Friedrichshafen/Stuttgart. At the ‘auto motor und sport’ congress, Mamatha Chamarthi, Chief Digital Officer of ZF Friedrichshafen AG, and Serkan Arslan, director of automotive in EMEA at Nvidia, showcased the importance of artificial intelligence for the future of the transportation industry, highlighting ZF ProAI, based on the Nvidia DRIVE PX 2 AI platform, as the first product to be ready for volume production of self-driving vehicles.
"Artificial intelligence will be a game changer across the board," said Mamatha Chamarthi, ZF's CDO. "ZF operates in many industries, from automotive to construction, agriculture, and industrial technology. So our advances in artificial intelligence offer a broad customer base substantial added value."
In general terms, the challenge is achieving real-time processing of huge data volumes. This stems from a growing number of sensors coupled with increasingly higher resolutions. Level 3 and 4 autonomous driving as well as many of the predictive functions found in future vehicles are not possible without the help of AI. ZF's vision of making vehicles see, think and act is also transferrable to other applications such as smart factories, autonomous harvesters, and wind farms.
"There's no doubt artificial intelligence and deep learning will define the automotive industry of the future," said Serkan Arslan, Nvidia Director of Automotive in EMEA. "AI will not only improve transportation, but will also enhance the efficiency in development and validation cycles."
Deep learning is crucial to the advancement of AI in the transportation industry and beyond. It is a form of machine learning that interprets signals not on the basis of mathematical rules, but by applying experience-based knowledge.
ZF ProAI – scalable supercomputing power
At the beginning of this year, ZF and Nvidia officially announced a collaboration and unveiled the ZF ProAI supercomputing self-driving system. It uses the scalable Nvidia DRIVE PX 2 AI supercomputing platform to process inputs from an array of cameras, LIDAR, radar and ultrasonic sensors. This enables ZF ProAI to monitor the surroundings of the vehicle in real time. Also, the system can precisely locate the vehicle on HD maps for advanced safe route planning. In addition, ZF ProAI follows all automotive-grade industry requirements, enabling it to function at extreme temperature ranges, deal with high humidity and dust levels, and withstand the rigors of commuting. Volume production is scheduled for 2018.
ZF will supply ZF ProAI as a vehicle-integrated system updated via the cloud. Extra functions and capabilities can be added throughout the lifetime of a vehicle. The system is designed for V2X applications and in turn can communicate with other vehicles as well as the surrounding infrastructure. Furthermore, the learning and networking capacity of ZF ProAI can use swarm intelligence to make an entire vehicle fleet safer and more efficient.