Autonomous driving must be safe and reliable. This requires trustworthy new electronic components that are highly energy-efficient with high performance and still guarantee basic functionality even in the event of a fault. In conjunction with artificial intelligence (AI) methods, these create the technological basis that allows the vehicle to react appropriately to all driving situations. In this way, autonomous driving will make traffic safer and more efficient and also open up new mobility prospects for people who are unable or unwilling to perform the driving task themselves.
Goals and approach
The goal is to develop a high-performance computing platform for sensor data processing in autonomous vehicles. For this task, novel neuromorphic electronic components will be developed in addition to classical digital ones, which will allow an efficient application of AI methods. In the process, so-called memristors are used to combine data-storing and data-processing components and interconnect them in a network. This enables fast and energy-efficient data processing for environment detection already at the sensors. This in turn reduces the load on the main processor, which controls the reliable and fault-tolerant data processing of the overall system. New interfaces are to be developed for connecting the AI accelerators to an open processor architecture.
Innovations and perspectives
By researching neuromorphic electronics and integrating them into an open processor architecture, KI-PRO is creating a basis for the future development of application-specific, high-performance and energy-efficient microelectronics solutions in Germany and Europe.