The special feature of this hardware platform is the use of Graphics Processing Units (GPUs), which makes the training and calculation of neural networks and deep learning applications faster than just using Central Processing Units (CPUs). The algorithms based on neural networks recognize in real-time just how other road users or traffic signs look, and even in complex traffic situations they enable orientation on the road. The simulation software CarMaker can now also be linked to the NVIDIA hardware platform and used to train neural networks in virtual driving tests. It also makes it possible to virtually develop and test deep learning sensor fusion algorithms for controlling driver assistance systems and automated driving functions.

More Information: press.ipg-automotive.com