June 18, 2024

The role of supercomputing in enabling autonomous cars

Highly automated and networked cars require reliable computer systems

A research project aims to ensure that autonomous cars can operate with a high degree of automation and connectivity on the road in the future, while maintaining safety and reliability. This effort also aims to bolster the competitiveness of the German automotive industry in the years to come.

To enable the safe operation of autonomous vehicles on the roads, powerful computer systems are needed in both the vehicles and the road infrastructure, as well as in data centers. The CeCaS (CentralCarServer) research project aims to address this challenge. Infineon Technologies Semiconductor is leading the project, which involves 30 partners from industry and research. Together, they are developing architectures, software engineering principles, and implementation methods for high-performance computers in cars. Bosch, Continental Automotive, ZF Friedrichshafen, and Hella are among the companies participating in the project.

KIT and TUM: Scientific Coordination for Connected Cars

The Karlsruhe Institute of Technology (KIT) and the Technical University of Munich (TUM) are responsible for scientific coordination of the CeCaS (CentralCarServer) research project. The aim of this project, with coordination by Infineon Technologies Semiconductor, is to develop energy-efficient and cost-effective high-performance computers that meet the demanding requirements of the automotive industry for computing power and scalability. The project includes involvement from industry leaders such as Bosch, Continental Automotive, ZF Friedrichshafen, and Hella. According to Jürgen Becker, head of the Institute for Information Processing Technology (ITIV) at KIT, the development of such high-performance computers is essential for the future competitiveness and technological sovereignty of the German automotive industry. CeCaS aims to create automotive supercomputing that meets the highest standards for safety and reliability.

As part of the CeCaS project, there is a development of processors, interfaces, and system architectures. They are working to create a flexible software environment that can cater to the latest algorithms in the vehicle. The project mainly focuses on the utilization of artificial intelligence (AI).

Hardware accelerator for future cars

The project also includes the development of novel multi-purpose hardware accelerators by KIT for efficient image processing and the reliable integration of AI in cars. These accelerators will be connected to high-speed interfaces and integrated into high-performance processors, with a particular focus on the AI components between the sensor nodes and central computer. Moreover, the researchers are working on new development tools to analyze and meet real-time criteria, as well as benchmarking software to assess hardware accelerators.

Autonomous cars: security against attacks

As a component of the project, the scientists from TUM’s Chair for Information Security are dedicated to developing a Safe&Secure memory controller that prioritizes safety and security. Specifically, the research team aims to protect both volatile and non-volatile memory from cold boot or memory manipulation attacks using encryption and integrity protection mechanisms. They also plan to examine the combination of encryption with error correction code.

Modern chip technologies for future cars

According to expert Becker from KIT, the progress of automotive technology is closely linked to the progress of computer and information technology, and particularly the ability of the automotive industry to leverage modern chip technologies.

The CeCaS research project, which has a project volume of nearly €90 million, is planned to continue for three years. The project is being funded by the Federal Ministry of Education and Research (BMBF) through its Mannheim initiative, named after the birthplace of the automobile, with around €46 million.

About The Author