Bosch and Daimler have been working together to roll out future autonomous cars and have announced that they will source a new AI platform in that effort. Both companies will source the Drive Pegasus platform AI processors and software from NVIDIA. The tech will be used to help autonomous and automated vehicle navigate the complex environment of congested city streets.
The agreement between Bosch, Daimler, and NVIDIA will see NVIDIA supply AI platform components needed. The Drive Pegasus platform will be provided under the contract that includes processors and software allowing the processing of vehicle-driving algorithms generated by Bosch and Daimler using machine learning methods.
The ECU network will reach a computing capacity that can support hundreds of trillions of operations per second. The firms note that this is akin to the performance delivered by at least six synchronized and highly advanced desktop computer workstations. NVIDIA will also participate in the development of the platform with the other companies.
This powerful AI system will be leveraged to handle all the information that is gathered by the different systems in the car including radar, video, lidar, and ultrasonic sensors. The video sensor alone in that scenario gathers 100 gigabytes of data in a single kilometer according to Daimler. The system is able to asses that information and plan the vehicle path in fractions of a second. Daimler says that this speed is on par with how long it takes the sensation of touch to reach the human brain.
That length of time is said to be between 20 and 500 milliseconds. The ECU network will be integrated into the battery cell cooling circuit inside the car. Mercedes is planning on EVs for autonomous urban driving vehicles and those vehicles will have a cooling system onboard. Bosch and Daimler have been working together since August of 2017 on autonomous vehicles for city streets.