same time a number of different sensors environment to process the data. ”
Under the terms of the agreement with Nvidia, the chip manufacturer from Santa Clara, USA, will provide a powerful processor and a graphics chip for managing the electronic control unit (ECU) – and microcontrollers equipped with a sensor that controls its transmission motor traffic, as well as software for the proper operation of the algorithms created by the autonomous driving simulations.
Additionally, the customized cooling system developed jointly by Bosch and Daimler will keep the ECU cool and carry up to 100 gigabytes of data per kilometer of mileage.
“Our collaboration with Daimler and Bosch will unite the strengths of each company involved in this project,” said Danny Shapiro, director of the Nvidia Automotive Division. “Bosch, the world’s largest automotive parts and parts supplier, brings the know-how it has in its field and, combined with Daimler’s experience, will achieve a great job that will stand out from the rest competing in this kind.”
Nvidia unveiled the Nvidia Drive PX Pegasus platform at GTC Europe in October 2017 in Munich, Germany, the latest generation platform for autonomous vehicle driving. The two Nvidia Xavier systems and dual graphics cards are capable of more than 320 trillion operations per second, transferring up to one terabyte of data per second, and receiving data from 16 cameras and six lidar sensors.
Its data centers, which utilize the Nvidia DGX-optimized architecture and run the AV Drive Constellation emulator to detect objects, process algorithms for route planning, and more.