Infineon develops supercomputers for highly automated, connected cars

     According to Infineon Technologies, high-performance computers built into the car that process all available data and information as quickly and reliably as possible so that the vehicle can be driven safely are key to an automated connected car. The Mannheim-Cecas research project aims to develop a corresponding supercomputing platform for vehicles. The project, which involves 30 research partners from the industry and universities, has been included in the German federal government's large-scale funding program aimed at promoting the digital transformation of the automotive and mobility industries. Infineon led and coordinated the project.

     Peter Schiefer, president of Automotive electronics at Infineon Technologies, said: "The popularity of driver assistance systems means that the future of autonomous driving is limitless. We need to combine our strengths to develop a reliable high-performance computing architecture for highly automated vehicles. As a global market leader in automotive semiconductors, Infineon will not only contribute its expertise and technical expertise in automotive systems, but will also be responsible for coordinating research efforts throughout the project."

     The Mannheim-Cecas research project is committed to researching and developing a fully functional central computing platform for the highly automated vehicles of the future, to fill the emerging market gaps in connected and electric vehicles. Even for everyday use, connected cars and electric vehicles will need to be equipped with cost-effective, energy-efficient high-end computers. These computers will have to cope not only with increasing computing demands and complexity, but also with stringent automotive standards. This requires a central computing platform to be developed with security, maximum performance, and automotive supercomputing in mind, including specially designed processors, interfaces, and system architecture.

     The central computing unit will use innovative, high-performance processors for automotive applications, as well as non-planar transistor technology (FinFET). Dedicated hardware accelerators and adaptive software platforms for autonomous vehicles will support these high-performance processors. This is the so-called "convolutional neural network" approach to acceleration and event-driven neuromorphic accelerators. In addition, the on-board power network and the vehicle integrated circuit package also need to make necessary changes. The project team's goal is to achieve full Vehicle Functional Safety Certification (ASIL-D) at the system level.

     The German Federal Ministry of Education and Research (BMBF) funded the Mannheim-Cecas project with around 46 million euros and included it in the government's "Mannheim" programme (named after the birthplace of the car). The plan aims to take automotive development to the next level through digital transformation. The research project is led by Infineon, with the participation of many component suppliers, experts, research institutions and universities. Members of the project include well-known companies such as Bosch, Continental and ZF Friedrichshafen, as well as the Fraunhofer Institute and the Technical University of Munich (TUM) and the Karlsruhe Institute of Technology (KIT). The 30 partners are proposing to raise a total budget of about 90 million euros to develop the concept of an electronic central computer for cars for the future. The three-year Mannheim-Cecas project is one of the most government-funded collaborations in Germany.