| Abstract: |
|
Automated driving functions are becoming more and more important to customers and OEMs. The more complex these functions are, the more computing power is required to cope with the requirements of this functions in respect to performance, safety and security. This opened recently the market for ADAS computing platforms, which enter now the automotive market in a bigger scale. However, there is a high dynamic in this market and ADAS platforms are developing much faster than expected one or two years ago. This is in line with the change of the E/E Architecture in cars that are getting more connected, automated and electrified. The talk will give a view on how ADAS computing platforms will develop within the next years, how to manage the volatility and uncertain partitioning of the system and how to monitor the high processing performance. Therefore the integration of specific HW-accelerators to compute graphics and Deep Neural Networks (DNNs) in such computing platforms will be discussed. |
|
;){kind=link}