Computing platform tackles automated driving

September 17, 2015 // By Christoph Hammerschmidt
Elektrobit, Infineon and Nvidia have jointly developed a computing platform that will facilitate the design of advanced driver assistance systems (ADAS) including controllers foe autonomous driving.

The platform features the high number crunching power required to handle the enormous complexity of automated driving. Systems based on the new platform will be capable of not only being aware of the current vehicle situation but even to interpret it. These findings are then forwarded to the driver or other critical systems in the vehicle, other vehicles in the near environment or to cloud-based downstream data systems.


The solution consists of the Nvidia Drive PX Self-Driving Car computing platform which contains two Nvidia Tegra processors, enhanced by an Aurix Tricore 32-bit microprocessor from Infineon. Elektrobit’s (EB’s) Autosar-compliant EB Tresos software suite running on these three processors is the third integral component of the system. The Aurix microcontroller stands out through its excellent real-time processing capabilities along with sophisticated safety and security functions, all well balanced for use in driver assistance systems. The Nvidia Drive PX platform is capable of handling and processing multiple camera and sensor data streams. In addition it is well suited to handle demanding graphics applications, image recognition and image processing as well as machine learning. The EB Tresos software seamlessly integrates Linux and Autosar applications; in addition, Nvidia Drive PX basic functionalities for monitoring and redundancy management functions can be integrated for the sake of functional safety. The software organises cross communication processes across multiple CPUs in a safe, secure and reliable execution environment which supports ASIL level D, the highest level of functional safety defined in ISO 26262.


“Based on our Autosar software we developed an architectural concept that enables carmakers and suppliers to implement the most advanced driver assistance functions”, advertises Björn Giesler, Senior Director Driver Assistance at EB. “A typical use case for such a platform could be, for instance, a platform for automated parking assistance solutions that needs to process enormous amounts of data and fulfils a high safety level”.


More information: