RepliCar - Referenzsensorik zur hochpräzisen Sensorvalidierung für das automatisierte Fahren

Safe perception of the environment is a crucial prerequisite for automated driving. The sensors and processing chains used must meet the highest requirements for reliability, precision and realism. There are currently no approaches for effective and efficient validation of sensors and processing chains. In order to create the missing data basis for validating sensors for automated driving, a reference system consisting of radar, camera, lidar, GNSS and inertial sensors should be built and integrated into a test vehicle. This reference sensor system is several years ahead of today's series sensors and represents the basis for recording precise sensor data. For this purpose, the Institute of Radio Frequency Technology and Electronics (IHE) at the Karlsruhe Institute of Technology (KIT), the Offenburg University of Applied Sciences (HSO) and Freudenberg FST GmbH develop a highly accurate radar system, which is an essential element in the reference system for environmental perception and is described at intervals by the Institute for Automatic Control Systems (IRS) at KIT. The consortium leader – ANavS GmbH – enables the creation of an accurate representation of reality, a so-called “ground truth”, for use in validation processes by developing a precise environment perception and a highly accurate sensor data fusion. As part of the project, the steps from simulative validation to the release of the sensors and functions are defined and implemented by the network partners AKKA, FZI, IAVF, IPG and GTÜ. The individual innovations in the validation and testing process for series sensors include AI methods for analyzing recorded scenarios, a “sensor-in-the-loop” test bench, integration into existing simulation tools and a modular verification and validation process. The partners HQS and RAC are building a high-performance data platform for managing, processing and providing reference data. With the help of the moving demonstrator from Dr. Ing. h.c. F. Porsche AG, the first data is recorded and used for the release process of an exemplary perception system.

The project runs from July 1, 2023 to June 30, 2026, has a planned cost of approximately 10 million euros and is funded by the Federal Ministry of Economic Affairs and Climate Protection (BMWK) with approximately 5 million euros from the federal budget.

Press release: https://www.kit.edu/kit/33251.php